QUALITY ASSURANCE IN DIALYSIS

DEVELOPMENTS IN NEPHROLOGY

Volume 36

The titles published in this series are listed at the end of this volume.
Quality Assurance
in Dialysis

Edited by

LEE W. HENDERSON, M.D., EA.C.P.

Professor, Department of Internal Medicine,
Section of Nephrology, Rush Medical College,
Chicago, illinois, U.S.A. and
Vice President Scientific Affairs,
Baxter Healthcare Corporation, Renal Division

and
RICHARD S. THUMA
Vice President Quality Assurance,
Baxter Healthcare Corporation, Renal Division

Springer-Science+Business Media, B.V.

Library of Congress Cataloging-in-Publication Data
Ou.l1ty assurance In dialysis / edited by Lee H. Henderson and Richard
S. Thuma.
p. cm. -- (Developments In nephrology; v. 36)
Inc I udes Index.

1. Hemodlalysls--Oualtly control. 1. Henderson, L. H. (Lee H.),

1930- II. Thuma, Richard S. III. Series, Developments In
nephrology; 36.
[DNLM, 1. Hemodlalysls--standards. 2. Ouality Assurance, Health
Care--standards. 3. Peritoneal Dlalysls--standards. HI DE998EB v.
36 1994 I HJ 378 011 1994J
RC901.7.H45033 1994
617.4'61059--dc20
DNLM/DLC
for Library of Congress 94-970

ISBN 978-94-015-8299-5 ISBN 978-94-015-8297-1 (eBook)

DOI 10.1007/978-94-015-8297-1

Printed on acid-free paper

All Rights Reserved

© 1994 Springer Science+Business Media Dordrecht

Originally published by Kluwer Academic Publishers in 1994.
Softcover reprint ofthe hardcover 1st edition 1994
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
Table of contents

Preface Vll

List of Contributors Xl

1. Quality Assurance: Starting a Program in Dialysis

John H. Sadler

2. Patient and Therapy Perspectives: Choosing the Patient: "Is Better

VVorse? 11
C. M. Kjellstrand

3. Quality Systems in the Dialysis Center: Peritoneal Dialysis 23

Barbara F. Prowant, Karl D. Nolph, Zbylut J. Twardowski,
Lois M. Schmidt, Leonor Ponferrada and Ramesh Khanna

4. Continuous Quality Improvement in Dialysis: Operations and

Controls for Multi-Center Systems 47
Edward E. Berger and Edmund G. Lowrie

5. Quality of Care in Home Dialysis 63

Christopher R. Blagg

6. VVater Treatment for Hemodialysis 85

Prakash Keshaviah

7. The Impact of Membrane Selection on Quality Assurance in

Dialysis 109
Lee W. Henderson

8. Quality of Life Assurance in Hemodialysis 133

A. Paul Heidenheim and Robert M. Lindsay

v
vi

9. Dialyser Reuse and the Quality of Therapy 151

Lee W. Henderson

10. Multicenter Trials as a Measure for Improving the Quality of

Clinical Decisions 155
David N. Churchill

11. Quality Assurance in Renal Transplantation 167

Robert W. Steiner

12. Quality Assurance in Dialysis Product Manufacturing 191

Richard S. Thuma

13. Human Resource Issues in Quality Management 213

Gail S. Wick

14. Quality Criteria for the Clinical Record 225

Cindy Jo Ping, Susan Gross and Corrine E. Algrim

15. Continuous Quality Improvement and the Best Demonstrated

Practices Program 239
Arthur Holden and Rosalie Villano

Subject Index 253

Preface

This is a time in history when the concept of Quality is reaching new highs
in terms of public awareness. Articles describing quality, CQI, quality tools,
critical success factors, failures, and lessons learned appear in local news-
papers, trade journals, scientific periodicals, and professional publications
on a daily basis, yet implementation of a quality system in many hospital
units is approached with caution and the basic tenants of quality systems and
CQI continue to be misunderstood.
In the United States, today, the public debate on healthcare issues rages
on. The application of strategies, such as cost-benefit analysis as a means
for evaluating addition of new technologies to the healthcare cost structure has
not succeeded in curbing the rise in costs of healthcare services.
Because of this focused attention by third-party payers, federal and state
governments, and insurance companies, healthcare organizations are being
pressured to change. One of the strategies for changing involves implementing
quality assurance practices. The focus on quality should produce improvements
in productivity, innovation, and profitability. But, most importantly, the desired
outcome of a quality assurance program is self-improvement.
In its drive to become more productive and more competitive, industry
looked to such Quality Leaders as W. Edwards Deming, and J.M. Juran for
ideas. According to Deming, the way to become competitive is to undergo a
top-to-bottom quality-based transformation of the organization. This quality
transformation will produce the productivity improvements, innovation, and
profitability increases needed. The Deming philosophy emphasizes that quality
is never fully achieved; process improvement is never ending.
But, what is quality? Without defining it, David Garvin makes the point that
"in its original form, quality activities were reactive and inspection-oriented;
today, quality related activities have broadened and are seen as essential for
strategic success."! How can the broad context of quality be applied to the
diverse aspects of ESRD?
Furthermore, although far from a new concept, Continuous Quality Improve-
ment (CQI) has taken its place as a dominant theme in many industries.CQI

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, vii-ix.

© 1994 Kluwer Academic Publishers.
viii

is more broadly applicable, both in concept and execution, to service as well

as manufacturing-based operations. How have the variously described elements
of Continuous Quality Improvement been linked to aspects of renal therapy
strategies?
Many industries are now concentrating on customer satisfaction as a means
for creating competitive advantage. What are the implications of the recent
focus on customer satisfaction and meeting customer requirements with respect
to the process of caring for patients with end stage renal disease?
How do the concepts of quality in a dialysis unit stack up against the
issues of quality of life?
The editors believe that this book will aid in addressing some of those ques-
tions. The contents, although quite diverse, attempt to address some aspects of
the question posed above.

Overview of Contents

Chapter one explores the basics of starting a quality program in dialysis. Dr.
Sadler makes the point, which is repeated many times throughout this volume,
that Inspection or top-down, retrospective strategies using a set of specifica-
tions and threshold limits on processes in order to detect failures or defect
levels that exceed the threshold is inappropriate in the dialysis setting. The
pursuit of progressively greater excellence through systematic analysis of
processes and their direct, immediate outcomes is central to CQI.
Dr. Kjellstrand introduces an interesting idea: that, contrary to popular
opinion, poor outcomes in dialysis may be indicative that the physician has
fulfilled his obligation to a larger number of truly sick patients much better
than one whose survival statistics look more favorable. Thus, the insistence
that morbidity and mortality represent indicators of quality of care may actually
be detrimental to quality of care.
In their description of quality systems in a dialysis center, Ms. Prowant
et at. organize their discussion around the eight characteristics of successful
companies identified by Peters and Waterman in their book In Search of
Excellence 2 and apply those ideas to center operation.
The advantages of multi-center operations and application of CQI princi-
ples is discussed in Dr. Berger's chapter. Berger asserts that the notion of
CQI has been so rapidly adopted as "gospel" that it is losing its meaning.
Berger and Lowrie describe critical success factors for implementation of
CQI in a multi -center setting.
Aspects of quality of care in the home dialysis setting is covered by Dr.
Christopher Blagg. In his discussion, he uses the JCAHO Accreditation Manual
for Home Care. 3 It is just as important there be a active QA assessments
of the care provided these patients as for patients dialyzing in a dialysis unit.
Dr. Lee Henderson discusses quality of life and quality of care and their
relationship to membrane selection. Dr. Henderson sees membrane selection
 ix

as the "single most important primary variable in the dialysis prescription

that can be selected by the dialysis unit director".
Following up on the theme of quality of life, Dr. Lindsay writes about the
measurement of quality of life for the dialysis patient. He believes that by
paying careful attention to the delivered therapy and the patient, medical
practitioners can significantly influence not only quantity of life but quality
of life.
Dr. Churchill applies quality assurance principles to multicenter clinical
trials. He believes that the results of clinical trials conducted using these quality
methods can be used to establish standards for clinical care.
Dr. Robert Steiner looks at both the best and worst features of quality assur-
ance as it is applied to renal transplantation programs. At its best, the quality
process helps the organization define broad goals and meet them by using
systematic methods to address activities and structures. At its worst, QA results
in needless activities, clerically oriented, that consume valuable resources
and produce almost no real improvement. The staff "focuses narrowly
on statistical goals ... [and] attempts to collect data to document lack of
deficiency" because of fear of criticism or regulatory sanctions.
Central to any discussion of QA systems and CQI is human resources.
Gail Wick addresses fundamental human resource issues in quality manage-
ment as applied to the dialysis center.
In a similar way, Ping, Gross, and Algrim apply quality principles to the
clinical record. They use a continuous quality improvement model to evaluate,
assess, and reduce risks, implement cost effective resource management
measures, and identify potential problem areas.
Finally, Arthur Holden and Rosemary Villano relate continuous quality
improvement to the Best Demonstrated Practices program (BOP). Imbedded
in the BOP program are several quality improvement concepts and tools,
including goal setting, eduction, communication, team-work, and so on.
Several technical discussions of quality applied to water systems, dialyzer
reuse and dialysis product manufacturing are also presented.

Notes

1. Garvin D.: Managing Quality, New York: The Free Press, 1988.
2. Peters TJ, Waterman RH Jr: In Search of Excellence. New York: Harper & Row, 1982.
3. Held PJ, Brunner P, Odaka M, Garcia JR, Port FK, Galin DS: Five-year survival for
end-stage renal disease patients in the United States, Europe, and Japan, 1982 to 1987.
Am J Kidney Dis 15: 451-457, 1990.
List of Contributors

Corrine E. Algrim, R.N., B.S.N. Lee W. Henderson, M.D., F.A.C.P.

Baxter Healthcare Corporation Baxter Healthcare Corporation
1620 Waukegan Road, MPD-Dl 1620 Waukegan Rd., MPR-A2N
Waukegan, IL 60085 McGaw Park, IL 60085
U.S.A. U.S.A.

Arthur Holden, Ph.D.

Edward E. Berger, Ph.D.
Baxter Healthcare Corporation
Vice President for Government Relations 1620 Waukegan Rd., MPR-AZN
National Medical Care, Inc.
Waukegan, IL 60085
Reservoir Place
U.S.A.
1601 Trapelo Road
Waltham, Massachusetts 02154 Prakash Keshaviah, Ph.D.
U.S.A. Suite 722
825 South 8th Street
Christopher R. Blagg, M.D. Minneapolis, MN 55404
Northwest Kidney Ceters U.S.A.
700 Broadway
Seattle, WA 98122 Ramesh Khanna, M.D.
U.S.A. School of Medicine
MA436 Health Sciences Center
University of Missouri-Columbia
D.N. Churchill, M.D., F.A.C.P., F.R.C.P.C. Columbia, Missouri 65212
Department of Medicine U.S.A.
McMaster University
50 Charlton Avenue, EAst Carl M. Kjellstrand, M.D., Ph.D., F.A.C.P.,
Hamilton, Ontario, Canada F.R.C.P.
L8N 4A6 2E3-3l W C Mackenzie
Department of Medicine
Susan Gross, R.M., M.N.A. University of Alberta Hospital
5784 South Lansing Way Edmonton, Alberta, Canada
Inglewood, Colorado 80111 T6G2B
U.S.A.
Robert Lindsay, M.D., ERC.P.F., F.R.C.P.C.,
F.A.C.P.
A. Paul Heidenheim, M.Soc. Director Renal Unit
Victoria Hospital Corp. Victoria Hospital Corp.
375 South Street 375 South Street
London, Ontario, Canada London, Ontario, Canada
N6A 4G5 N6A 405

xi
xii
Edmund G. Lowrie, M.D.
National Medical Care
1601 Trapelo Road
Waltham, MA 02154
U.S.A.
Karl D. Nolph, M.D.
Department of Medicine
University of Missouri Medical Center
807 Stadium
Columbia, MO 05212
U.S.A.
Cindy Jo Ping, R.N., B.S.N.
Baxter Healthcare Corporation
Rt. 120 and Wilson Road RLT-06
Round Lake, IL 60073
U.S.A.
Leonor Ponferrada, B.S., R.N.
Dialysis Clinic Inc.
3300 LeMone Industrial Boulevard
Columbia, Missouri 65201
U.S.A.
Barbara Prowant, R.N., M.S.N.
Dialysis Clinic Inc.
3300 LeMone Industrial Boulevard
Columbia, Missouri 65201
U.S.A.
John H., Sadler, Ph.D.
University of Hospital
22 S. Green St.
Baltimore, MD 21201
U.S.A.
Lois M. Schmidt, B.S., R.N.
2122 Fairfax Avenue Apt #9
Nashville, TN 37212
U.S.A.
Robert W. Steiner, M.D.
UCSD Medical Center
225 Dickinson Street
San Diego, CA 92103
U.S.A.
Richard S. Thuma
Baxter Healthcare Corporation
1620 Waukegan Rd., MPR-A2N
McGaw Park, IL 60085
U.S.A.
Zbylut J. Twardowski, M.D.
Professor of Medicine
MA436 Health Sciences Center
University of Missouri-Columbia
Columbia, Missouri 65212
U.S.A.
Rosalie B. Villano, M.S., M.P.P.
Baxter Healthcare Corporation
1620 Waukegan Road MPR-A2E
McGaw Park, IL 60085
U.S.A.
Gail Wick, B.S.N., R.N., C.N.N.
Gail Wick & Associates
5420 New Wellington Close
Altanta, Georgia 30327
U.S.A.
CHAPTER 1

Quality assurance: Starting a program In dialysis

JOHN H. SADLER

The words, "quality assurance" are familiar to most health professionals, but
not explicitly defined. Starting a formal QA program may seem redundant
(we already do that) or cumbersome (we don't have space or money for staff
who aren't productive) or inappropriate (the Network and surveyors do that).
In fact, good QA practices are not just a method to protect oneself from
critics and government review, but a systematic approach to reorganizing all
practices, assessing and improving them.
Program leaders such as the Medical Director, Nurse Manager, Admin-
istrator, CEO, must first become aware of the importance and utility of QA
throughout facility operations and commit to QA as an operating principle.
The federal structure surrounding ESRD therapy and the recurring and limited
range of dialysis services encourage that awareness and make serial, com-
parable observations accessible. The leaders must then encourage quality
assurance practices as an element of every operation. Properly done, these
practices are the prime instrument for staff education and improving perfor-
mance.
The most important function of quality assurance is self-improvement.
Fulfillment of that function will have byproducts which will document
activities to external agencies, establish a database for analysis and studies,
assist in staff evaluations and cost effectiveness analysis.
This approach to quality assurance is called Continuous Quality Improve-
ment (CQI). That contrasts with the widely practiced (by professional review.
organizations [PROs], federal surveyors and others) or external, "top-down"
QA review which imposes external standards and seeks to discover failures.
External review defines a threshold level of functioning which must, as a
minimum, be met. If it is met, no sanctions will be imposed. If it is exceeded,
no action is required. External review only seeks to protect from inadequate
care. It has no system for promoting good quality of-care, only for punishing
poor quality care. Because of its approach, which is at least suspicious if not
threatening, the response is often resistance, defensiveness, and suspicion in
return, which sours the atmosphere. This atmosphere has undermined many
well meaning efforts at QA and has led others to establish an internal staff,

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 1-9.

© 1994 Kluwer Academic Publishers.
2

outside operational clinicians, to attempt to practice QA in defense. Skepticism

is widespread.
Continuous quality improvement does not set thresholds but pursues
progressively greater excellence through systematic analysis of processes, their
direct, immediate (proximate) outcomes, and allowing the operator of that
process to examine its result and to make changes to improve the result
and/or the process. CQI permeates all practices and brings critical review to
the level of the person doing the practice, who ought to be most interested
in its quality. Collegial internal discussion rather than regulatory oversight
can stimulate each staff member to be his or her best. Simply stated, it is critical
scrutiny of each task and its product. The results obtained are data which
can be aggregated for further analysis to enable rational changes in practice.
The same technique will evaluate the effect of whatever change is made.
This internal, immediate review by those involved lets them take an active role
in practice decisions. Rather than being threatened from outside, they are
free to improve what they do as a part of doing it. Whether it is tabulating
complications or tracking chemistries, this review will leave a record demon-
strating concern for quality and action to improve it.
Other data from national or regional sources such as the United States Renal
Data System, Medicare claims data tabulation distributed by HCFA or the renal
Networks provide useful yardsticks for comparison of outcomes and effec-
tiveness. These are becoming more available in ESRD programs as Networks
are given more access to HCFA processed data from claims and registration
of ESRD patients. The Network can then send facilities their specific report
and comparable values aggregated nationally, by network, by state, or other
grouping. In this way a facility may compare hospitalizations, access revisions
and replacement, standardized mortality ratios, transplant rates, and other
events not yet reported but able to be derived from claims data.
All forms of QA practice require time and support; in other words, money.
The absence of money designated to QA is one way to undermine it. The
composite rate payment for ESRD services is all-inclusive, but has not included
any cost estimation for QA. Inflationary reduction in that payment leaves
little to cover costs outside direct therapy. While it is possible that quality
of outcomes will improve in the absence of QA measures, effectively applied
CQI practices can guarantee maximum achievement of excellent performance
in a given system. As experience is gained, the results can make clear whether
more or less frequent monitoring is needed, possibly increasing cost effec-
tiveness.
Since CQI should maximize productivity and efficiency, there may be saving
to offset some QA costs.
Having made the commitment to Continuous Quality Improvement in
principle, how does that become policy and practice? Each clinician at every
level of operations can define its steps and the proximate outcome of each step,
select monitoring appropriate to current practices, set goals for improving prac-
tices. Discussion groups can review each other's outline and make suggestions.
 3

Those staff directly involved then have a means to make adjustments for
improvement without bureaucratic constraints. Careful records are the source
of accountability. Control of the QA process passes to the person doing the
procedure and seeing its outcome: the one most gratified by improving it.
Aggregating outcomes allows insights which might not be realized as each
single event occurs. Internal critical scrutiny provides the basis for efficiently
making decisions following which the results are reviewed in the same way
as the observations leading to the change. If the change undertaken improves
outcomes or facilitates the process, assessment has led to greater quality of
performance. The practice is no longer a periodic oversight but a continuum
of critical scrutiny and data collection which is part of every activity.
Selection of specific items to track will improve with practice. Those
observations need to be explicit and informative rather than general statements
about the process. For instance, "post dialysis bleeding from (one) or (both)
puncture sites" rather than "access problem." The data needs to be useful by
itself, not a marker that there is a record of an incident to be reviewed. Check
off headings allow easy notation of events and unequivocal entry of which
event. It is better to have 100 specific items which may be noted with a
check or a date than to have 25 items pointing out the need to look back at
the record for details.
As an example, consider that simple but most anxiety producing step for
the patient, fistula puncture. Usually routine but potentially disruptive of
care, it is readily susceptible to analysis through aggregating experience to find
relationships to specific patients, staff members, or shifts. It allows factual
assessment of different products such as needles, their placement or method
of securing. Similarly, review of routine chemistries and of pressures in the
blood circuit should alert the team to spot the possibility of recirculation of
blood. Standing orders may be developed to allow clinical staff to proceed with
tests for recirculation when observations indicate it is likely. This kind of direct
involvement is helpful for learning, for responsibility, and ultimately, for
prompt intervention or understanding of the patient's condition, which results
in better care.
Such an analysis requires time but also assures that each involved clini-
cian shares the fundamental knowledge base on which greater expertise is built,
can assess personal performance as well as that of others, can understand
and perform tests such as those to determine recirculation, and has the aware-
ness that he or she may, by recording the justification, proceed independently
to get information for clarifying or resolving issues. Result: better staff, better
satisfaction, and better patient care.
Similar approaches are equally applicable to peritoneal dialysis. The patient
is in frequent contact and has repetitive visits for routine observation.
Noteworthy observations may easily be places in a format to allow checks
or entry of date of observation of "pain at the end of drainage" or purulent
drainage from exit" as tabulated events as well as a narrative record of the
circumstances. Tracking the patient's weight and blood pressure become part
4

of the matrix for scrutiny over time through being recorded in a format for
easy, repetitive review.
Decentralization of oversight creates opportunities for immediate recogni-
tion of problems and prompt response by the operator, diminishes risks to
that same operator, and allows self-management and pride in achievement.
Since the records are an output of the action, staff evaluation can improve with
good records. The clarity and usefulness of records often improve in such a
system. Hierarchical relationships - the chain of authority - ideally become
resources as well as constraints.
When the steps of each process are evaluated separately, redundant and
ineffective actions can be detected and eliminated. Idiosyncratic practices
may be recognized, corrected where appropriate, or turned into standard
procedure if the unique method offers advantages. Learning and communi-
cating increase, and perfunctory or casual practices which are likely sources
of inconsistent quality of outcomes can be found, and improved by those
employing them. The general practice of critical scrutiny can lead to improved
morale. CQI brings the scope of the review down to a manageable level
where a sense of control in QA is returned to the clinicians.
The possibility of such results from instilling CQI into every aspect of
clinical care is not a certainty. Leading examples, open and constructive
responses to scrutiny, and total fairness to all are elements needed to succeed.
Each person is a player in this QA scheme. It is internal, imbedded in all
practices as part of the process, its outcome, and its goals.
Arvidis Donabedian defined the elements of QA as structure, process, and
outcome. As practiced for years, structure has been assessed as credentials,
physical plant, instruments and devices, numbers of sites, staff, and mate-
rials; none of which gave assurance of how it is used. Process assessment
became routine through review of written procedures, defined responsibili-
ties and documentation of timing and completion of steps; none of which
gave assurance that the process was carried out effectively or efficiently. Lately
most discussion of QA has put down these measures and endorsed outcome
assessment as the gold standard, looking at measured results; complications,
costs, death, or recovery as examples. In the absence of case-mix or complexity
data (structure) and the path to the result (process) no comparisons could be
critically made. Raw outcomes, unadjusted by these factors, allow no rational
review of care on outcome analysis alone.
In health care, all three aspects of care are important in assessing the ultimate
or comparative merit of the service under review. That is true whether the
service is a diagnostic test, medication, or procedure.
With increasing demand for cost effectiveness as well as (or even instead
of) overall effectiveness, cost assessment sometimes threatens to become the
most important measurement of acceptability. Only an effective and realistic
assessment of the quality of care will give confidence to payers, health
professionals, and the public that the service is a good value, worth what it
costs.
 5

We should anticipate broadening the evaluation to include patient satis-

faction, functional assessment, and health status; measures earlier felt to be
social rather than medical or too "soft" for objective analysis. Health status
measures have now demonstrated validity in several studies. The ESRD com-
munity can learn from these reports and expand its assessment of outcomes.
In the final analysis, the outcome which best defines quality is for the patient
to be most vigorous, most independent, with minimum disturbance of usual
productive activities. Most medical assessments note problems and leave us
assuming that the absence of problems represents quality. The use of health
and functional status measures allow the patient's assessment of the overall
outcome or of specific functions to become a positive statement of benefi-
cial outcomes. These studies are still time consuming but are improving. The
analysis of their results is also being refined and with experience, made more
lucid and useful to clinicians.

Human aspects

In medical settings, the analysis of operations and outcomes is not as straight-

forward as in assessing variation in products from an assembly line. The
outcome desired is benefit to the patient; which is significantly, indeed totally
constrained by the conditions and capabilities with which that patient started.
No diagnostic assessment of a short person will make that person tall. No
therapy will achieve performance levels for a patient which were beyond his
capability if the disorder being treated had never existed.
Prolongation of life will be more successful if the life expectancy was
long without disease; rehabilitation will not likely exceed the pre-disease
capability of the patient; the nature of the disease or disorder and its accom-
panying (co-morbid) conditions will determine what goals may reasonably
be set. The full awareness of the starting point, sometimes called case-mix
or severity analysis, but perhaps best defined as patient complexity is a require-
ment for evaluating the outcomes at any point in the course. A manufacturer
may set specifications for the materials to be used in producing a product
and require tight limits on any variance. The clinician, on the other hand
must take what comes to the clinical setting. We can't insist that our patients
all be young, talented, educated and motivated. That leaves several aspects
of the outcome out of real control - such as compliance - and this must be
part of the analysis of care, its appropriateness and quality, and its relative
outcomes in different patients. Without adjusting for these factors, judgment
of relative outcomes is pointless.
Understanding patient complexity is the first step in assessing outcome
quality. The absence of complexity data is one of the critical weaknesses in
current ESRD data systems and in ESRD quality assessment. Such informa-
tion exits and is readily available at onset, but soon drifts into background if
not tabulated in an accessible and useful form. It is not practicable to go
6

back and find it; the detail needed to understand that patient is often not
recorded. Memory fails. Subsequent review can't distinguish between indi-
viduals with the same diagnosis but with widely different effects of the disease.
For example, there is usually a major difference between insulin dependent
diabetes mellitus and non-insulin dependent diabetes mellitus, and further great
variation in effects of the disease on individuals within each group. HCFA's
ESRD data does not yet even distinguish the groups, much less the effects
or complications. We need to know. Only the clinical team can get the
information and must get it at the beginning.
If outcomes of care are to be evaluated, first there must be characteriza-
tion of the patient on intake into the program or as soon as possible afterward.
Age, sex, race, primary diagnosis causing ESRD are usually recorded. These
have a bearing on what outcomes are possible. At the same time, secondary
diagnoses and co-morbid states are not routinely tabulated and make a dif-
ference too. For instance, the quality and security of access to the circulation
has a critical bearing on both success and the effort required to reach that
success. Tobacco and alcohol use may limit survival more than renal failure.
Loss of vision, hearing,an extremity or even a close family member may be
severely limiting or require extra measures to avoid deterioration. Such facts
about each patient should be tabulated in easily usable form as characteris-
tics for sorting, not only as part of a narrative.
Had we done a better job of analyzing patient variation we might have come
to objective criteria for adequacy of dialysis treatment which would be
generally acceptable. Including complexity in the description of clinical cir-
cumstances might help to make objective the efficient decisions which
experienced clinicians make through their practice "short cuts."

Implementation

Once the leadership of a facility or program is informed about continuous

quality improvement principles, then the stage is set. It must begin at the
top. Staff may then be taught those same principles and allowed to partici-
pate in deciding how to apply them in their setting. At the beginning, as
throughout, it is not a QA committee function but an idea permeating the entire
operation, producing assessment of quality of system, performance, and results
as a part of each activity, carried out by each individual. The example of the
leadership is central to success.
All staff must recognize that CQI presumes each of them wants to do the
best he or she can, and will trust them to rationally look at what they do and
thoughtfully seek to improve. Then the program must develop a plan for data
collection and analysis so that experience of multiple procedures can give
insights which are not clear to one or a few operations.
It helps to have a computer, but this kind of aggregation is possible in a
dialysis facility by designing forms to put the information in a style and
 7

position to be easily collected. Initial recording should be adequate for both

clinical record keeping and analysis for improvement. Duplicate notation will
discourage the operator with its extra effort before any benefit could be
achieved Once for all purposes should be the rule whenever possible.
The CQI gurus point out that there are many customers for each process
or outcome; sometimes you serve directly the patient, sometimes the institu-
tion, sometimes the colleague serving beside you, and sometimes a laboratory
or agency. Many times a service is more for collective benefit of a group
than any individual. What this idea tells us is simply that each "customer" is
a legitimate evaluator of what is done. That only reinforces the CQI idea
that quality is not imposed from above, but assessment of quality is part of
all that each one does.
The outcomes often examined are death, hospitalization, procedures,
untoward events, debilitation. In CQI we are to look more minutely at processes
and outcomes. Effective and uncomplicated placement of needles in the fistula,
good blood flow, and removing them without mishap or more than minimal
bleeding in a reasonable time are at least three outcomes easily noted, directly
related to the process, and very important to both the quality of the care and
quality of that portion of the patient's life.
Effective anticoagulation may be observed both from the perspective of
the individual patient's results or a general overview of how consistently the
practice is done and how consistently the results match the desired goal of
no clotting and no bleeding, or the monitored result of clotting studies in the
specified range.
Control of the patient's final weight has bearing on blood pressure,
symptoms both during and following dialysis, and reflects on the accuracy
of the process. That process may be seen as accurate weighing, appropriate
prescribed weight, adequate observation of pre-dialysis status to assess that
prescription, proper operation if the dialysis machine and its proper perfor-
mance to reach the goal. One outcome, measured as weight, but also reflected
in blood pressure and the patient's report of symptoms; five processes bearing
on the outcome achieved, each capable of assessment and improvement. When
added to similar observations on multiple patients, the practices of the facility
can be seen as effective or not.
Peritoneal dialysis connection technique is critical to effective treatment
without peritonitis, though there may be other mechanisms of infection.
Monitoring by staff alone is inadequate, so here the patient must be instructed
and reinforced in reporting any break in technique or any observation (cloudy
fluid, unusual discomfort, fever, etc.) which requires follow-up. In combina-
tion with regular examination during routine visits, a tabulation of factors
can assess the individual, the group of patients, and through the group, the
effectiveness of techniques and training.
Monitoring routine monthly chemistries takes note of several outcomes
reflecting nutrition, adequacy of urea removal, phosphate binding, calcium
balance, liver disease, blood lipids, and more. Each value may be separately
8

assessed for each patient where specific additional conditions influence the
interpretation, or collectively to see if overall practices are meeting the goal
or require improvement. Multiple outcomes, each reflecting processes directly
affecting the result. Many are amenable to improvement. This is a good
example of a point where forms to display serial values for each patient
reveal the individual's course as well as current status. Further, a spread-
sheet of patients down one side and headings for lab values across the top
allows quick visualization of where this population's metabolic problems are.
Anemia control is now usually readily achieved, but still requires moni-
toring and thought. Even with recombinant erythropoietin, blood losses during
dialysis must be kept to a minimum and unexplained changes require assess-
ment for too little iron, too much aluminum, blood loss or hemolysis. A
single outcome can have multiple interval observations and several procedures
which affect it. Communication of findings and collection of all reports for
effective analysis is important. All clinicians can help each other in quality
assessment by making it easy to get the data together for review.
Water quality is an important element of hemodialysis which has been a
source of many problems. Each water treatment system has multiple values
to be monitored, each of which is a separate output related to one or more
processes in water treatment. If the initial condition of the tap water is not
known, the result of water treatment may be uninterpretable. Data must be col-
lected and acted upon, the outcomes evaluated to see if the process is effective,
and decisions regarding any need for change based on standards derived by
national consensus, reviewed by informed members of the staff.
If the water is contaminated with bacteria, the system must be disinfected,
but he cause of the contamination must be sought. Until the mechanism is
understood, no plan to prevent recurrences is likely to be practical. Although
the standards to be met are determined by outside agencies, the method of
meeting those standards and what resources will be expended in meeting
them is up to the operating staff of the facility.
All of these examples of critical scrutiny in dialysis usually exist in most
facilities, and often not recognized as the central QA function. CQI incorpo-
rates common sense and practical measures to monitor and assess practices
and results. It is often referred to as the "see, think, plan, act" sequence. The
steps often run together quickly since they are part of doing the job. There
is no separation of the observation, its assessment, and implementing the
rational correction.
These measures of quality require that everyone be involved. Routine
recording of clinical observations is the substance of the assessment of the
quality of practice. critical scrutiny requires only the analysis of those obser-
vations with an eye to better results for the patient or more effective practices
for the clinicians. Good records of clinical care are the essential QA document
in a CQI system. Appropriate and timely communication with the physician
or other responsible person completes the loop to action.
All this emphasis on ordinary practices as important to measuring the
 9

improving quality is intended to emphasize two points: the integral role of

QA in every part of the process of clinical care: and the common sense logic
of improving that care. What has not been made explicit is that knowledge
in depth about the pathophysiology of renal failure must be combined with full
knowledge of the technology of dialysis, using internal and external criteria,
for the review to be genuinely critical and for the review process to lead to
improvement in quality of care. Unless you know enough to recognize which
direction is toward better and which is worse, you can't pursue excellence.
Success requires that there be reward for effort in any endeavor. In CQI
as well, the team must be gratified by the accomplishments of seeing their
actions and records in a new light to improve quality. If they are enabled to
take more control of the clinical work they do and to change it in demonstrable
ways for the better, much of the reward is already achieved. Other mechanisms
of recognition for effective change leading to improvement are easily devised
to promote the competitive spirit of seeking to be your best - and the best,
while you are at it. Again, leadership cannot impose Continuous Quality
Improvement nor can the leaders expect CQI to proceed without their
participation. It is part of everyone's job.
CHAPTER 2

Patient and therapy perspectives:

Choosing the patient "is better worse?"

C.M. KJELLSTRAND

I believe almost all the differences in mortality that exists between different
European regions and the United States, even when age is controlled, are
due to different acceptance criteria or transplantation activity. The reason
dialysis patient mortality is highest in U.S.A. and the Nordic countries is
because they have the highest acceptance to dialysis and the highest trans-
plant rates in the world. If this is true, a worse survival may reflect a better
fulfillment of a nephrologist's duties:
Quality assurance in this chapter is defined as a dialysis unit having a
morbidity and mortality that is comparable to an acceptable norm. The norm
will be derived from large data bases which are constantly updated, provide
much detail and also study the interaction and independence between dif-
ferent factors.
I believe quality comparison is one of the most important and neglected
areas in modern medicine. I will however in this chapter concentrate on the
pitfalls of quality assurance. This is not because of cynicism or a perverse
interest in the flaws of science. Rather, if the problems with quality assur-
ance are not considered the following may result:
1. The old and the poor masses may unnecessarily perish because large groups
of patients who can be successfully sustained may remain untreated and
die too early due to being considered "poor risk".
2. The best physician, humanely striving to take on the old and the sick may
be unjustly punished, while the greedy scoundrels, who ruthlessly select
out the "best cases" and then deny them transplantation, may be rewarded.
3. Propaganda by business looking hard at the bottom line may unneces-
sarily escalate the cost of treatment.
I am purposely avoiding rehabilitation as an outcome parameter because
dialysis rehabilitation status seems to be influenced more by cultural and eco-
nomic factors and by patient motivation, than any factor of dialysis itself [1].
Table I outlines various factors that have been thought to influence the
outcome of dialysis patients. In general, physicians have control and sole
responsibility only for Technical Treatment Related Factors and only some
control over Patient - Treatment Related Factors.

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 11-21.

© 1994 Kluwer Academic Publishers.
12

Table T. Risk factors for death in dialysis patients.

I. PRE-EXISTING
I. Demographic
Age XXX
Sex 0
Race XX
2. Social
Married ?
Family support ?
Area ?
Smoking ?
Alcohol ?
Income ?
3. Diagnosis
Diabetes XXX
Hypertensive nephrosclerosis XX
Systemic Disease XX
PCKD ++
4. Type and duration of renal failure
Acute ?
Intermediate ?
Chronic ?
Late start XX
5. Co-morbid conditions
Chronic heart failure XXX
Arteriosclerotic heart disease XXX
Stroke XXX
Peripheral vascular disease ?
Pulmonary XXX
Malignancy XXX
Gastro-intestinal ?
Hepatic ?
Hypertension XX

II. SELECTION
6. Generous acceptance XXX
High transplant rate XXX

III. TREATMENT RELATED

7. Technical
Late start XX
Insufficient (KTN < 0.90) XXX
IPD XXX
CAPD ?
FAST ?
Blood pressure control XXX
Biocompatibility ?
Water quality ?
Membrane type ?
Reuse ?
 13

Table I. (Continued).

8. Patient
Malnutrition (low BunlCr-ratio low Bun, low Chol, low Alb,
low BMI, low transferin) xxx
High CA x P04 product XXX
High interdialytic weight gain XXX
Inactivity XXX
Blood pressure control XXX

XXX - Leads to higher mortality.

o - Of no influence in mortality.
++ - Leads to lower mortality.
? - Influence unknown.

The purpose of quality assurance in dialysis is to ascertain whether the dif-

ferences in outcome are due to treatment related factors or due to differences
in the pre-existing factors as outlined in Table I. However, it is important to
understand that Selection and Transplant rates are also under the control of
physicians and these will be the most important factors discussed. Many of
these factors appear self-evident to common sense and some of them are also
scientifically proven. For example, advancing age leads to shorter survival
as do systemic diseases such as diabetes mellitus and multiple myeloma. In
rigorous multivariate analysis these two different factors, independent of each
other, shorten survival [2-11].
It must also be understood that the interaction of these many factors may
be incorrectly interpreted, even when scientifically studied. Almost all studies
can be criticized either because the materials presenting detailed patient data
were too small and therefore the complicated statistical analysis necessary was
not robust, or if the material was large, many patient details provided, were
invariably soft. For example, the European Dialysis Transplant Association
have no rigorous, good, prospective data on co-morbid conditions besides
the diagnosis of kidney disease. Finally there are poorly understood, and as
yet unanalyzable shifts or changes in risk factors that occur with time. As
patients were studied in one large centre, the presence at start of dialysis of
arteriosclerotic heart disease, stroke, COPD, malignancy or diabetes all
adversely affected survival [4] but when the presence of these factors was
studied in the same centre in a patient population over the age of 70, they all
disappeared as risk factors, advanced age overwhelmed all of them [12].
Finally, whatever the facts are, they will be the subject of wildly different
interpretation depending on prejudice and interest. Whatever the merits or
demerits of reuse are, the patient on dialysis will regard them with suspicion
if he believes it will give him AIDS, but will like it if he believes it causes
less damage to his blood. Both the greedy physician and the one who thinks
protein coating is advantageous will like it, and the manufacturer of great
quantities of cheap dialysers view it with disdain.
14

Influence of pre-existing risk factors

Demographic risk factors

Age is a very important risk factor. While life expectancy of both young and
middle-aged patients is measured in decades, the mean survival time for
patients above age 70 is only 5 years on dialysis [2-12J. However, the relative
risk of dialysis i.e. the chance of dying on dialysis over a 5 year period when
compared to non-dialysed age matched population decreases with age. Thus
young patients aged less than 45 years encounter a 20 times increased chance
of dying within 5 years when they go on dialysis compared to only a twofold
increase in patients over the age of 75. It is obvious that old age not only
overrides other risk factors but also the very risk of dialysis itself [7J. Sex does
not appear to be a risk factor, thus there is no difference in survival between
men and women in the U.S. [4, 7, 10, 11J. On the contrary, race is of
importance in that non-white patients survive better than white [10, IIJ. The
5 year probability of surviving for black patients was 41.2% versus 37.4%
for white patients, even when adjusted for age, sex and primary disease [11J.
This unexpected result may have to do with the low transplant rate in black
patients and points to an important role for selection in patient survival [13-15].

Social risk factors

There appear to be no good studies of the influence of marital status, family

support or on the influence of smoking or alcohol. Income may in a perverse
may influence survival on dialysis both in a positive and negative way. One
can expect a poor patient to be more often malnourished, an important
predictor of poor survival [16-18J, but also to be less often transplanted [15],
an important factor in improving survival on dialysis, as will be discussed.

Diagnosis

Certain diagnoses are associated with a higher death rate. This includes diabetes
and hypertensive nephrosclerosis and some other systemic diseases such as
myeloma and amyloid which appear to be associated with a shorter survival.
On the other hand polycystic kidney disease appears to be associated with a
good outome [2-12J.

Type and duration of renal failure

There appear to be no good studies of this. Intuitively one would guess that
a fairly short period of uremia may be associated with less chronic meta-
bolic effects and thus with a "stronger body" starting dialysis.
 15

Co-morbid conditions

There are many studies that prove that the presence of other diseases may
shorten survival on dialysis. Chronic heart disease, atherosclerotic heart disease,
strokes, peripheral vascular disease, COPD and malignancies have all been
associated with a poor outcome as has severe hypertension [2-11, 21]. Some
of these diseases appear to be additive in their ill effect. For example, in one
study, non-diabetic patients, younger than 45 years, without risk factor, had
a 10% six year mortality, and those with arteriosclerotic heart disease or stroke
had a 20% mortality rate but those with both diseases had a 40% mortality rate
[4]. It is obvious that to evaluate the influence of all these co-morbid factors
and their interrelationships with each other and age requires a large number
of very carefully prospectively examined patients. Such material does not exist
and may never be in existence.
Malnutrition is an extremely important factor in predicting outcome. thus,
patients who have a low BUN/creatinine ratio, or a low BUN, cholesterol,
triglycerides, albumin, BMI or transferrin value have a very high mortality rate
on dialysis [16-18].

The influence of selection and transplantation

It is quite clear that as patients present to dialysis many factors exist which
will predict their survival. It is then equally clear that by careful selection
one can greatly influence survival results. We hypothesized that physicians
who liberally accept many patients per population were likely to take on
many old patients and many patients with degenerative diseases, systemic
diseases, or malnourishment. The survival results of such a physician would
be "poor" when compared to a more selective and fastidious physicians who
treated only those who were young, and except for their kidney disease
otherwise healthy. While this is self-evident, it may also appear equally self-
evident that simple age matching could avoid many of these problems but
things are more complicated than this as will be discussed below. Secondly,
we hypothesized that someone who co-operated with a very active transplant
program would also have poor mortality rates. Over 80% of all patients, who
now start dialysis in the United States and in many other countries with a
high acceptance rate, will remain on dialysis and not be transplanted. The
chance of receiving a transplant is now falling all over the world as dialysis
acceptance rates continue to rise, while transplant rates have levelled or even
declined [19]. Kidneys for transplantation are thus a much more scarce resource
than machines for dialysis. Transplantation removes the very best patients, who
are the young, without other diseases, and who have a long life-expectancy
on dialysis. This results in a worse survival for centres with a high trans-
plant rate when compared to centres where transplantation rates are low.
To investigate this we correlated cumulative 4 year survival in age-matched
16

dialysis patients to acceptance rate for 5 European regions: Latin Europe,

(France, Spain and Italy) Benelux (Belgium, Netherlands, Luxembourg),
German countries (West Germany and Austria), the British Isles, Nordic
countries (Sweden, Norway and Denmark) and the United States. Similarly,
cumulative survival for these regions was correlated to the percent of patients
transplanted at 4 years [20]. The result appears in Figure 1. In 1985 nephrol-
ogists in U.S.A., where survival was lowest, accepted four times as many
patients as British nephrologists and twice as many as in Sweden, Germany
and Canada. The transplant rate was twice that in Latin and German coun-
tries. The data was also analyzed by stepwise and multiple regression analysis,
which gave a much better fit than similar simple linear aggression: R = 0.96,
P = 0.02 for the equation:

4 year cumulative dialysis survival =

107 - 0.7 x acceptance rate per million
- 0.4 x percent transplanted at 4 years [20].
Thus in this particular analysis over 90% of the differences in cumulative
survival between regions were explained by different acceptance and trans-
plant rates, even in age-matched patients. We have in a later model refined
this approach and studied it for different age groups and also studied the
influence of diabetes. In this newer model acceptance rates and percent patients
with diabetes are the most important predictors of long term survival of
young patients while transplant rates and acceptance rates appear more
important for the old patients [21].
It is evident that acceptance and transplant criteria are important determi-
nants when other factors are held constant, but beyond that probably also
represent other factors that may be very difficult to define and to quantitate
for a unit. Such factors include clinical impression that encompasses the
cumulative sum of or subtle grading of co-morbid conditions. These factors
are already playing a role in U.S.A. as indicated by an article investigating
case mix of patients with end-stage renal disease in profit and non-profit
dialysis units [22]. When 307 patients treated in 5 proprietary, free-standing,
profit based units were compared to 3135 patients treated in non-profit making
hospitals, it was found that when the patients were ranked in severity groups
based on age, race, primary renal disease and co-morbid conditions, the hospital
based facilities had a higher percentage of patients who were more ill. Thus
60% of the hospital treated patients were in the 3 groups with the highest
severity index but only 50% of the patients in free standing facilities had
such a high index. However, even within each severity group, hospital based
patients had a lower 5 year survival rate than patients in free-standing facil-
ities [8, 22]. One interpretation of these data is that the for-profit units offered
better treatment. My interpretation is that their fastidious selection was the
cause of the better survival and that the severity indexes used were simply very
crude compared to clinical impressions by physicians selecting patients to their
profit units. Patient dumping obviously has at its base economic considerations
 17

DIALYSIS MORTALITY INCREASES

WITH INCREASING ACCEPTANCE RATE
PATIENTS 15 • 44 YEARS
90 ,----------------------------------,

60 ~-4 __ ~~ __ ~~ __ ~~~~~~~ __ ~~

20 40 60 80 100 120 140

ACCEPTED PER MILLION AND YEAR

DIALYSIS MORTALITY INCREASES WITH

INCREASING TRANSPLANTATION RATE
PATIENTS 15 ·44 YEARS
90 r-~~--------------------------~
• BENELUX
a
•
FRG·AUSTRIA
BRIT ISL.

-USA

60 ~~--~--~~~~--~--~~--~~
20 30 40 so 60 70
PERCENT TRANSPLANTED AT 4 YEARS

Fig. 1. (Top) Linear regression analysis showing the relation between cumulative 4-year
dialysis survival for patients age 15 to 44 and the acceptance rate to dialysis. The more patients
that are accepted, the worse is cumulative survival. (Bottom) Cumulative 4-year dialysis survival
in relation to the percent of patients transplanted at 4 years. The cumulative survival becomes
worse as more patients receive transplants at 4 years. The transplant surgeon takes the young
and healthy who show good survival on dialysis and leaves only the old and those with many
other diseases who have poor survival. When both acceptance rate and transplant rates are
evaluated through multiple regression analysis, over 90% of the differences in cumulative survival
are explained. Thus, differences in survival are not dependent on geographical location, but on
generosity in acceptance and transplant rates.
(From reference 20 reproduced with permission of W. B. Saunders Co.)
18

[23]. The old, the sick and the malnourished will cost more, because they
are more difficult to care for. When they are dumped on public facilities not
only does the survival curve for the fastidious unit shift upwards but the
survival curve of the receiving hospital will shift downwards [22, 23]. This
gives the twice false impression, that quality is better, when in reality selec-
tion is the reason for the improved survival. In a perverse fashion then "better
may actually mean worse". Fastidious "good outcome" physicians have not
fulfilled their obligations to those who need their services particularly urgently.
One can of course argue that "sick" patients should not be treated at all
as the life expectancy is short and the cost relative high. Actually, when counted
in QUALY's hemodialysis for any patient fares very poorly. When the yearly
cost is divided by the expected quality of life and its duration the figures
become very high [24, 25]. The ultimate extension of this of course is that "the
only good patient is a dead patient" and although one can shrug that off as a
meaningless cynicism, when this is brought into the clinical arena it is indeed
a deadening reality for the old, as the easiest of all selection criteria is age,
in itself often used to exclude the old from treatment [26-28].

Treatment related factors

Technical

This group of factors is the true matter of investigation for quality assur-
ance;
Is the physician performing technically adequate dialysis?
Many factors enter into this equation, and over some the physician has only
partial control as for example late start of dialysis. It has been shown that
patients who come with severe uremia to dialysis may have incurable sec-
ondary metabolic effects that lead to complications and an increased mortality
[29], while those started early do particularly well [30]. To give a patient
insufficient dialysis, for example as expressed as KTIV of less than 0.9 has
been associated with a high morbidity [31, 32]. Similarly intermittent peri-
toneal dialysis is no longer regularly used because of the high complication
rate and the high mortality associated with this procedure [7]. Comparable data
do not exist for home hemodialysis and what differences exist, appear best
explained by different selection criteria [7, 9, 10]. CAPD has the advantage
of an even control of uremia and electrolytes but offers the patient KTIV values
that are insufficient for hemodialysis patients [31, 32]. Fast dialysis has also
been associated with an increased mortality [33] but much of this appears to
be due to the fact that fast sufficient dialysis and short insufficient dialysis
are not separated [34]. Another factor though is that fast dialysis, even when
thought to be adequate, may in reality more easily become inadequate than
slower and longer dialysis. Biocompatibility has been one factor associated
 19

with outcome. More biocompatible and open membranes have been thought
to lead to less dialysis arthropathy - amyloid. However, the largest study of
this problem, comparing the 10 year incidence of these diseases between
matched patients on poly acrylonitrile or cellophane membranes has failed to
show any statistical significance [35]. Still, many physicians will use the
open membrane for patients. The higher cost of this, in a system with a capped
budget, results in an erroneously perceived increased quality for some patients
but no dialysis for others.
Blood pressure control during dialysis is an important factor for longevity
as shown many years ago [36]. Reuse, that has resulted in lethal infection,
has invariably been associated with a higher survival or no difference in
survival when compared to no-reuse in rigorous studies [8].

Patient treatment related factors

These factors are only partially under a physician's control. For example, a
high calcium x phosphorus product, malnutrition, "giving up", a high intra-
dialytic weight gain and hypertension are factors over which physicians have
only partial control [8]. These factors are only partially influenced by a physi-
cian's actions and patient education, but are much influenced by compliance.
Inactivity in old patients has also been associated with a high mortality rate.
It is not known if inactivity only is the reflection of other underlying factors
that simultaneously shorten life and lead to inactivity, or if it is a risk factor
itself, that can be influenced by education, physical therapy and training
[12].

Conclusions

The real purpose of quality assurance in dialysis appears to me to be to detect

the insufficiencies in technical treatment related factors. Obvious other factors
which need to be considered are demographic, social, diagnostic and co-morbid
conditions. Other factors complicate the evaluation because of their subtle
interaction between technical treatment factors and patient compliance. Such
factors are blood pressure and Ca x P0 4 product, nutrition, weight gain and
inactivity. Overriding all the previous ones in influencing mortality are gen-
erosity in selection and a high transplant rate. Both of these will negatively
influence survival of dialysis patients. Therefore "worse may be better" and
it does not appear unreasonable to at least consider that a "poor outcome" is
indicative that the physician has fulfilled his obligation to a larger number
of truly sick patients much better than one whose survival curves look more
favorable. Ignoring this fact carries the risk that good physicians will be dis-
couraged, shytsters rewarded and large patient groups remain untreated and
die earlier.
20

References

1. Carlson DM, Johnson WJ, Kjellstrand CM: Functional status of patients with end-stage renal
disease. Mayo Clin Proc 62: 338-344, 1987.
2. Degoulet P, Legrain M, Reach I, Aime F, Devries C, Rojas P, Jacobs C: Mortality risk factors
in patients treated by chronic hemodialysis. Nephron 31: 103-110, 1982.
3. Hutchinson T, Thomas DC, Mac Gibbon B: Predicting survival in adults with end-stage renal
disease: An age equivalence index. Ann Int Med 96: 417-423, 1982.
4. Shapiro FL, Umen A: Risk factors in hemodialysis patient survival. ASAIO 6: 176-184,
1983.
5. Vollmer WM, Wahl PW, Blagg CR: Survival with dialysis and transplantation in patients
with end-stage renal disease. N Engl 1 Med 308: 26, 1553-1558, 1983.
6. Neff MS, Eiser AR, Slifkin RF, Baum M, Baez A, Gupta S, Amarga E: Patients surviving
10 years of hemodialysis. Am 1 Med 74: 996-1003, 1983.
7. Hellerstedt BA, Johnson WJ, Ascher N, Kjellstrand CM, Knutson R, Shapiro FL, Sterioff
S: Survival rates of 2,728 patients with end-stage renal disease. May Clin Proc 59: 776-783,
1984.
8. Held PJ, Pauly MV, Diamond L: Survival analysis of patients undergoing dialysis. lAMA
257: 645-650, 1987.
9. Mailloux LU, Bellucci AG, Mossey RT, Napolitano B, Moore T, Wilkes BM, Bluestone
PA: Predictors of survival in patients undergoing dialysis. Am 1 Med 84: 855-862, 1988.
10. Silins J, Fortier L, Mao Y, Posen G, Ugnat A-M, Brancker A, Gaudette L, Wilgle D:
Mortality rates among patients with end-stage renal disease in Canada, 1981-86. CMAl 141:
677-682, 1989.
11. Renal Data System U.S.A., The National Institutes of Health, The National Institute of
Diabetes and Digestive and Kidney Diseases, Division of Kidney, Urologic, and Hematologic
Diseases Bethesda MD. Annual Report 1989.
12. Westlie L, Umen A, Nestrud S, Kjellstrnd C: Mortality, morbidity and life satisfaction in
the old dialysis patient. Trans Am Soc for Artif Intern Organs 30: 21-30, 1984.
13. Kjellstrand CM: Giving life-giving death. Ethical problems with high technology medicine.
Acta Med Scand 725 (Supp): 1-80, 1988.
14. Kjellstrand CM: Racial, sexual and age discrimination in renal transplantation. Arch Intern
Med 148: 1305-1309, 1988.
15. Held PJ, Pauly MV, Bovbjerg RR, Newman J, Salvatierra 0 Jr: Access to kidney trans-
plantation - has the United States eliminated income and racial differences? Arch Intern
Med 148: 2594-2600, 1988.
16. Acchiardo SR, Moore LW, Latour P: Malnutrition as the main factor in morbidity and
mortality of hemodialysis patients. Kidney Int 24: 16:S-199-S-203, 1983.
17. Oksa H, Pasternack A, Pasanen M: Serum urea-creatinine ratio as a prognostic index in
hemodialysis patients. Clin Neph 27: 3:125-130, 1987.
18. Marckmann P: Nutritional status and mortality of patients in regular dialysis therapy.
lInt Med 226: 429-432, 1989.
19. Kjellstrand CM, Shideman J: On the impact of and need for chronic dialysis and renal
transplantation. Trans Am Soc Artif Intern Organs 34: 328-332, 1988
20. Kjellstrand CM, Hylander B, Collins A: Mortality on dialysis - on the influence of early
start, patient characteristics, and transplantation and acceptance rates. Am 1 Kidney Diseases
15: 483-490, 1990.
21. Collins A, Kjellstrand CM, Hull A, Parker T: Mortality on dialysis appears directly
dependent on the generosity of the nephrologist and the activity of the transplant surgeon.
ASATO 36 (Abstr): 16, 1990.
22. Plough AL, Salem SR, Shwartz M, Weller JM, Ferguson CW: Case mix in end-stage renal
disease. N Engl 1 Med, 1432-1436, 1984.
23. Ansell DA, Schiff RL: Patient Dumping: Status, Implications, and Policy Recommendations.
lAMA 257: 1500-1502, 1987.
 21

24. Williams A: Economics of coronary artery bypass grafting. Br Med 1 291: 326-329, 1985.
25. La Puma J, Lawlor EF: Quality-adjusted life-years: Ethical implications for physicians
and policymakers. lAMA 263: 2917-2921, 1990.
26. Wetle T: Age as a risk factor for inadequate treatment. lAMA 258: 516, 1987.
27. Lamm RD: Health care as economic cancer. Dialy & Transpi 16: 432-433, 1987.
28. Callahan D: Setting limits - medical goals in an aging society. New York: Simon & Schuster,
1987.
29. Kjellstrand CMN, Evzans RL, Petersen RJ, Shideman JR, von Hartitzsch B, Buselmeier
TJ: The "unphysiology" of dialysis: A major cause of dialysis side effects? Kidney Int 7:
S30-S34, 1975.
30. Bonomini V, Feletti C, Stefoni S, Vangelista A: Early dialysis and renal transplantation.
Nephron 44: 267-271, 1984.
31. Lowrie EG, Laird NM, Parker TF, Sargent JA: Effect of the hemodialysis prescription on
patient morbidity. N Engl 1 Med 305: 1176-1181, 1981.
32. Gotsch FA, Sargent JA: A mechanistic analysis of the National Cooperative Dialysis Study
(NCDS). Kidney Int 128: 526-534, 1985.
33. Kjellstrand CM: Short dialysis increases morbidity and mortality. Contr. Nephroi44: 65-77,
1985.
34. Levin NW: Mortality impacts of shorter dialysis procedures. Kidney Inti 35: 254, 1989.
35. Brunner FP, Brynger H, Ehrich JHH, Fassbinder W, Geerlings W, Rizzoni G, Selwood
NH, Tufveson G, Wing AJ: Case control study on dialysis arthropathy: The influence of two
different dialysis membranes: Data from the EDTA Registry. Nephroi Dial Trans 5: 432-436,
1990.
36. Lundin P, Adler AJ, Feinroth MV, Berlyne GM, Friedman EA: Maintenance hemo-
dialysis. lAMA 244: 38-40, 1980.
CHAPTER 3

Quality systems in the dialysis center:

Peritoneal dialysis

BARBARA F. PROWANT, KARL D. NOLPH,

ZBYLUT J. TWARDOWSKI, LOIS M. SCHMIDT,
LEONOR PONFERRADA and RAMESH KHANNA

The goal of this chapter is to discuss the characteristics of systems (struc-

ture) and activities (process) within a peritoneal dialysis program which
contribute to optimal outcomes (quality) for peritoneal dialysis patients.
The text is organized around the eight characteristics or attributes of
successful companies identified by Peters and Waterman in their book In
Search of Excellence [1].

Hands-on, value driven

Peters and Water found that successful companies have a sound set of beliefs
and values upon which they premise all policies and actions and summarizes
these as "hands on and value driven" [1].

Program integration

One value embraced by successful peritoneal dialysis programs is that it is

essential for the peritoneal dialysis program to be integrated with acute and
chronic hemodialysis units [2] and a renal transplantation program. There
are many advantages of a truly integrated program. Decisions concerning
optimal therapy can be made without the bias of how it will affect the income
to the program and/or physician. The patient can transfer back and forth
between therapies when necessary with relative comfort and ease.

Philosoph} of self-care

Another value imperative for achieving quality in a peritoneal dialysis program

is the conviction that chronic peritoneal dialysis is an acceptable treatment
for ESRD and that it can be successfully managed as a self-care therapy.
Although not all patients are totally responsible for self-care, almost all patients

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 23-45.

© 1994 Kluwer Academic Publishers.
24

dialyze at home either independently or with the assistance of a partner. It is

imperative that the administrative staff and all team members believe that
patients and their families can learn to dialyze safely and effectively at home,
and that self-care, home dialysis offers advantages to the patient in terms of
independence, control, and quality of life.
Appropriate goals for a home peritoneal dialysis program are: (a) to enable
patients to dialyze safely and effectively and; (b) to assist patients in main-
taining an optimal level of function at home.

Patient involvement in choosing a chronic dialysis therapy

Another value inherent to successful peritoneal dialysis programs is that the

patient and family should be encouraged to participate in the choice of a
therapy which best meets their needs and fits their lifestyles. Ninety-three
percent of the 32 centers of excellence for modality selection practices
identified by Baxter Healthcare allowed the patient to make the final choice
of chronic dialysis modality after receiving professional assessment, education
and recommendations. These centers had an impressive technique survival
of 85% at three years [3].
Data from 326 patients beginning dialysis therapy in our own program
over the past 2 years indicate that significantly more patients who received
predialysis education at least one month prior to the need for dialysis chose
a self-care home dialysis therapy compared to patients who presented with
uremic symptoms and in need of immediate dialysis [4].
A number of models for dialysis modality selection have been developed
[5, 6]. The Missouri Kidney Program has published a predialysis patient
education program [7] which as been adapted by Baxter Healthcare [8]. Key
features of successful predialysis education programs are listed in Table I.

Table I. Key components of successful predialysis patient education programs.

• Education is initiated 3-6 months prior to the need for chronic dialysis.
• There an unbiased presentation of all treatment options and pros and cons of each.
• Classes supplement one to one sessions.
• Families and significant others are included in the education process.
• Patients meet patients on various ESRD therapies.
• The patient is assessed by the renal team.
• Medical advantages and/or contraindications for a therapy are discussed with the patient.
• The patient is included in the decision-making process.

A recent review of63 patients who selected peritoneal dialysis [4] indicated
that the predominant reason (25%) was to maintain independence, activities
and flexibility in scheduling. Motivation to perform one's own dialysis and
maintain some control was the reason 15% chose peritoneal dialysis and an
additional 12% of the patients cited the ability to continue working. Although
 25

the majority of patients chose PD for positive reasons, fourteen percent selected
peritoneal dialysis because of long distances to a hemodialysis unit and an
additional 9% because they felt travel to center hemodialysis or the demands
of home hemodialysis would impose an unacceptable burden on their families.

Administrative support

A PD program cannot succeed and expand without strong and unified

administrative support. For a PD program to thrive both the administrator
and medical director must believe that peritoneal dialysis is a legitimate dialysis
therapy and a valid treatment option on an equitable status with hemo-
dialysis; and that PD is a revenue producing program. Only when there is
such a philosophy will the PD program be able to obtain adequate space,
personnel, equipment, budget and support.
Allowing designated nursing staff to work solely in the home dialysis
program is one indicator of administrative support. Of the 18 centers of
excellence for patient education and training practices identified by Baxter
Healthcare, 92% assigned nurses to the peritoneal dialysis outpatient program
only with no responsibilities for center hemodialysis or intermittent peri-
toneal dialysis [9].
Another example of strong administrative support is flexibility to choose
from more than one peritoneal dialysis system, so that each patient has access
to a system which will meet his or her unique needs. For example a handi-
capped or visually impaired patient may need an assist device, an active
individual concerned about body image might desire a disconnect system,
and a patient with recurring hernias might need overnight cycler dialysis.
Three examples of the lack of strong administrative support follow. The first
is a peritoneal dialysis program that chooses not to provide nursing back-up
evenings, nights and weekends in order to avoid paying nurses for call time.
Patients often do not call on weekends to report complications because of
difficulties communicating with the physicians, or because a physician may
not perceive their concerns to be a legitimate problem.
A second example is a unit that does not allow primary nurses to call their
patients between clinic visits because they are unwilling to pay the long
distance telephone charges. Consequently, patients visit the emergency room
or are admitted for problems which could have been prevented or easily
managed at home had they been identified early.
The third example is a peritoneal dialysis program with more than 30
patients which operates out of two small treatment rooms, one of which is
the only nurse's office. If clinic visits are scheduled during a training day
the nurse has no place to speak confidentially to a patient who calls for
assistance. Furthermore, patient records kept in the nurses office are not
adequately secured.
26

Philosophy of excellence

Finally, for a peritoneal dialysis program to achieve a consistent quality of care

the administrative, professional and support staff must share a commitment
to quality. they must believe that their program can and does provide a high
quality of care and caring. They must be willing to go above and beyond
the realm of routine activities, to try innovative approaches, and occasion-
ally step outside the bounds of the job description. The philosophy of
excellence includes a willingness to individualize dialysis prescriptions and
to provide an optimal dose of dialysis, and adequate support services. The
highest possible level of health and rehabilitation is truly the goal for each
patient. The patient is the staff's central focus and the difficult or complex
patient is seen as a challenge, not a nuisance or problem.

Productivity through people

Productive companies treat their employees with dignity and respect. Employ-
ees are partners, experts and team members [1].

The peritoneal dialysis team

Peritoneal dialysis as a subspecialty has emphasized the importance of inter-

disciplinary collaboration and a team approach to patient care [10, 11]. The
peritoneal dialysis team becomes a continuous quality circle responsible for
the quality of care provided to their patients. Typically the team responsible
for the care of peritoneal dialysis patients is composed of a physician, nurse,
dietitian and social worker.
The major physician responsibilities are to prescribe appropriate therapy for
end stage renal disease and other medical problems, to diagnose and treat
complications of ESRD and dialysis therapy. The physician can also facili-
tate the effectiveness of the team by clarifying and/or validating the roles of
the other team members, to patients, other physicians or other departments and
institutions.
Table II lists the areas in which we believe a physician must be knowl-
edgeable and competent to effectively manage peritoneal dialysis patients.
In order to prescribe appropriate therapy physicians caring for peritoneal
dialysis patients must understand peritoneal dialysis kinetics and how to assess
peritoneal membrane transport rate. Furthermore, the physician should be
able to assess the adequacy of dialysis based on: (a) quantifying dialysis;
(b) combined renal and dialysis urea and/or creatinine clearances; (c) inter-
pretation of laboratory values; (d) nutritional status, and (e) patient well-being.
Roles of the nurse are to provide the education, guidance, support, super-
vision and assistance patients require to perform PD at home, to adhere to
the treatment regimen and to experience the highest possible quality of life.
 27

Table II. Areas of physician knowledge and competence required for a successful peritoneal
dialysis practice.

General competence in hemodialysis

Peritoneal dialysis
Catheter insertion protocols
Diagnosis and treatment of catheter related problems
Types of peritoneal dialysis therapy
Various systems for peritoneal dialysis procedures
Recognition, prevention and treatment of complications
Diagnosis and treatment of peritonitis
Management of diabetes mellitus and regulation of blood sugar with intraperitoneal insulin

Because most nurses do not come to peritoneal dialysis programs with

experience in nephrology or peritoneal dialysis an extensive orientation and
education program is required in order to achieve a high level of nursing
care. One such program lasts for 6 weeks during which the nurse learns
theoretical information about peritoneal dialysis as well as learning to perform
peritoneal dialysis procedures. A competency based learning system which
allows the learner to demonstrate the requisite knowledge and/or skills in
each specific area can be used efficiently and effectively for initial orienta-
tion. The new nurse observes home dialysis education, clinic visits, home visits
and outpatient nursing management. The new nurse begins working with
patients with a preceptor so he or she has a readily available resource and
support. As the nurse learns or reviews ESRD and peritoneal dialysis content
and principles of adult education he or she begins teaching patients in familiar
topics such as monitoring blood pressure, then moves on to other topics. The
number of primary patients is gradually increased, and patients requiring
more complex care are added as the nurse gains experience. Nurses wait
three months before taking call, and then another nurse is available for
consultation and back up support.
The social worker's role is to assess the patient's (and family's) financial
and psycho-social status and to provide psychosocial support, counseling and
referrals as needed.
The dietitian's role is to assess the patient's nutritional status and make
recommendations regarding the diet regimen. The dietitian teaches the patient
and significant others about the therapeutic diet, assists them in meal planning,
and helps them incorporate the dietary regimen into their lifestyle. To achieve
a high quality of care for peritoneal dialysis patients the renal dietitian needs
to have a basic understanding of dialysis, and how it is related to nutrition,
the typical diet orders [12], and which parameters to routinely monitor. The
dietitian needs to understand basic principles of adult education and have
adequate time to teach patients and families and to evaluate their adherence
to the diet plan. The Council on Renal Nutrition provides guidelines for staff
patient ratios based upon the extent of services [13].
Finally, the patient, the focus of services provided by the team, may actively
28

participate in the team's decision making process. Some institutions recom-

mend that self-care or home dialysis patients participate in team care
conferences to develop the long term care plan. In other units the team's
recommendations and/or care plan are later reviewed with the patient.

Team approach and interactions

Assembling an interdisciplinary group to care for patients does not neces-

sarily ensure that they will function as a team. It is essential that the patient
and family be viewed as a whole and that there is a system which will prevent
fragmentation of care by promoting communication and collaboration [14].
Team interactions take place in a variety of settings; however, some
structured meetings are essential. Regularly scheduled team meetings provide
opportunity to review the patient's current status, to discuss problems, to
develop the team's long term plan of care, and for collaborative decision
making. Hospital rounds by an interdisciplinary team also facilitate coordi-
nated, continuous care.
Mutual respect, similar goals, effective communication, and techniques to
manage conflict among team members are essential for the team to function
effectively.
Physical proximity can enhance team function. Team members with
adjoining desks or offices (or even offices in the same building or facility)
are likely to have much more informal interaction with each other than with
physically distant team members [14].

Professional enhancementljob enrichment

The ultimate goal of professional enhancement and job enrichment is to attract

and retain staff. Short term goals are to improve morale, increase motivation
and job satisfaction, and to reduce stress and absenteeism.
Membership and active participation in professional organizations provides
an avenue for professional enhancement. Continuing education for all team
members is also essential for professional enhancement as well as optimal care.
Team meetings, clinics and rounds provide almost continuous opportunities for
informal teaching. Working with a partner or mentor is an effective way for
professionals new to peritoneal dialysis to acquire knowledge and learn to make
clinical assessments, diagnoses, and management decisions. Unit inservices
and professional education meetings provide more structured learning as does
formal education.
 29

Autonomy and entrepreneurship

Successful companies allow employees to be independent, and creative. They

encourage risk taking and support good tries with the attitude that a reason-
able number of mistakes or failed attempts are requisite for success [1].
Primary nursing is a system that assigns the nursing care of each indi-
vidual patient and family to one nurse. The primary nurse is responsible and
accountable for providing individual, comprehensive, and continuous nursing
care for a group of patients. The primary nurse may also coordinate health care
services provided by other disciplines. A number of studies have shown that
primary nursing enhances job satisfaction and professional development
[15-18]. Primary nursing provides high levels of attainment on job enrichment
criteria such as autonomy, direct feedback, identification with the whole
product and task variability [19].
Primary care seems to be the most appropriate nursing modality for
outpatient peritoneal dialysis and is the nursing modality most widely utilized
in peritoneal dialysis programs in North America. Ninety-four percent of the
centers of excellence for patient education and training practices utilized
primary nursing during home training and 81 % continued to utilize primary
nursing for outpatient followup [9].
Assigning additional projects and responsibilities to staff members who
are clinically competent and efficient also provides diversity, and an avenue
for personal growth and professional development. Table III lists a number
of such activities. Most of these tasks and activities are inherent components
of a home dialysis program, so assigning such projects does not actually add
additional work.
Peritoneal dialysis staff are frequently asked to teach others about this
therapy. These invitations provide opportunities to develop skills in planning
and providing professional education and in public speaking. Opportunities
to do technical or professional writing are also available to peritoneal dialysis
staff members. Beginning writers may start with simple in-house projects
such as policies and procedures or patient education materials.

Table III. Staff projects to enhance professional growth and development.

• Write or revise policies and procedures.

• Evaluate new products and peritoneal dialysis systems.
• Develop patient education modules.
• Develop patient education materials.
• Participate in quality improvement activities.
• Develop clinical expertise in related specialties, e.g. gerontology, diabetes.
• Serve as a liaison to long term care facility(s).
• Serve as a liaison to associated hospital(s).
• Collect and analyze peritonitis, exit site infection data.
• Participate in or direct research projects.
• Serve on institutional committees.
30

Simultaneous loose tight properties

Peters and Waterman noted that top companies are both centralized and
decentralized. They encourage autonomy at all levels, yet hold a core set of
values [1]. An example of this in the dialysis unit is the professional team
approach which encourages autonomy in patient care within a discipline, yet
strongly promotes the communication and continuity of care among team
members and in a number of settings (the home, outpatient clinic, hospital).
The shared value system discussed above is another example of the tight
properties which contribute to success.

Simple form, lean staff

The organization structure of successful companies is described as "elegantly

simple" with lean staffing at the top levels [1]. When the organization matrix
gets too complex, the priorities and accountability become confused. It might
be worthwhile to evaluate the corporate structure of dialysis programs to assess
whether the organization is simple and streamlined, and whether it truly
supports quality in patient care.

Stick to the knitting

The most successful companies stick close to the central skill or product and
this enables them to perform more effectively than the competition [1]. Because
of the highly specialized nature of peritoneal dialysis programs, few units
branch out into other types of care. This discussion will focus on what is
required to provide a high quality of peritoneal dialysis.
First of all, peritoneal dialysis programs have physical space requirements.
Eighteen centers of excellence for patient education and training practices
identified by Baxter Healthcare had significantly higher technique survival and
patient survival rates than their counterparts. All of these units had space
devoted solely to the home peritoneal dialysis program [9]. Eighty-two percent
of these centers had a PD training room, a separate PD clinic area, and a PD
nurses' office. All of the units also had separate storage and utility areas for
peritoneal dialysis [9].
One of the most basic requirements is a set of standards of clinical practice.
A standard is the yardstick of the quality of a service and Mason states that
nursing standards "define unequivocally what quality care is and provide
specific criteria that can be used to determine whether quality care has been
provided" [20]. A standard describes what should be done and how the patient
will benefit from the care. The American Nephrology Nurses Association
has published standards of clinical practice for nephrology nursing [21].
Appropriate standards can be selected and adapted for use in a particular
dialysis unit.
 31

A policy and procedure manual that guides safe practice is also essential.
Policies for all nursing procedures, machine and equipment maintenance,
emergencies, and ma.naging problems will help ensure consistent, safe care.
A unit's standards and policies and procedures are also used by inspectors
and surveyors to evaluate the program. Although clinical policies and proce-
dures are developed primarily by the nursing staff we recommend consultation
with administration regarding legal issues, with the medical director regarding
nursing protocols to manage complications, and with patients regarding the
self-care procedures. An annual review and update of policies and proce-
dures is appropriate. Examples of the types of policies and procedures required
for a peritoneal dialysis program are listed in Table IV.

Table IV. Selected Types of Peritoneal Dialysis Unit Policies and Procedures.

General Procedures
Handwashing
Measuring blood pressure
Quantitative urine collection
Peritoneal dialysis procedures
Exit site care procedures
Exchange procedure (for each system used)
Cycler procedures
Machine set up for closed drain
Machine set up for open drain
Administration of intraperitoneal medication
Catheter adapter change procedure
Peritoneal equilibration test
Quantitative dialysate collection
Protocols for managing problems and complications
Obstruction of flow
Fibrin in dialysate
Contamination of the system
Crack or hole in catheter
Hypervolemia
Peritonitis
Emergency procedures
Cardio-pulmonary arrest
Fire
Hurricane or Tornado

Medical protocols and procedures also contribute to quality. For teaching

institutions a manual for housestaff and renal fellows is essential for consis-
tent care. Examples of content are listed in Table V.

Catheter placement

Establishing a permanent access is a prerequisite for successful peritoneal

dialysis. Preoperative preparation for peritoneal catheter insertion typically
32

Table V. Selected Content from Manual for Housestaff and Fellows.

Peritoneal Dialysis
Acute catheter insertion procedure
Chronic catheter insertion
Current policies
Marking catheter exit site
Preoperative orders
Catheter placement video
Catheter Break-in
Chronic dialysis orders
Peritoneal equilibration test
Exit site infection protocol
Peritonitis protocol

includes the choice of catheter type, determination and marking of the exit site,
and cleansing the abdomen with a disinfectant scrub. Determining the exit
site is usually a joint decision and the surgeon, PD nurse, patient and/or
nephrologist may be involved. Factors to consider for optimal exit site
placement are: avoiding skin folds, the beltline, and scar tissue; and place-
ment where the patient can observe and manipulate that catheter for ease of
exit care [22, 23]. Broad spectrum prophylactic antibiotic therapy is gener-
ally recommended [23, 24].
Catheters can be medically inserted with a trocar or peritonesocopy or
surgically inserted. For optimal results a few general guidelines apply to
either procedure.
• Catheter placement should be limited to experienced surgeons or nephrol-
ogists [24].
• Local anesthesia combined with a sedative is adequate for uncomplicated
insertion [24].
• A lateral or paramedian insertion site is preferred [25].
• The catheter should be soaked in sterile solution prior to insertion to saturate
the cuffs and expel air [23, 24, 26].
• Sutures should not be used at the exit site [23, 26].
• Solution should be infused and drained prior to closure to evaluate catheter
function [24, 26].
• Nonreative, absorbable sutures should be used for the initial incision [23,
24].
• A sterile dressing is applied and the catheter is anchored at the exit site
to prevent movement [24].
The catheter should be immobilized well during healing to avoid tension
and torquing of the catheter. Tight clothing or other external pressure and
trauma to the exit site should be avoided [24].
Ambulatory dialysis should be delayed for 10 to 14 days. During this time
intermittent peritoneal dialysis can be administered with the patient resting
supine and with gradually increasing solution volumes, or the patient may
receive hemodialysis [27].
 33

Exit site care protocols

The impact of exit site care procedures on the incidence of exit site infec-
tion has been evaluated post catheter insertion [27-29] and an expert panel has
also made recommendations for post operative exit care [24]. Common
elements of these post operative exit site care procedures are listed in Table
VI.

Table VI. Common elements of post operative procedures for peritoneal catheter exit site care.

• Preoperative prophylactic antibiotic coverage

• Restrict dressing changes to PD staff
• Strict aseptic technique
• Immobilize catheter
• Nontoxic cleansing agent
• Exit dried after cleansing (air, 4 x 4)
• Sterile dressings
• Continued for ~ 7 days

The chronic exit site care procedures which have been recommended [24]
and studied [29-33] are more varied. Common elements include cleansing
and drying the exit site and securing the catheter. The ideal cleansing agent
is not known and recommended frequency varies from daily to several times
weekly.

Patient education

Nurses in home PD programs should be familiar with principles of learning,

and principles of adult and patient education. There are a number of excel-
lent texts available [34-37], as well as information specific to the ESRD patient
[21, 38] so this will not be discussed in detail here.
Each PD program needs to develop a generic curriculum for PD patient
education that can be modified for each individual patient and adapted for
patients with special needs. Teaching materials such as an instructor's manual,
printed information for patients or a patient education manual, patient proce-
dures, audio-visual aids (posters, slides, videos), practice supplies, a PD model
or "dummy tummy" for practicing exchanges, and patient record forms, need
to be developed. Use of these materials will be individualized based on the
characteristics of the teacher and learner.
The initial PD nursing assessment typically includes an education assess-
ment. Table VII lists items from a PD patient education assessment [39]. The
patient education process typically proceeds from assessment of the patient's
ability and readiness to learn to developing an individualized plan for the
patient's education. The plan includes developing behavioral objectives, out-
lining content, identifying specific teaching and learning activities, and
34

Table VII. Components of initial nursing assessment related to education.

• Educational background
• Work experience
• Previous involvement in self-care activities
• General level of health
• Level of cognitive function
• Psychiatric/emotional status
• Current knowledge of PD
• Concerns regarding ESRD and/or dialysis
• Current stressors and symptoms or stress
• Factors that interfere with health care or following the medical regimen
• Level of activity and independence
• Dialysis partner or backup support
• Physical disabilities which could affect learning
• Motivation to learn and perform PD
• Best way to learn
• Best time to learn
• Expectations of PD education program
• Reading test
• Memory test

planning for evaluation. The process continues through the actual implemen-
tation and evaluation phases. Most PD programs use a 1: 1 nurse patient ratio
for initial patient education.
Table VIII lists topics covered by over 90% of 18 facilities evaluated for
the best demonstrated practices in patient education [9]. Lecture and discus-
sion were the primary modes of patient teaching among the best demonstrated
practice centers. Demonstration, return demonstration and simulated problem
solving were also used by all of these facilities [9].
The patient education process is documented a number of ways. An account
of the assessment, goals and progress is recorded in the progress notes or nurses
notes. In addition, annotations are often made on the patient objectives or
education checklist to document a patient's mastery of the subject, that a
topic has been omitted, or that the routine approach or procedure has been
modified. At the completion of training there is a comprehensive evaluation
of the patient's (and/or partner's) knowledge and skills. A variety of testing
methods are used: verbal and written testing, return demonstration of proce-
dures, simulated problem solving.
It takes about 6 days for an average patient to complete PD training. Training
time varies from 5-8 hours per day and ranges from 5-10 days [9].
The process of patient education is ongoing, so review, reassessment of
learning needs and/or teaching take place at almost every patient contact.
 35

Table VIIl. Topics Required for Initial PD Training at 18 Facilities [9].

• Asepsis*
• Handwashing
• Exchange procedure*
• Exit site care*
• Recommended diet, meal planning
• Fluid balance
• Record keeping
• Procedure for system contamination
• Causes of peritonitis
• Peritonitis prevention
• Peritonitis symptoms*
• Reporting peritonitis to unit
• Peritonitis treatment
• Catheter complications
• Supply inventory
• Vital signs
• Laboratory values
• Patient responsibilities
• Communications and call system

* Minimum knowledge required for all patients

Followup care

PD patients require frequent monitoring, assessment, guidance and support

after completion of self dialysis education as they begin to dialyze indepen-
dently at home. This is most efficient and cost effective if the frequency and
type of followup is tailored to the patient's specific needs. Many programs
contact newly discharged patients two or three times during the first week
and then gradually increase the intervals between telephone calls. The first
clinic visit is usually scheduled a week or two post discharge and thereafter
the frequency is adjusted depending upon how well the patient is coping and
the number and type of problems. Clinic visits for nursing assessment and
further teaching are sometimes scheduled independently.
To our knowledge, most peritoneal dialysis programs require patients to
be seen in clinic every 4 to 8 weeks. Activities during a routine clinic visit
might include a review of home records or otherwise documenting the home
dialysis regimen, measurement of vital signs (including supine and upright
blood pressures), assessment of fluid balance, physical examination, evalua-
tion of the catheter exit site, selected blood chemistries and hematology, review
of medications, evaluation of activity level and rehabilitation status.
A more comprehensive assessment including more extensive bloodwork,
x-rays, EKG and evaluation of residual renal function is usually done bi-
annually.
Home visits are a valuable adjunct to followup from the center. Assessing
the patient and family in the home provides valuable insights about family
36

interactions, the degree of self care, supply inventory and storage, general
management of health, emotional adjustment and dietary practices [40].
Furthermore, home visits to both patients with perceived problems and those
doing well resulted in recommendations for change [40].

Liaison with hospitals

Outpatient peritoneal dialysis programs must work closely with affiliated

hospitals, in order to assure that hospitalized patients receive continuous,
coordinated care. Ideally, all PD patients would be admitted to a renal ward
where the entire nursing staff understands ESRD and could perform PD
procedures. Few programs enjoy such a luxury, or in fact, have a chronic
dialysis population large enough to support a renal ward. Other options are
a hospital PD staff; contractual arrangements allowing the outpatient PD nurses
to provide PD and exit site care for hospitalized patients; or a few medical-
surgical nurses trained to do PD procedures. It is also imperative that the
hospital(s) have supplies and PD systems compatible with those the patients
use at home and that PD procedures are consistent with the procedures used
and taught by the outpatient unit.
Staff nurses caring for PD patients in the hospital need to understand basic
principles of peritoneal dialysis and the routine medication regimens. Some
dialysis units have formal programs where ESRD is included in staff orien-
tation and continuing education programs are regularly scheduled. In large
hospitals a renal clinical nurse specialist can provide consultation and edu-
cation on an ongoing basis.
When catheters are inserted at the hospital (either on an inpatient or out-
patient basis) the nursing staff must be familiar with the preoperative care,
the operative procedure and postoperative catheter care and communicate with
the outpatient peritoneal dialysis staff at discharge [41].
Communication between the nurses responsible for outpatient and inpatient
dialysis is essential. The current dialysis regimen, current medications, fluid
status, PD related infections and other complications must be provided to
the staff taking over the patient's care. The use of a standard form for hospital
admission and discharge can facilitate effective communication of such details.

Liaison with extended term care facilities

As the age of dialysis patients increases, some chronic dialysis patients require
continuous skilled nursing care and reside in long term care facilities. Peritoneal
dialysis programs have successfully taught the staff at nursing homes to
perform PD procedures and manage fluid balance and other aspects of care.
The education outline is similar to that used for patient and family educa-
tion. The patient receiving chronic PD in a long term care facility still requires
 37

follow up by the PD staff, and continuing education and support are provided
to the staff.

Back up support or on call program

Patients performing dialysis and managing their ESRD at home require 24 hour
professional staff coverage to guide and assist them in identifying and
managing problems or complications. Support is also essential for newly
trained patients just beginning to dialyze independently at home. Both nursing
and physician must be on call 24 hours/day. Because most questions or
problems are related to dialysis procedures and/or are within the realm of
nursing practice, in most programs the nurse is the initial contact. For medical
problems the nurse consults with or refers to the physician on call. Although
we are aware of programs with only physicians on call, this is not optimal,
because problems other than medical emergencies are frequently not reported
by the patient, or are not dealt with until regular office hours when the nurse
returns.

Prescribing adequate dialysis

The traditional CAPD prescription (8-9 liter dialysis solution and 3-4 daily
exchanges) does not provide adequate dialysis for all patients, especially
after loss of residual renal function [43]. In order to determine the optimal
dialysis prescription for an individual patient the physician needs informa-
tion regarding the patient's peritoneal membrane characteristics. The peritoneal
equilibration test [42] measures the dialysate to plasma ratio of creatinine at
0, 2, and 4 hours dwell time, dialysate glucose/dialysate glucose at 0 dwell
time at 2 and 4 hours, and ultrafiltration volume. These values can be compared
to or plotted on published curves to determine if solute transport is average,
high or low. We recommend a peritoneal equilibration test at the time the
patient begins chronic peritoneal dialysis therapy, and repeated when there
are clinical indications that there may be a change in membrane transport
characteristics. Such indicators include an increase or decrease in ultrafiltra-
tion, and an unexplained change in serum chemistries.
Twardowski and colleagues [43, 44] and Diaz Buxo [45] have described
patterns of ultrafiltration and clearances in relation to solute transport. Table
IX indicates the most appropriate or preferred dialysis prescriptions based
on solute transport rates.
Obviously, if physicians are to prescribe the most appropriate therapy based
on peritoneal membrane transport rates, there must be administrative and
nursing support for CCPD, IPD and other cycler therapies. This includes
policies and procedures, availability of machine installation and maintenance,
nurses skilled in operating the cyclers, an educational curriculum and mate-
rials, and billing systems.
38

Table IX. Dialysis prescriptions based on peritoneal equilibration test results [43].

Response to standard CAPD (4 2 liter exchanges)

Solute transport Ultrafiltration Clearances Preferred prescriptions

High Poor Adequate Nightly IPD

Nightly TPD
DAPD
High average PD Adequate Adequate Standard dose
(any regimen)
Low average PD High Adequate Standard dose
Inadequate High dose PD
Low HD Excellent Inadequate High dose PD

IPD = intermittent peritoneal dialysis

TPD = tidal peritoneal dialysis
DAPD = dialy ambulatory peritoneal dialysis (no long overnight exchange)

Adequacy of peritoneal dialysis may be judged in a number of ways. Table

X indicates criteria for clinical assessment as summarized by Twardowski [44].
Determination of total (renal and dialysis) clearance of urea and/or
creatinine can also be used to evaluate the adequacy of the peritoneal dialysis
prescription. A minimum creatinine clearance of 40-50 liters per week (for
body surface area of 1.72 m2) has been recommended by Twardowski. For
anuric patients [44] 55.4 liters per week was recommended by Boen [46].
Urea kinetic modeling has also been applied to evaluate the adequacy of

Table X. Clinical assessment of adequacy of dialysis.

Clinical criteria
Patient "feels well and looks good"
Blood pressure controlled
Good fluid balance
Stable lean body mass
Stable nerve conduction velocities
Absence of uremic symptoms
Anorexia
Dysgeusia
Nausea
Vomiting
Asthenia
Insomnia
Laboratory criteria
Electrolytes within normal range
Serum creatinine < 20 (muscular persons)
< 15 (non muscular persons)
Hematocrit> 25% without EPO or steroids
 39

CAPD [47-50]. Applying urea kinetic modeling to clinical studies which

quantified dialysis has shown that patients with a weekly KtlV ~ 1.7 have better
clinical outcomes [51] and fewer deaths [52] than those with a weekly KtlV
urea < 1.2.

Prevention of peritonitis

There are a number of approaches to peritonitis prevention. Matching the

peritoneal dialysis system to the patient's abilities will reduce the risk of
contamination to the system. Reevaluation of the patient's abilities and
procedure technique after peritonitis has occurred enables the nurse to identify
a problem with technique or determine if the current system is no longer
appropriate for the patient. Even if no technique problems are identified,
changing to y-set [53, 54] or an ultraviolet light [55, 56] system may reduce
the incidence of peritonitis.
Effectively teaching the patient how to identify a break in technique or
contamination when they do occur and the appropriate response will reduce
the incidence of peritonitis. The use of prophylactic antibiotics for known
contamination of the system is recommended. Baxter Healthcare's best
demonstrated practices program found that ten of 15 centers with a 1986
peritonitis rate of 1 episode every 18.7 months routinely used prophylactic
antibiotics for a break in technique [57]. Thirteen of the fifteen centers
routinely prescribed prophylactic antibiotics for a known contamination of
the system. Ten of the same 15 centers also prescribed prophylactic anti-
biotic therapy for dental procedures to prevent hematogenous contamination
of the peritoneal cavity [57].

Effective diagnosis and treatment of peritonitis

Prompt diagnosis and effective treatment of peritonitis are essential for quality
care in a peritoneal dialysis program.
A high percentage of no growth on dialysate cultures [58, 59] during the
early years of CAPD stimulated research to improve the effectiveness of
laboratory cultures. There is abundant evidence that special procedures are
required to culture small numbers of bacteria diluted in dialysate. Culturing
large amounts of fluid [60, 61] using filtration [58, 62] or centrifugation [62,
63] to concentrate the sample, and removal of antibiotics present in the
specimen [61] have been shown to increase the proportion of positive cultures.
Bint et al. [64] suggest that the rate of positive cultures in clinical peritonitis
should exceed 90%.
Williams et at. recommended the use of a cytocentrifuge for the differen-
tial white cell count [65]. The cytocentrifuge concentrates the cells in a small
area on the slide so there are enough cells for an accurate differential count.
Two expert committees have issued recommendations for treatment of
40

peritonitis [61, 66] and reviews of peritonitis treatment has been published [67,
68]. In addition, Twardowski et at. [69] and the North American expert panel
[61] have developed decision trees for the medical management of peritonitis
episodes.

A bias for action

Peters and Waterman found that the successful companies were not para-
lyzed by formal communications and procedures [1]. Employees from the
top down communicated frequently and informally. Systems were simplified
and evaluation of progress was based on a few key numbers. In addition,
employees were encouraged to be innovative, and to experiment. Two of
these qualities particularly apply to peritoneal dialysis programs: monitoring
a few simple numbers, and encouraging innovation and experimentation.
Evaluation of the incidence of peritonitis and characteristics of peritonitis
episodes is a useful indicator of quality within a peritoneal dialysis program
and should be continuously monitored. A simple ratio of peritonitis episodes
over patient months exposure may be used to calculate the peritonitis rate or
life table analysis may be used to determine the probability of the first (or
subsequent) peritonitis episodes [70-72]. It is also of interest to compare
infection rates for each type of peritoneal dialysis therapy and each type of
peritoneal dialysis system which is used. Identifying the portal of entry (or
presumed etiology) may help in identifying trends and developing strategies
to reduce the incidence of peritonitis [59].
The incidence of exist site infection is also a useful indicator of the
effectiveness of catheter placement and exit site care procedures. Although
there is not a uniformly accepted definition of exit site infection in the
literature, infection rates in a single program can be compared over time if
the definition of exit site infection is consistent. Catheter removal and the
reasons for peritoneal catheter removal should also be monitored.
Patient survival and technique survival (the proportion of patients remaining
on peritoneal dialysis therapy) determined by actuarial techniques should
also be monitored as general indicators of program quality.
Innovation and experimentation is probably a hallmark of most successful
PD programs. We believe that all clinical research eventually results in
improved patient care, whether from increased knowledge of physiology,
improvements in patient assessment, development of new procedures, more
effective education, improved documentation and communication, develop-
ment of a new regimen or delivery system, development of new diagnostic
techniques, alternative treatment methods, improved cost effectiveness, or
identification of specific risk factors [73]. Participation in research can also
enhance professional growth and increase job satisfaction resulting in improved
staff retention for non-physician team members.
 41

Close to the customer

Successful companies provide unparalleled quality and service. They stay in

touch with their customers, and many of their best ideas and innovations
come from listening to the consumer [1]. These companies successfully satisfy
the needs of their clients and "anticipate their wants".
It seems to be a given that the members of a health care team will be
close to the customer; however, this may not always be the case. A new
nurse in our peritoneal dialysis program made the following observation after
a busy morning clinic:
"There's a difference between the staff physicians and the residents and
fellows in the way the communicate with patients. The residents and fellows
talk to the patients and tell them what they know; the staff physicians
listen to the patients and find out how they're really doing."
Although this example focuses on the physician, administrative staff and all
professional team members need to spend time with patients, listening as
they express their needs, and carefully evaluating their recommendations for
improvements.
• What are the patient's health-related goals?
• Which problems or symptoms distress them most or interfere with their
activities?
• What do they want to know?
• What do they need to want from us?
• Are there recommendations or procedures that they do not follow? Why
not?
At each patient visit we should be certain that we have fulfilled not only our
agendas, but have dealt with the perceived needs of the patient and family
as well.
A company also "stays close to the customers" by providing the services
they want. There are many ways a peritoneal program can enhance patient
services:
• Offer flexible appointment scheduling
• Return patient calls promptly
• Provide a toll free number for patient calls
• Arrange for repeat or additional laqoratory tests to be done at a labora-
tory near the patient
• Provide follow up clinics at satellites
• Communicate effectively at hospital admission and discharge
• Provide dialysis services in hospital
One way to "anticipate the wants" of our patients is through the use of
advance directives. At least one dialysis program has developed a formal
patient education program on advance directives [74, 75] and all patients are
offered the opportunity to complete the educational program and document
their choices regarding the type of procedures they wish to undergo and
designate an individual to make these decisions if they are unable.
42

Peters and Waterman also point out that problems with "product" or service
may be offset by caring [1]. On a recent patient satisfaction survey one of
our patients expressed dismay that the satellite unit near her home did not offer
CAPD support, and that she could not be admitted to the local hospital because
none of the nurses were trained to do peritoneal dialysis. In spite of this, she
was quite positive about the dialysis unit, largely due to the "caring" she
received. "I'm well pleased with all the services rendered to me by the nursing
staff and the [other] personnel. ... I couldn't have been treated better."
Measuring customer satisfaction is another hallmark of successful busi-
nesses. We recommend an annual evaluation of the quality of service using
a surveyor questionnaire. This must be written at an appropriate reading
level and printed with a large hold typeface so as many patients as possible
can read it.
Another principle borrowed from good business practice is to conduct loss
reviews. When a patient transfers to another dialysis program the reason
needs to be identified and then the team should evaluate whether the transfer
could have been prevented by changes in the structure or process or the level
of caring.

Summary

There are many facets to establishing and maintaining a peritoneal dialysis

program. A comprehensive and integrated approach to building in quality, as
recommended in this chapter, increases the chances for and degree of success.

References

1. Peters TJ, Waterman RH Jr: In Search of Excellence. New York: Harper & Row, 1982.
2. Boen ST: Integration of continuous ambulatory peritoneal dialysis into endstage renal failure
programmes: Present and future. In Atkins RC, Thomson NM, Farrell PC (eds) Peritoneal
Dialysis (pp 424-429). Edinburgh, Churchill Livingstone, 1981.
3. Baxter Healthcare Corporation: The Best Demonstrated Practices Program: Modality
Selection Practices. Deerfield, IL: Author, 1988.
4. Campbell A: Choosing an appropriate chronic dialysis therapy: A study of decisions by
nephrology staff and patients (abstract). Perit Dial Int 11 (Supp 1): 37, 1991.
5. Ross CJ, Rutsky EA: Dialysis modality selection in the elderly patient with end-stage
renal disease: Advantages and disadvantages of peritoneal dialysis. Peritoneal Dialysis in
the Geriatric Patient, A Supplement to the Advances in Perit Dialysis 6: 11-17, 1990.
6. Hamburger RJ, Mattern WD, Schreiber MJ Jr, Soderblom R, Sorkin M, Zimmerman SW:
A dialysis modality decision guide based on the experience of six dialysis centers. Dialysis
& Transplantation 19(2): 66-69, 84, 1990.
7. Shayman D, King K: Missouri Pre-Dialysis Patient Education Program. Columbia, MO:
Missouri Kidney Program, University of Missouri-Columbia, 1986.
8. Baxter Healthcare Corporation: Predialysis patient education program. Deerfield, IL: Author,
1990.
9. Baxter Healthcare Corporation: The Best Demonstrated Practices Program: Patient
Education and Training Practices. Deerfield, IL: Author, 1987.
 43

10. Nolph KD, Prowant BF, Webb J: National conference of continuous ambulatory peritoneal
dialysis. Perit Dial Bull 1: 65-66, 1981.
11. Oreopoulos DG: The peritoneal dialysis nurse: The key to success. Perit Dial Bull 1:
113-114, 1981.
12. Wilkens K (ed): Suggested Guidelines for Nutrition Care of Renal Patients. American
Dietetic Association, 1986.
13. Council on Renal Nutrition NKF, Inc: Guidelines for Estimating Renal Dietitian Staffing
Levels. New York: National Kidney Foundation, 1989.
14. Ducanis AJ, Golin AK: The Interdisciplinary Health Care Team. Germantown, MD: Aspen,
1979.
15. Marram GD, Barrett MW, Bevies EO: Primary Nursing, A Model for Individualized Care.
St. Louis: CV Mosby, 1979.
16. Giovanetti P: Evaluation of primary nursing. Annual Review of Nursing Research 4: 127-151,
1986.
17. Reed SE: A comparison of nurse-related behavior, philosophy of care and job satisfaction
in team and primary nursing. Journal of Advanced Nursing 13: 383-395, 1988.
18. Gardner KG: The Effects of Primary Versus Team Nursing on Quality of Patient Care
and Impact on Nursing Staff and Costs: A Five Year Study. Rochester NY: Rochester General
Hospital, 1989.
19. Joiner C, Van Servellen GM: Job Enrichment in Nursing: A Guide to Improving Morale,
Productivity and Retention. Rockville, MD: Aspen, 1984.
20. Mason EJ: How to Write Meaningful Nursing Standards, 2nd ed. New York: John Wiley
& Sons, 1984.
21. Burrows-Hudson S (ed): Standards of Clinical Practice for Nephrology Nursing. Pitman,
NJ: American Nephrology Nurses' Association, 1993.
22. Copley JB: Prevention of peritoneal dialysis catheter-related infections. American Journal
of Kidney Diseases 10: 401-407, 1987.
23. Suzuki T, Hasuo M, Suzuki T et al.: How to prevent exit site/tunnel infection in CAPD
patients? (abstract). Perit Dial Int 9 (Supp 1): 196, 1989.
24. Gokal R, Ash S, Helfrich BG et al.: Peritoneal catheters and exit-site practices: toward
optimum peritoneal access. Perit Dial Int 13: 29-39, 1993.
25. Helfrich GB, Pechan BW, Alijani MR, Barnard WF, Rakowski TA, Winchester JF:
Reduction of catheter complications with lateral placement. Perit Dial Bull 3 (Supp 4):
S2-S4, 1983.
26. Tenckhoff H: Chronic Peritoneal Dialysis Manual. Seattle: University of Washington, 1974.
27. Copley JB, Smith BJ, Koger DM, Rodgers DJ, Fowler M: Prevention of postoperative
peritoneal dialysis catheter-related infections. Perit Dial Int 8: 195-197, 1988.
28. Schmidt L, Prowant B, Schaefer R, Russ J, Kennedy J, Ryan L, Burrows L, Satalowich
R, Bartelt C: An evaluation of nursing intervention for prevention of postoperative
peritoneal catheter exit site infections (abstract). ANNA Journal 13: 98, 1986.
29. Jenson SR, Pomeroy M, Davidson M, Cox M: Evaluation of dressing protocols that reduce
peritoneal dialysis catheter exit site infections. ANNA Journal 16: 425-431, 1989.
30. Starzomski R: Three techniques for peritoneal catheter exit site dressings. ANNA Journal
11: 9-16, 1984.
31. Prowant BF, Schmidt LM, Twardowski ZJ, Griebel CK, Burrows LK, Ryan LP, Satalowich
RJ: A randomized prospective evaluation of three procedures for peritoneal dialysis catheter
exit site care. ANNA Journal 15: 219-223, 1988.
32. Luzar MA, Borwn C, Bals D, Hill L, Issad B, Monnier B: CAPD exit site care - results
of a randomized multicenter study. Perit Dial Int 10: 25-29, 1990.
33. Fuchs J, Gallagher ME, Jackson-Bey D, Krawtz D, Schreiber MJ: A prospective random-
ized study of peritoneal catheter exit-site care. Dialysis & Transplantation 19(2): 81-84, 1990.
34. Falvo DR: Effective patient education. Rockville, MD: Aspen, 1985.
35. Whitman NJ, Graham BA, Gleit CJ, Boyd MD: Teaching in Nursing Practice: A Professional
model. Norwalk, CT: Appleton-Centry-Crofts, 1986.
44

36. Smith CE (ed): Patient education: Nurses in partnership with other health professionals.
Orlando: Grune & Stratton, 1987.
37. Redman B: The Process of Patient Teaching in Nursing (5th ed.). St. Louis: Mosby, 1985.
38. Baer CL: Principles of patient education. In Lancaster LE (ed.) Core Curriculum for
Nephrology Nursing, 2nd ed. (pp 17-28). Pitman, NJ: American Nephrology Nurses'
Association, 1990.
39. Peritoneal dialysis health history and nursing assessment. Columbia, MO: Dialysis Clinic,
Inc.
40. Ryan L, Prowant B, Schmidt L et al.: Home visit effectiveness for peritoneal dialysis patients.
ANNA Journal 20: 333-6, 1993.
41. Thaler MK, Sasak C: Cooperative nursing care for patients using peritoneal dialysis. ANNA
Journal 15: 237-240, 1988.
42. Twardowski ZJ, Nolph KD, Khanna R, Prowant BF, Ryan LP, Moore HL, Nielsen MP:
Peritoneal equilibration test. Perit Dial Bull, 7: 138-147, 1987.
43. Twardowski ZJ, Khanna R., Nolph KD: Peritoneal dialysis modifications to avoid CAPD
drop-out. In Khanna R, Nolph KD, Prowant B, Twardowski ZJ, Oreopoulos DG (eds)
Advances in Continuous Ambulatory Peritoneal Dialysis/I987 (pp 171-178). Toronoto:
Peritoneal Dialysis Bulletin, Inc., 1987.
44. Twardowski ZI: PET - A simpler approach for determining prescriptions for adequate
dialysis therapy. In Khanna R, Nolph KD, Prowant BF, Twardowski ZJ, Oreopoulos DG
(eds) Advances in peritoneal dialysis, vol 6 (pp 186-191). Toronto: Peritoneal Dialysis
Bulletin, Inc., 1990.
45. Diaz-Buxo JA: Peritoneal permeability in selecting peritoneal dialysis modalities.
Perspectives in Peritoneal Dialysis 5(2): 6-10, 1988.
46. Boen ST, Haagsma-Schouten WQG, Birnie RJ: Longterm peritoneal dialysis and a
peritoneal dialysis index. Dialysis & Transplantation 7: 377-380, 1978.
47. Teehan BP, Schleifer CR, Sigler MH, Gilgore GS: A quantitative approach to the CAPD
prescription. Perit Dial Bull 5: 152-156, 1985.
48. Lysaght MJ, Pollock CA, Hallet MD, Ibels LS, Farrell PC: The relevance of urea kinetic
modeling to CAPD. Trans Am Soc Artif Intern Organs 35: 784-790, 1989.
49. Keshaviah PR, Nolph KD, Van Stone J: The peak concentration hypothesis: A urea kinetic
approach to comparing the adequacy of continuous ambulatory peritoneal dialysis (CAPD)
and hemodialysis. Perit DialInt 9: 257-260, 1989.
50. Gotch FA: Application of urea kinetic modeling to adequacy of CAPD therapy. In Khanna
R, Nolph KD, Prowant B, Twardowski ZJ, Oreopoulos DG (eds) Advances in continuous
ambulatory peritoneal dialysis, vol 6 (pp 178-180). Toronto: Peritoneal Dialysis Bulletin,
Inc., 1990.
51. Brandes IC, Piering WF, Beres JA: A method to assess efficacy of CAPD: Preliminary
results. In Khanna R, Nolph KD, Prowant B, Twardowski ZJ, Oreopoulos DG (eds) Advances
in Continuous Ambulatory Peritoneal Dialysis, vol 6 (pp 192-196). Toronto: Peritoneal
Dialysis Bulletin, Inc., 1990.
52. Teehan BP, Schleifer CR, Brown 1M, Sigler MJ, Raimondo J: Urea kinetic analysis and
clinical outcome on CAPD. A five year longitudinal study. In Khanna R, Nolph KD, Prowant
B, Twardowski ZJ, Oreopoulos DG (eds) Advances in Continuous Ambulatory Peritoneal
Dialysis, vol 6 (pp 181-185). Toronto: Peritoneal Dialysis Bulletin, Inc., 1990.
53. Maiorca R, Cantaluppi A, Cancarini GC et al.: Prospective controlled trial of a Y
connector and disinfectant to prevent peritonitis in continuous ambulatory peritoneal dialysis.
Lancet ii: 642-644, 1983.
54. Churchill DN, Taylor DW, Vas SI, Oreopoulos DG et al.: Peritonitis in continuous
ambulatory peritoneal dialysis: A multi-centre randomized clinical trial comparing the Y
connector disinfectant system to standard systems. Perit DialInt 9: 159-163, 1989.
55. Zappacosta AR, Perras ST: Reduction of CAPD peritonitis rate by ultraviolet light with
dialysate exchange assist device. Dial Transplantation 9: 483, 1988.
 45

56. Port FK, Held PJ, Nolph KD, Turenne MN, Wolfe RA: Risk of peritonitis and technique
Failure by CAPD connection technique: A national study. Kidney 1m 42: 967-74, 1992.
57. Baxter Healthcare Corporation. The Best Demonstrated Practices Program: Peritonitis
Management and Antibiotic Therapy Practices. Deerfield, IL: Author, 1987.
58. Rubin J, Rogers WA, Taylor HM, Everett ED, Prowant BF, Fruto LV, Nolph, KD: Peritonitis
during continuous ambulatory peritoneal dialysis. Annals of Internal Medicine 92: 7-13,
1980.
59. Prowant B, Nolph K, Ryan L, Twardowski Z, Khanna R: Peritonitis in continuous
ambulatory peritoneal dialysis: Analysis of an 8-year experience. Nephron 43: 105-109,
1986.
60. Vas SI: Microbiologic aspects of chronic ambulatory peritoneal dialysis. Kidney International
23: 83-92, 1983.
61. Keane WF, Everett ED, Fine RN, Golper TA, Vas SI, Peterson PK: CAPD related peritonitis
management and antibiotic therapy recommendations. Perit Dial Bull 7: 55-68, 1987.
62. Vas VI: Examination of fluids during peritoneal dialysis. Perit Dial Bull 1(1): 2-3, 1980.
63. Knight KR, Polak A: Laboratory diagnosis and oral treatment of CAPD peritonitis. Lancet
ii: 1301-1304, 1982.
64. Bint AJ, Finch RG, Gokal R, Goldsmith HI, Junor B, Oliver D: Diagnosis and manage-
ment of peritonitis in continuous ambulatory peritoneal dialysis. Lancet 1: 845-848, 1987.
65. Williams P, Pantalony D, Vas SI, Khanna R, Oreopoulos DG: The value of dialysate cell
count in the diagnosis of peritonitis in patients on continuous ambulatory peritoneal dialysis.
Perit Dial Bull 1: 59-62, 1981.
66. Keane WF, Everett ED, Fine RN, Golper TA, Vas S, Peterson PK, Gokal R, Matzke GR:
Continuous ambulatory peritoneal dialysis (CAPD) peritonitis treatment recommendations:
1989 update. Perit Dial Int 247-256, 1989.
67. Vas SI: Peritonitis. In Nolph KD (ed) Peritoneal Dialysis, 3rd ed. (pp 261-288). Dordrecht:
Kluwer, 1989.
68. Horton MW, Deeter RG, Sherman RA: Treatment of peritonitis in patients undergoing
continuous ambulatory peritoneal dialysis. Clinical Pharmacy 9: 102-118, 1990.
69. Twardowski ZJ, Nolph KD, Khanna R, Ryan LP, Prowant BF: Peritonitis management in
the CAPD program at the University of Missouri-Columbia: November 1983-0ctober 1984.
In Khanna R, Nolph KD, Prowant B, Twardowski ZJ, Oreopoulos DG (eds) Advances in
Continuous Ambulatory Peritoneal Dialysis (pp 61-65), Toronto: Peritoneal Dialysis
Bulletin, Inc., 1985.
70. D' Apice AJF, Atkins RC: Analysis of peritoneal dialysis data. In Atkins RC, Thomson
NM, Farrell PC (eds) Peritoneal Dialysis (pp 440-444). Edinburgh: Churchill Livingstone,
1981.
71. Corey P: An approach to the statistical analysis of peritonitis data from patients on CAPD.
Perit Dial Bull 1(6): S29-S32, 1981.
72. Pierratos A, Amair P, Corey P, Vas SI, Khanna R, Oreopoulos DG: Statistical analysis of
the incidence of peritonitis on continuous ambulatory peritoneal dialysis. Perit Dial Bull
2: 32-36, 1982.
73. Prowant B: Lack of research by non-physician health care professionals. Perit Dial Bull
8: 11-13, 1988.
74. Hammes BJ, Colvin E: If I Only Knew: A Patient Education Program on Advance Directives.
LaCrosse, WI: Lutheran Hospital Regional Dialysis Center.
75. Colvin E, Hammes BJ: If I Only Knew: A Patient Education Program on Advance Directives.
ANNA Journal 18: 557-60, 1991.
CHAPTER 4

Continuous quality improvement in dialysis:

Operations and controls for multi-center systems

EDWARD E. BERGER and EDMUND G. LOWRIE

Introduction

The multi-center system poses special challenges and offers special oppor-
tunities for dialysis quality assurance and improvement. Within the frame of
reference defined by the generic Continuous Quality Improvement (CQI)
model, for example, the multi-center system's large patient population and
breadth of clinical experience provide opportunities which are not available
to the single facility for developing and monitoring empirical indicators of
quality. At the same time, however, the implementation and operation of a
quality assurance system cannot be imposed by fiat upon a large and decen-
tralized system of facilities, each with its own physician staff responsible
for patient care. Relatively sophisticated communications and incentives are
required to achieve quality assurance goals in such a multi-center system.
The discussion here is based upon our experience with the ongoing
development and implementation of National Medical Care's (NMC's) dialysis
quality assurance system. As the largest U.S. provider of chronic renal dialysis,
with approximately 385 facilities caring for nearly 30,000 chronic dialysis
patients at the end of 1991, NMC has been able to bring a unique set of
resources to bear upon the problems of dialysis quality monitoring. Most
specifically, the company has maintained a number of closely related and
integrated patient and facility data systems designed to allow description and
analysis of patient characteristics, clinical laboratory values, treatment char-
acteristics, and a variety of outcome measures including morbidity and
mortality.
Periodic routine reports and special analyses derived from these data sources
are provided to practitioners at each individual facility in the NMC system.
Medical Directors and facility clinical staff thereby have the opportunity to
compare their own facility's performance on a variety of measures to that of
the system as a whole. These reports have proven to be powerful tools in
revealing clinical trends not obvious at the level of the individual facility,
provoking review at each facility of the quality of care being delivered, and
evoking facility-specific adaptations to improve outcomes.

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 47-62.

© 1994 Kluwer Academic Publishers.
48

By simple extension, a large system of facilities under common manage-

ment provides an interesting analogue to the problem of national imple-
mentation of a comprehensive quality system. NMC's experience thus contains
lessons, or at the very least perspective, for the development of public policy
initiatives for clinical quality assurance in dialysis and other institutional health
care settings. But these lessons are rooted deeply in the particular approach
to quality assurance selected, and it is to that general topic that we will next
tum.

Characteristics of an effective CQI system

The notion of CQI has so rapidly and completely been adopted as quality
assurance gospel in health care that it threatens, in its ubiquity and its gen-
erality, to lose its meaning. It is necessary therefore, in order to assure some
substance in our use of the term, to review briefly what we take to be the
key components of a CQI system. Our presentation is largely consistent with
that contained in the 1990 Institute of Medicine report "Medicare: A Strategy
for Quality Assurance" [1].
CQI, first and foremost, is predicated upon incremental progress toward
improved quality with no prescribed targets or limits. The process does not
end; there is no arbitrary division between acceptable and unacceptable states;
there is simply a continuing search for ways to make outcomes better.
A primary focus on outcomes and patient needs, rather than processes, is
a second critical characteristic of CQI. It is not that processes are unimpor-
tant. Undeniably, understanding of and control over the processes which
comprise the dialysis treatment are critical aspects of dialysis quality control.
An effective quality assurance system must promote such understanding and
control. But processes are not important for themselves. They become
important only because (or when) they are empirically linked to outcomes they
help to create and control. CQI encourages caregivers to review outcomes
and to make those outcomes the basis for a search for the adjustments in the
treatment process which will demonstrably yield clinical improvements.
Third, attention to systems of care rather than individual cases or providers
is equally important. CQI recognizes that results will always manifest them-
selves as statistical distributions, and that from a systemic perspective particular
outliers are far less interesting or important than broad measures of central
tendency and dispersion. If nothing else, focus on outliers is an inefficient
investment of scarce quality assurance resources. For an individual facility,
therefore, CQI would lead to statistical analyses of selected outcomes and their
determinants rather than review of a small sample of case records; for a system
of facilities, these same analyses can be extended to include summary measures
of the overall performance of each facility in the system.
Fourth, CQI focuses on the continuous feedback of new knowledge from
clinical practice as a source for clinical adaptation by practitioners. This
 49

contrasts to some other quality assurance models, which impose relatively static
standards of care based upon professional consensus, conventional wisdom,
or specific published doctrine. In CQI, a cycle is established which con-
stantly feeds itself: measured outcomes ~ exploratory dialogue ~ clinical
adaptations ~ improved outcomes, and so on.
Fifth and finally, CQI depends for its power and success on internal control
mechanisms driven by professional norms rather than external regulatory
monitoring. It assumes that caregivers have a commitment to providing care
of the highest possible quality, and works to empower them to improve clinical
standards by providing the best and most recent possible information upon
which to base clinical policies and decisions. It is always and essentially a
collegial rather than an adversarial process.
Externally imposed norms and the power to take punitive action neces-
sarily require defined and stable standards against which to measure specific
practitioners or facilities. They also virtually guarantee that practitioners and
facilities will be forced into a defensive rather than a cooperative stance in
dealing with the quality assurance authority. But those are qualities which
are antithetical to the nature of the CQI process. This truth becomes extra-
ordinarily important when considering public policy interventions to improve
clinical quality. It may be that the public regulatory enterprise necessarily
involves elements which are incompatible with CQI.
We need only look at the government's efforts to improve hospital care
by publishing hospital-specific annual risk-adjusted mortality data to see this
problem in graphic form. Publication of annual hospital mortality rankings
has lead to great public misunderstanding and outcry. As a result, hospital
administrators and practitioners have been forced to expend enormous effort
defending their institutions, through attacks on the government's general
competence and/or specific methodological shortcomings. The need to engage
in this public defense creates resistance to change rather than open partici-
pation in the kind of process of inquiry and adjustment which fosters real
quality improvements. For the vast majority of hospitals (and other types of
providers), private distribution of comparative mortality data would have
more beneficial effects on quality of care than the publication of rankings
by institution.

Adaptation of an industrial control model

CQI, which is at heart a set of goals and a set of supporting principles, does
not by itself provide the structure for an effective dialysis quality assurance
program. There is a further need for an operating model which can order the
variables encountered in the exploration of dialysis quality, provide guidance
in the development of the statistical analyses and information sharing processes
required, and help practitioners and facility managers to control those vari-
ables. For NMC, the general operating model of choice has been the one
50

defined for other applications by the Federal Food and Drug Administration's
Current Good Manufacturing Practices (CGMPs) [2], as concretized and
expressed in the form of the Quality Triangle.
For dialysis facilities, accustomed to dealing with Medicare Conditions
of Coverage as interpreted by state survey teams and possibly with State
licensure requirements, CGMPs have sometimes been viewed as extra-
ordinarily arduous and difficult; in 1986, for example, a short-lived proposal
that dialyzer reuse be made subject to CGMP-like requirements was greeted
by nothing short of horror. And it is certainly true that, in the hands of an
unrestrained and largely untrained external agency, CGMPs might in fact
become the excuse for unreasonable regulatory demands. But this is also true
of the chronic renal facility Conditions of Coverage, and some dialysis
facilities would argue that they have experienced unreasonable treatment at
the hands of state surveyors interpreting those familiar standards. The fact is
that CGMPs provide a proven model for ongoing product and process spec-
ification, monitoring and control which is extraordinarily well adapted to
quality assurance in dialysis facilities. The "production" of a dialysis treatment
and the generic manufacturing process addressed by CGMPs have many points
of contact, easily recognized and understood by persons with experience in
both settings.
CGMPs are predicated upon a rigorous systems approach which maintains
that carefully controlled "inputs" and standardized "processes" will yield
predictable and constant "outcomes". They require:
1. Clear specification of the character and quality of inputs to the produc-
tion process;
2. Clear description of the standardized process;
3. The definition and implementation of control systems to monitor inputs and
processes;
4. Carefully controlled documentation of all of these elements to allow for
internal and/or independent audit of process integrity;
5. A demonstration that the finished product meets predetermined performance
standards, either through universal testing or sampling as deemed appro-
priate; and
6. Tracking and investigation of production failures (represented for devices
by failure to pass release testing, customer complaints about released
devices, and/or reported incidents of failure in use) in order to under-
stand, isolate the causes of, and eliminate quality problems.
Each of these enumerated aspects of CGMP controls has a clear analogue
in dialysis. For the dialysis facility, inputs include the patient prescription,
the supplies and equipment used to provide a treatment, and the staff resources
and training available in the facility. Procedural manuals and related rules
define the constant elements of the process of providing a treatment, and the
dialysis flow sheet and patient medical record provide the documentation by
which to assess process control and such immediate physical parameters of
treatment as weight loss. Most facilities exercise careful control over incoming
 51

supplies, maintain equipment according to manufacturers' protocols, and imple-

ment standards for training and/or education of different classes of staff; they
also identify, track and attempt to eliminate procedure breaks and other
unexpected events which might impact on patient care. These activities are
important parts of a comprehensive facility quality assurance program but,
as we shall see, they are not adequate to assure quality.
There are different levels of outcomes in dialysis patient care. The proxi-
mate or immediate outcome is the character and quality of the dialysis treatment
itself, as indicated by successful and safe delivery of the prescribed therapy.
Proximate outcomes are those most directly controlled by the dialysis facility
and its staff, and therefore easiest to incorporate into a facility quality assur-
ance program. This aspect of dialysis is frequently assessed by sampling some
measure of delivered therapy, such as a monthly Urea Reduction Ratio (URR)
- the percentage reduction in Blood Urea Nitrogen (BUN) as calculated on
the basis of pre- and post-treatment values. URR is one simple way to
determine whether the treatment delivered was in fact the treatment prescribed.
The fact that the dialysis prescription is both an input to the treatment
and a standard for the assessment of a proximate indicator of the quality of
the treatment is important to keep in mind. It is one of the reasons why it is
wholly insufficient to deal only with the integrity of the dialysis process,
and absolutely necessary to go beyond process to a more sophisticated for-
mulation of outcomes. It may well be that changing the prescription will
generally be the most important thing that can be done to improve the quality
of an individual patient's care.
A variety of laboratory values thought to be indicative of therapeutic
adequacy are tracked monthly for every patient, and these values may be
viewed as intermediate indicators of treatment quality. But it should be
understood that as our concern broadens beyond the immediate dialysis
treatment, the ability to hold the dialysis facility solely responsible for quality
is diminished. Other factors, including comorbidities and patient compliance,
affect these intermediate measures. The facility can influence but cannot fully
control those measures.
For both proximate and intermediate outcomes, dialysis facility documen-
tation may not be as rigorous as the FDA would normally require in a medical
device establishment, but the essential character of the CGMP requirements
is already a part of the culture of the dialysis facility. It is only for more
long-term measurements of quality, such as morbidity and mortality rates, or
measures of rehabilitation, that the standard operating procedure of the dialysis
facility fails to incorporate a close analogue to the CGMP requirements. For
these long-term outcomes, where the facility's control over multiple con-
tributing factors is highly attenuated, the challenge of quality assurance, and
the potential benefits to patients, are greatest.
52

The Quality Triangle as a model for organizing clinical inquiry and

indicator development

The Quality Triangle (Exhibit 1) is a simple construct which can be used to

explain the linkages among the various levels of potential process and outcome
failures which must be observed and evaluated in a dialysis quality assur-
ance process. One could as easily construct a triangle based upon success rather
than failure. Here, a favorable outcome is simply the avoidance of a failure.
At one level a concrete representation of the logic and discipline imposed
by CGMPs, operationally the Quality Triangle is a useful tool for organizing
observations into meaningful categories and structuring analyses to shed light
on the determinants of quality. The triangle is easily adapted to other clinical
settings by substitution of the processes appropriate to those settings; it is
widely used in non-clinical settings to help monitor and control worker safety
or manufactured product quality.

Clinical
Quality
Morbidity

Clinical Failures

Clinical/lab

~~ Process!
Procedure
~-:.~;".".~ .. ~
Lf~~~~~~~~~~~ffi~~~~~~~~~~varlances
Exhibit 1.

As compared to higher levels of the triangle, variables specified at lower

levels are always:
1. Causally and (for the individual patient) temporally antecedent; they affect,
but are not affected by, the variables in levels above;
2. Less severe in their effect; at the lower levels, failures mayor may not lead
directly to adverse outcomes, but as you move up the levels of the triangle
the probability of causing an outcome failure increases; and
3. More numerous; there will be more procedure breaks than laboratory vari-
ances, more measurable clinical failures than deaths.
 53

Thus, the process or procedure variances found at the base of the triangle
are expected to account for the most numerous adverse observations; they
are antecedent to and by themselves less important than the observations or
events at higher levels, such as clinical laboratory variances, clinical failures,
hospitalizations, or deaths. Deaths and hospitalizations (the two highest levels
of the triangle) are, for the most part, caused by clinical failures, which are
themselves the product of clinical and/or process variances. The specific
variables included in each level of the triangle in Exhibit 1 are exemplary.
Others may readily be added or substituted.
Tracking of variables at the lower levels of the triangle is critical for two
reasons. First, being more numerous, they will be easier to observe; second,
being antecedent and of less severity, their reduction will have a profound
impact on the observed rate of more serious failures tracked at higher levels.
In other words, for example, early identification of a problem such as fistula
recirculation (perhaps through monitoring of the urea reduction ratio) will
prevent underdialysis of a number of patients, development of clinical com-
plications in some subset of those patients, and the need to hospitalize some
further subset.
As is apparent from the discussion above, the linkages between variables
in the lower levels of the triangle and variables in higher levels determines
the utility of the triangle for quality enhancement activities. While the linkages
may in some cases seem obvious, it is in fact the case that most have not
been empirically demonstrated. To a greater degree than is often admitted,
dialysis remains a treatment for which theory is well understood, practice is
often disassociated from that theory, and the links between practice and theory
are only recently being subjected to empirical review.
Intuitively, it seems obvious that a high rate of missed treatments will
eventually translate into some patients receiving inadequate therapy; that failure
to follow (within reasonable limits) a proper dietary regimen will result in
nutritional problems; and that a laboratory value out of the normal range for
a large group of patients will lead in time to predictable clinical complications.
But the nature and importance of those linkages are not well understood, .and
there is little empirical data available which can help us to define critical levels
for variables and/or the degree of association between different antecedent
variables and their probabilistic sequelae.
The National Cooperative Dialysis Study (NCDS) provided a carefully
controlled clinical trial which demonstrated the integrity of important elements
in the logic of the triangle, such as (1) the link between dialysis mechanics
(represented by treatment length) and adequacy of treatment (reflected in target
BUN value), and (2) the link between the adequacy of treatment (BUN control,
subject to adequate nutrition) and therapy failure (hospitalization, death, or
withdrawal from the trial for clinical reasons) [3]. But the degree of therapy
control exercised in that clinical trial cannot practically be duplicated in the
ongoing operations of a normal dialysis facility, and the specific lessons of
the NCDS for that setting are therefore not wholly clear. Dialysis quality
54

assurance needs ongoing analysis of the relationships among critical vari-

ables in the operational setting.
More recently, a series of studies performed at NMC as part of our ongoing
CQI effort have documented the relationships among a variety of patient
characteristics, clinical laboratory values, and patient mortality. We have been
able to exploit the size and comprehensiveness of our Patient Statistical Profile
system, and the linkage of that system to our clinical laboratory reporting
system, to explore the relative contributions of a variety of patient charac-
teristics and laboratory indicators to the determination of patient mortality
and morbidity.
These studies have helped to identify the centrality of clinical indicators
of patient nutritional status (serum albumin and serum creatinine) as well as
to provide empirical confirmation of the importance of treatment duration as
well as urea reduction in determining the success of therapy. They have
provided insight into the relative risk of adverse outcomes associated with
different levels of critical variables such as albumin [4], aluminum [5],
and urea reduction ratio (URR) [6], and have allowed exploration of the
hierarchy of causation among variables susceptible to clinical intervention
and those patient characteristics generally viewed as "case-mix" descriptors
and not susceptible to intervention [7].
Work of this type - detailed empirical analysis of the nature and degree
of the relationships among variables located at different levels of the causal
hierarchy represented by the Quality Triangle - is only possible with the
large patient base found in the multi-facility system, the ESRD Network
Area, the membership of a national trade association, or the entire national
dialysis system [8]. No individual facility, or even small group of facilities,
has a patient population sufficient to allow statistically valid multi-factorial
analyses of these complex interrelationships. But at present, no public or
quasi-public organization has made a full commitment to involving indi-
vidual facilities in an effective CQI information exchange.
Individual facilities not associated with large systems are therefore depen-
dent upon published reports for the most current clinically relevant information.
Even then, the CQI model is significantly compromised. Facilities have great
difficulty integrating their own experience with reports which do not contain
that experience as a component. Their level of engagement cannot be as great
when they are passive recipients of information rather than full partners in a
two-way information flow. While valuable, the process is something less
than the fully interactive exchange called for by the CQI mode.

Interactive communication with facilities

The flow of information to and from participating facilities is the most critical
operational element of a CQI program in the multi-facility system. For one
thing, as noted above, full involvement of the facility in a two-way flow of
 55

information is a critical determinant of the facility's commitment to the quality

improvement effort. If the facility reports its experience, it must then see
that experience reflected in descriptive and analytic reports based upon the
summed experience of the entire multi-facility system. Those reports should
be timely, so that the process is reinforced. They should be structured to
allow for easy comparison of each facility's performance to systemwide
benchmarks. And they must be recognized as useful in guiding clinical practice
or decision making. CQI program reports should, in summary, come to be
viewed as beneficial in and of themselves, a reward to the facility for its
data submission effort. Design of facility data collection instruments and system
reporting procedures must, therefore, consider these goals.
Other design goals for data collection and reporting speak to the integrity
and efficiency of the process. It goes without saying that data must be accurate,
and that accuracy will be affected by the design of data collection instru-
ments, techniques for transferring data to a unified data base, and controls over
the reporting process. Furthermore, it is obvious that system design can affect
the timeliness of data collection and reporting, and can be an important factor
in minimizing the cost of any desired level of system performance. It must
simply be remembered that any interesting multi-facility CQI program will
require extensive and sophisticated data handling and reporting capabilities,
and should not be attempted without professional systems design and imple-
mentation help.
NMC's data collection instruments and system reports are not provided
here because we wish to focus on the general structure of a CQI-based quality
assurance program for a multi-facility system, not the specifics of NMC's
program. There are many alternative ways of meeting the demands of the
CQI model, and NMC's is only one possibility.
NMC's quality assurance program obtains patient-specific data through
several related data collection instruments:
1. The Patient Statistical Profile (PSP) system registers every patient admitted
to a participating facility, encodes a variety of key descriptive demographic
and clinical variables, and records significant clinical events such as
hospitalizations (with discharge diagnoses and length of stay), transfers,
and discharges (to other facilities, to transplant, to death, etc.). Data forms
for the PSP system are at present completed in the facility and trans-
mitted by mail to the corporate office, where data are entered into the
computer. In time, facilities will have the capability to do direct data
entry, significantly reducing the total time required for this aspect of the
system.
2. NMC's Clinical Laboratory System, resident at our subsidiary LifeChem
laboratory, provides comprehensive historical patient-specific data on
clinical laboratory tests performed by LifeChem for all of its customers
(including non-NMC accounts). LifeChem reports test results to customers
electronically, and also writes those test results to a large data base which
can be linked to the PSP through common patient and facility identifiers.
56

3. From time to time, NMC undertakes a special time-limited clinical data

collection project on a topic of special interest. Typically, such efforts
require relatively brief data collection forms which are treated as supple-
ments to the standard PSP form. This mechanism provides the opportunity
to perform more specialized and targeted analyses than would be possible
with the PSP alone without the investment and ongoing expense associ-
ated with a permanent PSP system revision.
Reports back to facilities are of three general types:
1. Descriptive reports of facility experience compared to system-wide
empirical norms. There are, in this category, routine quarterly reports of
experience on key clinical quality indicators: mortality, both raw and
case-mix adjusted; overall hospitalization rates, with detailed analysis of
the 10 leading hospitalization discharge diagnoses; patient turnover for other
reasons; and lost treatments. There are also reports of the distribution of
clinical laboratory values in the facility and for the system as a whole.
These reports provide a snapshot of each facility's performance relative
to all other facilities, and provide a simple tool by which practitioners
can assess whether or not their results indicate the need to make changes
in any aspects of facility operation or clinical management. Reporting
against system mean performance is an important element in the CQI
model, as it allows for continuous progress without arbitrary limits.
2. Periodic evaluative reports on facility mortality experience. NMC gener-
ates a simple time series display of each facility's patient mortality
experience by calendar quarter, with mortality displayed as the ratio of
observed to expected (risk-adjusted) patient mortality (Exhibit 2). Risk
adjustment is made for basic demographic and comorbidity risk factors.
This presentation provides the facility Medical Director and medical staff
with one very clear indication of the quality of care being provided in
their facility, and directs their attention, in an unambiguous manner, when
observed mortality exceeds expected mortality for more than a single
reporting period, or when performance displays a negative trend (Exhibit
3).
It should be noted that NMC's system focuses on mortality experience
for two distinct reasons. First, patient mortality is inarguably the single
most elemental and unambiguous indicator of therapy failure. If we cannot
succeed in keeping patients alive in reasonable numbers, we cannot aspire
to going beyond that to making significant improvements in functionality,
life satisfaction, etc. Second, the system has been developing during a
time when excess mortality in the U.S. dialysis population [9], particu-
larly as it is associated with duration of dialysis treatment [10], has been
identified as a critical indicator of broad underlying quality of care
problems. In other words, our reports focus on mortality because of its cen-
trality to our enterprise, but also because of its topicality. It is a subject
which commands attention, and that is a necessary function of communi-
cations in a multi-facility quality assurance system. Risk-adjustment of
 57

PSP DATA SYSTEM

Observed/Expected Death Ratios
Quarterly OlE Ratios As Expected
6

o
~3

-
a:
o
W2
'- --- -. -.- ,--------------------- --- -----------.
..~-~............•.....~............•...........•...........•.... ~ ...................... .

o -.--.~
• ........................
_.-----, ::-.-.-.-•. ..................
,!--". • __ ._.__ ._._----_
,..----------_._._._.
..-:
• ..
01 02 03 04 01 02 03 04 01 02 03 04 01 02 03 04 01 02 03 04
1- 1986 -i 1- 1987 -I I- 1988 -I 1- 1989 -I 1- 1990 -i

• = Actual OlE Ratio

- - -= Expected OlE Rallo Umlts
QUARTER
- = Quarter Moving Average

Exhibit 2.

PSP DATA SYSTEM

Observed/Expected Death Ratios
Quarterly OlE Ratios In An Upward Trend
6 •.........•..•...••..•..••.••..••..•••••.•..•.•.•...••....•...••..•......•••.•..•...•......•.......••...••.•..........•.•.•••.•..•...•.•..•..•.•..

4 ...................•................•...•..................•..............•....................................•..................................

o
~3

-
a: •
...........................................................................•.....................................................................
W 2

01 02 03 04 01 02 03 04 01 02 03 04 01 02 03 04 01 02 03 04
I- 1986 - I I - 1987 --! I- 1988 --! I- 1989 --! I- 1990 - I
• =Actual OlE Rallo QUARTER
- - - = Expected OlE Ratio Umlls
- - =Quarter Moving Average
Exhibit 3.
58

mortality (and morbidity) analyses is done for two reasons. First, it is likely
that there are certain irreducible risks associated with some patient char-
acteristics, and failure to account for those risks would skew some facilities'
results. Second, risk-adjustment has become so commonplace that practi-
tioners would likely be tempted to take refuge behind their "sicker patient
group" to explain relatively poor performance if we reported non-adjusted
mortality. It is important to note, however, that risks are not always what
they seem at first look, and it may be that the assumption that a subgroup
of patients (e.g. diabetics) suffers from elevated risk may in fact obscure
the opportunity to realize significant improvements in treatment [11].
3. Analytical or exploratory reports on the relationships among clinical vari-
ables captured in the various data systems. These reports, which are
disseminated throughout the system at irregular intervals as analyses are
performed and refined, have proven to be the most important element in
raising questions about common clinical practice and in suggesting inno-
vations. Exhibit 4, for example, displays findings on the relationship
between URR and patient mortality which have caused reexamination and
upward revision of urea removal targets in many facilities. Findings such
as these, when reported to facilities, provide the challenges to which we
hope practitioners will respond.

Relative Death Risk By URR

In 13,473 NMC Patients
October 1990 - March 1991
2.0 i - - - - - - - - - - - - - - - - - - - - - - - - j . Case Mix Adj.
<.001 ~ CaseMix
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _- j + Lab Adj.
1.8

~ 1.6
a:
~ 1.4
:;::
~
a: Q)
1.2

1.0

0.8

<45 45-50 50-55 55-60 60-65 65-70 >=70

Values over bars = p values URR (%)
NS = not significant

Exhibit 4.
 59

These reports, by design, are never directive. They explore the clinical data,
report the statistical findings, and speculate about the implications. This stance
is critically important. While NMC does as a matter of policy require that
quality assurance system reports be reviewed at regular QIA team meetings,
and the minutes of those meetings are audited for compliance, that is the extent
of the control built into our system. We do not prescribe specific goals or
standards for facilities, or demand that practices be changed in any partic-
ular way in response to these reports. We cannot and do not desire to dictate
physicians' clinical behavior, or to pose as more expert in clinical decision
making than the audience. Efforts to do so would be counterproductive to
the interactive process that the CQI model demands. We do desire to provoke
the audience's interest and to provide stimuli to their critical and evaluative
processes. If physicians are motivated by these reports to explore the
applicability of the findings to their practice, or to investigate for themselves
whether an indicated relationship seems to hold for their patients, the system
is working, and that work will be reflected over time in changes in clinical
behavior and ultimately in the reported data.

Conclusion: Lessons for public policy

We have described the basic principles underlying a comprehensive CQI model

for dialysis facility quality assurance, and the basic structure by which NMC
has implemented that model for our large multi-facility system. Development
is ongoing, and new clinical initiatives are expected to emerge from the
interactive process we have described.
Does the process really work to improve quality of care? We believe so,
for several reasons:
1. It is very clear from formal and informal communications that caregivers
are paying close attention to much of the information being provided to
them, and grappling with the clinical issues raised by that information.
Attention varies, and it is safe to say that a larger portion of the audience
responds to the interpretive memoranda than responds to the quarterly
laboratory reports, but the program as a whole has effectively engaged
its public. For example, repeated observation of the very high proportion
of hospitalizations attributable to vascular access failure has engendered
tremendous interest in and attention to the problem of access manage-
ment; NMC has embarked on more detailed studies of the problem as
one response to these observations, and a number of our affiliated physi-
cians have already implemented aggressive and innovative access
management programs as a clinical response. Practitioner responses to
the weight of evidence demonstrating the centrality of nutrition in patient
survival have been equally vigorous.
2. It is clear that many Medical Directors care deeply about the relative
performance of their facility on quality indicators. Absent any external
60

enforcement or peer exposure, most nonetheless respond quickly to data

indicating less than median performance or higher than expected mor-
tality.
3. We have seen objective evidence at the facility level of clinical adapta-
tions to increase median URR in light of analyses demonstrating the
relationship between that measure of treatment effectiveness and mor-
tality. Physicians will change their prescribing practices to take account
of clear clinical findings.
4. NMC's overall mortality experience, representing a 20% sample of all
chronic dialysis patients, compares favorably to the nation's, despite the
absence of case-mix differences which would explain that fact. For the
years 1988-1990, HCFA's facility survey data reveal raw mortality rates
for patients in NMC's facilities of 22.4%, 22.3% and 21.1 %, respectively,
as compared to 24.8%, 24.1 % and 23.7% for non-NMC facilities.
It seems apparent that a similar program could be developed for all u.s.
facilities using organizational resources already in place. The United States
Renal Data System (USRDS), the Agency for Health Care Policy and Research
(AHCPR), the National Institutes of Health (NIH), the Renal Networks, and
HCFA's Bureau of Data Management and Strategy (BDMS) all house some
portion of the resources and capabilities required for a publicly sponsored
quality assurance program consistent with the CQI model, and may in
significant part be redundant. In terms of data resources and infrastructure,
the ESRD program is well situated to be at the leading edge of evolving
Medicare quality assurance strategy.
The critical unsolved problem upon which public policy efforts in dialysis
quality assurance have to date foundered is how to incorporate CQI's essen-
tial principles in a publicly mandated quality improvement program: Can a
mandatory program engage practitioners at the same level of intensity as a
voluntary program? Can a regulatory agency administer a quality assurance
program which does not rely upon a punitive element, as CQI would require?
Can a bureaucracy learn to depend upon the internalized norms of the provider
community rather than externally imposed norms? Is the concept of progress
without a defined end goal adaptable to a public program?
We can say with some certainty that the facility survey process as we
know it, while it can help to guarantee minimal standards of safety and care,
cannot be adapted to the goals of a CQI program. The public does in fact
require an authoritative regulatory assessment of certain minimum organiza-
tional, environmental and clinical aspects of dialysis facility operation. But
that critical policing function cannot realistically be entrusted to the same body
charged with implementing an ongoing quality assurance effort predicated upon
a cooperative, facilitative, and non-punitive stance. Furthermore, the HCFA
survey authority extends only to those aspects of operations which are under
the control of the facility management, and has not in any satisfactory way
addressed the problem of how to influence the clinical behavior of physi-
cians.
 61

In the Omnibus Budget Reconciliation Act of 1986 Congress charged the

ESRD Network Organizations with responsibility to develop standards and
criteria relating to the quality and appropriateness of patient care. After several
years characterized by little or no progress on this mandate by the Networks,
HCFA elected to take a leadership role and (with the Networks) has worked
to develop a set of criteria - Medical Review Criteria Screens - which can
be applied uniformly by all of the Networks. While there have been efforts
to wrap this Network Quality Indicator program in the cloak of the 10M quality
assurance project and CQI [12], it in fact is inconsistent with the basic
principles of continuous quality improvement in critical ways:
1. It compares facility performance to rigid standards defined by group
consensus; CQI demands that quality indicators be empirically demon-
strated, and that facilities be given information to evaluate themselves
against system-wide performance, with no arbitrary definition of what is
or is not acceptable;
2. It uses a small sample of individual case records to review performance,
rather than the overall pattern of care of the facility as favored by CQI;
3. It is in practice if not in intent overly prescriptive with regard to how
facilities must respond, and it relies upon the threat of punishment for
compliance; it thereby fosters compliance at the cost of progress, and
promotes defensiveness rather than openness; CQI allows for flexible and
creative response to problems, and upon iternalized professional norms
for compliance;
4. It gives nothing back to the facilities in the way of useful information,
and does not in any way increase the resources available to the facility
in its efforts to improve quality.
As an example of how the Networks' regulatory role and posture create
perverse incentives, the KTIV standard for adequacy of dialysis is interesting.
There are several different formulas for calculating KTIV, and the Networks
have not prescribed a formula, although they have prescribed a target KTIV
value. That value is based upon general consensus of an expert advisory panel,
but has never been validated empirically. NMC has had laboratory clients,
in response to Network reviews, request that we change the formula used by
our clinical laboratory in order to increase reported KTIVs. In other words,
when regulatory compliance becomes the issue, it is easier to fix the report
than to fix the treatment. The formula NMC uses for calculating KTIV had
originally been adopted in the belief that a conservative reported result would
encourage more aggressive dialysis prescriptions.
It may well be that the most appropriate public policy stance would be
for the government to facilitate - through training, funding and other mech-
anisms - the formation of voluntary provider consortia to share clinical data,
perform analyses, and feed results back to participating providers. Membership
in such a consortium, or implementation of a quality assurance program
which meets broad eligibility criteria, might be made a Medicare Condition
of Coverage. Alternatively, the government might create a single ESRD
62

program-wide quality assurance data bank reflecting the experience of all

dialysis facilities, modelled after the existing USRDS. HCFA's survey teams
could effectively assure that facilities review the data received from such a
data bank or from voluntary consortia, but it would be a grave mistake to
go beyond that function to prescribe specific quality assurance responses.
An effective public policy towards clinical quality assurance will facilitate
and empower health professionals by improving their access to relevant
information and analyses, and by removing rather than creating impediments
to the exercise of informed clinical judgement.

References

1. Lohr KN (ed): Medicare: A Strategy for Quality Assurance. Washington, DC: National
Academy Press, 1990.
2. Code of Federal Regulations, Title 21, Part 820.
3. Lowrie EG, Laird NM (eds): The national cooperative dialysis study. Kidney Int. Supplement
No. 13, April 1983.
4. Lowrie EG, Lew NL: Death risk in hemodialysis patients: The predictive value of commonly
measured variables and an evaluation of death rate differences between facilities. Am J
Kid Dis 15(5): 458-482, 1990.
5. Chazan lA, Lew NL, Lowrie EG: Increased serum aluminum: An independent risk factor
for patients undergoing long-term hemodialysis. Arch Intern Med 151: 319-322, 1991.
6. Lowrie EG, Lew NL: The urea reduction ration (URR). A simple method for evaluating
hemodialysis treatment. Contemporary Dialysis and Nephrology February: 11-20, 1991.
7. Lowrie EG, Lew NL, Huang WH: Race and diabetes as death risk predictors in hemodial-
ysis patients. Kidney Int 42 (Suppl 38): 5-22-5-31, 1992.
8. Berger EE, Lowrie EG: Mortality and the length of dialysis. JAMA 265: 909-910, 1991.
9. Held PI, Pauly MV, Diamond L: Survival analysis of patients undergoing dialysis. JAMA.
257: 645-650, 1987.
10. Held PI, Levin NW, Brovbjerg RR, Pauly MV, Diamond LH: Mortality and duration of
dialysis treatment. JAMA 265: 871-875, 1991.
11. Lowrie EG, Lew NL: Commonly measured laboratory variables in hemodialysis patients:
Relationships among them and to death risk. Seminars in Nephrology, 12(3): 276-283, 1992.
12. McClellan W: Applying the 10M perspective to the networks' approach to QA. Nephrology
News and Issues 6(1): 44-46, 1992.
CHAPTER 5

Quality of care in home dialysis

CHRISTOPHER R. BLAGG

Introduction

Home hemodialysis was first used in 1963 in the United States and Great
Britain to reduce the cost of long-term dialysis for chronic renal failure so
that the limited resources available could be used to treat more patients
[1, 2]. Shortly thereafter, home intermittent peritoneal dialysis (IPD) was
developed [3]. In recent years, continuous ambulatory peritoneal dialysis
(CAPD) [4] and various forms of continuous cycling peritoneal dialysis
(CCPD) [5] have generally replaced IPD.
With establishment of the Medicare End-Stage Renal Disease (ESRD)
Program in 1973, almost all U.S. residents who develop ESRD are eligible
for financial support for treatment by dialysis or kidney transplantation. As
of 1990, of 129,800 dialysis patients covered by the Medicare ESRD Program,
109,056 (84.0 percent) were on hemodialysis and 20,744 were on some form
of peritoneal dialysis [6]. Recent years have seen a significant increase in
the number of patients on peritoneal dialysis in the home, so that by 1990 these
accounted for 16 percent of all dialysis patients, and most of these (83.0
percent) were on CAPD. Simultaneously, the number of patients on home
hemodialysis has declined, so that in 1990 this was 2,483. This is 2.3 percent
of all hemodialysis patients, and 1.9 percent of all dialysis patients. Despite
the increase in the number of patients treated by CAPD and CCPD, from
12,189 in 1985 to 19,967 in 1990, at the same time the percentage of all patients
dialyzing at home declined from 19.3 to 17.4.
In 1990, of 22,640 patients dialyzing at home, 2,483 were home hemo-
dialysis patients (11.0 percent of all home dialysis patients), 16,969 were on
CAPD (75.0 percent), 2,998 were on CCPD (13.2 percent), and 190 were listed
as on "home peritoneal" dialysis (0.8 percent). Between 1985 and 1990 the
average annual increase in the number of patients treated by CAPD was 8.6
percent and for CCPD was 25.8 percent. This compared with an annual increase
of 8.9 percent in the total dialysis population, and a decline of 9.0 percent
in the home hemodialysis population. The average annual increase for all

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 63-83.

© 1994 Kluwer Academic Publishers.
64

modalities of in-center dialysis was 9.4 percent, while that for all modalities
of home dialysis was 6.7 percent.
Recent years have seen concern that mortality of United States hemo-
dialysis patients is significantly higher than that of patients in Europe, Japan,
and other developed countries [7, 8]. While this may in part reflect the sicker
patient population now being treated, it is also clearly related to the widespread
use of inadequate dialysis in the United States. For example, in 1986-1987,
hemodialysis patients reported to the European Dialysis and Transplant
Association (EDTA) registry on average received 30% more dialysis than
did patients in the United States [9]. This resulted from both shorter dialysis
times and the use of dialyzers with a lower surface area in the United States.
Because of this concern and growing interest in the evaluation and moni-
toring of health care, the Health Care Financing Administration (HCFA), the
Renal Physicians Association and other organizations and agencies have
become involved in assessing quality of care in dialysis and transplant patients.
So far there has been less interest in how this can be applied to patients
dialyzing at home. This chapter will discuss some of these issues.

Quality assurance and home care

Quality of care has been defined as "the performance of specific activities

in a manner that either increases or at least prevents the deterioration in
health status that would have occurred as a function of a disease or condi-
tion. Employing this definition, quality of care consists of two components:
(1) the selection of the right activity or task or contribution of activities, and
(2) the performance of those activities in a manner that produces the best
outcome" [10]. Donabedian first suggested that quality of care of patients could
be audited in the three dimensions of structure, process and outcome [11].
Structure refers to facilities, equipment and staff, and is not necessarily a
good guide to quality. Process refers to the activities of medical care, and
outcome refers to change in the patient's current or future health that can be
attributed to the medical intervention. Process and outcome are complemen-
tary, but because outcome other than mortality may be difficult to define,
process is often used as a surrogate for outcome. Process relates to appropriate,
effective, ethical, relevant, equitable, coordinated and socially acceptable care,
while outcome includes assessment of mortality, morbidity and other defined
end points as well as patient satisfaction [12].
Quality assurance aims to maintain quality of care and has been defined
as "a formal and systematic exercise in identifying problems in medical care
delivery, designing activities to overcome the problems, and carrying out
follow-up monitoring to ensure that no new problems have been introduced
and that corrective steps have been effective" [13].
With growth in the use of other forms of home care such as parenteral
and enteral nutrition; intravenous administration of antibiotics, anticancer
 65

agents, antihemophilic globulin, and other drugs; and various sophisticated

respiratory and cardiac care devices, quality assurance for care in the home
setting has become important. Between 1986 and 1988 the Joint Commission
on Accreditation of Healthcare Organizations (JCAHO) undertook a national
process to develop general standards for home care services. Services provided
by home health agencies, pediatric home care agencies, and private, mainly
for-profit, companies were considered, including home infusion and nutri-
tion services, home respiratory care services, and provision of durable medical
equipment and necessary supplies in the home. As a result of four national
field reviews and 10 national meetings involving more than 5,000 providers,
patients, payors and regulators, consensus was reached on the quality concerns
common to the home care industry generally. These concerns are:
• Recognition and respect for patient rights and responsibilities.
• An effective process for access to services, assessment of patient needs, plan-
ning to meet those needs, coordination of care, and appropriate discharge.
• Effective safety management and infection control in the delivery of home
care services.
• Adequate documentation of the care provided.
• Systematic monitoring and evaluation of the quality and appropriateness
of the care provided.
• Effective and efficient management of the organization responsible for
providing the home care.
• Appropriate governance of the organization responsible for providing the
home care.
• Delivery of services by home care professionals.
• Home pharmaceutical services, including dispensing, delivery and admin-
istration of drugs and training for self-administration of drugs or other
treatment, with ongoing clinical monitoring of patients.
• Personal care and support of activities in daily living.
• Selection, delivery and setup of medical equipment in the home, with
instruction in its safe and appropriate use and ongoing maintenance.
These 11 principles of quality home care served as the basis for chapters
in a manual entitled Standards for the Accreditation of Home Care published
by the JCAHO. This was revised in 1991 as the Accreditation Manual for Home
Care [14], and is being revised again for 1993.
Federal regulations developed for the Medicare ESRD program [15] and
various state and local regulations also have addressed many of these issues.
As a result, the annual ESRD Facility Survey carried out by state surveyors
for the Survey and Certification and Review Branch of the Division of Health
Standards and Quality of HCFA is responsible for ensuring compliance of
dialysis facilities with these regulations (42 CFR 405.2133-2184).
The final regulations implementing the Omnibus Budget Reconciliation Act
of 1986 required that the ESRD networks develop criteria and standards
relating to quality and appropriateness of patient care. As a result, in con-
sultation with several health care professionals, HCFA has developed case
66

review screens for use by staff of the ESRD networks in reviewing dialysis
facilities. These are described below. The screens are likely to be modified
with further experience, but this approach undoubtedly will continue to be used
to assess quality of care in dialysis patients.

Quality assurance in home dialysis

Quality Assurance (QA) for patients treated by either home hemodialysis or

peritoneal dialysis will be discussed under the general subject headings used
in the JCAHO Accreditation Manual for Home Care [8], with reference to
Federal Regulations as appropriate. In the remainder of this chapter, 'home
dialysis' refers to all modalities of dialysis in the home; where reference is
to a specific modality of dialysis this will be noted.

Patient rights and responsibilities

Federal regulations (42 CFR 405.2138) require the governing body of all
dialysis facilities to adopt written policies on the rights and responsibilities
of patients, and procedures for carrying these out. These must be made
available to patients, guardians, next of kin, sponsoring agencies, representa-
tive payees and the public. Facility staff must be trained and involved in
execution of these policies and procedures.
a. Informed patients. Patients must be fully informed:
• Of their rights and responsibilities and of all rules and regulations gov-
erning patient conduct and patient responsibilities.
• Of services available to them and related charges, including charges
for services not covered by Medicare. These services must include the
option to choose treatment by kidney transplantation and any of the
modalities of dialysis, including home dialysis.
• Of their medical condition by a physician unless medically contra-
indicated as documented in the patient record.
• Regarding the facility's reuse of dialysis supplies, including dialyzers.
• Regarding their suitability for transplantation and home dialysis.
b. Participation in planning. Patients must have the opportunity to partici-
pate in planning their own treatment, including selection of the modality
of care, and to refuse to participate in experimental research. Patients will
be transferred or discharged only for medical reasons, for their own welfare
or that of other patients, or for nonpayment of fees (except as prohibited
by the Medicare program). Patients must be given advance notice to ensure
orderly transfer or discharge.
c. Respect and dignity. Patients must be treated with consideration, respect
and full recognition of their individuality and personal needs, including the
need for privacy during treatment and provision of translation services if
required.
 67

d. Confidentiality. Patients must be ensured confidential treatment of their

personal and medical records, and may approve or refuse release of such
records outside the facility except on transfer to another health care insti-
tution, or as required by Federal or state law and the Secretary of Health
and Human Services.
e. Grievance mechanism. Patients must be encouraged and assisted in under-
standing and exercising their rights. These include the opportunity of
addressing grievances and recommending changes in policies to the dialysis
facility, the ESRD network, and regulatory agencies and bodies, without
restraint, interference, or fear of reprisal.

Patient care

Assessment of patient needs is an important part of the required process of

developing a patient long-term program and a patient care plan.
a. Patient long-term program. (42 CFR 405.2137(a» Each patient must
have a written long-term program representing the selection of a suitable
treatment modality (i.e., dialysis or transplantation) and dialysis setting
(e.g., home or in a center). This long-term program must be developed
by a team which includes the physician director of the dialysis unit where
the patient is being treated, a physician director of a unit which offers
self-dialysis training if this is not available at the location where the patient
is being treated, a transplant surgeon, a qualified nurse responsible for
nursing services, a qualified dietician and a qualified social worker. The
patient long-term program must be reviewed and revised by a similar
team at least annually, or more often as indicated by the patient's response
to treatment.
The patient, parent or legal guardian must be involved in development
of the patient's long-term program and their preferences must be given
consideration.
There is reason to believe many patients are not made aware of their
option to select some form of home dialysis. This applies particularly to
home hemodialysis because this is now only offered by a few facilities.
Unfortunately, if the patient is not informed of the advantages and disad-
vantages of all modalities of treatment they are unlikely to press for home
dialysis. Also, Medicare surveyors rarely, if ever, enquire whether a
physician from a program offering self-dialysis training is involved in
development of the long-term plan. In facilities not approved for home
dialysis training it is unusual for such an individual to be included in the
team. As of December 1991 there were 2,084 dialysis providers in the U.S.,
but only 1,271 (61 percent) were certified as training facilities [16], and
it is uncertain how many of these had an active home dialysis program.
One reason for the decline in home hemodialysis after introduction of
the Medicare ESRD program was the rapid proliferation of dialysis units
68

and nephrologists, most of whom had no exposure to home hemodialysis.

Consequently, when CAPD became available and required much less
training it rapidly filled the void that had resulted from the lack of home
hemodialysis programs in many parts of the country. This is unfortunate,
as home hemodialysis may be the best form of treatment for suitable
patients, and so far neither CAPD nor CCPD has proved to be a long-
term treatment in most patients.
b. Patient care plan. (42 CFR 405.2137(b». Each patient supervised by a
facility, including home dialysis patients, must have a written patient care
plan based on the nature of the patient's illness, the treatment prescribed,
and an assessment of the patient's needs. This must be developed by a team
consisting of at least the patient's physician, a qualified nurse respon-
sible for nursing services, a social worker and a dietician. The care plan
is reviewed by the team at least monthly until the patient's medical con-
dition has stabilized, six monthly thereafter, and is revised as necessary
to insure it provides for the patient's ongoing needs.
For a home dialysis patient, the care plan must also provide for periodic
monitoring of the patient's home adaptation, including visits to the home
by qualified facility personnel to the extent appropriate. This is usually
done by a nurse from the training staff.
Together, the patient long-term program and the care plan should ensure
the patient's needs are met and the appropriate care is coordinated. For
patients opting for home dialysis, this will mean referral to the facility's
own home hemodialysis or peritoneal dialysis training program or their
transfer to another facility for training. It will also include provision for
the supporting services required by the home dialysis patient, either directly
by the dialysis unit or by arrangement with a home dialysis program at
another unit or with a supplier.
c. Selection of patients for home dialysis. The most important factors in selec-
tion of patients for home dialysis are the patient's desire to be treated at
home, the availability of a suitable residence, and, in the case of home
hemodialysis and CCPD, the availability of someone to assist with dialysis.
In Seattle we have trained more than 2,500 patients for home hemodial-
ysis, IPD, CAPD and CCPD over the last 25 years. Our experience is
that anyone physically capable of doing some form of dialysis can be
trained to do this safely at home, provided they are motivated and are
encouraged and supported by training staff.
The only absolute medical contraindication to home dialysis is if the
patient requires the administration of intravenous medications during
dialysis. Medical problems such as angina or arrhythmias during dialysis
are not necessarily contraindications to home dialysis with a trained
assistant. Physical handicaps such as blindness, deafness or relative immo-
bility do not necessarily preclude some form of home dialysis [17, 18].
Perhaps the most important contraindication is a history of serious non-
compliance with the treatment regimen.
 69

As for age, home hemodialysis can be used for patients of all ages except
infants provided a family member or other assistant is available. In children,
adolescents and patients likely to be transplanted early, CAPD and CCPD
are particularly useful. At the other age extreme, in Seattle almost 50%
of home hemodialysis patients are aged 60 or older, and there have been
many successful patients in their 70s and 80s. Apart from a greater like-
lihood of medical complications, the most common problem for elderly
home hemodialysis and CCPD patients is lack of a suitable family member
to assist. This can be overcome by providing a trained dialysis assistant.
Patients should not do home hemodialysis on their own, although this
has been done successfully [19]. In contrast, CAPD should always be
done by the patient, making this particularly useful for the single patient.
Ability to learn to do safe home dialysis appears to correlate with the
patient's motivation rather than with their intelligence. Illiterate patients
or those who do not understand English will have more difficulties, but
with illustrations, videotapes and suitable translation such patients can be
trained successfully.
Another important psychological aspect of home dialysis is the degree
of anxiety of the patient and family. Some patients prefer CAPD or CCPD
because of their relative simplicity, but for all modalities of home dialysis
anxiety usually can be relieved by an appropriate approach before and at
the time the patient starts dialysis. Whenever possible the patient should
be introduced to the home dialysis options before they start treatment,
and have the opportunity to meet with and visit successful home dialysis
patients.

Safety management and infection control

Federal regulations require that a dialysis facility's patient care policies specify
the functions carried out by self-dialysis patients with respect to contamina-
tion prevention (42 CFR 405.2140(c)). Also, in the unlikely event dialyzer
reuse is done in the home, regulations require maintenance of records that
can be used to determine whether the procedures for rinsing, cleaning, disin-
fection, preparation and storage of reused items conform to Federal require-
ments (42 CFR 405.2150). This regulation details the conditions for reuse of
hemodialyzers and other dialysis supplies based on recommendations of the
Association for the Advancement of Medical Instrumentation (AAMI) [20].
a. Education. Safety management and infection control for home dialysis
patients require a program designed to educate staff, patients, family
members and, where applicable, dialysis assistants, in safety measures
necessary to minimize the hazards related to home dialysis. The home
dialysis training program should include written and verbal instruction in
basic home safety; safe and appropriate use of dialysis equipment; storage,
handling, delivery and access to supplies and drugs, including needles
70

and tubing sets, dialysate, dialysate concentrate and dialysate additives,

heparin, and erythropoietin; and identification, handling and disposal of
hazardous materials and wastes in a safe and sanitary manner in keeping
with applicable laws and regulations.
There must be documentation of ongoing education of staff, patients,
family members and dialysis assistants, and of their knowledge and
performance of safety measures and of the safe and appropriate use of
dialysis equipment and supplies.
All accidents, injuries and safety hazards must be reported and docu-
mented, and all accidents and injuries reported must be investigated and
the results of such investigations must be considered during review of
the quality of patient care. In addition, all equipment malfunctions and
serious injury or death associated with equipment or supplies must be
reported to the vendor and manufacturer.
The dialysis unit must have an emergency preparedness plan to ensure
continuing care and support in case of an emergency or other circumstances
making it impossible for the patient to continue dialysis in the home. This
should include a plan for providing backup dialysis for medical, psycho-
logical, social or technical problems in the individual patient, and a plan
for providing care and support for many home dialysis patients after a
natural disaster such as earthquake or flooding. These plans must be
communicated to the patient.
For all forms of home dialysis, the best way to ensure patient safety
is to have a specialized home dialysis training unit which does enough
training and supports enough patients regularly to maintain the skills of the
nursing and technical staff. This is best achieved by regionalization of home
dialysis to large training units.
Safety is dependent on selection of the appropriate fully monitored
equipment with home hemodialysis and CCPD, and on careful patient
training with all forms of home dialysis.
b. Infections. Measures must be taken to prevent, identify and control
infections. Infection control procedures should address personal hygiene,
aseptic procedures, transmission of infections, and appropriate cleaning and
sterilization of equipment. Patients, their families and dialysis assistants
must be advised of the precautions they should take. Staff must be instructed
in the importance of infection control and personal hygiene and their
responsibilities in seeing these are applied in the home dialysis program.
There must be a system for evaluating, reporting and maintaining records
of infections related to home dialysis occurring in patients, family members
and dialysis assistants, and, as appropriate, among staff.
In all patients, a follow-up program, including a home visit during
dialysis by a training nurse at least once a year, allows assessment of the
patient's adherence to the necessary standards.
 71

Home care records

Adequate patient records are essential prerequisites for quality assurance.

Federal regulations (42 CFR 405.2139) require a dialysis unit to maintain
complete medical records on all patients supervised by the unit, including home
dialysis patients. A staff member must serve as supervisor of medical records
services, with responsibility for seeing records are properly documented,
completed and preserved. Records must be readily available and systematically
organized to simplify compilation and retrieval of information. Federal regu-
lations (42 CFR 405.2163(e» also require a dialysis unit providing self-dialysis
services to have a record-keeping system which assures continuity of care.
Records should contain patient identification and name and telephone
number of a family member or other individual to be notified in case of
emergency or death. There must be documentation of the assessment of the
patient's needs, whether or not the patient is treated with reprocessed dialyzers,
the establishment of an appropriate treatment plan, and the care and services
provided to the patient (42 CFR 405.2139(a». There must be evidence the
patient was informed of the results of the assessment of their suitability for
transplantation and home dialysis as a part of the development of their long-
term care plan. The record should also include signed consent forms, referral
information with authentication of the diagnosis, medical and nursLlg history
reports of physician examinations, diagnostic and therapeutic orders, obser-
vations and progress notes, reports of treatments and clinical findings, reports
of laboratory and other diagnostic tests and procedures, and, if appropriate,
a discharge summary including final diagnosis and prognosis.
During home dialysis training there should be regular assessment of the
patient and any assistant's skills. No patient should be allowed to dialyze at
home until they have consistently displayed their ability to do this safely.
Records of home dialysis patients should also include a description of any
special safety measures required, notes on the suitability or adaptability of
the home for dialysis, notes of services provided by staff members, and the
names and telephone numbers of the home dialysis nurse, social worker and
nutritionist responsible for the patient's care. There should also be the name
and telephone number of the patient's physician, diagnoses, notes on blood
access and other surgeries, a current medication profile, information on
allergies, current dietary information, and, for home hemodialysis and CCPD,
information on the individual responsible for assisting with dialysis.
Ongoing progress notes must include reports of all patient interactions
with nursing, social work, dietetic or technical staff, and documentation of
all incidents occurring at home related to dialysis. Entries regarding a dialysis
may be completed by staff, or by the patient or assistant, and countersigned
by staff. Home hemodialysis and CCPD patients must complete and return
monthly log sheets documenting each dialysis, including equipment used, pump
setting, negative pressure, duration of dialysis, weight change, blood pressure
measurements, and any problems not already reported directly to the training
72

unit. In addition, patients must send a monthly blood specimen to a clinical

laboratory for the usual chemical analyses. The monthly log sheets and chem-
istry results are reviewed by the patient's physician and the training staff.
When care or a service is provided by other individuals or organizations,
this must be recorded in the patient record. Records of care and services
provided must be maintained by the facility or the supplier directly respon-
sible for billing for these, according to applicable law and regulation.
The dialysis unit must maintain home dialysis patient records in the same
way that records are kept on patients treated by outpatient dialysis. Format
of the records should be standardized, documenting all care and services
provided, maintained so all pertinent information from outside sources can
be routinely assembled in a timely manner, and there must be a system for
ready location of records and retrieval of information.
The facility must have a policy delineating who has authority to make entries
in and review the patient record. A physician's verbal or telephone orders
can be recorded by a qualified individual, and must be authenticated by the
physician within a period specified by the facility and according to applic-
able law and regulation. All entries must be authenticated by the individual
making the entry, including date and title, and the policy should state when
a signature, initials or a computer key may be used to authenticate an entry.
Records must be reviewed to assure that documentation is entered in a timely
manner and following facility policy.
When a home dialysis patient is discharged from the ongoing care of a
facility the record must be completed within a reasonable time, as spelled
out in facility policy. Upon discharge, a summary will be written in the
record, including date and reason for discharge, name of the facility to which
the patient is transferring, status of any medical problems, status of the patient,
and a summary of the care and services provided to the patient.
Home dialysis patient medical records will be retained in accordance with
applicable State statute or, in the absence of a State statute, for five years
from the date of discharge in adults, or for three years after the patient comes
of age under State law, whichever is longest (42 CFR 40S.2139(e)).
Reasonable security measures will be taken to safeguard both the home
dialysis record and its content, whether in hard copy, on film, or in computer
form, against loss, defacement, tampering, and unauthorized disclosure or
use (42 CFR 40S.2139(b».
The facility will have policies regarding confidentiality of records and
release of health information. Written consent of the patient or an authorized
person acting for the patient is required for release of information not provided
by law.
A representative sample of records should be reviewed at least quarterly
to assure these reflect the care and services provided, and the condition and
progress of the patient.
 73

Quality assurance

A facility providing home dialysis services should have an ongoing quality

assurance (QA) program designed to:
• Monitor objectively and systematically and evaluate the quality of patient
care in the home.
• Pursue opportunities for improving patient care.
• Resolve any problems identified.
There should be a written plan delineating the objectives and scope of the
QA program, activities to be monitored and evaluated and how this will be
done, methods for reporting results and taking follow-up action, and who
has responsibility for monitoring each activity. The program should address all
care and services provided to home dialysis patients, whether directly or
through contract with another institution or supplier.
Services monitored and evaluated should include those directly related to
home dialysis, including adequacy of dialysis and pharmaceutical services;
personal care and support services; and equipment management services. There
also should be ongoing identification of important aspects of care associated
with risks and recognized problems.
Measurably indicators should be defined which reflect activities, events,
occurrences, outcomes and patient satisfaction. The level of performance that
represents the extent to which care is effective, adequate, and provided in
the setting best suited to the patient's needs should be identified for each
indicator, and also the level of care which represents the threshold at which
further evaluation of quality and appropriateness of care must be initiated.
The patient's physician, the medical director of the home dialysis facility,
and a qualified home dialysis nurse, social worker and nutritionist should
develop these indicators and thresholds. They should evaluate the informa-
tion collected in order to pursue opportunities for improvement, identify
problems, and assess staff performance and patient satisfaction. Actions should
be taken to resolve identified problems, and the effectiveness of these actions
should be evaluated and reported.
The findings of the QA program may require action by the home dialysis
training staff and the facility administration. Relevant findings can be used
to evaluate the performance of staff and may require in-service training or
continuing education. Findings, conclusions, recommendations, actions taken
and the results of these, should be documented and reported. When oppor-
tunities occur to improve home dialysis patient care and services or when
problems are found which involve other organizations or suppliers, the
appropriate findings, conclusions and recommendations should be communi-
cated to them for appropriate follow-up and action.
The objectives, scope, organization, and effectiveness of the QA program
should be reevaluated at least annually, and revised as necessary.
a. Specific requirements for home hemodialysis. Good blood access is essen-
tial so the patient or assistant can do fistula puncture. The dialyzer used,
74

duration of dialysis, and ultrafiltration rate should be selected to minimize

symptoms during and after hemodialysis, and hypotensive drugs should
be used with caution. The patient's home must have a suitable water source,
electrical supply and drainage. Dialysis equipment should be chosen for
reliability, adequacy of monitoring, ease of use, and ready availability of
maintenance and repair services. Water treatment must be based on analysis
of the patient's own water supply.
b. Home dialysis training. This should begin as soon as possible after the
patient starts dialysis so maximum independence can be encouraged.
Training for CAPD and CCPD can start as soon as the catheter is estab-
lished, and the patient can be dialyzing at home within a few weeks of
starting treatment. Blood access is a limiting factor for home hemodial-
ysis training, and whenever possible should be created early, before the
need to start dialysis. The first step in training is for the patient or
assistant to learn fistula puncture as this is often the greatest psycholog-
ical barrier to overcome. Most patients and assistants can be trained to
do safe home hemodialysis in three to six weeks once fistula puncture is
established, and in Seattle the average number of training dialyses is
fewer than nine.
Whenever possible training should be done in a room away from the
main dialysis area. In Seattle we have a separate orientation unit where new
patients dialyze in a private room for the first two months or so. This
gives the opportunity to provide information on medical issues and avail-
able social work, nutritional and other supporting services. Patients are
assisted in choosing which modality of treatment to undertake. As appro-
priate, these include transplantation and all forms of home dialysis [21].
Training materials should be available to provide information about
kidney disease, blood or peritoneal access, diet, medications, treatment
options, the dialysis procedure, and the availability of supporting services.
Printed materials, slides, posters and videotapes may be particularly helpful
[22].
Home dialysis contributes to rehabilitation as patients learn new skills
and develop self-confidence in their ability to have control over their own
treatment and eventually over their life in general. Every effort should
be made to show that dialysis is simple and safe if carried out properly.
The patient must be trained to take responsibility for as much of their
own care as they can so as to avoid overdependence on their spouse or
family. CAPD patients should always take total responsibility for their own
care. For home hemodialysis and CCPD patients, a paid assistant helps
reduce the impact of dialysis on the family, but always raises the danger
the patient will be become too dependent on the assistant. Consequently,
successful home dialysis requires careful assessment of patient and family,
careful selection of a compatible assistant if one is required, and ongoing
support and follow-up by facility staff.
During home hemodialysis and CCPD training the progress of the patient
 75

and assistant should be assessed at regular intervals to ensure acquisition

of the necessary skills and knowledge. There is no need for them to learn
the intricacies of the equipment, but rather how to do dialysis safely, what
actions to take in case of an emergency or other problem during dialysis,
and when and how to call for assistance and advice.
c. Home dialysis. Once a patient has gone home, monitoring of care is
dependent on information returned to the facility by the patient or their
assistant, home visits by training staff, and regular visits by the patient
to their physician's office. The patient or assistant must complete a log
sheet for each dialysis and return this to the facility monthly. Also, for
patients self-administering erythropoietin, a log sheet recording doses and
dates must also be returned. These log sheets enable facility staff and the
physician to ensure the patient reports getting enough dialysis and the
appropriate dose of erythropoietin. Dialysis technique should be assessed
at intervals during a home visit from a training nurse. This enables
correction of problems, updating on modifications of technique, or referral
for formal retraining. The patient must send a monthly blood sample to
the laboratory for testing, the results being distributed to the patient,
physician and facility. Finally, all home dialysis patient should be seen
in the physician's office at least monthly for review of their general health
and any problems related to dialysis.
d. HCFA medical case review criteria screens. HCFA recently developed
several medical case review screens for use by networks to review dialysis
facilities. There are intended to apply to home dialysis patients as well
as to in-centre dialysis patients. The following are the criteria screens for
adult patients. Those for pediatric patients are similar.
(1) Anemia: The hematocrit will be checked and recorded monthly, and
will be 24% or greater with no decrease of five or more points in
the last three months, and the patient will not have been transfused
in the last six months. Exceptions include following an acute bleeding
episode or when there is a documented explanation for why the patient
is not on erythropoietin.
(2) Blood pressure control: For hemodialysis patients, predialysis and
postdialysis diastolic blood pressure will be 100 mm Hg or less on
at least six of the last 12 dialyses, unless there is evidence of
appropriate interventions or documentation of noncompliance. For
home hemodialysis patients this information will generally be obtained
from the patient's logsheets. For peritoneal dialysis patients, dias-
tolic blood pressure will be 100 mm Hg or less on at least five of
the most recent clinic visits.
(3) Metabolic control: Other than in exceptional circumstances hemo-
dialysis patients should have a predialysis serum potassium concen-
tration of less than 6.5 mEq/1 for two of the last three monthly
determinations. As for adequacy of dialysis, the patient must have a
pre-/postdialysis BUN reduction of greater than 60 percent, and/or a
76

KtlV urea equal to or greater than 1.0, and/or a pre-/postdialysis serum

creatinine concentration reduction of greater than 50 percent, measured
at least once in the preceding three months. For peritoneal dialysis
patients, there should be a serum potassium concentration of less
than 6.0 Meq/l on the last three measurements. As for adequacy, the
peritoneal dialysis patient should have a total (i.e. residual plus
dialysis) creatinine clearance of at least 40 l/week/1.73 m2 measured
within the last six months.
(4) Nutritional status: With certain exceptions, hemodialysis patients
should have a serum albumin concentration of greater than 3.5 g/dl
on at least two of the last three monthly measurements. For peri-
toneal dialysis patients, this figure is greater than 3.0 g/dl. Also, for
peritoneal dialysis patients, there should not have been a weight loss
of more than 5 percent of postdialysis weight over the most recent
three months.
(5) Renal osteodystrophy: With certain exceptions, patients should have
serum phosphorus concentration between 3.5 and 6.0 mg/dl and a
serum calcium concentration of between 8.5 and 11.5 mg/dl on at least
two of the three last monthly measurements, and an alkaline phos-
phatase of less than twice the upper limit of normal over the previous
six months.
(6) Transplantation as a treatment modality: The medical record must
show the patient and/or family or caregiver has been informed of
transplantation as a treatment option, and there must be documenta-
tion by a member of the transplant team that this has been done. The
record must also show that the patient was considered for trans-
plantation at the onset of end-stage renal disease and at least annually
thereafter.
There are problems with some of the screens currently in use. In
the case of the serum albumin concentration while the levels proposed
by HCFA are appropriate where the laboratory uses the bromcresol
green method to measure this, if the laboratory uses the more specific
bromcresol blue method all values should be 0.5 g/dl lower [23].
Similarly, KtlV can be calculated in several different ways, each with
different pitfalls in technique, and a KtlV of 1.0 is now known to
represent inadequate dialysis. This should be at least 1.2. These
problems point out the difficulties when the bureaucracy develops
criteria for application to clinical providers with insufficiently broad
input and without adequate preliminary testing. Undoubtedly the
screens will be modified with time, but a facility providing home
dialysis services should also develop its own internal QA program.
 77

Management and administration

Federal regulations describe the qualifications and role of the chief execu-
tive officer (42) CFR 405.2135(c)), and the physician director (42 CFR
405.2161) of a dialysis facility. The former is responsible to the governing
body for operation of the facility, for assuring compliance with applicable laws
and regulations, and for taking action on reports and recommendations of
authorized planning, regulatory and inspection agencies. He or she is also
responsible for carrying out policies and procedures established by the
governing body, including those addressing delivery of care and services and
financial management.
The physician director, who may also be the chief executive, is respon-
sible for the quality of services provided, including home dialysis services,
their availability, and appropriate staffing. He or she is also responsible for
participation in selection of a suitable modality of treatment for all patients,
assuring adequate training of nurses and technicians in dialysis techniques,
assuring adequate monitoring of patients and the dialysis process, assuring
development, availability and implementation of a patient care policy and
procedures manual, and ensuring that teaching materials are continuously
available for patient use whenever self-dialysis training is offered.
Lines of responsibility and accountability in the facility must be clearly
established, including the relationship of the home dialysis program to other
departments and services, the governing body, and, where applicable, the
medical staff. Facility staff must participate in orientation, in-service training,
and continuing education programs.
When any home dialysis-related services are provided by an outside
organization or supplier, there must be a written agreement defining the nature
and scope of the services.

Governing body

Federal regulations (42 CFR 405.2136) require there be an identifiable

governing body or designated persons with full legal authority and responsi-
bility for managing and operating a dialysis facility, and full disclosure of
ownership to the State survey agency. The governing body must establish
operational objectives for the facility and adopt rules and regulations regarding
health care and safety of patients and protection of patients' personal and
property rights. They must appoint a qualified chief executive officer who is
responsible to them for operation of the facility and development and imple-
mentation of personnel policies and procedures that support sound patient
care and promote good personnel practices.
The governing body is also responsible for approval of the facility's patient
care policies, including those related to the provision of home dialysis and
necessary support services. In particular, they must assure that access to self-
78

care is available to all patients for whom the facility maintains patient care
plans.

Home health services

Successful home dialysis requires carefully supervised patient training and

follow up. This is best provided by nurses experienced in both dialysis and
patient education. Not all nurses are able or willing to adopt a "hands off"
approach and allow patients to learn by making mistakes during training. Thus,
selection of appropriate staff for home dialysis training is most important.
Federal regulations (42 CFR 405.2102) require that a home dialysis unit be
directed by a nurse having at least three months' experience with self-care
dialysis. In practice, this should be the absolute minimum.
The need for experienced staff is a good reason to consider regionaliza-
tion of home dialysis training and support programs. This ensures sufficient
patients to provide continuing experience for staff. This staff should serve
also as support staff, providing 24-hour telephone support, advice and direc-
tion for home dialysis patients with problems, and making home visits on a
routine and an as-need basis.
The training staff must contact the patient's physician whenever warranted
by the patient's condition. Written reports on the care and progress of their
patients should be provided to physicians at least monthly, and there should
be procedures to ensure ready access to medical consultation whenever
necessary.
The home dialysis program should have written policies and procedures
regarding resuscitation of patients, and provision for documentation in the
patient record of relevant orders by the patient's physician. Patient and
assistant must be aware of how to contact local emergency medical services.
There should be a written agreement with the clinical laboratory that
provides services to the home dialysis patients.

Pharmaceutical services

The use of drugs by home hemodialysis patients should be carefully moni-

tored. This applies particularly to antihypertensive agents. State law may
delineate who can dispense drugs used during home dialysis and which drugs
may be self-administered by the patient. Usually the facility can an dispense
and arrange delivery of dialysate concentrate and appropriate additives such
as potassium chloride if required, and heparin for use during dialysis. Other
drugs should be dispensed by a pharmacy or the physician's office, and not
by the dialysis facility, and it is the physician's responsibility to ensure
patients understand their drug regimen and are trained in self-administration
of any drugs they may require. The one exception is self-administration of
 79

erythropoietin by home dialysis patients. Training for this can be done in

the physician's office or by the home dialysis training staff. Both need to
review the log sheets reporting the dose and dates of erythropoietin adminis-
tration.
The requirements for CAPD and CCPD patients are similar, although it is
much more common for these patients to receive their supplies and related
drugs from an outside supplier.
All drugs, equipment and supplies provided to home dialysis patients should
be selected and procured based on evaluation of the needs of the patient
population and the prescribing habits of their physicians. There must be a
process for labeling and dispensing dialysis-related drugs and supplies and
records must be maintained as required by law and that ensure adequate control,
accountability and appropriate use. The records should include the physi-
cian's order; patient name, address and telephone number; date of the
prescription; supplies or drugs provided and their number, concentration or
dosage; name of the prescribing physician; and name or initials of the indi-
vidual responsible for providing the supplies or drugs. The label on the supplies
or drugs should include the name, address and telephone number of the home
dialysis program or the supplier; the full name of the patient; the name of
the product, its composition or strength and the amount provided; the
expiration date; directions for use; and any other pertinent information.
There must be a system to ensure supplies and drugs are delivered to the
right patient, and essential drugs and dialysis supplies should be available
on an emergency, 24-hour a day basis. In case a problem should be dis-
covered with a supply or drug there must be a system for recall.
The patient, family member or assistant must be trained in the safe and
appropriate use of dialysate concentrate and additives, of heparin during
dialysis, and of erythropoietin. This should include instruction on storage in
the home, disposition of unused supplies, and disposition of used supplies,
including syringes, tubing sets, dialyzers, intravenous fluid bags and dialysate
containers. Teaching regarding self-administration of erythropoietin should
include aseptic technique for injecting erythropoietin either intravenously in
association with hemodialysis or subcutaneously. Appropriate storage methods
including storage duration, handling and disposition of unused erythropoietin,
and proper disposition of used supplies such as syringes, vials, and ampules
also must be taught.
Patients must also be provided with clear up-to-date information enabling
them to contact an appropriate individual in the event of problems or
emergencies related to dialysis or to drugs administered in the home.

Personal care and support services

Federal regulation (42 CFR 40S.2163(c» require that a dialysis facility have
either a licensed social worker with a masters degree in social work, or an
80

individual with two years experience as a social worker, one of these years
being in a dialysis or transplant unit, and who has a consultative relation-
ship with a social worker with a masters degree. Such a person, with the support
of the training staff and the patient's physician, can provide support for personal
care and the activities of daily living. For example, with erythropoietin and
coordinated social work, vocational counseling and exercise training, many
patients can be helped improve their daily lives. In those trained early for home
dialysis there is often opportunity for the patient to continue to work.
Unfortunately, some facilities provide much less support than is desirable
and few patients are referred for vocational counseling.
While HCFA sponsored a conference on rehabilitation of ESRD patients
some years ago, this did not lead to any actions. It is to be hoped that current
and proposed efforts by the Institute of Medicine and the American Association
of Kidney Patients will be more successful. Clearly, home dialysis has an
important role to play in providing the best quality of life and greatest
opportunity for rehabilitation for dialysis patients.

Equipment management

A facility providing home dialysis must provide surveillance of home

adaptation, installation and maintenance of the patient's equipment; testing and
appropriate treatment of the water supply; and ordering of supplies on an
ongoing basis (42 CFR 40S.2163(e)).
Staff providing equipment and related services must have current education,
training and experience appropriate to the equipment and services provided.
They must have competence in assessing electrical and plumbing needs at
the patient's home; equipment safety checks and troubleshooting; equipment
setup; specific structural and environmental requirements for safe and effec-
tive use of equipment in the home setting; and communication and teaching
skills.
There should be an established procedure for delivery of equipment and
supplies to the patient's home. This should include the frequency and method
of delivery appropriate to the treatment and the patient's needs, and a method
of delivery designed to assure that equipment and supplies are sanitary and
in working order.
Once equipment is in the patient's home there should be a process to
assure that it is safe, sanitary, and working as intended. If dialyzer reuse is
done there must be an established procedure for cleaning and disinfected
dialyzers. There must be documentation of the manufacturer, model, serial
number, and maintenance instructions of all equipment, and a process for
obtaining timely and accurate information about equipment hazards, defects
and recalls. Qualified individuals should be used to maintain the equipment
according to the manufacturer's standards. The safety and working condition
of the equipment should be monitored regularly, and the results of this
 81

monitoring documented. Emergency maintenance, replacement or backup

equipment must be available when necessary.
Training and education appropriate to the equipment and supplies must
be provided to the patient and assistant in a language or form they can under-
stand by a qualified health care professional. This must include a description
of their responsibilities for the care and use of the equipment and supplies.
Instruction is the responsibility of the training program, and there should be
documentation that the patient and assistant have been effectively trained in
the safe use of the equipment. Only the essentials of equipment maintenance
should be taught, and ongoing maintenance should be the responsibility of
the technical services staff or of the supplier. Twenty four-hour availability
of advice by telephone regarding any equipment problems is needed, and
service or maintenance should be available within no more than two days.
Patients should not be expected to deal with any but the simplest equipment
problems. They must be given written information on when and how they
can contact the facility or supplier for equipment maintenance and repair.
There must be initial and ongoing assessments designed to assure that
equipment and supplies are appropriate for use by the individual patient in their
home environment and are provided according to the orders of the patient's
physician. Selection of the equipment, including selection of the dialyzer,
the dialysis machine for hemodialysis or CCPD, and the appropriate formula
for dialysate concentrate or CAPD or CCPD fluids, is the responsibility of
the patient's physician together with the facility. It is the home dialysis training
unit's responsibility either to deliver and set up the equipment in the home
or to contract for this with a supplier. For all home hemodialysis and CCPD
patients, the home should be visited ahead of time to ensure the availability
of appropriate electrical and water supplies and drainage. If necessary, an
electrician or plumber should be used to make any necessary alterations. The
patient can also be given advice on how to make the best use of the space
available for dialysis and for storage of supplies [24].

Patient survival and quality of life with home dialysis

Data from the United States Renal Data System (USRDS) shows that survival
on dialysis is related to several factors [25]. These include age, sex, race and
the primary cause of the renal disease. In the past, Cox analyses of data from
individual programs have shown no significant differences in survival with the
different modalities of dialysis [26,27]. More recently, USRDS data have been
used to examine one-year survival, adjusted for age, sex and race, in non-
diabetic and in diabetic patients treated by outpatient hemodialysis or CAPD.
While there was no obvious difference in the nondiabetic patients, in diabetic
patients survival at one year in those treated by hemodialysis was 69.6
percent, and in those treated by CAPD was 65.7 percent. However, the results
in the nondiabetic patients were not adjusted for race, nondiabetic specific
diagnosis or five year age difference. As the CAPD patients were younger,
82

with a median age of 48 years versus 52 years in the hemodialysis patients,

they would have been expected to have a lower mortality. Clearly, more
detailed comparisons, adjusted for other co-morbid factors, are needed before
it can be concluded that the one-year mortality is higher with CAPD than
with out-patient hemodialysis. This is under investigation. Unfortunately, com-
parable data are not available for home hemodialysis patients.
Two major studies have looked at quality of life in dialysis and transplant
patients. One examined quality of life and rehabilitation in 859 patients
randomly selected from 11 dialysis and transplant centers across the United
States [28]. The second studied a similar number of patients from dialysis units
in one Northeastern geographic area [29]. Both studies found the best quality
of life and rehabilitation in patients with a successful transplant. Among
dialysis patients, even when adjustment was made for medical and demographic
differences between the various patient popUlations, home hemodialysis
patients had a better quality of life and were more frequently rehabilitated than
patients treated by in-center hemodialysis. Patients on CAPD were interme-
diate between the two hemodialysis groups. These studies clearly showed
one of the major advantages of home dialysis.

Conclusions
Home dialysis, whether home hemodialysis or one of the modalities of
peritoneal dialysis, is an important option that should be available to all suitable
patients. Unfortunately, this is not always the case. However, it is just as
important there be active QA assessments of the care provided these patients
as for patients dialyzing in a dialysis unit. The problems are increased by
the logistical difficulties associated with dialysis remote from the center.
Nevertheless, the results of treatment with home dialysis are reported to be
as good or better than with center dialysis. Careful QA should help to maintain
this record. This is important as we move into what ReIman has called the
"Era of Assessment and Accountability" in medical care [30].

References
1. Merrill JP, Schupak E, Cameron E, Hampers CL: Hemodialysis in the home. lAMA 190:
468-470, 1964.
2. Baillod RA, Comty C, Ilahi M, Konotey-Ahu1u FID, Sevitt L, Shaldon S: Overnight
haemodialysis in the home. Proc Eur Dial Transpl Assoc 2: 99-103, 1965.
3. Tenckhoff H, Shilipetar G, van Paasschen WJ, Swanson E: A home peritoneal dialysate
delivery system. Trans ASAIO IS: 103-107, 1969.
4. Popovich RP, Moncrieff JW, Dechard JB, Bomar JB, Pyle WK: The definition of a novel
portable/wearable equilibrium peritoneal dialysis technique. Abstracts, ASAIO 5: 65, 1976.
5. Diaz-Buxo JA: Continuous cyclic peritoneal dialysis. In Nolph KD (ed) Peritoneal Dialysis
3rd ed. (pp 169-173). Dordrecht, The Netherlands: Kluwer Academic Publishers, 1989.
6. Health Care Financing: Research Report: End Stage Renal Disease, 1990. Department of
Health and Human Services, Health Care Financing Administration, Bureau of Data
Management and Strategy; Office of Research and Demonstration, Baltimore, MD, 1992,
HCFA Pub. No. 03328.
 83

7. Hull AR, Parker TF: Introduction and summary: Proceedings from the morbidity,
mortality and prescription of dialysis symposium, Dallas, TX, September 15 to 17, 1989.
Am J Kidney Dis 15: 375-383, 1989.
8. Held PJ, Brunner F, Odaka M, Garcia JR, Port FK, Gaylin OS: Five-year survival for
end-stage renal disease patients in the United States, Europe, and Japan, 1982 to 1987.
Am J Kidney Dis 15: 451-457, 1990.
9. Held PJ, Blagg CR, Liska OW, Port FK, Hakim R, Levin N: The dose of hemodialysis
according to dialysis prescription in Europe and the U.S., 1986-87. Kidney Int 42: Suppl
38, 516-521, 1992.
10. Brook RH, Kosecoff JB: Commentary: competition and quality. Health Affairs, 7: 150-161,
1988.
11. Donabedian A: Evaluating the quality of medical care. Millbank Memorial Fund Quarterly
44 (Suppl): 166-206, 1966.
12. Hopkins A: Measuring the Quality of Care. Royal College of Physicians of London, 1990.
13. Lohr K, Brook R: Quality assurance in medicine. American Behavioral Scientist 27, no 5,
1984.
14. The 1991 Joint Commission. AMHC Accreditation Manual for Home Care. Joint
Commission on Accreditation of Healthcare Organizations, Oakbrook Terrace, IL, 1991.
15. Code of Federal Regulations: Public Health: 42, Parts 400-429. Revised as of October 1,
1990, pp 137-151.
16. 1992 National Listing of Medicare Providers Furnishing Kidney Dialysis and Transplant
Services. Government Printing Office, Washington DC, 1992.
17. Roberts CM: Home dialysis in the haemophilic paraplegic patient. Proc Eur Dial Transpl
Nurses Assoc 7: 62-66, 1979.
18. Roberts CM, Davis B, Pavitt Let al.: Home dialysis training in a blind patient and a deaf
patient. Proc Eur Dial Transpl Nurses Assoc 1: 41, 1973.
19. Baillod RA, Moorhead JF: Review of ten years' home dialysis. Proc Eur Dial Transpl Assoc
11: 68-75, 1974.
20. AAMI Recommended Practice: Reuse of Hemodialyzers. Association for the Advancement
of Medical Instrumentation, Arlington VA, 1986.
21. Eschbach JW, Seymour M, Potts A, Clark M, Blagg CR: A hemodialysis orientation unit.
Nephron 33: 106-110, 1983.
22. Stinson GW, Clark MF, Sawyer TK, Blagg CR: Home hemodialysis training in three
weeks. Trans ASAIO 18: 66-69, 1972.
23. Blagg CR, Liedtke RJ, Batjer JD, Racoosin B, Sawyer TK, Wict MJ, Lawson L, Wilkens
K: Serum albumin concentration-related HCFA quality assurance criterion is method-depen-
dent: revision is necessary. Am J Kidney Dis 21: 138-144, 1993.
24. Sawyer TK, Blagg CR: Home preparation and installation for home hemodialysis. In
Nissenson AR, Fine RN (eds) Dialysis Therapy, 2nd ed, Philadelphia: Hanley & Belfus,
Inc., pp. 52-54, 1993.
25. Held PJ, Garcia JR, Blagg CR, Wolfe RA, Eggers PW, Port FK, Fitzsimmons SC: Trends
in mortality rates among dialysis and transplant patients in the U.S. Abstracts ASN, 51a,
1989,
26. Vollmer WM, Wahl P, Blagg CR: Survival with dialysis and transplantation in patients
with end-stage renal disease. N Engl J Med 308: 1553-1558, 1983.
27. Burton PR, Walls J: Selection-adjusted comparison of life-expectancy of patients on
continuous ambulatory peritoneal dialysis, haemodialysis, and renal transplantation. Lancet
1: 1115-1119, 1987.
28. Evans RW, Manninen DL, Garrison LP, Hart G, Blagg CR, Gutman RA, Hull AR, Lowrie
EG: The quality of life of patients with end-stage renal disease. N Engl J Med 312: 553-559,
1985.
29. Bremer BA, McCauley CR, Rona RM, Johnson JP: Quality of life in end-stage renal disease:
a reexamination. Am J Kidney Dis 13: 200-209, 1989.
30. Reiman AS: Assessment and accountability: the third revolution. N Engl J Med 319:
1220-1222, 1988.
CHAPTER 6

Water treatment for hemodialysis

PRAKASH KESHAVIAH

Introduction

According to an Italian proverb "Aqua torbido non lava" i.e. "Dirty water
does not wash clean". This can be most aptly applied to the water used for
hemodialysis. If the water used for hemodialysis is "dirty", the blood of the
hemodialysis patient will not be "washed clean". The hemodialysis patient
is exposed to more water in one year than the normal population in 20 years.
Further, the hemodialysis patient's blood is exposed to this large quantity of
water across a thin non-selective membrane that has none of the "wisdom"
of the gastrointestinal tract of a normal individual. Also, the hemodialysis
patient's ability to excrete harmful toxins from such exposure is limited by
compromised renal function. The combination of large volume of exposure,
non-selective transport across the dialyzer membrane, and compromised renal
function make for a situation that is potentially hazardous to the health and
well-being of the patient on hemodialysis. Unless the water used for hemodial-
ysis is analyzed periodically and subjected to appropriate treatment processes
whose efficacy is monitored regularly, there may be serious risks to patient
well-being that have been well documented in the dialysis literature.

Contaminants with documented toxicity in hemodialysis

The risks and hazards associated with inadequately treated water are sum-
marized in Table I along with the lowest contaminant levels at which toxicity
has been observed. It should be noted that many of the contaminants asso-
ciated with toxicity in the hemodialysis setting are not regulated by the
Safe Drinking Water Act of 1974 [1]. Also, even for contaminants that are
regulated by the Safe Drinking Water Act, the levels at which toxicity is
observed in hemodialysis are lower than the safe limits established for drinking
water.
Aluminum. Aluminum may not be naturally occurring in the source water
but may be added to the water by the municipal supplier as alum (aluminum

L. W. Henderson and R.S. Thuma (eds.J, Quality Assurance in Dialysis, 85-107.

© 1994 Kluwer Academic Publishers.
86

Table I. Water contaminants associated with toxicity in the hemodialysis setting.

Contaminants Lowest concentration associated with toxicity (mgll)

Aluminum 0.06
Copper 0.49
Zinc 0.2
Calcium/Magnesium 88 (Ca++)
Sodium 300
Fluoride 1.0
Nitrate 21 (as N)
Sulphate 200
Chloramines 0.25
Microbiological Contaminants

sulphate) to treat water with a high content of suspended colloids. Exposure

of the hemodialysis patient to aluminum levels greater than 0.06 mg/l has been
associated with a progressive syndrome of neurological deterioration and
encephalopathy [2-5]. This syndrome, called dialysis dementia, is often fatal
and is characterized by personality changes, reduced short term memory, speech
disturbances, muscle spasms, seizures, and dementia. Bone diseases such as
osteomalacia and osteodystrophy have also been associated with a prolonged
exposure to high levels of aluminum [6, 7]. A type of anemia called micro-
cytic hypochromic anemia may also result from aluminum exposure [8, 9].
Patients on hemodialysis may be exposed to aluminum from two sources -
the water used for preparing dialysate and from ingestion of aluminum-
containing phosphate binders. Water as the source of exposure has been well
established by epidemiological studies [10].
Copper. Copper may be leached from the water distributing system or from
the dialysate delivery system, by acidic water causing chills, nausea and
headache, liver damage, and fatal hemolysis [11-13]. Water may become acidic
if a dual bed deionizer is in use with exhaustion of the anionic resin bed.
Zinc. Zinc may be leached into the water from galvanized iron components
used in the water distribution system or the dialysate delivery system. The toxic
effects of zinc exposure include nausea, vomiting, fever, and anemia [14].
Calcium or magnesium. Calcium or magnesium levels are elevated in water
supplies that are hard. If the water is not softened, exposure to high levels
of calcium or magnesium may result in a syndrome characterized by nausea,
vomiting, muscular weakness, skin flushing, and hypertension or hypoten-
sion [15-17].
Sodium. Sodium levels may be naturally elevated in the water supply or
may be the result of softening extremely hard water. Also, softener malfunc-
tion may result in high sodium levels [18]. Hypernatremia, increased thirst,
and excessive water intake are consequences of a high sodium level in the
water supply.
Fluoride. Fluoride is often added to the municipal water supply to prevent
 87

dental caries. Even at the Safe Drinking Water level of 1 mg/l, prolonged
hemodialysis exposure may result in bone diseases such as osteomalacia and
osteoporosis [19, 20]. Accidental over-fluoridation of the water by the supplier
has been associated with adverse symptoms and one fatality [21].
Nitrate. Contamination of the water supply by fertilizers or excessive bac-
terial contaminants may result in high nitrate levels with consequences of
methemoglobinemia, cyanosis, hypotension, and nausea [22, 23]. Very high
nitrate levels may also cause hemolysis [23].
Sulphate. High sulphate levels in dialysis water have been associated with
nausea, vomiting, and metabolic acidosis [24].
Chloramines. Chloramines are commonly used bactericidal agents in munic-
ipal water treatment. Because they are powerful oxidants, inadequate removal
of chloramines may result in hemolysis, hemolytic anemia (characterized by
Heinz body formation) and methemoglobinemia [25-29].
Microbiological contaminants. There have been documented clusters of
pyrogen reactions associated with high levels of gram negative bacteria in
the water supply [30-33]. Symptoms noted include shaking chills, fever,
hypotension, headaches, myalgia, nausea and vomiting.

Water quality at the source

The various sources of water may be broadly classified into two major cate-
gories - ground waters and surface waters. In both cases, the water is derived
from the natural water cycle, water evaporating from oceans, rivers, and other
reservoirs into the atmosphere, the vapor condensing and returning to earth
as rain, snow, or hail. Some of this water flows on the earth's surface as streams
and rivers and collects in closed bodies of water such as lakes and ponds.
Precipitated water may also seep into the earth, collecting in underground
reservoirs and aquifers, being classified as ground water.
The quality of surface water depends on the location. In agricultural
locations, surface water may be contaminated by fertilizers and pesticides.
In urban locations, surface water may be contaminated by industrial wastes
and sewage. Unless surface water is appropriately treated for hemodialysis use,
it may pose a grave risk to the well-being of the hemodialysis patient.
Ground water percolates through various layers of soil and its quality may
depend on the geology of that area. Organic contaminants are usually removed
by the percolation of water through the various soil layers. However, inorganic
contaminant levels may increase. Ground water percolated through limestone
is usually hard containing high levels of calcium and magnesium compared
to water percolated through granite.
In 1974, landmark legislation was passed by Congress called the Safe
Drinking Water Act [1] which mandated the surveillance of drinking water
supplies by federal and state governments with the EPA assigned the respon-
sibility of enforcing these regulations. However, these regulations only apply
88

to public water systems defined as "any system, publicly or privately owned,

that serves an average of 25 persons daily or has at least 15 service connec-
tions for at least 60 days of the year" [34]. Even with public water systems
that are subject to EPA regulations, the level of compliance is of the order
of 90% so that approximately 10% of public water systems may be supplying
unsafe water. It should not be assumed, a priori, that the water supply for a
hemodialysis unit meets the EPA standards. It is essential that contact be
established between the staff of the dialysis unit and the municipal water
supplier regarding water quality and treatment practices. There should also
be an on-going dialogue to ensure that changes in water treatment pro-
cedures by the supplier are communicated to the dialysis unit. This
communication is particularly important with water supplies that are fluori-
dated, situations where large amounts of aluminum may be used to flocculate
organic contaminants, and with the use of chlorine and chloramines for
disinfection of the water. Also, if the supply source (ground/surface) changes
during the year or the mix of these sources varies, there may be large seasonal
variations. Appropriate plans need to be implemented at the dialysis unit to
ensure appropriate quality of product water used for dialysis despite large
seasonal variations.
The National Interim Primary Drinking Water Regulations [35] established
by the EPA set maximum contaminant levels for inorganic and organic
contaminants in drinking water. These maximum levels are summarized in
Table II. While water meeting these levels is safe for drinking and for cooking

Table II. National Interim Primary Drinking Water Standards.

Chemical Maximum Level (mg/I)

Inorganic
Arsenic 0.05
Barium 1.00
Cadmium 0.01
Chromium 0.05
Lead 0.05
Mercury 0.002
Nitrate (as N) 10.0
Selenium 0.01
Silver 0.05
Fluoride* 1.4-2.4
Organic
Endrin 0.0002
Lindane 0.004
Methoxychlor 0.1
Toxaphene 0.005
2,4-D 0.1
2, 4, 5 - TP (Sivex) 0.01

* Varies according to annual average air temperature.

 89

purposes, it is not, as previously noted, safe for dialysis because of the larger
volume of exposure, exposure to blood across a thin non-selective membrane,
and compromised renal function. More stringent limits than the EPA stan-
dards must be met, and these have been established by the Association for
the Advancement of Medical Instrumentation [36].

AAMI water quality standards

The contaminants occurring in municipal water supplies may be categorized

into inorganic, organic, radioactive, and microbiological contaminants. The
water quality standard in hemodialysis established by AAMI addresses
inorganic and microbiological contaminants. Organic and radioactive con-
taminants are addressed by the Safe Drinking Water Act and no further limits
have been set by AAMI, as the potential toxicity of organic and radioactive
contaminants in the hemodialysis setting is not well established.
The AAMI Water Quality Standards for hemodialysis were established in
1982 and were based in part on recommendations made in an FDA spon-
sored study of the risks and hazards of hemodialysis [37]. The AAMI standards
have been adopted by the American National Standards Institute. The AAMI
standard is summarized in Table III.
To understand the rationale underlying the AAMI standard, it is con-

Table III. AAMI water quality standard.

Category Chemical Maximum level (mg/I)

Normal constituents of dialysate Calcium 2 (0.1 mEq/l)

Magnesium 4 (0.3 mEq/l)
Potassium 8 (0.2 mEq/l)
Sodium 70 (3.0 mEq/I)
Toxic contaminants regulated Arsenic 0.005
by national interim primary Barium 0.01
drinking water standards Cadmium 0.001
Chromium 0.014
Lead 0.005
Mercury 0.0002
Selenium 0.09
Silver 0.005
Other contaminants with documented Aluminum 0.D1
toxicity in hemodialysis Chloramines 0.10
Free Chlorine 0.5
Copper 0.10
Fluoride 0.20
Nitrate (as N) 2.0
Sulphate 100
Zinc 0.10
90

venient to divide the inorganic contaminants into three categories: normal

constituents of dialysate, toxic contaminants regulated by National Interim
Primary Drinking Water Standards, and other contaminants with documented
toxicity in hemodialysis.
The maximum acceptable level for substances normally included in dialysate
was established based on clinically acceptable variations of these substances
in dialysate. As an example, the AAMI has set a maximum level of 2 mg/l
(0.1 mEq/l) for calcium. Variations of this magnitude in the calcium content
of dialysate (typically 2.5-4.0 mEq/l) are both hard to measure and unlikely
to have adverse patient consequences. For contaminants regulated by the
Drinking Water Standards, a maximum level was chosen that was one-tenth
of the drinking water level, unless the limits of detection or the no-transfer
level for that contaminant were higher than the one-tenth level. The no-transfer
level is the contaminant level in dialysate at which no diffusive transfer
occurs from dialysate to blood. For contaminants with documented toxicity
in hemodialysis, the AAMI level was set at the lowest level at which toxicity
had been documented.
The AAMI standard of 200 cfu/ml for bacterial contamination was based
on studies conducted by the Centers for Disease Control [38, 39] which showed
that at levels higher than 200 cfu/ml, there would be amplification of
bacterial contamination in dialysate with the potential for pyrogenic reac-
tions. In addition to the 200 cfu/ml level for water used to prepare dialysate,
AAMI has also set limits of 2000 cfu/ml for final dialysate and 1 ng/ml
(Limulus amebocyte lysate (LAL) assay) for the bacterial lipopolysaccharide
level in water used for the reprocessing of dialyzers for multiple use, with a
requirement that this water should have a bacterial count of less than 200 cfu/ml
and should be filtered through a nominal 5 11m filter.
With high flux hemodialysis, back filtration from dialysate to blood occurs
in the distal section of the dialyzer. The potential for biologically active
endotoxin fragments or dissociated endotoxin to cross over from dialysate to
blood is enhanced in the presence of back filtration [40, 41]. According to
the interleukin hypothesis [42, 43], even small amounts of endotoxin, unde-
tected by the LAL test, may stimulate cytokine production by monocytes. These
cytokines have been associated with a wide variety of adverse effects. Whether
endotoxin-induced cytokine production occurs with high flux dialysis is a
subject of lively debate, but there are investigators who are recommending
the use of filtered endotoxin-free dialysate for high flux dialysis.
The concentration of chlorinated hydrocarbons (pesticides, chlorophenoxys),
herbicides, and radioactive substances are limited by the Drinking Water
Standards. However, no further reduction of levels of these contaminants for
hemodialysis applications has been recommended as the potential risks from
these contaminants are not known. The combined use of activated carbon
and reverse osmosis to treat water for hemodialysis use provides some measure
of protection against these contaminants. Activated carbon is effective in
adsorbing organic contaminants in a molecular weight range of 60-300 daltons
 91

whereas reverse osmosis is capable of removing larger organic contaminants

(> 200-300 daltons). Treatment processes that are effective for non-radio-
active contaminants are expected to be effective for removal of their
radioactive isotopes.

Achieving the water quality standards

A variety of water treatment technologies are available that need to be used

alone or in combination, depending on the source water quality, to achieve
the water quality standard for hemodialysis. These various methods are listed
in Table IV along with the type of contaminants that they are capable of
removing. We will describe here some of the commonly used treatment
methodologies along with a description of the operating principles, guide-
lines for selecting particular treatment methods, and how these various methods
can be integrated into a suitable water treatment system.

Table IV. Water treatment processes used in hemodialysis.

Process Contaminants removed

Filtration Particulates, bacteria, viruses, endotoxins

Activated carbon adsorption Small organics, chlorine, chloramines
Reverse osmosis Large organics (-300 daltons), ionic species, & bacteria
Softening Calcium and magnesium
Deionization Ionic species (cationic and anionic)

Filtration

Filters playa key role in water treatment with a variety of uses from removal
of large particulates to removal of bacteria and viruses. There are basically two
filtration mechanisms - depth filtration and surface filtration. Depth filtra-
tion is typically used for removal of coarse particulate material from water
by allowing the water to permeate the filter matrix and by trapping the
particulates mechanically in channels that are smaller than the particulate or
by adsorptive capture. Some depth filters (e.g. bed filters) consist of multiple
layers of filter media, each layer capable of retaining progressively smaller
particles. The largest particles are removed by the first layers and the smallest
in the final layers of the bed. Bed filters are suited for removing particles up
to 10 j.Lm in size. Surface filters exclude solute at the surface of the filter
primarily by size exclusion, although charge and hydrophobic and hydrophilic
interactions may also be involved. Surface filters are suited to removal of
both dissolved and suspended solutes.
Cartridge filters typically consist of a filter medium with a central drainage
core. Cartridge filters may employ both depth and surface filtration. Cartridges
92

of cellulosic materials and synthetic polymers are available with solid, mesh,
pleated, wound and woven configurations. The cartridge is contained in a
filter having both inlet and outlet ports. Cartridge filters capable of removing
particles as small as 1 11m are available.
Membrane filters employ surface filtration and are usually in the con-
figuration of a thin sheet of porous material with the filter being capable of
retaining solutes that are larger than the pores. Water passes freely through
the membrane pores. As the pores of the filter may not all be of the same
size, the size cutoff of the membrane is not sharply defined and is usually
quoted as the size at which 90% retention is achieved. Cellulosic and poly-
meric membranes are used in water treatment applications. Membrane filters
can be classified as microporous filters, ultrafiltration, and reverse osmosis
filters. Microporous filters are suited to removing particulates, bacteria, and
viruses up to 0.1 11m in size. Ultrafilters are capable of removing endotoxins,
proteins, and solutes as small as 0.001 11m (-200-300 daltons). Reverse
osmosis filters exclude larger organic solutes (-200 daltons) and ionic species
as small as 0.001 11m. Reverse osmosis filters are discussed in greater detail
below.
Filters may be operated in a dead end configuration or with cross flow as
shown in Figure 1. Depth filters are usually used in the dead end configura-
tion whereas membrane filters are commonly used with cross flow. In cross
flow, the tangential flow sweeps away solutes from the surface of the
membrane and reduces filter clogging.
Filters should be sized on the basis of the desired flow rate of water through
the filter. Filters have a finite capacity and need to be replaced or regener-

A. Dead end Filtration B. Cross-flow Filtration

Feed Stream Feed Stream Reject Scream

Permeate Scream
1
Permeate Stream

Fig. 1. Schematic representation of dead end and cross-flow filtration.

 93

ated once this capacity is exceeded. The pressure drop across the filter is
monitored and used to determine when the filtration capacity has been reached.
If used beyond this capacity, pressure drops will increase, flow rates decrease,
and filter performance (solute retention) will deteriorate. Depth filters may
be regenerated by back washing, and the use of regenerative chemicals.
Cartridge and membrane filters are usually replaced when the specified
maximum pressure drop is exceeded. In a water treatment system, filters may
become potential sites for bacterial colonization and proliferation. This
potential should be considered in choosing a filter and determining its location
in the water treatment system. Water treatment systems may also require
frequent disinfection or filter replacement to control bacterial proliferation.

Activated carbon adsorption

Activated carbon adsorption is suited to the removal of a wide variety of

organic contaminants especially of a size too small to be removed by reverse
osmosis filters. Activated carbon has an extremely porous internal structure
with a very high internal surface area. This porous structure is achieved
by pyrolysis of materials such as coal, peat, wood, and bone in regulated
atmospheres. Porous carbon granules thus produced are contained in a bed,
water percolating through the granular bed, solutes being absorbed within
the internal porous structure of the carbon granules. Once the adsorption
capacity of the bed is exceeded, solute breakthrough occurs and the maximum
acceptable outlet concentration of solute is exceeded. The carbon bed then
needs to be replaced. Because of the porous structure of the granules, there
is a tendency for fragmentation of the granules by mechanical attrition, and
the particles or 'fines' so released need to be trapped by a suitable downstream
filter to prevent them from blocking orifices and clogging equipment down-
stream.
In the hemodialysis application, activated carbon is used to remove con-
taminants such as chlorine, chloramines, hypochlorites, and chloroform.
Chloramines are replacing chlorine as bactericidal agents in municipal water
treatment because free chlorine has a tendency to combine with organic
chemicals to form carcinogenic compounds called trihalomethanes. Trihalo-
methanes are not found with choramine usage. However, chloramines are
powerful oxidants and are toxic in the hemodialysis setting causing hemolysis,
hemolytic anemia, and methemoglobinemia.
In hemodialysis, the breakthrough of chloramine or total chlorine (free
chlorine + chloramines) should be used as a criterion for detecting exhaus-
tion of the carbon bed. The monitoring of free chlorine alone may result in
an overestimation of the remaining adsorptive capacity. The life of a carbon
bed can be extended by operating at low water flow rates. Such low flow
rates utilize the adsorptive capacity of the bed more completely. Often two
carbon beds in series are used in a water treatment system. When the upstream
94

bed is exhausted, the second bed is moved upstream and a new bed added
downstream. This arrangement provides an adequate margin of safety.
Because of the high surface area of carbon beds, they provide a good site
for bacterial proliferation, the organics adsorbed by the carbon serving as nutri-
ents for the colonized bacteria. Also the removal of chlorine and chloramines
by the carbon exacerbates the proliferation of bacteria. Suitable water treat-
ment equipment must be used downstream to remove bacteria and endotoxins.

Reverse osmosis

Reverse osmosis is a membrane separation process that forces flow of solvent

across a semipermeable membrane under hydrostatic pressure. As shown in
Figure 2, when a semipermeable membrane separates a dilute solution from
a concentrated solution, there is a tendency for solvent flow from the dilute
solution to the concentrated solution across the osmotic gradient. In order to
impede the flow of solvent across the membrane, a pressure equal to the
osmotic pressure needs to be applied to the concentrated solution side of the
membrane. If the pressure applied to the concentrated solution side of a
membrane exceeds the osmotic pressure, there will be a reverse flow of solvent
i.e. flow from the concentrated to the diluted side. This is the rationale behind
the name 'reverse osmosis' for this water treatment process.
Reverse osmosis membranes are operated in a cross flow configuration. The

Hydrostatic
Pressure

:water::':':':'il~~

Semipermeable +
Semipermeable
Membrane Membrane

OSMOSIS REVERSE OSMOSIS

Fig. 2. Illustration of the principle of reverse osmosis.
 95

permeate is called product water and the cross flow containing the concen-
trated contaminants is called the reject stream. Typically the product water
to feed water ratio is of the order of 25 to 50% for hemodialysis applica-
tions.
There are 3 major types of reverse osmosis membranes: cellulosic, fully
aromatic polyamide, and thin film composite membranes. Reverse osmosis
as a water treatment process was first demonstrated in the 1950's with
cellulosic membranes (cellulose acetate). These membranes are asymmetric
having a thin, dense layer that provides the solute rejection capability and a
thick porous substructure that provides structural support and strength. It is
because of this substructure that the membrane can withstand high hydro-
static pressures. However, under very high pressures, compaction occurs with
the thin dense layer merging into the porous structure with a reduction in
product water flux. Compaction is accelerated under high temperatures.
Cellulose acetate membranes can only be used with a limited pH range of
4-8 because of their susceptibility to hydrolysis. They are also susceptible
to bacterial degradation. Cellulose acetate membranes have a high water
permeability but poor solute rejection. Aromatic polyamide membranes are
also asymmetric but are resistant to hydrolysis and bacterial degradation. They
can withstand higher temperatures and a wider pH range (4-11), but at
the extremes of this range they are more prone to irreversible membrane
degradation. They have better solute rejection characteristics than cellulosic
membranes but are extremely sensitive to oxidant damage. Water containing
chlorine or chloramine must be treated with activated carbon before being
fed to reverse osmosis modules with polyamide membranes. Thin film com-
posites are made by casting a thin, dense film onto a porous substrate, so
that the materials and manufacturing process for the two layers can be different
and can be optimized for providing the best characteristics of water flux, solute
rejection, pH and temperature tolerance and compaction resistance. The sup-
porting structure is usually polysulphone and the thin dense layer materials
include aromatic polyamide, alkylarylpolymer/polyamide, and polyfurane
cyanate. These membranes are more stable than the aromatic polyamide mem-
branes to oxidant exposure, but exposure to oxidants must still be limited.
The reverse osmosis module contains the reverse osmosis membrane with
provisions for access to feed, product, and reject streams. There are 2 common
module configurations used in hemodialysis - spiral wound modules and
hollow fiber modules. In the spiral wound configuration, a membrane fabric
sandwich is created by attaching two layers of membrane back to back to a
woven fabric like nylon or dacron. This membrane-fabric sandwich is wound
around a central perforated hollow tube in a jelly roll configuration with plastic
mesh used to separate adjacent layers (Fig. 3). This configuration is not
unlike the coil dialyzer configuration used in the early days of dialysis. Product
water flows through the membrane into the woven fabric and from the fabric
into the central hollow core. The plastic mesh controls the feed stream channel
height and improves mixing of the feed stream with 'quasi' turbulence. The
96

o
Perforated Product
Water Tube

Fig. 3. Representation of the spiral wound reverse osmosis configuration.

spiral wound module is contained in a pressure vessel designed for high

operating pressures. The spiral wound configuration provides a high membrane
packing density and is low in manufacturing cost. However, it provides
stagnation areas in the plastic mesh and in the dead spaces between the module
and the pressure vessel and may be hard to clean and disinfect completely.
In the hollow fiber configuration, the membrane is formed into fine hollow
fibers with inside diameters of 80-250 11m. The fiber bundle is potted in a
polyurethane or epoxy tube sheet and may be a straight or U-shaped bundle
(Fig. 4). The feed water stream flows on the outside of the fiber, and the
product stream through the fiber lumen. The fiber bundle is contained in an
appropriate pressure vessel. No supporting structures are required, making this
configuration even more compact than the spiral wound configuration.
However, the hollow fibers are susceptible to fouling and plugging and may
be hard to clean so that appropriate pretreatment is required upstream of the
hollow fiber reverse osmosis module. The water flux of the hollow fibers is
lower than that of flat sheet membranes, but because of the higher packing
density, more membrane surface area can be accomodated in the same volume
as the spiral wound design.
Reverse osmosis membranes have a broad spectrum of solute rejection from
particulates, bacteria and viruses, to larger organics and small dissolved inor-
ganic solutes. Ionic species are rejected more readily then non-charged species
and the rejection increases with valence. pH may influence solute rejection.
Typical rejection rate for inorganic contaminants are shown in Table V. While
larger organics (200-300 daltons) are readily rejected by reverse osmosis
 97

Close"up of hollow fibers

Oose-up of hollow fibers

Fig. 4. Representation of the hollow fiber reverse osmosis configuration.

Table V. Typical reverse osmosis membrane rejection of inorganic contaminants.

Contaminant % rejection*

Cations Ca++ 98
MgH 98
FeH , Fe+++ 98
Na+ 95
K+ 95
Anions HC0 3- 85-95
CI- 95
S04= 98
N0 3- 85-90

* Cellulose acetate membrane with 1500 mgtl TDS feed water, 75% recovery, 400 psi

membranes, smaller, non-polar, dissolved organics like methanol, ethanol,

and ethylene glycol are poorly rejected. While bacteria and viruses are rejected
by most reverse osmosis membranes, because of membrane defects and minute
seal leaks, reverse osmosis devices generally cannot be relied upon to produce
sterile, endotoxin-free water.
Performance of reverse osmosis devices is assessed by the rejection of ionic
contaminants measured as the ratio of product water to feed water conduc-
tivity. Product water flux rates depend not only on membrane characteristics,
but also on operating pressures (typically 200-400 psi), temperature
(-25°C), and feed water quality (ionic concentration). As the total dissolved
98

solids content in the feed water increases, product water recovery (ratio of
product water to feed water flow) decreases. In addition to conductivity
monitoring of rejection, product water and reject stream flow meters, pressure
gauges (pre and post pump as well as product and reject stream pressures)
and low and high pressure switches (loss of water supply and flow path obstruc-
tions) are used to monitor the performance of the reverse osmosis unit. Means
for safe cleaning and disinfection with an interlock switch to safeguard against
accidental institution of these procedures contribute to good unit design. A
periodic high velocity auto-flush feature for removing foulants from the
membrane surface is also a convenient feature.
When water flux or solute rejection begin to diminish, it may be time to
clean the reverse osmosis membrane. The module can be flushed from the feed
to the reject side at a low feed pressure so as to have almost no product flow.
In addition chemical cleaning with cleaning agents such as hydrochloric
acid, citric acid, sodium hydroxide, sodium EDTA, and detergents may be
necessary to restore membrane performance. The chemical stability of a
membrane, type of foul ants in the feed water and materials of construction
dictate the choice of cleaning agent. The manufacturer of the reverse osmosis
device has the responsibility for providing complete instructions for cleaning
and restoring membrane function. Though the capital cost of reverse osmosis
modules is high, with appropriate water pretreatment, suitable operating
conditions, and regular maintenance, cleaning and disinfection, a module life
of several years is feasible making reverse osmosis a highly effective water
treatment process from the solute rejection and cost-effectiveness points of
view.

Ion exchange

In this water treatment process, water flows through a resin bed, the resin
having ion-exchange properties, i.e. able to exchange certain ions on the
resin for certain ions in the feed water. Softeners and deionizers are both
examples of ion exchange devices. In a softener, the resin exchanges sodium
ions for calcium and magnesium ions in the hard water (Fig. 5). Deionizers
may contain cationic or anionic resins or a mixture of the two. The cationic
exchange resin exchanges hydrogen ions for cationic species in the feed
water. Anionic exchange resins exchange hydrogen ions for anionic species
in the feed water. In a mixed bed deionizer, the hydrogen and hydroxyl ions
released from the resin combine to form neutral water (Fig. 6). Ion-exchange
resins are usually in a bead configuration contained in a cylindrical column
(Fig. 7). These resins have a finite ion-exchange capacity and when exhausted,
they need to be regenerated. With softener resins, regeneration is accomplished
using a saturated sodium chloride solution (brine). Softener resin regenera-
tion is usually done on site at periodic intervals in an automated fashion under
timer control. A brine tank containing sodium chloride pellets or crystals is
 99

.. 8 8
;x;>...
'f' ~ 'I' ~

Na+ Na+
~

l;;. Na~ ~
Na+ J. ~
Na+
;y 8
.. 98
Na+
~ ~~ ~
Na+ Na+
~ .+'

. 98
~ ~ Ca++
~
Na+ ~ ~
~A.
~
~'"

Fig. 5. Representation of the ion-exchange process used for softening hard water.

Feed Water

Cationic Resin Anionic Resin

Fig. 6. Principle of mixed bed deionizers.

100

rr=::::;:==:=" Resistivity monitor

with low alarm at
I----...-------l 1 megohm-cm

Outlet Inlet
~--~--

8ed contains mixture

of anion & cation resins

Fig. 7. Physical configuration of resin bed deionizer.

located adjacent to the resin bed and at appropriate regeneration times, the
saturated brine solution is drawn from the brine tank by a venturi mecha-
nism. The volume of brine drawn depends on the regeneration frequency and
exchange capacity of the resin bed. The frequency of regeneration in turn
depends upon the volume of feed water processed and the hardness of the
feed water.
With deionizer resins, it is more common to have regeneration performed
off-site in a central facility because caustic acids (cationic resins) and alkalis
(anionic resins) are used for regeneration. The central facility will usually pick
up the exhausted exchange tank, leaving a regenerated tank in its place. For
hemodialysis applications, it should be ensured that the central regeneration
facility uses only food-grade chemicals and does not mix resins from indus-
trial users with hemodialysis resins. In industrial applications, deionizers may
be used for recovery of heavy metals or other hazardous contaminants which
may pose a serious threat to the hemodialysis patient. Similarly, for soft-
 101

eners, it is recommended that a purer form of sodium chloride than crude

rock salt be used for regeneration.
In the hemodialysis setting, a common configuration for deionizers is to
have two mixed bed deionizer tanks in service. As the first one gets exhausted,
the second downstream tank is moved up and a new tank placed in its stead.
The performance of deionizers is usually monitored using a resistivity meter
in units of ohm-cm. The minimum resistivity for hemodialysis water is
1 million ohm-cm or 1 meg ohm-cm. When the resistivity falls below this
value, the deionizer tank needs to be regenerated. As resistivity depends on
temperature, the resistivity meter needs to be temperature compensated. Instead
of a meter, a monitor light may be provided.
Because of the large resin surface area and bed configuration with stagnant
areas, deionizers are susceptible to bacterial colonization. Water treatment
processes such as ultrafilters or ultraviolet irradiators may be required down-
stream to limit bacterial levels in the final water used for hemodialysis.
In situations where dual bed deionizers are used, it must be ensured that
the cationic and anionic beds are appropriately sized so that they reach
exhaustion together. If the anionic resin gets exhausted first, the cationic
resin bed will continue to release hydrogen ions resulting in acidic product
water. As stated earlier, acidic water may leach copper or other toxic mate-
rials from the water distribution system or from the dialysis equipment creating
a potentially hazardous situation. Deionizer capacity is usually rated in terms
of grains of total dissolved solids as calcium carbonate. This term includes
all dissolved ionized species. In the case of softeners, the hardness of the water
is also measured as grains of calcium carbonate and in cases where the water
is extremely hard, allowance should be made for the increased sodium content
of the product water following softening. For example, if reverse osmosis is
used downstream, the increased sodium burden will need to be removed by
an appropriately sized reverse osmosis module.

System design

In the previous section, we have described briefly the major water treatment
processes used in hemodialysis applications. A water treatment system may
combine some or all of these processes depending on feed water quality,
seasonal variations, product water flow rates required, and economic factors.
It is beyond the scope of this chapter to get into the details of system design.
However, some broad guidelines for selecting water treatment process com-
binations will be delineated along with some requirements for the water
distribution system used to distribute the product water.
The steps involved in the design of the system area:
1. Determine product water requirements (flow rates, usage factors, appli-
cation (e.g. reuse vs dialysate water).
2. Define produce water quality (e.g. AAMI standards)
102

3. Assess feed water quality including seasonal variations and worst case
scenarios.
4. Determine the concentration reduction ratio for various contaminants for
various water treatment processes based on manufacturers' specifications.
5. Select combinations of water treatment processes taking into considera-
tion economics, product flow rates, and impact of various combinations
of pretreatment requirements and final bacteriological quality.
We will consider an illustrative case. The use of reverse osmosis will yield
a concentration reduction ratio of 0.1 for most inorganic contaminants but is
not as effective in removing small organics, free chlorine, and chloramines.
Further, oxidants like free chlorine may damage aromatic reverse osmosis
membranes. The use of activated carbon may, therefore, be a necessary pre-
treatment to remove free chlorine and chloramines as well as small organics.
As carbon filters may release fines, a sediment filter may be necessary down-
stream of the carbon filter. Also carbon filters and sediment filters may promote
bacterial proliferation. The reverse osmosis membrane may be quite capable
of handling the bacterial burden imposed by the use of activated carbon and
sediment filters. If either because of product flow requirements or economics,
this approach to limiting bacterial contaminants is not feasible, an ultrafilter
may also be required. Further, if the level of an inorganic contaminant like
aluminum is so high as to require a concentration reduction ratio of better
than 0.1, a deionizer may be necessary. From a bacteriological point of view,
locating a deionizer upstream of the reverse osmosis device is preferred.
However, economic reasons may dictate that using deionization to 'polish'
the reverse osmosis product water is the preferred approach. In such cases,
ultrafiltration may be necessary as the last step for maintaining bacteriolog-
ical quality of the product water. If usage requirements are such that the reverse
osmosis device is incapable of satisfying peak product water requirements, a
holding tank may be necessary to meet peak flow requirements, the tank
being filled during periods of lower product flow demand. In such situations,
the bacteriological quality of the product water may suffer because of stagnant
conditions in the holding tank and exposure to the atmosphere. The holding
tank may need to be designed to reduce such exposure with a recirculation
loop through an ultraviolet irradiator to limit bacteriological quality.
As the brief example demonstrates, many factors dictate the design of the
total system and many of these factors may have contradictory requirements
and consequences. Skill and experience with juggling these varying factors
is therefore necessary in designing a safe, reliable, and cost effective system.

Distribution system requirements

The product water distribution system consists of piping and associated fittings
such as valves, pressure and flow gauges, pumps, pressure regulators and seals.
Care must be exercised in the choice of these various components to ensure
 103

that the product water quality does not deteriorate during distribution and
that various product water locations can be served at the desired flow rates.
Materials of contamination such as brass, copper, aluminum, and zinc should
be avoided because they are contaminants of known toxicity in hemodialysis.
The piping should be made of more inert materials such as polyvinyl chloride
(PVC), non-pigmented polypropylene, stainless steel, or glass. PVC is widely
available and inexpensive. However, according to some authorities PVC may
not be acceptable as the surface of PVC is considered to be conducive to
bacterial growth. Not only should the materials of construction be relatively
inert, they should also not degenerate with exposure to commonly used dis-
infectants such as bleach. If PVC is used, it should be Type 1 (non-plasticized)
and meet the requirements of the National Sanitation Foundation for potable
water. If solvent welding is used for joints, enough curing time should be
allowed before use, and vigorous flushing of the distribution system to remove
residual solvent is recommended.
In designing the distribution system, care should be taken to ensure that
stagnant areas are avoided because such areas promote bacterial prolifera-
tion. Also, the system should be designed to ensure that all parts of the
system are exposed to cleaning agents and disinfectants where necessary.
The system should be designed to promote high flow velocity, avoiding dead
ends, and long, multiple branching layouts.

Water treatment and quality assurance

An obvious area of quality assurance with regard to water treatment is the

monitoring of final product water quality. While we will also consider this
aspect of quality assurance, it is important to point out that the role of quality
assurance in water treatment begins almost as soon as the decision is made
to establish a dialysis unit. At this early planning stage, it is critical that feed
water quality be evaluated and communications established with the munic-
ipal water supplier regarding source of the water, seasonal variations in water
quality at the source, seasonal variations in the blending of water from dif-
ferent sources at the municipal water treatment facility, and details of municipal
water treatment practices and seasonal variations in these practices to cope
with source quality fluctuations. Armed with this information, one can begin
to design the appropriate water treatment system at the hemodialysis unit.
While the area of system design is a specialized one requiring the appro-
priate knowledge and experience, the clinical staff must at least be well
informed consumers who can study the proposals of various water treatment
system vendors to ensure that an appropriate system is being proposed with
appropriate safety margins and maintenance requirements that are within the
technical abilities at the hemodialysis unit and are not so labor intensive or
expensive as to adversely affect compliance with maintenance requirements.
If an alternate design is available that has less stringent requirements, it may
104

be preferred, even if initial capital costs are higher, because of the long term
goal of uniform and appropriate product water quality to safeguard patient
well-being. Prospective vendors should be informed that upon installation of
the system, formal acceptance of the equipment is contingent upon valida-
tion of the entire system including the distribution system in terms of both
product water quality and quantity. Vendors should be responsible for com-
pliance with local plumbing and electrical codes in addition to meeting AAMI
product water requirements. Complete operating manuals should be provided
by the vendor including operator training requirements, monitoring require-
ments, recommended cleaning and maintenance schedules and procedures, and
troubleshooting guidelines.

Monitoring product water quality

Even if the water treatment system meets product water requirements at the
time of installation, deviations from these requirements may occur as a
consequence of deterioration of system performance, source water quality
variations, and changes in municipal water treatment practices. It is there-
fore, essential that appropriate monitors of water quality be installed at
appropriate locations of the water treatment system and that samples be drawn
at specified intervals for more detailed laboratory analyses to ensure compli-
ance with the AAMI Water Quality Standard. Appropriate documentation
procedures should be established for long term surveillance of water quality.
In 1988, many municipal suppliers in California switched from free chlorine
to chloramine. Despite all of the advance information provided regarding the
change and its impact on hemodialysis facilities, failure to check product water
for chloramines resulted in an outbreak of hemolytic anemia [29].
System performance may deteriorate with scaling and fouling of reverse
osmosis membranes, exhaustion of components such as activated carbon or
deionizer resin beds, bacterial contamination of the system, etc. Maintenance
schedules and practices should be designed to prevent such deterioration of
system performance. Monitoring will also alert the staff to system performance
deviations from desired levels due to unforeseen circumstances or will indicate
that the monitoring schedules may be inappropriate and may require some
tightening and fine tuning.
In describing the various water treatment processes, we have already
outlined the monitoring requirements for each process. Many of these monitors
are active, on-line monitors such as resistivity monitors, flow, pressure, and
temperature gauges, rejection ratio indicators, etc. However, the final deter-
minant of the adequacy of the water treatment system is the concentration
of various key contaminants in the final product water. These contaminant
levels should be determined at regular intervals using a certified laboratory.
Also other tests such as free chlorine/chloramine levels and bacterial count
 105

may need to be performed on site within minutes or hours of drawing the

appropriate samples. Dialysis unit personnel should have the appropriate
training to perform these procedures and the appropriate assay tests should
be available on site as needed. Standard clinical laboratory techniques avail-
able in a hospital laboratory for measuring bacterial contamination may not
be appropriate for water samples. Nutrient-poor media and lower culturing
temperatures may be required.
Unfortunately no specific guidelines regarding frequency of monitoring can
be laid down as they vary with feed water quality and system design. The
AAMI Water Quality Standard does make some recommendations regarding
the frequency of monitoring certain water treatment processes and additional
guidelines may be imposed by the Public Health authority in each state. For
example, California requires that chloramines levels be checked before every
patient shift. However, as stated by AAMI, "The monitoring of water purity
levels is considered the sole responsibility of the physician in charge of
hemodialysis or the medical professional designated by the physician as the
person in charge".
Some general guidelines apply. Parameters such as pressures, flows, rejec-
tion ratios that are measured continuously on-line should be logged at least
daily. For off-line monitors, daily monitoring may not be necessary. Initially,
a conservative approach with very frequent monitoring is advisable and as time
trends are established, less frequent monitoring may be suitable. Any time there
is a change in system design or maintenance, a new schedule of monitoring
may have to be established. Monthly testing of bacteria and endotoxin levels
is recommended. A complete chemical analysis of the product water using a
certified laboratory should be performed at least annually.

Concluding remarks

Inadequate or inappropriate water treatment poses one of the gravest risks to

the well-being of the dialysis patient because of the large volume of exposure
across a relatively non-selective membrane. Even if the water is appropri-
ately treated by the municipal supplier and is safe for drinking, it may be
hazardous to the health of the dialysis patient. The literature is replete with
many catastrophic consequences of inadequately treated water, and unless staff
at dialysis units are well informed about these consequences, history may repeat
itself. While water treatment technology is complex and many choices are
available relative to system design and configurations, clinical staff at dialysis
programs must become well informed consumers in order to deal effectively
with vendors of water treatment systems. Ultimately, the responsibility for
appropriate system performance, adequate water quality, and patient well-being
rests with the clinical staff in charge of the dialysis program.
106

References

1. The Safe Drinking Water Act of 1974, PL 523, 93rd Congress, Dec 16, 1984.
2. Alfrey AC, Mishell JM, Burks J, Contiguglia SR, Rudolph H, Lewin E, Holmes JH:
Syndrome of dyspraxia and multifocal seizures associates with chronic hemodialysis. Trans
Am Soc Artif Intern Organs 18: 257-261, 1972.
3. Alfrey AC: Dialysis encephalopathy syndrome. Ann Rev Med 29: 93-98, 1978
4. Dunea G, Mahurkar SD, Mamdani B, Smith EC: Role of aluminum in dialysis dementia.
Ann Intern Med 88: 502-504, 1978.
5. McDermott JR, Smith AI, Ward MK, Parkinson IS, Kerr DNS: Brain-aluminum concen-
tration in dialysis encephalopathy. Lancet 1: 901-903, 1978.
6. Elliot HL, Dryburgh F, Fell GS, Sabet S, MacDougall AI: Aluminum toxicity during regular
haemodialysis. Br Med J 1: 1101-1103, 1978.
7. Pierides AM, Edwards WG, Cullum UX, McCall JT, Ellis HA: Hemodialysis encephalopathy
with osteomalacic fractures and muscle weakness. Kidney Int 18: 115-124, 1980.
8. Short AI, Winney RJ, Robson JS: Reversible microcytic hypochromic anaemia in dialysis
patients due to aluminum intoxication. Proc Eur Dial Transpl Assoc 17: 226-233, 1980.
9. McGonigle RJS, Parsons V: Aluminum-induced anaemia in haemodialysis patients. Nephron
39: 1-9, 1985.
10. Schreeder MT, Favero MS, Hughes JR, Petersen NJ, Bennett PH, Maynard IE: Dialysis
encephalopathy and aluminum exposure: An epidemiological analysis. J Chronic Dis 36:
581-593, 1983.
11. Matter BJ, Pederson J, Psimenos G, Lindeman RD: Lethal copper intoxication in hemo-
dialysis. Trans An Soc Artif Intern Organs 15: 309-315, 1969.
12. Ivanovich P, Manzler A, Drake R: Acute hemolysis following hemodialysis. Trans Am
Soc Artif Intern Organs 15: 316-318, 1969.
13. Manzler AD, Schreiner AW: Copper-induced hemolytic anemia. A new complication of
hemodialysis. Ann Intern Med 73: 409-412, 1970.
14. Gallery EDM, Blomfield J, Dixon SR: Acute zinc toxicity in haemodialysis: Br Med J 4:
331-333, 1972.
15. Freeman RM, Lawton RL, Chamberlain MA: Hard-water syndrome. N Engl J Med 276:
1113-1118, 1967.
16. Evans DB, Slapak M: Pancreatitis in the hard water syndrome. Br Med J 3: 748, 1975.
17. Drukker W: The hard water syndrome: A potential hazard during regular dialysis
treatment. Proc Eur Dial Transpl Assoc 5: 284-287, 1969.
18. Nickey WA, Chinitz VL, Kim KE, Onesti G, Swartz C: Hypernatrernia from water softener
malfunction during home dialysis (letter). JAMA 214: 915, 1970.
19. Jowsey J, Johnson WJ, Taves DR, Kelly PJ: Effects of dialysate calcium and fluoride on
bone disease during regular hemodialysis. J Lab Clin Med 79: 204-214, 1972.
20. Lough J, Noonan R, Gagnon R, Kaye M: Effects of fluoride on bone in chronic renal failure.
Arch Pathol 99: 484-487, 1975.
21. Anderson R, Beard JH, Sorley D: Fluoride intoxication in a dialysis unit - Maryland.
Morbidity and Mortality Weekly Report 29: 134-136, 1980.
22. Carlson DJ, Shapiro FL: Methemoglobinemia from well water nitrates: A complication of
home dialysis. Ann Intern Med 73: 757-759, 1970.
23. Salvadori M, Martinelli F, Comparini L, Bandini S, Sodi A: Nitrate induced anemia in home
dialysis patients. Proc Eur Dial Transpl Assoc 21: 321-325, 1984.
24. Comty C, Luehmann D, Wathen R, Shapiro F: Prescription water for chronic hemo-
dialysis. Trans Am Soc Artif Intern Organs 20: 189-196, 1974.
25. Yawata Y, Howe R, Jacob HS: Abnormal red cell metabolism causing hemolysis in uremia.
A defect potentiated by tap water hemodialysis. Ann Intern Med 79: 362-367, 1973.
26. Eaton JW, Kolpin CF, Swofford HS, Kjellstrand CM, Jacob HS: Chlorinated urban water:
A cause of dialysis-induced hemolytic anemia. Sciences 181: 463-464, 1973.
 107

27. Kjellstrand CM, Eaton JW, Yawata Y, Swofford H, Kolpin CF, Buselmeier TJ, von
Hartitzsch B, Jacob HS: Hemolysis in dialyzed patients caused by chloramines. Nephron 13:
427-433, 1974.
28. Botella J, Traver JA, Sanz-Guajardo D, Torres MT, Sanjuan I, Zabala P: Chloramines, an
aggravating factor in the anemia of patients on regular dialysis treatment. Proc Eur Dial
Transpl Assoc 14: 192-199, 1977.
29. Safety alert: Chloramine contamination of hemodialysis water suppliers. Food and Drug
Administration, Department of Health and Human Services, February 19, 1988.
30. Favero MS, Petersen NJ, Carson LA, Bond WW, Hindman SH: Gram-negative water bacteria
in hemodialysis systems. Health Lab Sci 12: 321-334, 1975.
31. Lauer J, Streifel A, Kjellstrand C, DeRoos R: The bacteriological quality of hemodialysis
solution as related to several environmental factors. Nephron 15: 87-97, 1975.
32. Blagg CR, Tenckhoff H: Microbial contamination of water used for hemodialysis. Nephron
15: 81-86, 1975.
33. Robinson PJA, Rosen SM: Pyrexial reactions during haemodialysis. Br Med J 1: 528-530,
1971.
34. Stacha JH, Pontius FW: An overview of water treatment practices in the United States.
J Am Water Works Assoc 76 (Oct): 73-85, 1984.
35. Environmental Protection Agency, Office of Water Supply - National Interim Primary
Drinking Water Regulations, U.S. Government Printing Office, Washington D.C. 1978.
36. American National Standard for Hemodialysis Systems. (RD-5) Association for the
Advancement of Medical Instrumentation, Arlington, VA, 1982.
37. Keshaviah P, Luehmann D, Shapiro F, Comty C: Investigation of the Risks and Hazards
Associated with Hemodialysis Systems (Technical Report, Contract 223-78-5046). U.S.
Department of Health and Human Services, Public Health Service, Food and Drug
Administration, Bureau of Medical Devices, Silver Spring, MD, June 1980.
38. Favero MS, Carson LA, Bond WW, Petersen NJ: Factors that influence microbial
contamination of fluids associated with hemodialysis machines. Appl Microbiol28: 822-830,
1974.
39. Favero MS, Petersen NJ: Microbiological guidelines for hemodialysis systems. Dial
Transplant 6: 34, 1977.
40. Man NK, Ciancioni C, Faivre JM, Diab N, London G, Maret J, Wambergue FP: Dialysis-
associated adverse reactions with high-flux membranes and microbial contamination of liquid
bicarbonate concentrate. Contr Nephrol 62: 24-34, 1988.
41. Klinkman H, Falkenhagen D, Smollich BP: Investigation of the permeability of highly
permeable poly sulfone membranes for pyrogens. Contr Nephrol 46: 174-183, 185.
42. Dinarello C: The biology of interleukin 1 and its relevance to hemodialysis. Blood Purif
1: 197-224, 1983.
43. Lonnemann G, Koch KM, Shaldon S: Induction of interleukin 1 from human monocytes
adhering to hemodialysis membranes. Kidney Int 31: 238, 1987.
CHAPTER 7

The impact of membrane selection on

quality assurance in dialysis

LEE W. HENDERSON

Introduction

Broadly considered, the selection of a membrane for treatment of the patient

with kidney failure determines whether transplantation (glomerular basement
membrane), peritoneal dialysis or treatment with the artificial kidney occurs.
I will narrow the topic by excluding kidney transplantation but will explore
the various options within the latter two modes of treatment. I recognize,
however, that receipt and retention of a good quality kidney graft offers the
best duration and quality of life and must be considered as the gold standard
against which all other modes of treatment must be judged.
There is still much that remains unknown about membrane selection in spite
of a professional lifetime's worth of investigation of this topic by many
researchers, myself included. The wealth of choices today, as contrasted with
the early 1960's when the only choice lay between cellulosic and peritoneal
membrane, is both a tribute to the skill of the membrane manufacturer and
an indictment of the nephrologist for failing to understand the pathophysiology
of uremia with sufficient clarity to define the requirements for an ideal
treatment membrane.
In the subsequent pages I will describe reported differences in membrane
properties and how they may impact on quality of care. I will try to be clear
about where data ends and informed speculation begins. I will describe
categories or classes of membrane and typical properties rather than specific
membranes, i.e. a "generic" approach.
Quality assurance in the treatment of end stage renal disease has at least
two dimensions; quantity and quality of life. Offering an "adequate" dialysis
prescription is of course central to both. I will make the case that membrane
selection is the single most important primary variable in the dialysis
prescription that can be selected by the dialysis unit director as it drives the
decisions on treatment time and blood flow rates and impacts on both quantity
and quality of life. What follows is a discussion aimed at illuminating that
choice. The importance of this choice is underscored by government reim-
bursement agencies and their recent requirement to report presumed parameters

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 109-131.

© 1994 Kluwer Academic Publishers.
110

of adequacy such as KtIV, for urea and plasma albumen concentration, treat-
ment time, etc.

How may we define adequate dialysis?

At present we turn to urea kinetic modeling and the results of the National
Cooperative Dialysis Study (NCDS) [1] to provide a clinically qualified
quantitative definition of adequate hemodialysis! (there is no comparable study
addressing peritoneal dialysis). The derivative analysis by Gotch and Sargent
[3] has provided us with the now familiar "Domain Map" (Fig. 1) in which
normalized dietary protein intake is plotted against predialysis blood urea
nitrogen concentration with isopleths drawn out of the origin like spokes
from the hub of a wheel that show the amount of therapy rendered in terms

I
~KtIV
0.7 0.8
I I ,
140
- 0.9
120

- 1.5

20 ""inadequate excessive
I
...
protein protein
intake I intake
o ~ I
0.5 0.7 0.9 1.1 1.3 1.5 1.7

Protein Catabolic Rate

(g/kg/day)

Fig. 1. A Domain Map showing the interralationship between predialysis blood urea nitrogen
concentration, protein catabolic rate and the dimensionless parameter KtlV (modified from
reference 3).
 111

of the dimensionless parameter for urea of Kt/V. A patient consuming 1 gram

of proteinlkilogram lean body mass with a value for Kt/V of ~ 1.0 is, by NCDS
standards now widely considered to be adequately hemodialyzed. K repre-
sents the diffusive dialyzer clearance (in mllmin.) for urea; a value that is
directly related to the overall mass transfer resistance of the membrane for urea
as well as the area of the membrane employed i.e. membrane selection; t is
the length of the treatment in minutes and V, the total body water in milli-
liters. Before exploring the relevance of this ratio further, it will be important
to understand some of the underlying assumptions of the NCDS and the
limitations that they impose upon interpretation of the results.
The NCDS study was designed to determine whether urea kinetic modeling
could predict clinical outcome with sufficient accuracy to make it a useful
clinical tool for dialysis prescription. The answer by those conducting the study
and for many others, myself among them, was an unequivocal yes. But what
of the limitations? Little has been written on this subject and many unwar-
ranted extrapolations exist in the literature.
Urea as Surrogate for Uremic Toxins: it is generally conceded that urea,
itself, is not the most significant uremic toxin but rather that it is surrogate
for other more toxic solutes [4]. All that is required of urea to achieve this
surrogate status is that it maintain an "orderly" relationship in terms of its
concentration in the plasma water with the concentration of solutes for which
it is considered surrogate. Comparability of molecular size, charge, space of
distribution, polarity and rate of generation would likely assure such an orderly
relationship but are by no means required. For example, a solute of larger
molecular weight with a purely extracellular space of distribution could sustain
a proportional change in concentration to urea during a given weekly treat-
ment prescription in a reproducible manner that would qualify it as having
an orderly relationship to urea. As will be explored subsequently, this latter
surrogacy for solutes of larger size and/or differing space of distribution will
be more easily perturbed by variations in the treatment prescription than would
that for solutes with a molecular size and space of distribution that is com-
parable to urea. Note that there have been several promising candidates
identified since Scribner's original postulate that middle molecules were
pathophysiologically important in uremia, e.g. beta-2-microglobulin [5] the
polypeptide inhibitor of polymorph activity described by Horl et at. [6].
Inferential evidence for their presence and pathophysiologic importance is
strong [7, 8].
Time as Surrogate for Uremic Toxins: a secondary outcome measure of
the NCDS that was less rigorously addressed by the study protocol and almost
ignored in the published interpretation [1] is the impact of treatment time
(in minutes) on adequacy as assessed by the incidence of morbid events [8]
Time enters the numerator of the parameter Kt/V. As the numerator is a
clearance, time product one might simplistically assume shortening dialysis
time could be completely compensated for, clinically, by an increase in the
clearance of urea that holds this product constant, i.e. use of a larger or more
112

permeable membrane. While this is perfectly correct mathmatically, this is

not the case clinically. For example, doubling the clearance of urea so that
treatment time may be cut in half shifts the limiting mass transfer resistance
area product for urea away from the dialyzer membrane and places it upon
the cell membrane (averaged across the body) resulting in a significant
disequilibrium between plasma water and cell water [9]. This may be
quantitatively appreciated by the increase in urea rebound that occurs at the
conclusion of an high efficiency treatment [10] as urea moves across the cell
wall into the plasma discharging the gradient caused by the dialysis treat-
ment, i.e. urea is not distributed in a single pool. A clue to this event would
be noting that the volume of distribution for urea calculated from the drop
in plasma water urea concentration measured at the end of treatment was
identifiably less than the commonly accepted value computed as 58 to 60%
of the body weight in kilograms. For the moment, however, let us assume
that this disequilibrium effect is ameliorated by dialyzing for 20-30 minutes
longer at the doubled clearance rate so that comparable amounts of urea are
removed when compared with regular efficiency HD. Urea Kt/V for short
high efficiency HD then would be maintained equal to the conventional
clearance, time product. One must now ask how the increase in clearance
was obtained. This would most commonly be accomplished by using a
somewhat larger membrane and increasing blood flow rate to utilize this
augmented area. This, of course, is the strategy used in group III of the
NCDS in which low (comparable to control) pretreatment urea concentration
and short treatment time were employed. Figure 2, modified from Laird
et ai. [11] points out that this strategy is associated with a significant increase
in morbid events. As previously pointed out [8], this is a surprising finding
in that the increase in membrane area should have worked toward preserving
the surrogate status for urea for toxins that diffuse more slowly than urea,
making the therapy more nearly comparable to that offered in the control group.
The increase in morbid events with this strategy supports the interpretation that
the shortened time outweighs, in importance, the increase in membrane area,
resulting in a build up of pathophysiologically important slowly diffusing
solutes.
The recent reports of an increase in mortality and its association with
shortened treatment time (low t, no change in K) from Lowrie, et at. [12] using
a large database from a private dialysis chain and of Held [13, 14] using the
powerful USRDS database, provide experimental support for the importance
of solutes for which urea is surrogate in determining the quality of treatment
provided. It is important to note that these retrospectively examined data of
Lowrie and Held were not associated with any effort to sustain a satisfac-
tory dietary protein intake, nor to provide a Kt/V ~ 1. In the NCDS, of course,
a major component of this prospectively conducted protocol was aimed at
sustaining nutritional status with the maintenance of a dietary protein intake
of 1 gram/day/kg body weight. The comparability of the pretreatment blood
urea nitrogen concentrations noted by Held et ai. [13] between his short and
 113

LongTd '\ LongTd

LowTAC urea High TAC urea
0.8 -
\\
0.6 - \
\

...
~ ""- ...... ---- ..
.E
.;
.....
.....0
0.0
~ '\ LowTAC
:.0 0.8
~ '\
urea
,.Q 'I.
0
~ 0.6 \
'I.

0.4

0.2

0.0 ......-'-..I....I....I.~..-.I...I......I.....-.I...I.............I...I.-'-...&..II....I.-'-......I....I.-'-......
0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Normalized protein catabolic rate, gldaylkg of body wt

Fig. 2. The probability of failure is here defined as any patient who died, withdrew for medical
reasons or was hospitalized prior to 24 weeks in the experimental phase of the protocol. The
date of failure was the date of death, withdrawal or first hospitalization, whichever occurred first.
The solid curves were calculated from NCDS data. The dashed lines represent 95% confidence
limits. The impact of treatment time (Td) may be read vertically for low and high time average
concentration of urea (T AC urea) and horizontally for the impact of T AC urea on probability of
failure. The probability of failure approximately doubles when going from long treatment time
to short treatment time for both low T AC ure• (0.1 to 0.2) or high TAC ure• (0.3 to 0.5).

standard treatment time patient groups, points up the reduced dietary protein
intake in the short treatment group. This raises the question of malnutrition
and its role in determining a satisfactory treatment outcome. It remains to
be determined what the respective contributions are of those factors that
cause death in association with malnutrition vs. those that do so in asso-
ciation with retained, slowly diffusing toxins (i.e. middle molecules) and
whether these factors interrelate.
The importance of treatment time as being a significant prescription variable,
of course reaches back to the work of Scribner and Babb on middle
molecules and the square meter hour hypothesis [15]. Treatment time taken
in the context of a constant Kt product is surrogate for middle molecule
clearance, i.e. clearance of molecular weight species that are larger than
the conventionally considered uremic toxins and smaller than albumen
114

(> 500 < 60,000 daltons). The line of reasoning that makes these delibera-
tions relevant to our topic of membrane selection is as follows. If time may
be considered surrogate for middle molecule clearance, then membrane selec-
tion for a more open transport structure will also enhance middle molecule
removal and may translate into the potential for shorter treatment time or more
adequate (read lower mortality) treatment. In considering the limitations placed
on extrapolations from the NCDS, both urea and its surrogate solutes and
time and its surrogate solutes must be carefully considered.
Nutritional Status as an Index of Adequacy: It is important to note a
recent observation by Lowrie (Fig. 3), that identifies a powerful inverse
correlation between plasma albumen concentration and death rate in the ESRD
population under treatment with conventional hemodialysis [12]. Teehan
confirms this observation for patients on CAPD [16]. Albumin concentration
is widely believed to be one objective, if not very sensitive, measure of nutri-

• Crude I2l Adjusted

18~-------------------------------------------y-
17~------------------------------------------~
16 I----------------------------------~/.//J----_v
15 ~----------------------------------¥h/~----~
14 ~----------------------------------¥
13 1---______________________________--/
12 I---------________________________ ~

11 I---------------------------------~//A----_f
10
9
8
7
6
5
4
3
2
1
o
GT4.5 4.0-4.5 3.5-4.0 3.0-3.5 2.5-3.0 LT2.5

Serum Albumin Concentration (gldl)

Fig. 3. Relative risk of death is plotted again sed serum albumin concentration with the
reference concentration of 1 being the 4 to 4.5 g/dl. Those with serum albumin concentrations
of greater than 4.5 (OT 4.5) were not significantly different from the reference standard. The
powerful inverse correlation for the crude data is even more significant when the ratios are
adjusted for the case mix predictors of death.
 115

tional status. A low or falling serum albumin is consistent with negative

nitrogen balance and a diagnosis of malnutrition. This is by no means a new
observation, only a recent rigorous correlation in a large ESRD population
on treatment with hemodialysis. The importance of sustaining a satisfactory
protein intake is noted in the recommendations of the NCDS that intake of
1 gram of protein/kg of lean body wt. per day is necessary for an adequate
treatment prescription.
Lindsay [17] offers an hypothesis that suggests that the conventional wisdom
as drawn from the NCDS is backwards with regard to the interpretation of
this nutritional information i.e. that we must prescribe enough dialysis to
remove the metabolites generated by the breakdown of the ingested protein.
He suggests that the amount of dialysis and metabolite removal is the primary
driver of protein intake through a direct modulation of the appetite. He has
recently reported a prospective study in 18 patients supporting this hypoth-
esis [18]. This places the patient's state of nutrition at the very center of any
judgment on dialysis adequacy. Studies of patients on both peritoneal dialysis
and hemodialysis show that mild to severe malnutrition may be documented
in some 30-40% of study subjects [12, 19, 20]. The marginal nature from
the nutritional perspective of the therapy we now offer is pointed up by these
distressing figures. If Lindsay et al. [17, 18] is right, and I believe he is,
selection of a larger and/or more open membrane in the treatment prescrip-
tion will enhance appetite and reduce both morbid and mortal events.

Treatment strategies and quality assurance

Table 1 lists the experimental parameters and their test ranges used in the
NCDS [1]. I will specifically explore the following treatment prescriptions
which employ different membranes that fall outside the test ranges noted in
Table 1 and indicate how these prescriptions may affect the quality of the
treatment rendered:
A. Large area and/or highly permeable membrane
B. High flux membrane
C. The peritoneal "membrane"

A. Large area and/or highly permeable membrane:

High efficiency hemodialysis is commonly conducted using a large membrane
area (1.5-2.0 m2) and high flow rates of blood (> 400 ml/min.) and dialysate
~ 500 mllmin.) to effect a more efficient removal of uremic toxins. In the
U.S. this form of therapy has usually been employed in conjunction with UKM
and shortened treatment time (2.5-3 hours/treatment 3 times per week). In
its pure form, cuprophane or its equivalent rather than more open high flux
membrane is utilized. Studies by Collins et al. are fine examples of this
technique [21, 22] .
By expanding membrane area and flow rates beyond that employed in the
 0\
-
Table I. Data describing therapy control for the NCDS [I].

Longtime Short time

Group I Group II Group III Group IV

Control Exp. Control Exp. Control Exp. Control Exp.

Dialysis length, min 274 271 274 268 278 200 275 190
±1O ±8 ±1O ±6 ±13 ±15 ±15 ±l7
TAC urea, mg/dl 52.0 51.1 51.4 87.8 50.8 54.7 52.0 93.6
±3.9 ±4.1 ±4.2 ±121.6 ±3.7 ±5.3 ±3.1 ±7.2
Midweek BUN, mg·dl 74 76 75 108 72 77 75 ll5
±6 ±6 ±6 ±14 ±6 ±7 ±6 ±9
Dialyzer urea clearance, 158 171 161 77 162 206 165 107
mUmin ±41 ±41 ±36 ±20 ±32 ±30 ±30 ±21
PCR,glkg 1.04 1.04 0.96 1.08

TACurea =time averaged concentration of urea

PCR = protein catabolic rate
Membranes: Cuprophan, cellulose acetate
Area 1.2-2.5 m2
 117

NCDS, we encounter the previously described circumstance where urea clear-

ance would increase rather dramatically. The NCDS qualified the use of the
variable volume single pool urea kinetic model which presupposes that the rate
of mass transfer across the cell wall significantly exceeds that across the
dialyzer membrane, i.e. the cell wall does not significantly limit transport.
Figure 4 is taken from Frost and Kerr [9] and is one of the few attempts to
examine the overall area mass transfer resistance product across the cell,
taken on average across all of the body cells. They recognize, of course, that
the red cell membrane will have a different value than the cells dividing
blood from brain. This average value is only a mathematical construct but
it offers us useful insight, none the less. The shaded area of Figure 4

o Popovich et aI, 1975 [10]

V Christopher et aI, 1972 [8]
Il Bell et aI, 1965 [5]
Creatinine
o Frost, 1975 [9]
1,000

1"11
...cu
1"11
>< 100

:3
....
,Q =
e
I'll ....
cu
S;::;
:uc.. s
-
-;
~
10
~
U

10 10,000

Mol wt, dalton!

Fig. 4. The log of the average cell wall mass transfer coefficient (permeability times area) is
plotted against the log of solute molecular weight. Data points are represented by symbols
and have been taken from a variety of workers. The shaded area represents the domain of
hemodialyzer clearance for 1 sqare meter cuprophan membrane. Clearance across the cell wall
exceeds that for the dialyzer for solutes small in molecular weight, such as urea and creati-
nine. Uric acid, B12 and inulin are measured to fall significantly below dialyzer clearance.
118

represents the realm of dialyzer clearance. By doubling membrane area and

increasing flows it is apparent that the assumption of a single pool for urea
will be violated. As previously noted, increased membrane area (less so blood
flow rate) increases the clearance of middle molecules but the major impact
is to increase small molecular weight solute clearance. By shortening treat-
ment time the ratio of the mass of middle molecules that are removed per
mg. of urea is reduced even as the increase in membrane area increases middle
molecule clearance. Short of measuring the clearances of both urea and a
middle molecule such as beta-2-microglobulin (11,800 daltons) or vitamin
B-12 (1,355 daltons), it is not possible to say whether the treatment rendered
is adequate as defined by the NCDS or not, as removal of both large and
small toxins is clinically important. Recent work by Collins et al. [21] in
more than 1000 patients over a 4 year time frame (1986-1990) where both
urea and modelled vitamin B 12 clearances were sustained shows that survival
with short « 3 hours/treatment) large surface area bicarbonate hemodialysis
is statistically significantly improved in overall mortality if the Kt/V for urea
is maintained ~ 1.2, a value comparable to the value of 1.0 to 1.19 for the
age and risk factor adjusted control subjects. This work indicates that by
using a large area dialyzer to increase urea clearance to values that sustain Kt/V
comparable to the control group and concomitantly increasing middle molecule
clearance, there is an improved survival rate. The possible role of the use of
bicarbonate in this improved survival rate needs to be further evaluated. This
work is of great importance as it shows that shortening treatment time can
be safely done with high efficiency hemodialysis, i.e. proper selection of the
membrane to sustain the Kt product for both urea and vitamin B 12, at or
above NCDS standards.

B. High flux membrane:

By this I mean the use of membranes that have a far higher hydraulic
permeability than the more commonly used cellulosic membranes, cuprophane
or cellulose acetate. Use of such a membrane (e.g. polyacrylonitrile, cellu-
lose triacetate or polysulphone) requires special fluid cycling equipment to
permit control of ultrafiltration. Significant ultrafiltration may occur even in
the absence of net ultrafiltration rates that are high, simply by a form of
"Starling capillary flow" that produces ultrafiltrate at the arterial end of the
membrane and backfiltration of fluid (dialysate mixed with ultrafiltered plasma
water) at the venous end. Hemodiafiltration fits within this definition i.e. the
deliberate enhancement of ultrafiltration during a dialysis treatment to the limit
set by the area, hydraulic permeability product for the membrane used and then
replacing ultrafiltrate volume that exceeds the requisite fluid loss to restore
excess body water volume to normal. The convective element to transport
perturbs the relationship of urea to larger solutes as it is not size discrimina-
tory (i.e. urea's surrogate status as qualified by the NCDS for diffusive transport
is perturbed). If urea kinetic modeling (single or multiple pool) is applied to
this form of therapy and the Kt/V value is adjusted equal to NCDS standards
 119

for adequacy, one can assume a larger clearance of middle molecules and hence
a dialysis prescription that is "more adequate" than the same KtlV obtained
using cuprophane and solely diffusive transport. The fraction of overall mass
transport that is convective in mechanism may be considered to be surrogate
for treatment time. I speculate that a suitable mix of shortening treatment
time sustaining urea removal and increasing convective transport would result
in a treatment that is comparably adequate to conventional full time hemo-
dialysis. If adequacy is comparable, the shorter treatment time by definition
offers a higher quality of therapy.
At present there are only a few prospective studies that give clear qualifi-
cation of this treatment strategy. See, for example, studies such as that of
Channard et al. [18] using the more open complement kind PAN membrane
in which patients (n = 31) treated for 9.3 ± 0.2 hrs/wk are contrasted with a
control group (n = 31) treated for 16.2 ± 0.3 hrs/wk with cuprophane. A
significantly lower hospitalization rate and number of hospitalization days
for dialysis related complications was achieved with the PAN membrane in
spite of carrying a higher pretreatment blood urea nitrogen concentration
1.16 ± 0.05 mmolll for short treatment vs. 1.01 ± 0.03 mmolll for conventional
treatment. There are likely 2 mechanisms at work here, i.e. differences of
complement activation (this will subsequently be discussed) and the aug-
mentation of middle molecule clearances achieved with the PAN membrane.
Von Albertini et al. [24] have clinically applied the bench study of Cheung
et al. [25] in a small number of patients (n = 18) to deliberately exploit the
relationship (therapeutic equivalency) of time and middle molecules. They have
achieved an average of 2 hours in treatment 3 times per week for a year or
more, using two modules of high flux membrane and flow rates that are
nearly twice the common values (Fig. 5). His patients have not shown the
cardiovascular instability, considered limiting by most, for removal of excess
body water (3.6 ± 1.2 1) in a 2 hour treatment time frame with hemodialysis.
His rehabilitation rate for these patients was very high at > 90%. This extra-
ordinarily high flux treatment has shown the expected urea rebound that is
predicted when the limiting mass transfer resistance is the cell wall and not
the dialysis membrane [9]. His modeling studies support a more than adequate
status of clearance for not only urea but for the index solute Inulin (5,200
daltons) as well.
At present the only safe recommendation for this form of treatment, as noted
by Von Albertini, is to use a KtlV urea of 1.2 as a minimum value. This will
ensure that the mass of urea removed in spite of urea disequilibrium at high
clearance rates is adequate and larger and/or more slowly diffusing solutes that
are removed is comparable to, or greater than with conventional therapy [26].
An additional point of note, commented on above, may be drawn from
the recent work of Guttierez et al. [27, 28]. Their work implicates comple-
ment activation and cytokine release in the increase of catabolism noted by
Borah et al. [29], and Farrell et al. [30], in response to hemodialysis. The
extra catabolism that occurs on the days of dialysis with cuprophane results
120

~
Blood Pump

~-
500
1025
----.. Drain
Net UF Pump 1000
I

-
900 Differential
375 Pressure
[ Control

900

--
1000
Pyrogen
-
1000
Dialysate

-... 475
Filter

Fig. 5. Flow diagram and flow rates in mUmin for a hemodiafiltration circuit (von Albertini).
A pair of high flux membranes is used in series with blood and dialysis fluid flowing in
countercurrent manner. High flux membrane I is driven by the differential pressure controller and
blood pump to ultrafilter from blood to dialysate in the amount of 125 mUmin. Sterile pyrogen
free diluting fluid from the pyrogen filter and fluid cycling device is delivered to high flux
membrane II and restores all but 25 ml/min of backfiltration to the blood path. Diffusive
transport occurs in both membranes I and II.

in a calculated urea generation rate approximately 20% higher than occurs

on non dialysis days. As most high flux membranes are synthetic in formu-
lation, e.g. polyachrilonitrile, polysulphone and polyamide and, as such, are
complement kind, unlike cellulosic membranes that activate complement
(cellulose triacetate is a notable exception being both high flux and comple-
ment kind), one may expect a lower need for treatment with these membranes.
Hemofiltration may be considered a limiting case of hemodiafiltration in
which all solute transport is by convection. This technique clearly falls outside
the study parameters of the NCDS and in so doing, provides some fasci-
nating insights into the pathophysiology of uremia and has significant
implications for treatment quality through membrane selection. The most
substantial clinical experience with this technique may be found in the work
of Quellhorst and colleagues [31]. They report on more than 100 patients
maintained with post dilution hemofiltration with some having been treated for
up to ten years. The majority of the study sample have been on treatment
for over five years and as such, may be satisfactorily compared with a com-
parable population of patients that he has followed on "routine" hemodialysis.
Routine hemodialysis in this instance comprised 5 hours of hemodialysis thrice
 121

weekly, using 1.0 to 1.5 m2 cuprophane membrane and conventional flow rates
of blood (250-350 mllrnin.) and dialysis fluid (500 mllrnin.). The amount of
hemofiltration given was calculated with the formula that one third of the
total body water needed to be ultrafiltered at each treatment [26]. This
would mean that a 70 kg man with 42 I of total body water would exchange
25 liters of ultrafiltrate for diluting fluid, assuming a 2 I removal of excess
body water. Urea clearance in post dilution hemofiltration equates to the
volume of ultrafiltrate i.e. in this example the total volume of plasma water
cleared of urea would be 25 1. This of course, provides a KtlV for urea of
only 0.6; a figure that by NCDS standards should produce major morbidity
promptly. Figure 6 is a plot of mortality for these patients over the ten year
follow up. This may be compared with similar mortality curves for his
hemodialysis population and for that reported by the EDTA and French Dialysis
Registries [26].
His hemodialysis population was dialyzed for 5-6 hours thrice weekly, with
1.0 m2 cuprophane. This would likely ensure an adequate KtlV by NCDS
standards although weight data and dietary protein intake for these populations
is lacking. The patients in Figure 6 were randomly assigned to hemodialysis

100
....... (132)

~ " '''(115)
\ \ (115) ....
1974-1978

--
':\ (91) , ... (108)
\\ ......
90 ~104)
-.....
\\ .........
~
\. \ . ........... (72)
!IS \ \. ...........
>
"-
\ \ ........ (46)

t:
::s
\\
\ \ ..... (3) HF
"
CJ)
80
'\.
"\ \DP \
'- EDTA \
... HD
72 (11)

T
i i i
1 2 3 4 5 6 7 8

Time in Years
Fig. 6. Survival data for patients starting regular hemofiltration or hemodialysis treatment between
1974 and 1978. For comparison, data taken from both the EDTA and Diaphane registries for
the same time period are shown. See text for further discussion.
122

or hemofiltration and "poor risk" patients were excluded from the study, i.e.
patients with "complications such as, diabetes mellitus, cancer, and systemic
or severe cerebrovascular or cardiovascular disease" [31]. The message here
is that quality assurance quidelines for treatment with hemodialysis cannot
be mindlessly applied to a therapy where the selection of a more hydrauli-
cally permeable membrane alters urea's surrogate status for uremic solute
removal.

C. The Peritoneal "Membrane":

The vast majority of the literature regards the barrier to transport that divides
body water and peritoneal dialysate as a semipermeable membrane. That is
a pore-containing membrane that permits the ready passage of water and
small solutes but restrains or blocks the passage of proteins and cell elements
from entering the dialysate. Some models offer the ability to calculate mean
pore radius/pore length as well as the size distribution of pore radii, assuming
that these pores are right circular cylinders that traverse the membrane [32].
Other writers describe serial resistances to transport each with its own physical
and transport properties that sum to the earlier "unitary" description [33]. Even
more recent writings offer a description of the functional properties of this
barrier to transport not as a membrane at all but rather as a tissue with an
"homogeneously distributed" blood vessel (this is a mathematical not a physical
construct) which acts as a source for uremic solutes that find their way into
the dialysate [34]. While all of this work is fascinating and will permit
a more incisive understanding about intimate transport mechanism, these
mechanisms are not crucial to our understanding of the comparative proper-
ties on which to judge the treatment qualities offered by peritoneal dialysis.
For simplicity, I shall cast what follows in terms of a simple semipermeable
membrane that may be a serial or single transport resistance and/or a homo
or heteroporous structure as you would wish to conceive it.
The most glaringly inappropriate "leap of faith" in extrapolating NCDS
results beyond the tested parameters occurs when the mathematics of variable
volume single pool urea kinetic modeling as used in the NCDS are applied
without modification to judge adequacy when the peritoneal membrane has
been selected for use, i.e. chronic ambulatory peritoneal dialysis (CAPD).
The NCDS used only cellulosic membrane with a range of membrane area
between 1.2-2.5 m2• An informed guess about the area of the peritoneal
membrane that participates in solute and water transport (functional area)
places it at less than 0.8 m2 (i.e. 0.5-0.8m 2) [35]. The degree of openness to
diffusive solute transport between cellulosic and peritoneal membranes has
been studied by several workers and the result of an early such study is
shown in Figure 7 [36]. It is cast in terms of comparative clearances vs.
solute molecular weight for cellulosic dialyzers vs. the peritoneal membrane.
Small solutes such as urea « 60d) that are swiftly diffusing are more sensi-
tive to membrane area differences than are larger molecular weight species
such as the test solute, inulin (5200 d) i.e. small solutes are more blood flow
 123

1000 DowHFAK
Model 3
1M 2

100 ~------~~~------~~----~
"-
" "-
/\
Pet\oneal membrane
Inulin

10

t Sucrose
1 Uric Acid
ere thune

10 100 1000 10,000 100,000

Solute Molecular Volume (cm Igmole)

Fig. 7. Log log plot of the permeability area product ("clearance") versus solute molecular
volume. See text for further discussion of clearance. Note the crossover point between the
peritoneal membrane and the cellulosic membrane for solutes in the 5,000-6,000 dalton range.

dependent than large ones when clearance is being examined. Hence, urea
clearance for the peritoneal membrane falls well below that for cellulosic
membrane. It is important to note that peritoneal clearance as it is commonly
measured and described is different from the clearance term for the native
kidney or that for the artificial kidney that operates by diffusion (hemodial-
ysis) or that for convective transfer (hemofiltration) or any combination of
the two (hemodiafiltration). The common formula for peritoneal clearance

pD Vlt == C

where
C == clearance (mllmin.)
D == dialysate concentration (mg/dl)
P == plasma water concentration (mg/dl)
Vlt == volume (ml) of spent dialysate per exchange time in (minutes)
has in the numerator a mass of solute (DV) that has been removed and is
present in the spent dialysate over a given exchange time. (Note: Urea in the
dialysate is directly measured and this is not subject to errors that occur in
124

measuring hemodialysis clearance that result from problems of determining

flow rates accurately or the presence of access recirculation.) The driving
gradient for diffusion is the highest at the time when the dialysate concen-
tration for urea is zero and progressively declines with time as equilibration
between blood water and dialysate occurs (Fig. 8). The clearance calculated
with the common formula is a time averaged rate of clearance and as such
will always be somewhat lower than the "instantaneous" clearance measured
in a manner analogous to either the artificial or native kidney [37]. The
instantaneous clearance remains constant whenever measured during the course
of treatment whereas peritoneal "minute to minute" clearance falls during
the course of an exchange and this time averaged clearance falls as the
exchange time is lengthened. An instructive limiting case is that for the swiftly
equilibrating solute urea. For a patient with an average or high permeability
peritoneal membrane, urea may well be at or near equilibrium in the last
hour or two of a CAPD exchange, i.e. DIP = 1 for the final 60-120 minutes
and the time average urea clearance drops swiftly as a result. Furthermore,
no more mass is removed even through the equilibration of other solutes
continues, thereby perturbing the relationship between urea clearance and
the more slowly equilibrating solutes for which it is surrogate.

10

09

08

·s
:::
07
] 06
~

-~
u
::: 05
...
r;I
r;I
~
04
P
"-
<:0 03
p...
02

01
00
20 60 90 120 160 200 240 300 360 420 480

Time in minutes

Fig. 8. Stylized curves for the equilibration ratio between dialysate and plasma plotted against
time for a CAPD patient with average transport characteristics.
 125

For larger solutes the transport rate across the peritoneal membrane may
be comparable to or higher than that for cellulosic membrane. It is the rela-
tionship or relative concentration that urea bares to other more toxic solutes
that establishes its surrogate status. Furthermore, the NCDS employed a
schedule of three times weekly treatment with each treatment ranging (by study
group) from 3.0 to 4.5 hours (Table I). CAPD is, of course, a continuously
applied therapy. This having been said, I note for you that patients consid-
ered to be adequately treated on CAPD using clinical criteria (we have no
NCDS for CAPD) have a measured KtlVof 1.6-1.7 per week [38]. Dividing
this by 3 to equate it to the amount of hemodialysis offered on a thrice weekly
schedule identifies that CAPD at 0.5-0.6 is dramatically lower than, i.e. half,
that required for adequate hemodialysis (1.0-1.2).
Dr. Keshaviah emphasizes this point (i.e. the relevance of treatment
application time in a recent publication [39]. Table II is taken from that pub-
lication and casts clearance in terms of liters per week for solutes of different
molecular weight contrasting CAPD with two commonly used hemodialysis
membranes. CAPD, because of its continuous nature, showed higher net weekly
clearance for solutes as low in molecular weight as vitamin B 12 when con-
trasted with 1.5 m2 cuprophane (8 micron) applied for four hours, three times
weekly.

Table II. Weekly clearances of peritoneal and hemodialysis membranes (units = liters per week).

Solute Mol. wt. CAPD Cuprophane Cellulose triacetate

(8 microns) (high flux)

Urea 60 64 119 139

Creatinine 113 57 96 126
Vitamin BI2 1355 37 27 86
Inulin 5200 17 14 51
Beta-2 Mcgb. 11800 8 0 38

Cuprophane: CF 1511 (Baxter Healthcare Corporation), Qb/Qd = 200/500, 12 hr/wk.

Cellulose triacetate: CT 190 (Baxter Healthcare Corporation), Qb/Qd = 300/500, 9 hr/wk.

At least four points of concern make it incorrect to extrapolate NCDS

interpretations of urea kinetic modeling to CAPD:
- First, the impact of "saw tooth" chemistries vs steady state values has
been commented on by Keshaviah et ai. [39, 40]. The hypothesis that is
currently being tested by these investigators states that uremic toxicity is
more dependent on the predialysis peak chemistries than on trough or time
average values and that CAPD with its steady state chemistries may well
be exempt from some of the uremic toxicity (e.g. acidemia) that occurs
intermittently in the hemodialysis population undergoing thrice weekly
treatment.
- Second, there are well understood kinetic principles that link duration of
therapy with the removal of larger molecular weight solutes, i.e. time on
126

treatment is surrogate for middle molecule removal [6, 13]. With refer-
ence to Table II, one notes that with the selection of the peritoneal membrane
that is more open to the diffusive transport of larger solutes coupled with
the continuous nature of the CAPD prescription that urea's surrogate status
with regard certainly to the larger molecular weight toxins would be
radically different than for hemodialysis, i.e. more middle molecules
removed/gram of urea removed. I have noted previously the surprising
lack of attention paid to this relationship in the published results of the
NCDS. Recent analysis of data from the USRDS registry provides powerful
evidence that short prescription time correlates with early mortality and
likely offers one explanation of the higher death rate present in the U.S.
dialysis population than that in Europe [2, 14].
- Thirdly, as stated above, for a patient with a reasonably permeable
peritoneal membrane, one may well see that the dialysate to plasma con-
centration ratio for urea may achieve a value of unity for the last hour or
two of a six hour dwell time, i.e. no more movement of urea from blood
water to dialysate. This equilibration of urea does not preclude the continued
loss into the dialysate of other less swiftly diffusing toxic solutes. Again,
the surrogate status that urea holds for more toxic solutes of slower
diffusivity is abrogated by this event.
- Fourthly, as previously noted, for non complement activating synthetic
membranes, the choice of the peritoneal membrane for treatment exempts
the patient from a thrice weekly activation of complement with it's release
of the powerful inflammatory mediators, interleukin-1 and tumor necrosis
factor which appears to result in a burst of protein catabolism with hemo-
dialysis that is not present with CAPD [29, 30]. Recent work by the
Bergstrom Group [41] points out further that the relationship between
protein catabolic rate plotted on the axis and Kt/V urea on the ordinate
showed a markedly steeper slope than that for hemodialysis. This is
interpreted to show that increasing dialysis dose, as measured by Kt/V
urea has a "more salutory" effect on appetite in peritoneal dialysis than in
hemodialysis.
There are some common features between CAPD and hemofiltration that are
instructive and may explain a clinically satisfactory outcome at remarkably
low Kt/V values:
- Both employ a complement kind membrane and hence a reduced catabo-
lism of protein on dialysis days.
- Both have a disproportionately high clearance of middle molecules when
compared with urea.
- Both techniques are conducted using only sterile pyrogen free solutions -
unlike routine hemodialysis.
- Both techniques show less reduction in residual renal function over time
[31,42].
There are interesting points of difference that will require further study before
fully understanding their implications:
 127

- Hemofiltration shares with hemodialysis the common saw-toothed pattern

of chemistries and fluid overload that is the "hallmark" of a three times
weekly treatment schedule as contrasted with CAPD. This would argue
against the peak/valley hypothesis of Keshaviah et al. [40].
- Hemofiltration membranes are usually noncellulosic. The two in common
use (polyacrylonitrile and polysulphone) are active adsorbers of plasma
constituents, unlike the peritoneal membrane.
- Protein (5-15 gms/day) is lost in significant continuing quantities in CAPD
but not hemofiltration or hemodialysis.

Clearance vs. mass removed

I have a technical concern about the limitation of the clearance concept as a

preferred measurement parameter for characterizing the clinical performance
of a dialysis membrane. This concern was partially explored in relation to
peritoneal dialysis in the prior section.
With reference to Figure 8, I again note for you that clearance was designed
algebraically to be constant, I.e. not change with a changing concentration
of solute in the plasma water. A.V. Wolf et al. [43] by mimicking Homer
Smith [44] in this trait for the equation describing clearance in hemodial-
ysis, permitted an "apples to apples" kind of comparison for the clinician
between, the then familiar native kidney performance and that of the new
artificial kidney. This was useful. The disservice that results from the use of
clearance, today, is that it obscures the (clinically critical) rate and/or quantity
of mass removed. Take, for example, the high clearance rates for urea
(500-600 ml/min)that result from high flux hemodiafiltration as practiced
by von Albertini et al. [26]. While the calculated clearance is constant for
the entire length of the procedure, the mass of urea removed per minute
declines sharply due to the rapid fall in BUN with time. This fall is accen-
tuated by the limited availability of intracellular urea (as previously commented
upon). Said another way, the apparent distribution volume for urea is con-
siderably less than the total body water. This, then would mean an artifactually
low volume of distribution (V) for urea as computed by urea kinetic modeling.2
As clearance (K) is not reflective of changing plasma levels, the KtlV ratio
will be artifactually high. What we really wish to track from the adequacy
perspective of the NCDS, is the actual amount of urea that is removed by a
given treatment. For example, a device that continuously measured the
concentration of urea in the effluent stream of the dialysate could be used to
provide the clinician with a minute to minute amount of urea removed. The
amount (mass) of urea present pretreatment, is reasonably well determined
by using a nomogram derived space of distribution for urea and the product
of this volume and the pretreatment BUN. One may then examine what fraction
of urea mass is removed (FR urea) rather than the "fractional clearance" ,3
i.e. KtlV urea.
128

Urea mass removed 100

FR urea = Body content of urea pretreatment
x

It is easier for me to develop my clinical intuition around the fraction of

solute removed than it is to have any "gestalt" about KtlV - an hieroglyph
that I consider analogous to judging hydrogen ion concentration from pH.

Quality assurance principles

Use of synthetic membrane

From the above discussion it is apparent that one may not isolate selection
of the membrane from the other components of the treatment prescription when
examining the quality of therapy rendered. At present we have no hard data
on the comparative importance of:
A. The two physician variable components of the treatment prescription,
namely, treatment time and the elements comprising artificial kidney
clearance rate (i.e., membrane area, and permeability, the flow rates of
blood and dialysate, and the respective components of convection and
diffusion) and
B. Removal of the two broad solute domains i.e. conventional size « 500 d)
or middle molecule (> 500 d and < 60,000 d) size, which are affected
differently, depending upon the choice of membrane and technique.
One may only guide patient therapy by offering, at minimum, the amount
of therapy given by the best treatment group of the NCDS. This is not to
say that more treatment might not be desirable. Urea kinetic modeling coupled
with an understanding of the influence of treatment time on solute clearance
profile is required in order to know when this minimum is achieved. Where
uncertainty exist for a given prescription about middle molecule clearance it
is well to make actual measurements of a test solute such as inulin or vitamin
B 12 in order to resolve that uncertainty.
One may approach urea kinetic modeling using any of a number of sim-
plifications of the original approach that vary more in style than substance.
Not all patients will need to be modeled as empirical wisdom about average
patients requiring an average prescription is quit strong. It is only where
significant variations in patient parameters of height, weight protein cata-
bolic rate or the treatment parameters that determine clearance, that is, A.
and B. above, that depart from the common experience that formal modeling
will be required to assure the delivery of an high quality treatment.

Use of the peritoneal membrane

There is even more uncertainty here as there has, as yet,4 been no prospec-
tive clinical qualification of urea kinetic modeling or for that matter any
 129

other form of quantification for peritoneal dialysis. Nolph has recommended,

on the basis of his impressive clinical experience with CAPD, that a total
(urinary + peritoneal) creatinine clearance of at least 50 liters per week should
be offered to sustain good health [38]. Blake and Oreopoulos, on the other
hand, disclaim this as a useful guide on the basis of a 76 patient, 18 month
study [46]. More study is needed before a firm recommendation can be made.
At present there is a brisk upswing in the use of automatic fluid cycling
equipment for peritoneal dialysis, i.e. Automated Peritoneal Dialysis (APD).
As APD can be used to obtain more efficient removal of small, swiftly dif-
fusing solutes (e.g., urea, creatinine, uric acid, hydrogen ion, etc.), and some
have suggested that nightly application of APD could be used to obtain the
desired 50 liters of creatinine clearance per week at night, and permit a "dry
belly" during the day, recognize that this strategy will reduce the weekly
clearance obtained across the peritoneal membrane as time spent in treat-
ment will be reduced when contrasted with CAPD. While reduction in "face
time" with the therapy is highly desirable from the quality of life perspec-
tive, I am concerned that biochemical deterioration may shorten patient life.
I must recommend a "wet belly" continuously until prospective trials are
carried out to determine whether it is safe to shorten therapy time.

Notes

1. Clinical assessment such as changes of lean body mass, subjective well being, complica-
tion rate and the like, coupled with serial measurements of blood urea nitrogen, creatinine,
calcium phosphate, CO 2 content, plasma proteins, hemoglobin, etc. are still more commonly
used to assess treatment adequacy worldwide than is urea kinetic modeling. The quality
of treatment rendered using these parameters is sufficiently high to make subtle
distinctions in quality impossible to detect at the clinic level simply because the number
of study subjects required to show significant trends likely exceeds that available to even
the director of a large clinic. I note, for your persuasion to this point, the surprise
engendered in the U.S. dialysis community by the recent identification of a substantially
increased mortality rate in the U.S. dialysis patient population as contrasted with those of
Europe and Japan when natinal population bases were examined [2].
2. The popular use of 58-60% of body weight in kg or nomogram values involving height,
weight, sex, to compute total body water will obviate this problem but has problems of
its own. A malnourished patient will be assigned a lower body water volume and hence
make his calculated KtlV artifactually high.
3. I put fractional clearance in quotes as it is not strictly that. Note that fractional clearance
cannot exceed. KtlV can, of course, exceed unity, but this is because the same volume
of blood passes through the dialyzer repeatedly and is cleared again and again. Said another
way, if KtlV is truly a fractional clearance, a value of unity indicates that the entire
space of distribution for urea has been emptied. This of course does not happen with a
KtlV = 1.
4. A North American Multicenter trial to prospectively qualify urea or creatinine kinetic
modeling as a suitable predictor of morbid/mortal events is now being conducted under
the guidance of David Churchill, M.D. [39].
130

References

1. National Cooperative Dialysis Study. Kidney Int 23 (Suppl 13), 1983.

2. Held PJ, Brunner F, Odaka M, Garcia J, Port FK, Gaylin DS: Five year survival for end
stage renal disease patients in the U.S., Europe and Japan, 1982-1987. Am 1 Kidney
Disease 15: 451-457, 1990.
3. Gotch FA, Sargent JA: A mechanistic analysis of the National Cooperative Dialysis Study.
Kidney Int 28: 526-534, 1985.
4. Johnson WJ, Hagge WW, Wagoner RD, Dinapoli RP, Rosevear JW: Effects of urea loading
in patients with far advanced renal failure. Mayo Clinic Proc 47: 21-29, 1972.
5. Gorevic PD, Casey TT, Stone WJ, DiRiaimondo CR, Presli FC, Fragione B: Beta-2
microglobulin is an amyloidogenic protein in man. 1 Clin Invest 76: 2425-2429, 1985.
6. Hor! WH, Haag-Weber M, Georgopoulos A, Block LH: Physicochemical characterization
of a polypeptide present in uremic serum that inhibits the activity of polymorphonuclear
cells. Proc Natl Acad Sci 87: 6353-6357, 1990.
7. Bergstrom J, Furst P, Zimmerman L: Uremic middle molecules exist and are biologically
active. Clin Nephrol 11: 229-238, 1979.
8. Henderson LW: Of time, TAC urea and treatment schedules. Kidney Int 24 (Supp): S105-S106,
1988.
9. Frost TH, Kerr DNS: Kinetics of hemodialysis: A theoretical study of the removal of solutes
in chronic renal failure compared to normal health. Kidney Int 12: 41-50, 1977.
10. Miller JH, von Albertini B, Gardiner PW, Shinaberger JH: Technical aspects of high-flux
hemodiafiltration for adequate short (under 2 hours) treatment. Trans Am Soc for Art lnt
Organs 30: 377-381, 1984.
11. Laird NM, Berkey CS, Lowrie EG: Modeling success or failure of dialysis therapy: The
National Cooperative Dialysis Study. Kidney lnt 13 (Suppl): S101-S106, 1983.
12. Lowrie EG, Lew NL: Death risk in hemodialysis patients: The predictive value of commonly
measured variables and an evaluation of death-rate differences betwen facilities. Am 1 Kidney
Disease 15: 458-482, 1990.
13. Held PJ, Levin NW, Bovbjerg RR, Pauly MV, Diamond LH: Mortality and duration of
hemodialysis treatment. lAMA 265(7): 871-875, 1992.
14. Held PJ, Blagg CR, Liska DW, Port FK, Hakim R, Levin N: The dose of hemodialysis
according to dialysis prescription in Europe and the United States. Kidney lnt 42: SI6-S21,
1992.
15. Babb AL, Popovich RP, Christopher TG, Scribner BH: The genesis of the square meter hour
hypothesis. Trans Am Soc for Art lnt Organs 17: 81-91, 1971.
16. Teehan BP, Schleifer CR, Brown JM, Sigler MH, Raimondo J: Urea Kinetic Analysis and
Clinical Outcome on CAPD. A Five Year Longitudinal Study Adv in PD 6: 181-185,
1990.
17. Lindsay RM, Spanner E: A hypothesis: The protein catabolic rate is dependent upon the
type and amount of treatment in dialyzed uremic patients. Am 1 Kidney Disease 13(5):
382-389, 1989.
18. Lindsay RM, Spanner E, Heidenheim RP, LeFebvre JM, Hodsman A, Baird J, Allison
EM: Which comes first KtIV or PCR - Chicken or Egg. Kidney lnt 42: S32-S36, 1992.
19. Young A, Kopple J, Lindholm G, Vonesh E et al.: Nutritional assessment of CAPD patients:
An international study. Am J Kidney Disease 17: 462-471, 1991.
20. Held PJ, Port FK, Gaylin DS, Wolfe RA, Levin NW, Blagg CR, Garcia J, Agadoa L:
Evaluation of initial predictors of mortality among 4387 new ESRD patients: The USRDS
case mix study. JASN 2 (abstract): 328, 1991.
21. Collins A, Liao M, Umen A, Hanson G, Keshaviah P: High efficiency bicarbonate hemodial-
ysis has a lower risk of death in standard acetate dialysis. J Am Soc Nephrol 2: 318, 1991.
22. Collins A, Liao M, Umen A, Hanson G, Keshaviah P: Diabetic hemodialysis patients treated
with a high KtlV have a lower risk of death than standard KtIV. J Am Soc Nephrol 2: 318,
1991.
 131

23. Channard J, Brunois JP, Melin JP, Lavaud S, Toupance 0: Long term results of dialysis
therapy with a highly permeable membrane. Artif Organs 6: 261-266, 1982.
24. von Albertini B, Barlee V, Bosch JP: High flux hemodialfiltration: Long term results.
J Am Soc Nephrol 2 (abstract): 354, 1991.
25. Cheung AC, Kato Y, Leypoldt IK, Henderson LW: Hemodiafiltration using a hybrid
membrane system for self - generation of diluting fluid. Trans Am Soc for Art Int Organs
28: 61-65, 1982.
26. von Albertini B: High-efficiency hemodialysis: An overview. Contrib to Nephrology 61:
37-45, 1988.
27. Gutierrez A, Alvestrand A, Wahren J, Bergstrom 1: Effective in vivo contact between
blood and dialysis membranes on protein catabolism in humans. Kidney Int 38: 487-494,
1990
28. Gutierrez A, Alvestrand A, Bergstrom 1: Membrane selection and muscle protein catabolism.
Kidney Int 42: S86-S90, 1992.
29. Borah MF, Shoenfeld P, Gotch FA, Sargent JA, Wolfsom M, Humphreys MH: Nitrogen
balance during intermittent dialysis therapy of uremia. Kidney Int 14: 491-500, 1978
30. Farrell PC, Hone PW: Dialysis induced catabolism. Am J Clin Nutr 33: 1417-1422, 1980.
31. Quellhorst E: Long-term survival. In Henderson LW, Quelhorst EA, Baldamus CA, Lysaght
MI (eds). Berlin: Springer-Verlag, p 221, 1986.
32. Rippe B, Stelin G: Simulations of peritoneal solute transport during CAPO. Application
of the two-pore formalism. Kidney Int 35: 1234-1244, 1989.
33. Nolph KD, Twardowski Z: The peritoneal dialysis system, in Nolph KD (ed) Peritoneal
Dialysis (pp 13-27). Kluwer Academic Publishers, 3rd Edition, 1989.
34. Dedrick RL, Flessner MF, Collins 1M, Schultz IS: Is the peritoneum a membrane? ASAIO
J 5: 1-8, 1982.
35. Henderson LW: The problem of peritoneal area and permeability. Kidney Int 3: 409-410,
1973.
36. Babb AL, Johansen PI, Strand MI, Tenckhoff H, Scribner BH: Bi-directional permeability
of the human peritoneum to middle molecules. Proc of the European Dial and Trans Assoc
10: 247-262, 1973.
37. Henderson LW, Nolph KD: Altered permeability of the peritoneal membrane after using
hypertonic peritoneal dialysis fluid. J of Clin Invest 48: 992-1001, 1969.
38. Nolph KD: Quantitating peritoneal dialysis delivery: A required standard care. Seminars
in Dial 4(3): 139-141, 1991.
39. Keshaviah P: Urea kinetic and middle molecule approaches to assessing the adequacy of
hemodialysis and CAPO. Kidney Int 43: S28-S38, 1992.
40. Keshaviah PR, Nolph KD, VanStone JC: The peak concentration hypothesis: A urea kinetic
approach to comparing the adequacy of continuous ambulatory peritoneal dialysis and
hemodialysis. Peritoneal Dialy Bul 9: 257-260, 1989.
41. Bergstrom 1, Alvestrand B, Lindholm B, Tranaeus A: Relationship between KtlV and
protein catabolic rate (PCR) is different in continuous peritoneal dialysis (CPO) and
haemodialysis (HD) patients. JASN 2 (abstract): 358, 1991.
42. Lysaght MJ, Vonesh E, Gotch F, Ibels L, Keen M, Lindholm B, Nolph KD, Pollock CA,
Prowant B, Farrell PC: The influence of dialysis treatment modality on the decline of
remaining renal function. Trans Am Soc Artif Intern Organs 27: 598-604, 1991.
43. Wolf AV, Remp DG, Kiley IE, Currie GO: Artificial kidney function: Kinetics of hemo-
dialysis. J Clin Invest 30: 1062-1070, 1951.
44. Smith HW: Principles of renal physiology. Oxford University Press (pp 25-35), 1956.
45. Churchill ON, Taylor W, Members of the CANUSA Peritoneal Dialysis Study Group:
Canada-USA (CANUSA) Multicentre Study of Peritoneal Dialysis Adequacy; Description
of the study population and preliminary results. Adv Peril Dial 8: 88-92, 1992.
46. Blake PG, Oreopoulos DG: Total creatinine clearance is not a good predictor of patient
outcome in CAPO JASN 2 (abstract): 358, 1991.
CHAPTER 8

Quality of life assurance in hemodialysis

A. PAUL HEIDENHEIM and ROBERT M. LINDSAY

Introduction

To physicians who are concerned with the impact that end-stage renal disease
and hemodialysis have on the day-to-day life and psychological well-being
of patients and their families, a major part of the "quality" in quality assur-
ance consists of patient quality of life.
The patient with end-stage renal disease faces a debilitating and disrup-
tive chronic illness, and a treatment regimen that is itself a complex, demanding
and incessant intrusion into one's personal and social life. Dialysis patients are
required, at the very least, to stoically endure pain, discomfort, fatigue and
deprivation on a daily basis, year after year. And most dialysis patients
heroically struggle to continue to fulfil their roles as spouses, parents, friends,
workers and community members even as they are nudged by the healthy
around them into a twilight world of the " ... marginal person - here today,
gone tomorrow [1]." Thus it is not surprising to find that aspects of quality
of life, often referred to as 'psychosocial factors', have been of concern in
nephrology for many years. 1
The measurement of quality of life has become an integral part of clinical
research in many areas. As technological and pharmacological advances have
increased the efficacy of modern medicine, concomitant research efforts have
gone beyond the assessment of mortality and morbidity to include the dimen-
sions of physiological, psychological and social functioning. Likewise, as a
growing proportion of the health care system is involved with the management
of long-term chronic diseases and conditions, the focus of research has widened
to include the patient's physical comfort, psychological health, social and
vocational activity, and the impact on the patient's family. In short, the extent
to which medical treatment is consciously directed to help patients live com-
fortable, productive and satisfying lives is the extent to which clinical research
into that treatment has included measures of patient quality of life.
In addition, there is a growing awareness among medical practitioners that
knowledge of, and sensitivity to, quality of life issues can contribute to
successful treatment outcome. For example, minimizing and ameliorating

L.w. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 133-149.
© 1994 Kluwer Academic Publishers.
134

negative iatrogenic treatment effects promotes patient compliance with diffi-

cult therapeutic regimens that require sacrifice and commitment. Likewise,
in chronic treatments such as dialysis, a downturn in some aspect of a patient's
quality of life may signal the beginning of a coping problem which, if iden-
tified early enough, can be defused with a minor strategically-focused
intervention before it escalates into a major problem requiring a significant
investment of professional time and resources. Apart from its practical
importance, our growing understanding of the dynamics of patient quality of
life within the context of disease and treatment is also contributing to our
theoretical knowledge of the interaction between the mind, body and social
environment. Consideration of patient quality of life leads into a holistic
conceptual framework where medical treatments per se are seen as inter-
woven into the web of the patient's entire life. Changes in any single area
of life have repercussions in many other areas; thus medical decisions will
be made not in isolation, but with an ecological view as to how they will affect
other aspects of the patient's life, and how they will be impacted in turn by
psychological factors and social forces.
The promotion of quality assurance with respect to patient quality of life,
then, is called forth by the humanistic ethic of the helping professions, and
is justified by its practical contribution to treatment efficacy and its poten-
tial contribution to scientific knowledge.

Achieving quality of life assurance in dialysis

It may be helpful to briefly sketch the outlines of how the issue of quality
assurance, from the point of view of patient quality of life, could be addressed
in a pragmatic way within a busy dialysis unit. The foundation would be the
ongoing collection of psychosocial data from patients in a regular and
systematic fashion, in much the same way that medical and physiological
data is routinely collected today. In order to minimize the burden on medical
staff, the bulk of this information could be collected through patient-completed
questionnaires. In centre dialysis units, for example, most questionnaires could
be filled out before or during dialysis sessions. The role of unit staff would
be limited, in most cases, to the distribution and collection of questionnaires
with a quick check of completion rates. In order to minimize the burden on
the patients, questionnaires need to be short, clear, and easy to read and answer.
Data collection could occur relatively infrequently, perhaps quarterly, with
the option of larger-scale assessments on an even less frequent cycle.
Scheduling would likely be done on an individual patient basis, keyed to the
data of admission to the unit. In order to maximize the utility of the infor-
mation gathered, the measures chosen should meet high standards of validity
and reliability; as enumerated later, a substantial number of accredited quality
of life instruments are, in fact, currently available.
The routine collection of quality of life information will generate a data base
 135

with several valuable applications. The individual dialysis unit will instantly
have an empirical snapshot of their caseload, a descriptive cross-section
captured in the range of scores, typical values, characteristics of particular
subgroups, and values associated with individual patients of interest. By
matching measurements from questionnaires with personal knowledge of
individuals, unit staff will immediately begin to make the rather esoteric
numbers derived from the quality of life scales more meaningful, with certain
types of problems, for example, becoming associated with specific ranges of
values on particular scales. In addition, through relatively simple statistics
the inter-relationships between the different aspects of quality of life, and
the relationships between quality of life and other characteristics, can be
investigated at the local level, in response to local interests and concerns. Once
a series of measurements have been taken, it becomes possible to track indi-
viduals, subgroups, and the caseload as a whole over time. With experience,
the emergence of individual problems can be followed and even anticipated
through change in a patient's scale scores, thus facilitating early preventa-
tive interventions. Likewise the effects of therapeutic interventions over time
can be assessed. As long-term data builds up, it will be possible to trace the
course of psychosocial adaptation to treatment from initial adjustments
onwards, and to relate individual or subgroup patterns to other characteris-
tics, therapeutic milestones, and events in the unit. In this way, high-calibre
quality of life data, collected systematically and regularly over time, could
provide valuable supplemental information useful in diagnostic and therapeutic
decision-making in the dialysis unit.
When many dialysis units are routinely collecting quality of life data, their
utility is further enhanced. Individual units can compare patient and perfor-
mance data on a regional, national and international basis. Data on patients
with rare conditions or problems can be combined for more powerful analyses.
Comparisons between different types of units - rural versus urban, short-hours
versus long hours, reusers versus single use units, for example - become
possible on an ad hoc basis, facilitating answers to questions that today can
only be addressed through expensive "extracurricular" research projects.
All of these potential benefits from quality of life assurance, however,
assume that patient quality of life can be assessed in meaningful, valid, reliable
and efficient ways. It is to these points that we now tum, beginning with matters
of conceptual definition, followed· by measurement issues.

Conceptual definitions

Quality of life

The term "quality of life" strikes a chord of familiarity; we all have an

intuitive sense of what is implied in terms of the expectations, satisfactions
and frustrations in our own personal lives. The generality and all-inclusive-
136

ness of the concept, however, means that each individual will define the quality
of their own life in a unique way, with different priorities and different weights
differentially assigned to the various aspects of personal existence. In order
for the concept to have utility in scientific research and clinical practice, this
multiplicity of personal meanings must be transcended through conceptual and
operational definitions which have universal applicability2 at the same time
as they accommodate the richness of individual variation.
The first step in delimiting the concept is to note that we are concerned here
with patient quality of life, as distinguished from the quality of life of a general
population. The latter notion comes from within the tradition of the general
social survey in the field of sociology, and is concerned with the broad
assessment of the nature and grading of life along such dimensions as attitudes
towards self and others, racial and ethnic groups, women, family life, work,
social and leisure activities, personal economic outlook, national economic
outlook, and various political issues [2-4]. While the notion of patient quality
of life takes a more functional orientation within a narrower focus, a great deal
of theory and methodology is imported directly from these large scale social
surveys.
The concept of patient quality of life can be derived almost directly from
the World Health Organization's definition of health as "[a] state of complete
physical, mental and social well-being and not merely the absence of disease
or infirmity [5]." When we consider the quality of life of a dialysis patient,
by definition disease is present. The central focus is thus the level of physical,
mental and social functioning of the individual in the presence of kidney
disease. Patient quality of life " ... represents the functional effect of an illness
and its consequent therapy upon a patient, as perceived by the patient [6]."

Domains and components of quality of life

The main dimensions of patient quality of life are embedded in the WHO
definition of health: physical functioning, social functioning and psycho-
logical functioning. While specific components of these dimensions may be
chosen to fit particular research questions or clinical concerns, there are several
aspects of functioning that are of primary interest and importance with dialysis
patients (see Fig. 1).

Physical functioning

The physical health, well-being and level of functioning of patients consti-

tutes the traditional purview of medicine, and a great deal of physiological and
treatment related data is generally available to the physician in the patient's
chart and dialysis records. However, there are several items that bear directly
on quality of life; some of which is found in traditional medical data, and some
of which is unique to the quality of life perspective. Of primary importance
 137

IQUALITY OF LIFE I
PHYSICAL II PSYCHOLOGICAL I SOCIAL I
- Physiological - Depression - Psychosocial
Functioning - Denial Stress
- Disease Stress - Anxiety - Social-leisure
- Dialysis Stress - Social Activities
- Sleep Adequacy Introversion - Social Support
- Fatigue - Self Depreciation - Marital/Family
- Hypochondriasis Functioning
- Vocational
Functioning

Fig. 1. Quality of life, dimensions and components.

is a measure of physiological functioning that can provide, through a single

index number, a sense of the overall health of the patient. In addition,
depending on the severity and chronicity of the ESRD and accompanying
conditions, an assessment of functional capacity may be helpful. This is
particularly so with respect to such factors as mobility and visual acuity
which are directly relevant to the self-care and home-care settings. Disease-
related stress and dialysis-related distress are concerned with the impact
of ESRD and its concomitant therapy on the life of the patient. The former
emphasizes the incidence and severity of uraemic and other symptoms, while
the latter concentrates on the occurrence of physical problems and the
concomitant level of discomfort during the dialysis procedure. Other aspects
of treatment regimen stress, such as dietary and fluid restrictions, many also
be of interest. Finally, systematic assessment of the adequacy of sleep, and
the level of fatigue experienced by the patient round out the picture of the
physical domain of quality of life.

Psychological functioning

The psychological dimension of patient quality of life reflects the individual's

inner experience of the illness and treatment. This will include the direct impact
of the physical symptoms on subjective well-being, the meaning and signifi-
cance to the person's self-concept and self-esteem of the many losses entailed
in a debilitating chronic disease, adjustments to the increased immanence of
the possibility of death, and the mobilization of defense and coping mecha-
nisms. Psychological functioning is thus both influenced by, and simultaneously
exerts an influence over, physiological functioning and health.
138

Depression is widely recognized as one potential psychological outcome

of uraemia and dialysis. Likewise, the importance of denial as a coping
mechanism of dialysis patients is generally acknowledged. Anxiety, particu-
larly when related to treatment issues, is also an aspect of psychological
functioning with practical significance. Finally, social introversion and self-
depreciation may provide important links between physical and social
functioning.

Social functioning

The strength and pervasiveness of uraemic symptomatology, and the com-

mitment of time and energy demanded by dialysis and other treatment
considerations, means that virtually every aspect of a patient's lifestyle changes
after kidney failure. Continuing crises in the roles of marriage partner and
parent will precipitate shifts in family responsibility and decision-making.
Many patients must change jobs, work reduced hours, or leave their employ-
ment altogether, placing further burdens on spouse and family. And there
are further changes to patterns of housework, diet, recreation and social
activities.
On the other hand, social support from family and friends is one of the
key resources that the dialysis patient can draw on for help in coping with
the demands and disruptions of ESRD and its treatment. To the extent that
the patient's social network can be protected, nurtured and strengthened, the
individual can be buffered from the more extreme exigencies of their
situation. One of the important aspects of social functioning will be
psychosocial stress, the stress experienced in day-to-day personal and family
life related to the symptoms of kidney disease and the demands of dialysis.
The level of marital and family functioning, including cohesion, conflict,
communication, affection, organization and control, is both a primary resource
for the patient and is itself vulnerable to disruption as a result of illness and
treatment. General social support, the patient's perceptions of the degree to
which he/she is loved, valued and nurtured by his/her overall social network,
gives an indication of the availability of this resource. A reduction in the quality
and/or quantity of social-leisure activities can signal the emergence of a deeper
problem in the patient's life, as can a significant reduction in the patient's level
of vocational functioning.
While the dimensions and components of quality of life outlined above
are certainly not exhaustive, they do define the boundaries of the concept
and give an idea of the depth of material that is relevant. As the interest in
quality of life assurance spreads to more and more dialysis units, it will be
important to build a consensus around the choice of components to be targeted
for measurement so that a meaningful sharing and comparing of data will be
possible. Equally important will be a consensus on measurement strategies.
 139

~easurennentissues

Measurement approach

There are two basic, complementary approaches to the measurement of patient

quality of life: global and disease-specific. Global assessments tend to rate
levels of overall functioning and/or satisfaction in general, abstract dimensions.
In contrast, disease-specific assessment tools focus on concrete symptoms,
stressors, dysfunctions and discomfort related to a particular disease and its
treatment [7].
Within the global approach there are both "utility measures" and "health
profiles". Utility measurements provide a single number representing an
measure of overall quality of life. For example, the QL-Index [8] scores five
items (activity; living; health; support; outlook on life) via 3-point scales (range
0-2) for an overall APGAR-type rating of from 0 to to. Likewise the time
trade-off approach generates a single number measure of quality of life by
dividing the patient's current life expectancy in years into the shortest life
expectancy for which the patient would settle in return for full health [9].
As health profiles, global instruments may also be long and detailed, such
as the Sickness Impact Profile [10] and the Psychological Adjustment to Illness
Scale [11].
The global method is most useful for assessments of a large, heteroge-
neous population, or when comparing different illnesses. The general approach,
however, is usually too insensitive to detect small but clinically significant
changes and/or differences, and thus is of limited value in evaluation research.
Disease-specific assessment is generally more suitable in clinical trials of
alternative interventions. Since questionnaire items have been chosen specif-
ically to fit the research questions at hand, the scales tend to be sensitive to
change and able to detect subgroup differences. On the other hand, since
different illnesses affect different bodily functions and tend to contribute to
different physical and emotional problems, disease-specific instruments usually
have a narrow range of applicability limited to the illness and treatment for
which they were designed [12].
In terms of ongoing patient quality of life assurance, both the global and
disease-specific approach to measurement would be valuable. Widely used
global measures would provide the basis for broad comparisons with other
treatment modalities or patient popUlations, while a selection of disease-specific
scales are required to facilitate preventative interventions, treatment evalua-
tions, and other projects narrowly focused on issues in hemodialysis.

Disease-specific instrumentation

The authors have developed and/or employed a number of disease-specific

instruments over the past 15 years. These have been utilized in various research
140

projects, including a study of adaptation to home dialysis in the province of

Ontario [13-18], a clinical evaluation of short-hours dialysis with the
polyacrilonile membrane [19-20], and an investigation into the effects of
rHuEPO and dialyser geometry on dialysis efficacy and patient quality of
life [21-22]. A selection of these instruments are briefly described below to
give the reader an idea of the kinds of tools that are available. With the
exception of the Physiological Index, which is compiled by medical staff using
unit records, the questionnaires are all completed by patients.
(1) Physiological index: The physiological index was developed form the
"Standardized List of Somatic Criteria" of M.R. Strauch et al. and consists
of 18 weighted biochemical measures selected for their relevance to an
ESRD population: hepatitis; renal osteodystrophy; urine volume; serum
urea, creatinine, blood pressure; haematocrit; weight increment between
dialysis; potassium; diabetes; myocardial infarct; fundi; congestive heart
failure; angina pectoris; pericarditis; GI ulcers; anephric. Original items
and weights were validated through feedback from nephrologists from 100
West German dialysis centres [23]. Physiological Index scores range from
7 to 94, with higher values indicating poorer physiological functioning.
(2) Psychosocial stress scale: This nine item index measures the amount
of stress experienced in day-to-day personal and family life related to
the symptoms of ESRD and the demands of dialysis: financial problems;
marriage strain; dependency; altered sexuality; interference with vacations,
vocation, childcare and social life; and loss of family role. The scale yields
a stressor count (range 0 to 9) of the number of individual stressors
experienced, and a total stress score (0 to 81). An "average per-item" stress
score can also be computed by dividing the total stress by the number
of stressors.
(3) Disease stress scale: A 14-item scale of distress experienced due to the
symptoms of chronic uraemia and ESRD: being weak and tired; short-
ness of breath; blood pressure problems; up/down health; headaches; being
physically ill; fluid retention; cramps; itch; fear of death; inability to sleep;
dizziness; sexual problems; and nausea. Additional items can be included
for individual research projects. For example, in a clinical trial involving
erythropoietin, three extra questions were appended: easily feel cold or
chilly; inability to perspire; aching in arms/legs. The stressor count (0
to 14) indicates the number of items affecting the person, and the total
stress score (0 to 126) represents the sum total of disease related stress
from all 12 sources. Average per-item stress can also be computed.
(4) Sleep adequacy: The sleep questionnaire contains 4 subscales, computed
by summing the component items: Interference with Sleep (range 0 to 27),
Sleep Disruptions (range 0 to 27), Fatigue upon Awakening (range 0 to
24), and Fatigue during the Day (range 0 to 30).
(5) Dialysis somatic symptoms distress scale: This an index of 11 physical
symptoms that commonly occur during hemodialysis: nausea; vomiting;
headache; muscle cramps; dizziness; tingling of extremities; itching;
 141

shivering; back pain; chest pain; and hypotension. Four additional items
were included for the EPO project: dyspnea; wheezing; perspiration; hives,
rash. Patients are scored via a Symptom Count (0 to 11) or the number
of incidents that occurred during the dialysis session, and via a Total
Discomfort score (0 to 99) equalling the sum total of distress reported
from all items. Average per-item stress can also be computed.
(6) Fatigue rating scale: This scale measures the mean level of fatigue
experienced by patients in the time period between dialysis sessions.
Scores can range from 0 to 9.00.
(7) Basic personality inventory: The full Basic Personality Inventory is a
psychological profile composed of 12 scales, each consisting of 20
True-False items, with a corresponding score range of from 0 to 20 [24].
While the orientation of the inventory as a whole is towards the
measurement of psychopathology, we have found six of the scales to be
appropriate for use with populations that are psychologically normal:
depression, denial, anxiety, social introversion, self-depreciation and
hypochondriasis.
The Somatic Symptoms Distress Scale is included as Figure 2 in order to
illustrate some of the characteristics of these instruments. The most efficient
format is the self-completed questionnaire: data can be gathered with minimal
additional work by clinical staff. In order to maximize the return rates and
completion rates it is important to word the instructions and questions in
simple, concrete terms, and to utilize a layout that patients with mild visual
impairments find easy to read and respond to. As can be seen in the SSDS,
we prefer to separate questions on the occurrence of symptoms/stressors from
questions on the level of discomfort/stress, as this seems to minimize missing
data. The 9-point response scale which we use for level of discomfort or
stress allows for a wide range of variability within the between patients, and
facilitates the sensitivity of the instrument to treatment differences and change
over time.
It has been our experience that when patients report a reduction, for example,
in the overall level of dialysis-related discomfort, this may be due to the fact
that they are affected by a fewer number of symptoms, or due to a lessening
of the level of discomfort for certain symptoms, or to some combination of
the two. Hence scoring the SSDS involves taking a symptom count (the number
of Yes's), computing a total discomfort score (by summing the level of
discomfort for individual items with No-symptom not present weighted 0), and
calculating an intensity score by dividing total discomfort by symptom count.
Subsequent analysis will involve an assessment of change/difference in these
three scale scores, as well as an examination of each individual items to
locate the individual aspects of dialysis affected.
142

Circle NO if you did not experience the symptom while dialysing.

Circle YES if you did experience the symptom and indicate the level of
discomfort.

DISCOMFORT SCALE
very
low low moderate
f---+---+I--+---II
2 3 4 5 6 7 8 9

Symptom Level of Symptom Level of

Present Discomfort Present Discomfort

Nausea No Yes
D Itching No Yes
D
Vomiting No Yes
D Shivering No Yes
D
Headache No Yes
D Back Pain No Yes
D
Muscle Cramps No Yes
D Chest Pain No Yes
D
Dizziness No Yes
D Hypotension
(Low blood pressure)
No Yes
D
Tingling of No Yes
D
extremities Other (specify):
D
Fig. 2. Somatic symptoms distress scale (SSDS).

Reliability and validity

The reliability of a scale refers to its consistency internally and its stability
over time. In a scale with high internal consistency, all component items are
tapping slightly different aspects of the same underlying phenomenon and
are thus highly intercorrelated. Internal consistency reliability is generally
expressed by Cronbach's Alpha, a conservative measure that ranges from 0
to +1. A value of 0.60 is taken by convention as the minimum level for research
purposes, while a value of 0.90 may by required for diagnostic applications.
Stability over time is evaluated by test-retest reliability using the correlation
 143

between scores obtained over a short period during which the measured
phenomenon is assumed not to have changed [25-29]. In a recent assess-
ment of the instruments listed earlier [22], all scales except the Physiological
Index (an inventory of largely unrelated items), attained Cronbach's alphas
between 0.60 and 0.88, while test-retest coefficients ranged between 0.38
and 0.92.
Validity centres on the question of whether a scale actually measures the
concept it is intended to measure. Validity assessment tends to be a contin-
uous process as new data and experience are accumulated under different
conditions. The validity of a scale is usually determined relative to the specific
use(s) for which it is intended; hence different types of validity assessments
correspond to different research applications. The concept "quality of life",
when used in clinical research, often takes the role of a secondary treatment
outcome supplemental to morbidity and mortality [30]. The validity assess-
ment of a quality of life instrument is thus normally concerned with
construct validity: the adequacy with which the scale represents the underlying
theoretical concept, namely a specific domain of quality of life. Generally
this is demonstrated by showing that the scale scores correlate in logical
directions with other variables. Again, the instruments listed above have been
shown to be correlated in expected directions with related variables, and have
successfully discriminated between treatment groups, at statistically signifi-
cant levels [22].
It should be noted that the use of quality of life measures to assess quality
assurance in dialysis units represents a new application for these instruments,
for purposes other than those typically associated with clinical research. Hence
the validity and reliability of the scales in the ongoing monitoring of patients'
coping with ESRD and dialysis, for establishing local, regional, national and
even international norms of patient functioning, and for providing supplemental
input into diagnostic and treatment decisions, remains to be established through
actual clinical applications.

"Subjectivity" in quality of life measurement

The measurement of patient quality of life is often criticized as being "unsci-

entific" on the grounds that it is subjective. The subjective component, present
in both global and disease-specific assessment, consists of having the patient
report his/her perception of the degree of discomfort, stress, palliation,
interference, etc. that he/she is experiencing. The concern is that because
different individuals will make this judgement on the basis of different
thresholds of discomfort and stress, and different expectations regarding
palliation and interference, the exact same external stimulus will prompt a
variety of different ratings. This, in turn, introduces a degree of random
variation into the data which reduces the reliability and consistency of the
measurements.
144

On the other hand, it is this subjective perception which is at the root of

the individual's quality of life. A person who perceives himlherself as suffering
will experience their treatment and make compliance-related decision accord-
ingly, regardless of any objective, external determination. Thus, while the
variability of individual self-ratings is an important consideration during
research design, the subjective element is the target of measurement in quality
of life assessment, and not something to be avoided. By focusing analyses
on gain scores, i.e. the amount of change within individuals over time, we
can reduce the interference due to variability between individuals while
preserving the essential subjective nature of the target concept.

Preliminary population norms

This section provides descriptive statistics on several of the disease-specific

scales described earlier, in order to illustrate the format of the actual hard data,
and to make an initial contribution to the development of population norms.
The data are drawn from two separate research studies [19-22] involving a
total of 300 hemodialysis patients from 27 dialysis centres in 8 countries in
Europe and North America (see Table I). The 173 males and 127 females
ranged in age from 18 Through 84 years, with a mean of 55.1, a median of
58, and a standard deviation of 14.9. years. Glomerulonephritis was the most
common primary renal disease (31.3%), followed by pyelonephritis (12.7%),
hypertensive nephrosclerosis (12.0%), polycystic kidneys (10.7%); and diabetic
nephropathy (5.7%). Time on dialysis ranged from one month to over 21
years, with a mean of 42.9 months, a median of 22 months, and a standard
deviation of 48.2. One hundred and thirteen patients (37.7%) were medically
high-risk, including 88 with cardiovascular problems, 23 insulin-dependent
diabetics, and 31 patients with other systemic conditions. The data reported
below comes from the initial baseline assessments of patients as they entered
their particular study, prior to any experimental treatment.

Table 1. Country of origin.

# patients %

Canada 69 23.0
Denmark 5 1.7
England 30 10.0
France 21 7.0
Germany 33 11.0
Italy 72 24.0
Sweden 26 8.7
U.S.A. 44 14.7
Total 300 100.0
 145

Table II reveals that interference due to disease and the demands of dialysis
with work and housework (vocation) and vacations, are the most common
complaints affecting 39% of respondents. Interference with social life, and
financial and sexuality problems are also relatively common. The mean stress
for an item is a measure of the seriousness of the complaint. Interference
with vacation, sexuality issues, vocational interference, social interference,
financial problems and childcare interference all generated average levels of
stress, for those affected by the item, at levels towards the high end of the
scale. Data on the scale scores indicates that 78.4% of these patients were
affected by at least one item of the scale, with a sample-wide average of
2.37 items; the mean total stress level is 12.32 in a possible range of 0
through 81.

Table II. Psychosocial stress.

Item % affected Mean stress Std dey

Financial problems 29.3 5.12 2.01

Marriage strain 12.3 4.50 2.33.
Dependency 28.9 4.95 1.96.
Sexuality 27.5 5.55 2.48
Vacation interference 39.4 5.58 2.33
Vocation interference 39.4 5.25 2.17
Loss of family role 14.3 4.80 2.26
Childcare interference 12.2 5.06 2.57
Social interference 34.1 5.13 2.00
Scales

Stressor count 78.4 2.37 2.08

Total Stress 12.32 12.26
Per-item Stress 3.94 2.57

The primary disease-related stressor is feeling weak and tied, which affects
three quarters of the sample, although the actual amount of stress this causes
is moderate (Table III). The general health items (up and down health, being
physically ill) affect half the sample, with itch, blood pressure problems,
fluid retention and sleep disturbance also relatively common. For those
affected, the worst stressor is the impairment of sexual functioning; the mean
stress level for sleep problems, being weak and tired, fear of death, fluid reten-
tion and up/down health also fall towards the high end of the response scale.
Almost every patient (97%) reports at least one disease-related stressor, with
the average being 6 symptoms; the overall mean total stress is approximately
30 in a range of 0 to 126.
Meaningful Interpretation of BPI scores requires some standards from
known groups. Several population norms are presented in the scale manual
[24]. In general, dialysis patients tend to score higher than the population at
large on denial (a constructive defense mechanism), depression and hypo-
146

Table lll. Disease related stress.

Item % affected Mean stress Std dev

Weak and tired 75.4 5.34 2.01

Shortness of breath 37.4 4.86 1.94
Blood pressure problems 42.6 4.96 2.02
Up and down health 50.2 5.14 1.84
Headache 28.8 4.02 2.22
Being physically ill 50.0 4.76 2.17
Fluid retention 42.8 5.16 1.92
Cramps 40.9 4.18 2.18
Itchiness 47.0 4.85 2.48
Fear of death 25.6 5.21 2.30
Sleep problems 43.5 5.37 2.35
Dizziness 22.4 3.67 1.98
Sexual functioning 25.0 6.28 2.19
Nausea 19.6 4.29 2.53
Scales

Stressor count 96.9 6.05 3.24

Total Stress 29.97 20.80
Per-item Stress 4.60 1.69

chondriasis, and close to the general population on the other measures (Table
IV).
The most commonly reported symptoms during dialysis are itch, hypoten-
sion and muscle cramps, while itch and back pain cause the most severe
discomfort; however, the mean level of stress for all items are towards
the low end of the response scale (Table V). Ninety percent of patients
experience at least one physical symptom during dialysis, with an average
of 2 symptoms causing a mean total stress of 8.15 within a potential range
of 0-99.
It is interesting to compare the three stress/distress scales via the mean
per-item score for the scale as a whole: disease stress seems to be the most
severe intrusion into patient's lives (4.60) followed by psychosocial stress
(3.94) and dialysis distress (3.80). It should also be noted that the standard

Table IV. Basic personality inventory.

Range Mean Median Std dev

Depression 0-18 6.41 5 4.58

Denial 1-18 10.25 11 3.21
Anxiety 0-20 6.99 6 4.10
Social introversion 0-18 5.96 5 3.96
Self depreciation 0-15 3.62 3 2.94
Hypochondriasis 0-18 7.89 8 3.92
 147

Table v. Somatic symptoms distress.

Item % affected Mean Std dev

Nausea 21.3 2.27 1.73

Vomiting 13.9 2.03 1.61
Headache 34.4 2.72 1.94
Muscle cramps 44.0 3.06 2.08
Dizziness 23.1 2.30 1.87
Tingling extremities 19.9 2.48 2.07
Itching 45.7 3.37 2.18
Shivering 14.6 1.83 1.50
Back pain 23.4 3.32 2.05
Chest pain 11.7 1.88 1.44
Hypotension 44.7 2.87 2.01
Symptom count 90.4 2.04 1.77
Total distress 8.15 8.71
Per-item distress 3.80 1.60

deviations in the quality of life scales tend to be large, an indication of the

high degree of individual variation. These examples demonstrate first of all
that it is possible to meaningfully quantify different aspects of patient quality
of life, and secondly that a variety of comparative analyses can be under-
taken to furnish the clinician with helpful information. Comparisons of an
individual with other patients, comparisons of an individual's values with
values at other points in time, comparisons among groups of patients, and
comparisons of a unit's average values with these from other units or the
region/country are all possible when quality of life data is routinely col-
lected. Potentially, such analyses can inform diagnostic decisions, provide data
for hypothesis testing, and facilitate the evaluation of new equipment and
alternative therapies.

Conclusions

The chronic nature of end-stage renal disease, the extensive symptomatology

of uraemia, and the invasive character of the hemodialysis treatment regimen
constitute an enormous obstacle to the maintenance of a high quality of life
for renal patients. The challenge of quality of life assurance in dialysis is to
overcome this obstacle. The course of action we propose is to take our already
considerable skill and expertise at research, which is usually cloistered in
the specialized world of clinical trials, and to integrate it with ongoing clinical
practice in a synergistic methodology.
Step one involves identifying a selection of global and disease-specific
measures of patient quality of life that meet high standards of efficiency,
reliability and validity; many proven tools exist today, some may have to be
developed. Step two is the systematic collection of quality of life data on a
148

regular basis. This data collection is not part of an external research project,
but is added to the routine record keeping of the dialysis unit. Step three
consists of integrating the analysis of quality of life data into the flow of
therapeutic decision making within the unit - the screening of patients into
appropriate treatment modalities, the identification and diagnosis of
psychosocial problems, the choice of intervention, and the deployment of
preventative measures. The sharing of data between centres constitutes
step four. Through this interchange, individual units can compare caseload
characteristics and experiences at an empirical level. Of even greater poten-
tial is the creation of a growing quality of life database that can facilitate
the investigation of unusual problems and special populations, and can support
general research into the interrelationship between physical, psychological and
social functioning, and the determinants of high quality of life for dialysis
patients.
Quality assurance in dialysis represents the consolidation of ethical, prac-
tical and scientific goals in a unified perspective. The promise of quality of
life assurance is this: through the careful and systematic attention to dialysis
as delivered and dialysis as received, medical practitioners have a signifi-
cant contribution to make not only to the quantity of years remaining for the
renal patient, but also to the quality of that time.

Notes

1. The earliest paper known to the authors is: Schreiner GE. Mental and personality changes
in the uraemic syndrome. Med Ann DC 28: 316-323, 362, 1959.
2. Universally applicable in the sense of generally meaningful within the bounds of western
industrial culture; the term assumes a shared world-view and basic commonality of values
that would not be transferrable between major cultures.

References

1. Calland CH: Iatrogenic problems in end-stage renal failure. New Engl J Med 287: 334,
1972.
2. Campbell A, Converse PE, Rodgers WL: The Quality of American Life: Perceptions,
Evaluations, and Satisfactions (New York: Russell Sage Foundation), 1976.
3. Converse PE, Dotson JD, Hoag WJ, McGee WH III: American Social Attitudes Data
Sourcebook, 1947-1978 (Cambridge, Mass: Harvard University Press), 1980.
4. Campbell A: The Sense of Well-Being in America: Recent Patterns and Trends (New York:
McGraw-Hill), 1981.
5. Fries JF, Spitz PW: The hierarchy of patient outcomes. In Spilker B (ed) Quality of Life
Assessments in Clinical Trials (New York: Raven Press, Ltd.), pp 25-35, 1990.
6. Schipper H, Clinch J, Powell V: Definitions and conceptual issues. In Spilker B (ed) Quality
of Life Assessments in Clinical Trials (New York: Reven Press, Ltd.), pp 11-24, 1990.
7. Guyatt GH, Bombardier C, Tugwell PX: Measuring disease-specific quality of life in clinical
trials. CMAJ 134: 889-895, 1986.
8. Spitzer WO, Dobson AJ, Hall J et al.: The QL-index. J Chronic Dis 34: 585-597, 1981.
9. Churchill DN, Torrance D, Taylor DW, Barnes CC, Ludwin D, Shimizu A, Smith EKM:
 149

Measurement of quality of life in end-stage renal disease: The time trade-off approach.
Clin Invest Med 10: 14-20, 1987.
10. Bergner M, Bobbit RA, Carter WB et al.: The sickness impact profile. Med Care 19:
787-805, 1981.
11. Morrow GR, Chiarello RJ, Derogatis LR: A new scale for assessing patients' adjustment
to medical illness. Psychol Med 8: 605-610, 1978.
12. Guyatt GH, Jaeschke R: Measurements in clinical trials: Choosing the appropriate approach.
In Spilker B (ed) Quality of Life Assessments in Clinical Trials (New York: Raven Press,
Ltd.), pp 37-46, 1990.
13. Lindsay RM: A comparison of CAPO and haemodialysis in adaptation to home dialysis.
In Moncreith J, Popovich R (eds) CAPD Update - Continuous Ambulatory Peritoneal
Dialysis (Masson, NY: Modern Problems in Kidney Disease), pp 171-179, 1981.
14. Lindsay RM, Oreopoulos 0, Burton H, Conley J, Richmond J, Wells G: The effect of
treatment modality (CAPO vs. haemodialysis) in adaptation to home dialysis. In Atkins
RC, Thomson NM, Farrel PC (eds) Proceedings of the Pan Pacific Conference on Peritoneal
Dialysis (Edinburgh: Churchill Livingstone Press), pp 385-394, 1981.
15. Lindsay RM, Richmond J, Burton H, Conley J, Clark WF, Linton AL: Is home dialysis
too stressful? Controversies in Nephrology III, 395-405, 1981.
16. Wai L, Richmond J, Burton HJ, Lindsay RM: The influence of psychosocial factors on
survival of home dialysis patients. Lancet 2: 1155-1156, 1981.
17. Richmond JM, Linday RM, Burton HJ, Conley J, Wai L: Psychological and physiological
factors predicting the outcome on home haemodialysis. In Cluthe R (ed) Clinical Nephrology
(Frieburg, West Germany), pp 109-113, 1982.
18. Burton HJ, Kline SA, Lindsay RM, Heidenheim AP: The relationship of depression to
survival in chronic renal failure. Psychosomatic Medicine 48(3/4): 261-269, 1986.
19. Lindsay RM: Comparison of patient quality of life in short dialysis and conventional dialysis.
In Man NK, Mion C, Henderson LW (eds) Blood Purification in Perspective: New Insights
and Future Trends. (ISAIO Press), pp 203-204, 1987.
20. Lindsay RM, Heidenheim AP, Spanner E, Burton HJ, Lindsay S, LeFebvre JMJ: A multi-
centre study of short hours (SH) dialysis using AN69S: Preliminary results. ASAIO
Transactions 37(3): M465-M467, 1991.
21. Heidenheim AP, Lindsay RM: The impact of erythropoietin and Dialyzer geometry on patient
quality of life. Submitted to Clinical Nephrology
22. Lindsay RM, Heidenheim AP: Measurement of quality of life in international trials involving
dialysis patients and the use of recombinant human erythropoietin. In Bauer, Koch, Scigalla,
Wieczorek (eds) Clinical Applications of Erythropoietin (New York: Marcel Dekker Inc.),
pp 81-921, 1993.
23. Strauch MR, Lipke R, Schafheutle R, Nachbauer B, Stauch-Rahaiiser G: A standardized
list of somatic criteria for comparative assessment of regular dialysis therapy patients.
Artificial Organs 2 (Suppl): 370-372, 1978.
24. Jackson OJ: The Basic Personality Inventory (Port Huron, Michigan: Research Psychologists
Press), 1976.
25. Nunally JC: Psychometric Theory 2nd ed. (New York: McGraw-Hill), 1978.
26. Carmines EG, Zeller RA: Reliability and validity assessment (Beverly Hills: Sage
Publications), 1979.
27. Selltiz C, Wrightsman LS, Cook SW: Research methods in social relations 3rd ed. (New
York: Holt, Rinehart, Winston), 1976.
28. Helmstadter GC: Principles of Psychological Measurements (New York: Appleton-Centory-
Crofts), 1964.
29. Lemke E, Wiersma W: Principles of Psychological Measurement (Chicago: Rand McNally),
1976.
30. Spilker B: Introduction. In Spilker B (ed) Quality of Life Assessments in Clinical Trials New
(York: Raven Press), p 3, 1990.
CHAPTER 9

Dialyzer reuse and the quality of therapy

L.W. HENDERSON

Introduction

Reuse, since its early introduction by Shaldon [1] as a way to make treat-
ment more economical, has been at the center of controversy. As of this writing
there is a good deal more emotion than fact available on this matter. The issues
that enter the debate are both economic and scientific. I shall not provide a
detailed exploration of them all but rather reference the interested reader to
the source material. I do this as it is clear that one cannot, with the present
incomplete (and conflicting) information, make a clear statement about the
most fundamental concern about reuse; namely, the impact on treatment as
measured by comparative mortality rate. This is a sad commentary on both
U.S. regulatory agencies and North American nephrology, given that in 1990
reuse in one form or another was employed in 75% of the approximately
110,000 patients on treatment with hemodialysis in North America. About half
the reused dialyzers were disinfected manually and half with automated
systems. Formaldehyde was used in 54%, peracetic acid-hydrogen peroxide
in 39%, and glutaraldehyde in 7%.
There are many questions about-reuse. For example, when considering reuse
of cellulosic membrane (Le., the cheapest and most commonly used membrane):
- Does the transport performance of the dialyzer deteriorate with reuse and
if so how may it best be addressed quantitatively? [2]
- Does a reused dialyzer that is "scrubbed fully clean" with bleach have meta-
bolic disadvantages to the user due to restoration of complement activation?
[3]
- Does the new dialyzer contain sufficient manufacturing "by products" e.g.
particulates, cotton linters, residual ethylene oxide, etc. to warrant some
form of pre treatment rinse and, if so, should it contain protein to reduce
compliment activation? [4, 5, 6]
- If a small amount of denatured protein is allowed to remain on the
membrane following reuse in order to reduce complement activation, will
it act as an antigen when it finds its way into the body with subsequent
untoward effects? The reported "reuse syndrome." [7]

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 151-154.

© 1994 Kluwer Academic Publishers.
152

For the synthetic membranes that in general are far more active adsorbers of
plasma constituents than the cellulosic membranes:
- Does bleach, the preferred agent to remove all adsorbed protein, etch the
very thin discriminating skin of an asymmetric membrane like polysulphone
so that protein leak results? [8]
- Does failure to use bleach in these protein adsorbing membranes put the
patient at risk from repeated small infusions of potentially antigenic
("foreign") protein denatured by whatever other sterilant is selected?
- Does adsorption of useful plasma proteins that are in short supply (e.g.
erythropoietin, insulin, other?) trade off positively or negatively with the
concern about not using a bleach reprocessing solution? [9]
For all membranes:
- Does the repeated minor re infusion of the chosen sterilant have a long term
morbid or mortal outcome? e.g. cancer.
- Does the patient prevail in his or her wish not to participate in a reuse
program if the economic circumstances of the unit deteriorate as a result?
These many questions become moot if the practice of reuse results in higher
mortality. For this central question we have reports from the Urban Institute
using U.S. Department of Health and Human Services, Health Care Financing
Agency Medicare (HCFA) records and the National Institutes of Health (NIH)
identifying a reduced survival with certain forms of reuse. These are power-
fully large data bases that must command our attention. As these studies
have not, as yet, been published, I am going to offer a summary taken from
recent discussions, presentations and draft documents on the subject.

The HCFA study

For the Urban Institute study data from 1,121 free-standing dialysis units (859)
from units using membranes with low hydraulic permeability like curpo-
phane membrane) representing 85% of all free-standing units in 1990 were
examined (i.e., approximately 45,000 patients for 1990 (some 34,000 using
conventional low-flux dialyzers) and a similar number for 1989.
• Patients treated in free-standing dialysis units that employ conventional
non-reused dialyzers had lower mortality on average than did patients treated
in units that used selected germicides for dialyzer reprocessing (analysis
comparing units).
• Patients treated in dialysis units that used peracetic acid-peroxide or
glutaraldehyde as a disinfectant experienced higher mortality, on average
than did those patients treated in dialysis units that used low flux dia-
lyzers that were not reused. For peracetic acid-peroxide manual systems,
the mortality was higher by 15 percent; for Renalin automated systems,
the mortality was higher by 11 percent; for glutaraldehyde systems (both
manual and automated) the mortality was higher by 15 percent, on average
(analysis comparing patients).
 153

• Patients treated in dialysis units that used formaldehyde as a disinfectant

had a mortality rate that was similar to the reference group of no reuse.
• While there were differences across the mortality averages by germicide
subgroup, there was substantial variation in mortality within each of the
germicide subgroups. In each of the germicide subgroups that were some
units that had mortality estimates that were lower than the average of the
reference group (no reuse conventional dialyzers).
• There were no systematic increases in mortality associated with manual
of automated systems regardless of the disinfectant used. In other words,
the method of reuse was not statistically different than no reuse, other factors
held constant.

The NIH study

This study was a "case mix" study which matched approximately 3,100 reuse
and non-reuse patients from 280 dialysis units and "arrayed them according
to a number of medical and demographic characteristics." They showed
increased mortality only when dialyzers were disinfected manually with
peracetic acid-hydrogen peroxide. Automated use did not result in higher
mortality, nor did use of formaldehyde.
In summary, there are internal conflicts in the two studies. Both are retro-
spective and draw from a large sample size. The relative risk of death citation
is sufficiently small so as to be transparent to any given unit director. There
is sufficient variability in the data between units reporting their results to
raise questions in my mind as to the clinical significance of these small average
differences. To further add to the confusion, as earlier study from the USRDS
data base showed a reduced relative risk of death (0.88) for large dialysis
units that had been long term reusers, as contrasted with smaller units that
did not reuse (patient number = 4661) [9].

Conclusions

We must conclude that there is a crying need for more information in the
reuse domain in order to upgrade the quality of the therapy we offer. A specific
set of recommendations for safe reuse is not appropriate at this time, nor is
a recommendation for or against this practice.

References

1. Shaldon SR, Silva R, Rosen SM: Technique of a refrigerated coil of preservation hemo-
dialysis with femoral venous catheterization. Brit Med J 2: 411, 1964.
2. AAMI: Recommended practices for reuse of hemodialyzers (Arlington, VA, Association for
the Advancement Instrumentation), 1986.
154

3. Henderson LW, Chenoweth DE: Immune response to reuse: anaphylatoxins and IgE. In
Deane N, Wineman RJ, Bemis JA (eds) Guide to Reprocessing of Hemodialyzers (Boston:
Martinus Nijhoff Publishers), p 151, 1986.
4. Ogden DA: New dialyzer syndrome. New England Med. 302: 1262, 1980.
5. Ogden DA: Clinical response to new and reprocessed dialyzers. In Deane N, Wineman
RJ, Bemis JA (eds) Guide to Reprocessing of Hemodialyzers (Boston: Martinus Nijhoff
Publishers), p 87, 1986.
6. Pearson FC, Bohon J, Lee, W et 01.: Comparison of chemical analyses of hollow-fiber
dialyzer extracts. Artif Organs 9: 291-298, 1984.
7. Yudis M, Sirota RA, Stein HD: Dialyzer hypersensitivity reactions associated with reuse.
Am Soc Artif lnt Organs 14 (Abstr): 59, 1985.
8. Donahue PR, Ahmad S: Dialyzer permeability alterations by reuse. 1 Am Soc Neph 3 (Abstr):
363, 1992.
9. Cheung AK, Hohnholt M, Leypoldt JK: Hemodialysis membrane biocompatibility: The case
of erythropoietin. 1 Blood Purif9: 153-164, 1991.
10. Held PJ, Pauly MV, Diamond L: Survival analysis of patients undergoing hemodialysis.
lAMA 257: 645-650, 1987.
CHAPTER 10

Multicenter trials as a measure for improving

the quality of clinical decisions

DAVID N. CHURCHILL

The terms quality assurance and multicenter clinical trials encompass many
elements and the relationships between these elements are complex. The
principles of quality assurance can be applied to the research methodology
used in clinical trials. In turn, the results of these clinical trials can be used
to establish standards for clinical care. Finally, the efficacy of the applica-
tion of those standards in improving clinical care can be evaluated in the
context of a clinical trial.
The term quality assurance has been variously defined. Two definitions
appear to capture the spirit of this concept [1]. The first states that "quality
assurance is a program of systematic evaluation to ensure excellence in health
care". The second is that "the essence of quality assurance consists of the
establishment of standards, evaluation of those standards and implementa-
tion of actions that ensure adherence to those standards".
Use of the term standard, in turn, requires definition. Again, there are several
definitions. The American Nursing Association defines a standard as "an
agreed upon level of excellence; an established norm". Another definition is
"the level of performance considered acceptable by one having authority in the
situation or by those having authority in maintaining such performance levels".
Dependant on the objectives of the group developing the standards, the standard
may represent minimal expectations or it may reflect optimal performance.
A variety of similar but not synonymous terms have been used to describe
the concept of standards. These include standards, guidelines and clinical
practice parameters. These terms each have an inherent sense of flexibility
or, conversely, rigidity.

Quality assurance applied to clinical trials

The title of this section implies that standards exist and can be applied to
clinical trials. A recent analysis of the process by which guidelines or
standards are generated illustrates the heterogenous nature of that process
[2]. Differences exist in the definition of objectives, methods of guideline
development, implementation and evaluation.
L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 155-166.
© 1994 Kluwer Academic Publishers.
156

Rather than replicate the efforts of Audet et al. [2] and review the process
of guideline or standards development for clinical trials, I have selected one
representative example. The members of the Department of Clinical Epidemi-
ology and Biostatistics at McMaster University in Hamilton have developed
a set of guidelines for the critical appraisal of clinical articles advocating a
specific treatment [3].
The objective defined was to provide, for the reader of the results of clinical
trials, a set of guidelines for the critical appraisal of data published in the
medical literature. The objective, applying the criteria of Audet et al. [2], is
broad rather than specific. A broad objective makes evaluation of the process
difficult.
The method of guideline development was similar to the second model
described by Audet et al. [2]. The process relied on internal group judge-
ment with little use of outside experts to review and evaluate the pertainent
literature. The group consisted of the members of the Department of Clinical
Epidemiology and Biostatistics (clinicians, research methodologists and
biostatistians). The guidelines developed were submitted for peer review
followed by publication.
Implementation involves dissemination, training of individuals in the proper
use of the guidelines and monitoring. Dissemination has been internal and
external. Internal dissemination has been accomplished by incorporation of the
principles of critical appraisal in both the undergraduate and postgraduate
programs in the Faculty of Health Sciences at McMaster University. External
dissemination has been through the Annual Workshop on How to Teach Critical
Appraisal offered since 1981.
Linton and Peachey [4] stated that "Physicians like to believe that they make
logical clinical decisions based on analysis of known data, but this is often
not the case. Clinical policies have been developed on an ad hoc basis from
the considerations and decisions of hundreds of physicians acting individually.
Textbooks and medical journals publish this fragmentary information, and other
contributions come from continuing medical education, institutional policies,
the example of influential physicians and the often biased information provided
by pharmaceutical detailing". They also state that "honest self-evaluation would
suggest that the individual physician data base and analytic abilities are strained
even by apparently trivial clinical problems".
The standard, defined as an informed critical appraisal of published medical
literature pertaining to a specified clinical problem, incorporates the concept
that the published article must be scientifically credible and that the results
be applicable to the clinical problem which stimulated the physician to evaluate
this article. The former deals with the internal validity of the article, the
latter with external validity or generalizability. There are 6 questions or guides
for the physician appraising the article [3]. These evaluable components of
the standard are the criteria (Table I).
The first criterion is that the study deal with human subjects who have
been truly randomly allocated to one of several specified interventions. This
 157

Table I. Criteria to evaluate a publication addressing the efficacy of a clinical intervention [3].

Scientific credibility
1. Were the subjects truly randomly allocated to the specified treatments?
2. Were all clinically relevant outcomes reported?
3. Were both clinical and statistical significance considered?
Clinical applicability
4. Were the study patients recognizable?
5. Was the clinical intervention described in sufficient detail to permit replication in clinical
practice?
6. Were all study patients accounted for at the conclusion of the clinical trial?

is considered the gold standard for clinical trials. The methodologic weak-
nesses inherent in less rigorous designs should relegate them to use in two
circumstances. The first is when a randomized controlled trial is not possible
and with a clear statement of the biases in the design. The second is to generate
hypotheses to be tested using the randomized clinical trial methodology. With
the exceptions stated, the randomized clinical trial design is necessary but may
not be sufficient in that the subsequent 5 criteria must also be satisfied.
Before proceeding to the other criteria, the alternative research designs and
their weaknesses will be considered.
The clinical trial hierarchy starts with the case report and the case series.
Observations from these studies may generate hypotheses for more rigorous
examination. Cross-sectional surveys permit examination of relationships
between variables and permit hypothesis generation. The case-control study
design is highly subject to bias [5] with different studies often reaching
different conclusions for the same research question. A more common research
design compares clinical outcomes in a treatment group after an intervention
to a group which has not had the intervention. These can be further divided
into a one group design in which the outcomes in the treatment group are
compared to historical controls, a one group design in which outcomes before
and after the iintervention are compared and finally a design in which the
outcomes before and after the intervention are compared to outcomes in a
contemporaneous group which has not received the intervention [6, 7].
The one group before-after intervention design has been used extensively
in the evaluation of pharmacologic therapy for prevention of recurrent kidney
stones. The biases inherent in this design include co-intervention, regression
to the mean, systematic measurement error and natural history of the disease.
These biases will almost always favour the intervention under evaluation and
might lead the investigator to conclude that the intervention was effective when
it was not [6, 7]. The use of a contemporaneous but non-equivalent compar-
ison (i.e. not randomly allocated) group strengthens the research design but
introduces a selection bias [6, 7]. One can attempt to statistically adjust for
prognostic factors which are potential determinants of clinical outcomes and
which are unequally distributed between groups. Unfortunately, there may
158

be unrecognized confounders. This problem is best prevented by random

allocation to treatment groups.
The second criterion is that "all clinically relevant outcomes be reported".
In a randomized trial comparing clofibrate to placebo in the prevention of
coronary artery disease, clofibrate was effective in reducing both serum
cholesterol levels and myocardial infarction. Failure to determine the detri-
mental effect on total mortality would have lead to an erroneous conclusion
regarding the effectiveness of this drug [8]. The recent and important emphasis
on the quality of life as well as the quantity of life is well illustrated in the
example of women receiving adjuvant therapy for breast cancer [9].
The third criterion is that both clinical and statistical significance be
considered. Clinical significance refers to the importance of the change in
clinical outcome. The magnitude of the change which is considered clini-
cally important is one which is large enough to affect clinical behaviour. There
may not be agreement among clinicians regarding the change considered
clinically important but the range and distribution of those opinions can be
determined. The statistical significance refers to the probability that the results
are true. The probability of a false positive result (i.e. finding a difference
or effect of an intervention when there is no effect) is referred to as a Type
I error. The conventionally acceptable Type I error is 5% or p < 0.05. The
probability of a false negative result (i.e. finding no difference or effect of
an intervention when there is an effect) is referred to as a Type II error. The
conventionally acceptable Type II error is 0.10-0.20. The complement of the
Type II error is statistical power and this refers to the probability that the study
would detect a clinically important difference if such a difference existed.
The methods section of the study should define the Type I and II error
probabilities, the clinically important difference to be detected and the expected
outcome rate for the placebo or conventional treatment group. Provision of
these data permits an evaluation of the planning for the study. Simply reporting
statistical significance is no longer considered adequate. An efficient manner
of presenting both clinical and statistical importance is to use 95% confi-
dence limits. This approach has been recommended by the Annals of Internal
Medicine and other journals [10].
The fourth criterion is that the study patients be recognizably similar to
the patient(s) about whom the question has been asked. There are 2 compo-
nents to this criterion. The patients in the study must be well described and
they must be similar to the patients in question.
The fifth criterion is that the clinical intervention be available and be
described in sufficient detail to permit replication in clinical practice.
The sixth criterion is that all the patients who enter a study should be ac-
counted for at the conclusion of the study. If there are missing patients, this
might reflect general sloppiness on the part of the investigators or it might rep-
resent systematic error or bias. An arbitrary solution is to assume a bad outcome
for all missing members of a group and to recalculate the results with this
assumption. If the conclusion is not altered, the study results are acceptable.
 159

These criteria were developed to allow an individual clinician to critically

appraise an article addressing a specific clinical problem. However, there
will usually be more than one publication addressing a given clinical problem.
To be comprehensive, the clinician should perform a literature search using
a computerized data base. If there are several randomized clinical trials, the
physician should read all critically. If there are differing conclusions, the reader
must apply appropriate weights to each publication in order to reach an overall
conclusion. If there are no randomized clinical trials, the reader must apply
critical appraisal criteria to these articles and come to a weighted decision in
the same manner as for the randomized clinical trials. The obvious logistic
difficulties for the active clinician will prevent widespread use of this approach.
An alternative is to have standards or guidelines for clinical care recom-
mended by groups perceived by clinicians to be credible. Among the 3 models
described by Audet et ai. [2] was the Clinical Efficacy Assessment Project
developed by the American College of Physicians [11]. This consisted of a
thorough review of the relevant literature by "methods experts" followed by
the development of a position paper with the help of "content" experts. After
an extensive external review process, the position paper might be published
in the Annals of Internal Medicine. It would be reasonable to consider a similar
process for the establishment of standards in the field of Nephrology.

Standards established by clinical trials

There are few formally developed clinical standards in the field of Nephrology.
There are, however, generally accepted treatment practices and these gener-
ally accepted approaches may be considered informal standards or guidelines
for medical practice. Although an explicit application of the criteria evaluating
clinical trials addressing these generally accepted treatment practices was
not incorporated into development of standards, an implicit evaluation process
probably did occur.
What are some generally acceptable treatments? For renal transplantation,
the use of cyclosporine is an effective and widely accepted intervention. For
hemodialysis, the use of prescriptive dialysis based on urea kinetic model-
ling is widely accepted. In the area of continuous ambulatory peritoneal dialysis
(CAPD), the concept of flush before fill either with a disposable disconnect
system or with a disinfectant-multiple use system is considered superior to
conventional systems.
For the transplantation example, a randomized clinical trial of cyclosporine
in cadaveric renal transplantation was performed by the Canadian Multicentre
Transplant Study Group [12]. In that multicentre trial, 209 recipients of
cadaveric renal transplants were treated either with cyclosporine and pred-
nisone or with standard therapy which included prednisone and azothiaprine.
Graft survival at 1 year was 80.4% in patients receiving cyclosporine and
64.0% in patients receiving standard therapy. The authors concluded that
160

cyclosporine was preferable to azothiaprine in preventing renal transplant

rejection. The standard of scientific credibility and clinical applicability can
be evaluated by the application of the 6 criteria described earlier.
Were the subjects truly randomly allocated to the specified treatments? There
was a balanced restricted randomization according to treatment center. A
randomized block of varying size was used. There was no difference between
the groups at randomization with respect to variables considered of possible
prognostic significance. These were age, sex, race, diabetes mellitus, time
on dialysis, number of transfusions, levels of cytotoxic antibodies, time since
last transfusion, rehabilitation status, blood type, HLA matches and mis-
matches, donor age, donor sex, donor blood type, total kidney storage time,
rewarming time, perfusion time and transfusions during operation. The only
variable which was not balanced between the groups was the number of patients
receiving second or later transplants.
Were all clinically relevant outcomes reported? Both patient and graft
survival were reported. They also reported the number of rejection episodes,
graft function at 30 and 90 days post transplantation, incidence of post
transplantation acute tubular necrosis and complications (deaths, cancer,
infections and non-infectious complications). The criticism that these repre-
sented short term results was met by a follow-up publication in which the
results at 3 years were presented [13].
Were both clinical and statistical significance considered? The authors
explicitly define the clinically important difference as 20% with the expected
graft survival rate in the standard treatment group predicted to be 55% at
1 year. The actual graft survival rate in the conventional treatment group
was 64.0% compared to 80.4% in the cyclosporine group, a difference of
16.4%. This was clearly considered important by the readers of the publica-
tion in that it changed clinical behaviour. The Type I error probability was
set at 0.05 (two-tailed) and the Type II error probability at 0.10 for a statis-
tical power of 90% to detect a difference of 20% in one year graft survival.
The analysis examines the effect of cyclosporine as compared to standard
therapy on graft survival but also considers the possible effect of other
prognostic factors which might be determinants of outcome. Among the factors
chosen for examination, long perfusion times and long re-warm times had
a clinically important and statistically significant detrimental effect on graft
survival among the cyclosporine treated patients. On the other hand, receiving
5 or more blood transfusions improved graft survival among patients receiving
standard therapy. However, after multivariate adjustment for these risk factors,
the effect of cyclosporine was still statistically significant.
Were the study patients recognizable? The exclusion criteria were well
described in the methods section. Additionally, the numbers excluded for
each of the causes were reported.
Was the clinical intervention described in sufficient detail to permit repli-
cation in clinical practice? The cyclosporine group intervention included
azothiaprine and prednisone given in accordance with policy in each of the
 161

clinical centers. Other therapy used for the standard therapy group included
anti-lymphocyte globulin, plasmapheresis, leukopheresis, graft irradiation and
cyclophosphamide. This was not a blinded study and one cannot exclude
differential general care favoring the cyclosporine group.
The final criterion that all patients be accounted for at the conclusion of
the trial was met.
Prior to this publication, single center uncontrolled trials had suggested
that cyclosporine was superior to standard therapy but suffered from the
methodologic problems of all studies without random allocation to treatment
group. Multicenter studies permit enrollment of adequate numbers of patients
within a reasonable time to execute a study with acceptable statistical power
to detect clinically important effects as well as evaluate the effect of other
prognostic factors and their interactions. An additional strength is that the
generalizability of the results is enhanced if the treatment effect is constant
across centers. In this example, 11 of the 12 centers demonstrated better results
with cyclosporine.
This multicenter clinical trial satisfies the criteria for a methodologically
sound study of an intervention. In addition, the multicenter nature of the
study enhances generalizability.
For hemodialysis, the use of urea kinetic modelling is based on the report
of the National Cooperative Dialysis Study (NCDS) in 1981 [14]. The authors
concluded that if the dietary intake of proteins and other nutrients were
adequate, a mid-week pre-dialysis BUN value of 71-73 mg/dl was asso-
ciated with less morbidity than if this value were 105-109 mg/dl. Does this
publication meet the criteria for scientific credibility and clinical applicability?
Were the subjects truly randomly allocated to the specified treatments? There
was random allocation to one of 4 treatment groups, two producing mid-
week pre-dialysis BUN values of 120 ± 10 mg/dl and two producing values
of 70 ± 10 mg/dl. The randomization procedure resulted in no statistically
significant differences among groups with respect to age, history of heart
disease, hypertension, peripheral vascular disease, pulmonary disease, gastro-
intestinal disease or pre-randomization rates of hospitalization.
Were all clinically relevant outcomes reported? One outcome evaluated was
removal from the study for death or medical reasons, the latter defined by
any medical diagnosis on the exit form. A second outcome was hospitaliza-
tion related to uremia. Those not due to uremia (e.g. vascular access problems)
were excluded prospectively. There were only 3 deaths among the 151 patients
randomized. There were 32 patients withdrawn for medical reasons (i.e. due
to uremia). There were 66 withdrawn for other reasons (completed protocol,
patient preference and transplantation). The distribution of these other reasons
was not provided. This was an unblinded study. One might expect a bias in
attributing symptoms to uremia if the BUN value were known to the inves-
tigatory making that decision. For example, if 2 patients had anorexia, one with
a pre-dialysis urea of 70 mg/dl and another with a value of 120 mg/dl, one
would predict that the patient with the higher BUN would be more likely to
162

have that symptom attributed to uremia and be withdrawn form the study. A
similar bias could affect definition of hospitalization due to uremia despite
the authors attempts to prevent bias. An additional outcome of interest which
was not addressed was quality of life &/or uremic symptoms using a reliable
validated questionnaire.
Were both clinical and statistical significance considered? The statistical
significance was evaluated by comparing survival to the event of interest
(i.e. withdrawal or hospitalization) among the 4 groups. In the 2 high BUN
groups, 45 and 62% had been withdrawn by 1 year follow-up compared to
18 and 6% in the 2 low BUN groups. A similar effect was seen for hospital-
ization. The clinical effects were dramatic.
Were the study patients recognizable? The patients were reasonably well
described. They were between the ages of 18 and 70, were treated in a dialysis
center and had a maximum creatinine clearance of 3 mllmin. Those with cancer,
diabetes mellitus, uncontrolled hypertension, revers able renal failure, systemic
disease or unstable cerebro or cardiovascular disease, clinically important
pulmonary or hepatic disease were excluded. There was a stabilization period
of 3 months on standard therapy followed by random allocation to the defined
treatment groups. The duration of ESRD therapy was not provided nor was
there data about race and underlying renal disease. If these were long-term
hemodialysis patients, the prevalence-incidence bias [5] would be operative
and the results less applicable to new dialysis patients. Information regarding
baseline nutritional status would have been helpful.
Was the clinical intervention described in sufficient detail to permit repli-
cation in clinical practice? This criterion was satisfied. The dialysis membranes
used were cuprophane. Although not explicitly stated, the dialysate was
presumably acetate. The dialysis times were 4.5-5.0 hours for the long and
2.5-3.5 hours for the short dialysis. The target BUN values pre mid-week
dialysis were defined. The mathematical model used to achieve these targets
was referenced. All patients were accounted for at the conclusion of the
study.
This was, like the cyclosporine study, a multicenter effort. There were 8
centers in the United States in this study. Although not commented upon in
the results, center was a co-variate in the time to event analyses and was
presumably not a significant factor. There were subsequent analyses of these
data which re-inforced the importance of a minimally acceptable dialytic
removal of small molecules, more commonly expressed as KTIV [15]. The
changes in dialysis technology, particulary with respect to membranes with the
ability to remove potentially toxic middle molecular weight molecules, has
introduced some doubt about the applicability of the urea kinetic modelling
derived from use of cuprophane membranes. Nevertheless, the contribution
of this study to the development of quantitative dialysis prescription is a major
one.
For CAPD, a major clinical problem has been infection, particularly peri-
tonitis. For standard CAPD systems similar to that described by Oreopoulos
 163

et al. [16], about 2/3 of peritonitis episodes are due to organisms thought to
gain access to the peritoneal cavity through the lumen of the tubing used to
connect the bags of dialysate to the peritoneal dialysis catheter. The use of a
flush before fill sequence appears to have gained general acceptance as an
effective mode of preventing these infections. There are 2 multicenter ran-
domized clinical trials which have demonstrated the efficacy of the use of a
flush before fill technique with disinfection of the re-usable connecting set
[17, 18].
The first study was reported by Maiorca and colleagues in 1983 [17]. They
found that use of a Y connector disinfectant set was associated with a peri-
tonitis rate of 1 episode per 33 patient-months compared to 1 episode per
11.3 patient months for those using the conventional system.
Were the subjects truly randomly allocated to the specified treatments? There
was random allocation of 30 new CAPD patients to a standard method
and 32 to the Y connector disinfectant set. The randomization produced no
imbalances with respect to age or sex.
Were all clinically relevant outcomes reported? The outcomes of interest
were months of follow-up per episode of peritonitis and time to first episode
of peritonitis. There were only 2 deaths, 1 in each group. There were 11
episodes in 8 patients of accidental infusion of disinfectant into the abdom-
inal cavity, apparently without serious clinical sequelae. Technique survival
is commented upon but the small numbers preluded any statistical evalua-
tion. There was no comment regarding impact on hospitalization rates.
Were both clinical and statistical significance reported? The probability
of remaining peritonitis free for one year was 60% for the Y connector dis-
infectant set compared to 20% for the standard set. Expressed as months of
follow-up for episode of peritonitis, the results were 1 per 33 patient-months
for the former and 1 per 11.3 patient-months for the latter. Both were
statistically significant and clinically important. These findings were dramatic
and far exceeded previous reports in the literature. There are no data in the
methods indicating the reason for selection such a small sample size and this
weakened the impact of this study.
Two of the 3 criteria addressing clinical applicability of the results were
met. The intervention was well described and all patients were accounted
for at the conclusion of the study. However, the patients were not described
other than for sex ratio and mean age. The results section suggests that addition
of drugs to the dialysate was an exclusion criterion.
In this 2 centre study, there was random allocation to alternative forms of
treatment, appropriate determination of relevant outcomes and consideration
of both statistical and clinical importance. Failure to describe the expected
differences between the treatment alternatives, the Type I and Type II error
probabilities in sample size calculation decreased the credibility of the dramatic
difference reported. The patients were not described in sufficient detail to
justify unreserved application to another population of CAPD patients. The
proportion of dialysis patients treated with CAPD in Italy was small compared
164

to that in North America and this introduced the possibility of a selection

bias [5].
The second randomized clinical trial evaluating the effect of the Y connector-
disinfectant system was performed in Canada [18]. These investigators reported
that the probability of a new CAPD patient remaining peritonitis free for
1 year was 47% for the Y connector disinfectant system compared to 26%
for standard systems.
Were the subjects truly randomly allocated to the specified treatments? This
was accomplished using a variable blocking factor within each hospital. There
were 124 consenting patients eligible for the study. Sixty-one were allocated
to the Y connector-disinfectant system and 63 to the standard system. There
was no difference between the groups with respect to age, sex, marital status,
home environment, visual acuity, dialysis assistance, peritoneal dialysis
catheter, type of exit site care, dialysate additives, diabetic status or training
time. There was a statistically significant difference between groups favouring
the Y connector disinfectant set with respect to education.
Were all clinically relevant outcomes reported? The main outcomes were
peritonitis probability and peritonitis rates. The rate of exit site infections
did not differ between the groups nor was there a difference in technique
survival. Hospitalization days were quite different with those using the Y
connector disinfectant system hospitalized for 75 days for peritonitis compared
to 257 days for those using the standard system. Accidental infusion of the
sodium hypochlorite disinfectant occurred at least once in 25% of the patients
or 1 in 2500 exchanges.
Were both clinical and statistical significance considered? Comparison of
the survival to first peritonitis curves demonstrated a 61 % reduction in the risk
of peritonitis for those using the Y connector disinfectant system. This risk
reduction was present in all clinical centers. The clinical importance of this
reduction in peritonitis risk was supported by the reduction in hospitaliza-
tion days due to peritonitis. The development of an exit site infection doubled
the risk of peritonitis. The sample size calculation was provided in the methods
section and this included the clinically important difference, the expected
peritonitis probability in the standard treatment group and the Type I and II
error probabilities. The statistical considerations for the planned interim
analysis were explicitly stated.
The 3 criteria addressing the clinical applicability of the study were
satisfied. The inclusion and exclusion criteria were clearly stated. Additionally,
the number of patients excluded for various reasons were provided. The
intervention was referenced rather than described. All patients were accounted
for at the conclusion of the study.
This was a multicenter study carried out in 8 Canadian hospitals. It
confirmed and extended the observations reported in the Italian study [17]. The
consistent risk reduction in the 8 centers was a particularly convincing feature
of the study. Performance in a North American setting with a more detailed
 165

description of the patients, both eligible and ineligible, made the results more
generalizable to patients in North America.
Selection of these 3 generally accepted treatments in transplantation,
hemodialysis and peritoneal dialysis has been arbitrary. There are other
generally accepted treatments and there may be some who would not agree
that those selected represent treatment standards. The selection of the 4
randomized clinical trials which appear to have played an important role in the
acceptance of these treatments was to illustrate the degree to which these
trials meet the criteria for a standard of excellence in the execution of clinical
research.

Evaluation of quality assurance

If we are to incorporate the results of clinical trials in the development of

standards of treatment, we should evaluate the impact of use of these standards
on both the process of patient care and the outcomes affected by use of these
standards. The principles of clinical trial methodology can be used in the
scientific evaluation of the impact of implementation of a standard or set of
standards of clinical care. The group at McMaster University has published
a set of criteria for the reader of articles addressing the quality of clinical
care and is a useful reference for those considering evaluation of such
interventions [19].

References

1. Meisenheimer CG: Quality Assurance; A Complete Guide to Effective Programs (Rockville,

Maryland: Aspen Systems Corp), 1985.
2. Audet A-M, Greenfield S, Field M: Medical practice guidelines; current activities and future
directions. Ann Intern Med 113: 709-714, 1990.
3. Department of Clinical Epidemiology and Biostatistics, McMaster University Health Sciences
Center: How to read clinical journals; V: To distinguish useful from useless or even
harmful therapy. Can Med Assoc J 124: 1156-1162, 1981.
4. Linton AL, Peachey DK: Guidelines for medical practice; the reason why. Can Med Assoc
J 143: 485-490, 1993.
5. Sackett DL: Bias in analytic research. J Chron Dis 32: 51-63, 1979.
6. Churchill DN, Taylor DW: Thiazides for patients with recurrent calcium stones; still an open
question. J Urology 133: 749-751, 1985.
7. Churchill DN: Medical treatment to prevent recurrent calcium urolithiasis; a guide to critical
appraisal. Mineral Electrolyte Metab 13: 294-304, 1987.
8. Oliver MF, Heady JA, Morris IN, Cooper J: WHO cooperative trial on primary preven-
tion of ischemic heart disease using clofibrate to lower serum cholesterol; mortality
follow-up. Lancet 2: 379-385, 1980.
9. Gelber RD, Goldhirsch A, Cavalli F for the International Breast Cancer Study Group: Quality
of life adjusted evaluation of adjuvant therapies for operable breast cancer. Ann Intern
Med 114: 621-628, 1991.
166

10. Braitman LE: Confidence intervals assess both clinical significance and statistical signifi-
cance. Ann Intern Med 114: 515-517,1991.
II. Clinical Efficacy Reports. Philadelphia: American College of Physicians; 1987 and ongoing.
Available from the American College of Physicians, Independance Mall West, Sixth Street
at Race, Philadelphia, PA 19106-1572.
12. Canadian Multicentre Transplant Study Group: A randomized clinical trial of cyclosporine
in cadaveric renal transplantation. N Engl J Med 309: 809-815, 1983.
13. Canadian Multicentre Transplant Study Group: A randomized clinical trial of cyclosporine
in cadaveric renal transplantation; analysis at 3 years. N Engl J Med 314: 1219-1225,
1986.
14. Lowrie EG, Laird NM, Parker TF, Sargent JA: Effect of the hemodialysis prescription on
patient morbidity. N Engl J Med 305: 1176-1181, 1981.
15. Gotch FA, Sargent JA: A mechanistic analysis of the National Cooperative Dialysis Study
(NCDS). Kidney Int 28: 526-534, 1985.
16. Oreopoulos DG, Khanna R, Williams P: Continuous ambulatory peritoneal dialysis - 1981.
Nephron 30: 293-303, 1982.
17. Maiorca R, Cantaluppi A, Cancarini GC, Scalamonga A, Broccoli R, Graziani G, Brasa
S, Ponticelli C: Lancet 2: 642-644, 1983.
18. Canadian CAPD Clinical Trials Group: Peritonitis in continuous ambulatory peritoneal
dialysis; a multi-centre randomized clinical trial comparing the Y connector disinfectant
system to standard systems. Peritoneal Dial Int 9: 159-163, 1989.
19. Department of Clinical Epidemiology and Biostatistics, McMaster University Health Sciences
Centre: How to read clinical journals; VI. To learn about the quality of clinical care. Can
Med Assoc J 130: 377-381, 1984.
CHAPTER 11

Quality assurance in renal transplantation

ROBERT W. STEINER

Introduction

The basics of quality assurance for renal transplant programs are as follows:
(1) regular meetings of relevant transplant program personnel; (2) identifica-
tion of areas of concern; (3) monitoring of transplant program performance
in those areas; (4) initiation of changes to improve performance; (5) evalua-
tion of changes made in response to quality assurance concerns. Confidentiality
of proceedings is a practical requirement for transplant quality assurance
meetings to insure that program deficiencies will receive full and frank analysis.
Any quality assurance activity must be focused on selected issues and rely
on the judgment of the transplant team to direct and prioritize its efforts. As
the purpose of all quality assurance activities is patient benefit, it should be
relatively easy to integrate meaningful quality assurance activities into a renal
transplant program.
The best in quality assurance feature recognition of each individual's desire
to contribute to a better medical outcome, thereby improving morale and
factually educating caregivers as performance is constructively evaluated.
The process defines the broad goals of the organization and helps to meet
those goals by systematically addressing the activities and structure of the
organization with the goal of improving outcome.
The worst in quality assurance results in paperwork and other clerical efforts
which consume designated resources and produce little improvement. The staff
focuses narrowly on statistical goals or other descriptive activity, e.g., moti-
vated by fear of criticism or of regulatory sanctions; the staff primarily attempts
to collect data to document lack of deficiency - overall system improvement
and innovation are not directly addressed. Program changes which are made
are not well evaluated prospectively or retrospectively for efficacy.

Current approaches to quality assurance

A number of approaches to quality assurance in renal transplantation are

presented below in ways which illustrate potential pitfalls of each process.
L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 167-190.
© 1994 Kluwer Academic Publishers.
168

1. Quality assurance is addressed through medical or surgical case dis-

cussion conferences. The case discussion format provides an easy way to place
quality assurance on the hospital agenda. Unfortunately, these conferences
are often focused on medical education rather than critical, systematic eval-
uation of transplant program activities. Many potentially important individuals
(nurses, administrators) do not attend. There is often no closure - no ultimate
formalized consensus as to how program activities might change in response
to perceived goals, and there is no method of evaluating a change in policy
or procedure.
A similar problem exists when quality assurance activities are confined to
reports in the medical literature of transplant program experience, as sys-
tematic program review is not mandated, particularly to identify operational
problems or to propose or evaluate solutions. These two formats (medical M
& M, and reports to the literature) also do not provide the confidentiality
needed to address honestly many quality assurance issues.
2. Quality assurance is incident-driven. A key incident prompts an inves-
tigation and a change in policy based on that incident. For example, a patient
develops pulmonary edema and requires intubation after OKT3 therapy.
Guidelines are then established for control of volume status prior to OKT3
administration. Incident-driven quality assurance has the advantage of dealing
with a motivating, often serious occurrence. It does not involve systematic
review of program activities to detect less obvious areas where change would
be beneficial. At their worst, changes in response to incident-driven quality
assurance can be impulsive and may cause further unanticipated problems.
3. Quality assurance is concerned with collection of data on program
activities which are then presented regularly at meetings of the transplant staff.
This approach is prospective, but not likely to contribute to an improvement
in quality of care, as data tabulation is an end in itself. For best results, data
collection is selective and meaningful and only the beginning of productive
quality assurance activity.
4. Quality assurance is concerned with meeting statistical performance
standards which may be internally or externally imposed. For example, a
program has a goal of a 5% one year transplant postoperative mortality. A
higher observed mortality might produce efforts to show that patient mix
was biased in the direction of more expected complications. This is a defen-
sive approach to quality assurance, which can deal with important patient
care issues but often does not directly address systematic program improve-
ment. Externally imposed standards of care also may not be the most productive
for program improvement. As is discussed in later sections, available baseline
standards for clinical complications and expected results are not precise and
are difficult to apply.
5. Quality assurance can be approached via cooperative identification of
problem areas by experienced personnel, then data collected prospectively
or retrospectively relative to these areas. For example, several observed cases
of inadequate delivery of aerosol pentamidine for pneumocystis carinii
 169

pneumonia (PCP) prophylaxis, and observation of two cases of post trans-

plant PCP prompts a tabulation which suggests ineffective delivery of
pentamadine and then consideration of a change to trimethoprim/sulfa-
methoxazole prophylaxis, which is then evaluated for efficacy. This approach
is likely to produce relevant data as it derives from consensus of experienced
personnel about important program goals. It focuses on making explicit a
consensus on a detected deficiency in care, rather than on improving care in
general.
None of the above approaches fully conforms to current quality assurance
theory, although some of the features of each are necessary in any quality
assurance effort. All of these approaches are presented to illustrate potential
limitations or deficiencies in the quality assurance process.
6. The process of continuous quality improvement as it is currently for-
mulated [1] is intended to be an efficient, productive approach to program
evaluation and change. It goes beyond the consensus-on-deficiency approach
just described above. In this process, relevant members of the transplant team
meet to identify various goals of the program. Meeting these designated goals
becomes the focus of quality assurance efforts. Individuals who are served
by various aspects of the program are also identified. For example, trans-
plant coordinators serve transplant nephrologists and surgeons by providing
information on patients admitted for transplantation. Transplant coordinators
serve referring nephrologists in facilitating entry of their patients on to the
cadaveric waiting list. All aspects of the program ultimately serve the patients
who are in various phases of transplantation. The emphasis on goal identifi-
cation enables all concerned to focus on general program improvement.
Continuous quality improvement postulates that most highly functioning
transplant personnel are motivated and responsible and will become even more
so when given a chance to contribute overall program improvement. However,
the posture of the quality assurance team is to attempt to improve all aspects
of the program, even when "acceptable" performance is documented.
As previously stated, continuous quality improvement shares features of the
several previously described quality assurance approaches. Consideration of
goals of the quality improvement process is still driven in part by past
incidents, which involve problems as well as successes. Data collection is
necessary to assess program deficiencies, and the experience of others can
be used to suggest ways to improve care and to define the general standard.
All quality assurance activity in the final analysis must still be selectively
focused and rely on the experience and consensus of the transplant team to
prioritize goals.

Evaluation of current program experience

Published transplant experience provides a background against which a

program can identify possible areas of deficiency and suggestions for overall
170

improvement in quality, even in areas where a reasonable standard of care

has been achieved. The large body of experience in medical journals forms
an obvious basis for program evaluation but has the following inadequacies:
(a) A good experience is more often published that are poor results. A program
that perceives a high rate of certain postoperative surgical complications
may address the problem internally but not report these data to the medical
community. (b) Quality assurance is not directly addressed in may reports.
Detailed analysis of complications and possible improvements in care may
not concern many clinical investigators. (c) Patient care protocols (and
therefore categories for data comparison) are not uniform. For example, the
incidence of postoperative cardiac events can be expected to vary inversely
with the rigor of a program's preoperative cardiac evaluation. Liberal use
of OKT3 for postoperative acute rejection and lower dose postoperative
maintenance immunosuppression may be factors which increase the number
of reversibly mild rejections and increase OKT3 success rates. (d) Published
experience can be quickly out of data. For example, changes in immunosup-
pression (cyclosporin, OKT3) and new antiviral agents (acyclovir, gancyclovir)
make some earlier experience in post transplant infection less applicable to
present practice. (e) Statistical comparisons of published experience from
different programs can be difficult. In 1988, only 10% of the renal trans-
plant programs in the United States performed more than 100 transplants per
year. One-third of the programs in the United States performed 50 or less renal
transplants per year [2]. Utilizing the conventional definition of statistical
significance, an incidence of one postoperative ureteral obstruction in a
program doing fifty transplants per year is not different than an incidence of
five occurrences in fifty transplants (p value by chi square> 0.05), nor is
an incidence of zero versus three in a program doing twenty-five transplants
a year. The variability inherent in using published experience is illustrated
in Table I, which summarizes several studies on transplantation in older
patients. There is a marked variation in the ages studied and the calendar
years over which experience is reported, with many reports including patients
from the pre-cyclosporin and/or pre OKT3 eras.
A more fundamental criticism of statistical analysis of quality assurance
data in this era of medical malpractice awareness is that the conventional
scientific standard of statistical significance (p < 0.05) is arbitrary and does
not necessarily conform to a lay, "common sense" appreciation of probable
cause and effect. If a program has a problem identified with odds of only
nine of 10 (p = 0.1) or four out of five (p = 0.2) that a numerical difference
does not occur by chance alone, this may be productive to investigate. Thus,
statistical comparisons for assessing program performance have limitations.
 171

Table I. Renal transplantation in older patients: variability of reported data.

Number Age Years Perioperative 1 year

patients studied mortality* survival*
(4 months) patient/graft

Milwaukee (28) 17/920 65+ '67-'88 6% 88%/81%

Sweden (29) 55/- 65-75 '75-'87 14% 71 %/63%
Madisone (30) 34/- 60+ '83-'88 9% 91 %/83%
Cleveland (31) 26/- 60+ '85-'89 21% 79%176%
Washington (32) 24/- 55-67 '84-'87 14% 82%170%
Oxford (33) 63/- 55+ '75-'87 3% 87%173%
Toronto (34) 68/- 50+ '81-'87 88%/94%
Cincinnati (35) 53/365 50+ '80-'86 12% 83%174%
Florida (36) 67/- 50+ '83-'89 5% 90%175%
15/- 50+ '83-'89 0% 100%/100%

* Estimated from reported data if not explicitly provided.

Present regulation and monitoring

Several regulatory bodies require that renal transplant programs participate

in certain quality assurance activities or report selected data on a regular
basis. The Joint Commission on Accreditation of Healthcare Organizations
(JCAHO) requires quality assurance via a hospital wide process which is
required to (a) establish the most important aspects of care (b) identify
indicators for monitoring care, both trends and important single events, (c)
evaluate care when problems are suggested, and (d) implement and evaluate
corrective action in response to those problems. In addition to inpatient care,
some of the relevant areas which are to be monitored are diagnostic radi-
ology, diabetic services, ambulatory care services, nuclear medicine, nursing,
pathology, pharmacy, other hospital patient care support services, and infec-
tion control. The JCAHD requires monthly meetings of clinical departments
or services, medical record review, surgical case review, blood and drug
usage review, and monitoring of the clinical performance of all individuals
who are involved in patient care, including those not permitted by the hospital
to practice independently [3]. There is no specific quality assurance require-
ment for a renal transplant program as such, but the requirement for monthly
meetings of clinical services appears to apply to the transplant team as a clinical
entity.
The United States is divided into sixteen regional End State Renal Disease
Networks, administered by the Health Care Finance Administration. Reporting
requirements to one's network involve a patient status report signed by the
nephrologist, surgeon, and other members of the health care team, which
indicates dialysis modality, transplant status (active, inactive, under evalua-
tion) and rationale for same. The network also requires a report at the time
of transplantation on numerous recipient characteristics, including HLA typing,
172

MLC response, PRA, history of previous transplants, nephrectomy, splenec-

tomy, HBAG and CMV status, transfusion history, and early graft function.
Donor (living related or cadaver) information includes presence of
cytomegalovirus (CMV) antibody, hepatitis B surface antigen (HBAG),
neoplasm, warm and cold allograft ischemia time, cadaver donor pretreatment,
and renal function preharvest.
The United Network for Organ Sharing (UNOS) requires followup data from
each renal transplant recipient at the time of transplant, at six months, and then
annually. Historical data which are requested pertain to (a) health status (e.g.,
working, homebound, hospitalized, HIV status), (b) medication (cyclosporin,
azathiaprine, prednisone, intravenous Solu-Medrol, ALG, OKT3), (c) graft
status (quality of function, graft nephrectomy, need for dialysis, biopsy, cause
of graft loss), and (d) patient status (alive or dead, autopsy).
Data which are collected and reported by these various regulatory bodies
can be useful as a general standard to which to compare program performance,
but these data are not sufficient for detailed analysis of program performance.
Selected data from the 1989 annual report of the United Network for Organ
Sharing (UNOS) are presented in Figures 1 and 2 [4V

Scope of possible program evaluation

The following sections are written from the perspective of continuous quality
improvement, that is, they attempt to set out a schema for program evalua-
tion in terms of program goals.
The overall goal of a renal transplant program can be stated simply: to
maximize long term patient and allograft survival with a minimum of patient
morbidity. The scope of related quality assurance activity can be divided into
eight areas: (a) preoperative recipient screening, (b) management of the living
donor, (c) cadaver kidney quality, (d) surgical care, (e) postoperative medical
care, (f) immunosuppression, (g) long term graft survival, (h) long term medical
care, and (i) patient education and psychological well-being. At present, good
programs often differ significantly in their protocols and practices relative
to the above categories. The following sections will attempt to set out
important issues in quality assurance. To document fully or take a position
on the numerous and important controversies which are addressed is beyond
the scope of this chapter. In many cases, the precise risks and benefits
associated with many of the alternative approaches which are discussed below
have not been quantified.

Preoperative screening

The number of possible screening procedures and protocols for preoperative

disease is almost unlimited, and preoperative screening must be selective and
 PATIENT SURVIVAL FOR
CADAVERIC KIDNEY TRANSPLANTS IN 1988
Categorized by Recipient Age

100,........ _______________ _
-MM"'......."',,"""'.:.......................................
=---
--------------- ,
.--.. ----
c: ........
Q)
-
~ 80
--
a.. < 5 years •••••••••••• n = 41
Q) 5-18 years . - - - - - n = 320
19-45 years _ _ _ n=3740
-( lj
a: 46-65 years n = 2342
(lj > 65 years - - - - n = 169
>
.~ 60
:::J
Cf)

40~1~--~~--~~--L--L--L--L--L-~~
o 1 2 3 4 5 6 7 8 9 10 11 12
Month After Transplant
Fig. 1. UNOS patient survival in the first year post transplant. Survival of patients older than 65 is significantly less than survival of patients
19-45 years of age. Most deaths occur in the first 6 months after transplantation. (From reference [4]) - .)
V.l
-
 .....
-.l
.j::1..

GRAFT SURVIVAL FOR CADAVERIC

KIDNEY TRANSPLANTS IN 1988
Categorized by Recipient Peak Panel Reactive Antibody
100
.--.. ~.::
+-"
C ~..:
... ...........
Q) , ............. .........
~ ....................
........................................................................................
rf. 80 _--------------------------------------
" '- .........._---
Q)
1U
(( -~-----------------
0-20 ............ n = 4373
ctl 60 21-80 . - - - - - n = 1173
> > 80 _ _ _ n = 552
.~
:::J
CI)

40
o 2 3 4 5 6 7 8 9 10 11 12
Month After Transplant
Fig. 2. UNOS graft survival at one year based on percent panel reactive antibody (PRA). The outcome for PRA 0-20% differs from that for
PRA 80-100% (p < 0.01). (From reference [4])
 175

oriented to procedures which are relatively likely to promote patient well being.
Requirements for extensive preoperative screening may discourage the patient
from seeking transplantation or may result in long delays. Screening proce-
dures such as gastroesophageoduodenoscopy, colonoscopy, cystoscopy, and
arteriography have a low but definite morbidity and add to patient inconve-
nience. These procedures also add to the cost of transplantation. It is difficult
to evaluate the utility of admittedly low yield screening procedures which
are recommended to rule out unlikely and/or inapparent problems in dialysis
patients. For example, the possible occurrence of an undetected preoperative
bladder carcinoma in a transplant recipient may not justify a protocol to
cystoscope candidates routinely. Often, dialysis patients and referring nephrol-
ogists have the choice of several competing local renal transplant programs
and may not choose a program with extensive preoperative requirements. In
this way, competition among competing programs and the proclivity of many
potential recipients to avoid lengthy screening also may limit extensive
screening protocols.
The fundamentals of preoperative screening can be divided with some
overlap into: (a) achieving a complete and current history and physical
examination, (b) cardiac evaluation, (c) detecting chronic infection, (d)
detecting occult neoplasm, (e) characterizing patients immunologically and
identifying immunologically high risk patients, (t) anticipating anatomic
problems, (g) anticipating special medical problems, and (h) assessment of
patient compliance.
A physician knowledgeable in renal transplantation (usually a member of
the transplant team) should generate and review the basic intake patient profile:
medical history, physical examination, and laboratory testing, which should
include detailed head and neck examination, pelvic and rectal examination,
EKG, chest x-ray, and frequently PAP testing and mammography. Referring
physicians should be educated to keep their patient information current at
the transplant center. The discovery of, e.g., a previously unreported history
of active duodenal ulcer, an unexplained change in a cardiogram, or a new
deep-seated foot infection in a diabetic potential recipient just prior to
cadaveric transplantation will often disqualify that recipient, add to graft
cold ischemia time, and unnecessarily strain staff resources or contribute to
inappropriate preoperative decisions. These last minute surprises may be
unavoidable to some extent, but their occurrence should prompt a review of
preoperative evaluation and communication.
Cardiac disease is a major cause of morbidity and mortality for dialysis
patients [5, 6] and for transplant recipients [7]. Many patients may require
no further screening beyond a routine EKG. Asymptomatic patients with
cardiac risk factors such as age, a smoking history, hypertension, and hyper-
lipidemia should have noninvasive testing with a stress, stress thallium or
persantine-thallium study. Many dialysis patients do not have the endurance
to complete cardiac exercise testing [8, 9]. The thallium-persantine cardiac
scan, which does not require exercise, is helpful by most reports [10, 11],
176

but may miss significant disease [12]. Usually patients with positive nonin-
vasive testing go on to angiography and possible coronary bypass grafting
or angioplasty. The strongest indications for thorough cardiac evaluation are
diabetic status and history of MI or angina, but experienced programs may
differ on indications for detailed cardiac evaluation, as discussed in the last
section of this chapter. Selective testing for diabetic recipients based on other
risk factors has been proposed [13, 14].
Patients with certain cardiac risk factors (e.g., arrhythmias or coronary artery
disease) will require special postoperative care. The occurrence of signifi-
cant cardiac morbidity post-operatively should suggest a review of intake
protocols and physician communication. Cardiac events such as myocardial
infarction or sudden death in the first post transplant year may also be a
reflection of failure of pre transplant cardiac screening.
Immunosuppression will exacerbate preexisting chronic infections such as
otitis media, dental abscesses, prostatitis, chronic pyelonephritis, occult
tuberculosis, and osteomyelitis. These abnormalities are usually detected by
intake physical examination and must be corrected before transplantation.
Preoperative cholecystectomy for asymptomatic cholelithiasis or sigmoidec-
tomy for diverticulosis are not required in most programs. The identification
of infection post-operatively that could have been a chronic pretransplant
condition should be viewed as a possible failure of pretransplant screening.
Chronic hepatitis Band C may be exacerbated by immunosuppression but
this condition is not generally considered to be a contraindication to trans-
plantation [15, 16]. Individual pretransplant counseling in patients with
hepatitis may be indicated.
The presence of malignant neoplasm in the recipient is an absolute con-
traindication to renal transplantation. Basic screening for neoplasm includes
chest X-ray and stool guaiac test. Abnormalities in the basic intake profile such
as a thyroid or prostatic nodule, abnormal liver function tests (even at times
in patients with "hepatitis"), or a positive stool guaiac test should be pursued.
A variety of neoplasms may develop in long term dialysis patients [17],
including renal neoplasms [18] but screening ultrasound or CT studies of native
kidneys are not required by many programs. Neoplasms arise de novo with
increased frequency in rental transplant recipients [19], but the emergence
of solid tissue neoplasm in the first year post transplantation suggests a failure
in preoperative screening.
Patients must be accurately tissue typed for HLA and DR antigens, and
the amount of preformed antibody to those antigens quantified (% panel
reactive antibody or PRA). Patients who are at increased immunologic risk
for acute or accelerated rejection are those with PRA greater than 50-80%
and candidates for repeat transplants, especially when previous transplant
survival has been brief [20, 21, 22]. These patients should be identified
prospectively. Many programs believe that these patients should receive induc-
tion therapy with OKT3 or ALG, as discussed later.
Preoperative assessment of anatomic problems must be carried out by
 177

transplant surgeons. Factors include previous surgery in the iliac fossa

(including vascular procedures), large polycystic kidneys that preclude seating
of the allograft, and severe iliac vascular disease that may make arterial
implantation difficult or may produce a steal syndrome and exacerbate distal
arterial insufficiency. Prostatic obstruction should be identified preopera-
tively, but surgery is often best deferred until after transplantation due to the
need for urine flow to prevent urethral stenosis. Grade III or grade IV ureteral
reflux requires preventive repair or nephrectomy if accompanied by recur-
rent infection.
Another goal of the intake procedure is anticipating special postoperative
medical problems. These may occur with obesity [23], hypercoagulable states,
psychological sensitivity to corticosteroids, marked hypertension, drug aller-
gies, aversion to blood products (Jehovah's Witnesses), and the occasional case
of marked hypercalcemia with tertiary hyperparathyroidism [24], to name
several. Many significant postoperative medical complications are to some
extent foreseeable and their occurrence should prompt an evaluation of
preoperative screening and preparation.
Prospective assessment of patient compliance can be extremely difficult.
Poorly compliant dialysis patients may be highly motivated to avoid a return
to that modality, so a past history of noncompliance may not be important. The
opinions of medical and psychosocial personnel who know the patient well
are of primary importance. The two major compliance questions are (a) will
the patient take his medicines?, and (b) will the patient keep his medical
appointments and follow basic medical advice? Patients who are homeless
or otherwise lack basic support systems are at high risk for noncompliance,
as are patients who cannot communicate adequately with medical personnel
e.g., due to a language barrier. Patients who cannot afford medicines pose
obvious risks. A demented patient presents special compliance and ethical
questions. Graft failure post discharge - whether in the first year or later -
is significantly influenced by patient compliance [25]. Compliance problems
postoperatively therefore should prompt review of preoperative psychosocial
evaluation.

Management of the living donor

Living donorship has become an even more attractive alternative for transplant
programs with growing waiting lists and a shortage of cadaveric donors
[26, 27]. The goals of living donorship (related or unrelated) are (a) a safe
nephrectomy, (b) the absence of current and future renal disease in the donor,
(c) no disease communicated via the allograft, and (d) living kidney donation
which is informed and uncoerced. Donor evaluation protocols differ among
different programs but include a thorough history and physical, laboratory
screening (including HIV, CMV, hepatitis C and HBAG testing), IVP, and renal
arteriography. Cardiac stress testing should be undertaken in older donors
178

and should be at least as complete as for recipients with similar risk factors.
In general, the presence of any renal disease rules out donorship. The presence
of low grade hematuria which has been completely evaluated in a young
potential donor also can present a difficult decision, as it often has a benign
prognosis [28]. The finding of borderline hypertension also presents a diffi-
cult decision as to donor acceptability. Within limits, potential donors can
elect to donate in borderline cases after detailed discussion. Occasionally,
psychosocial evaluation can uncover unacceptable family dynamics involving
coercion or other unusual incentives to a potential donor. An attempt should
be made to follow living donors postoperatively for several years to assess
the adequacy of preoperative evaluation.

Cadaver kidney quality

Cadaver kidney quality will become more of an issue as cadaveric waiting lists
lengthen. Cadaver kidneys will increasingly be taken from older donors, donors
with hypertension, and donors with possible acute ATN [29]. UNOS and
regional organ procurement agencies may at times offer cadaveric kidneys from
donors of marginal quality, to be accepted for transplantation at the discre-
tion of the individual program. Poorer cadaveric kidney quality may result
in a higher postoperative dialysis rate, a lower one year graft survival, and poor
quality of graft function long term [29, 30, 31]. These same complications may
of course arise from factors which are independent of cadaver kidney quality:
deficient preservation techniques, increased warm or cold ischemia time [32]
cyclosporine toxicity, postoperative ATN, early allograft rejection, inade-
quate immunosuppression, anatomic problems from the surgery itself, and poor
patient compliance. Complications from allograft harvest which are not strictly
donor dependent include shortened or traumatized renal arteries, veins, or
ureters, inadequate preservation of ureteral blood supply, failure to identify
or preserve multiple renal arteries, and prolonged warm ischemia time.

Surgical care

Postoperative surgical complications in renal transplantation are generally

uncommon and include bleeding, allograft arterial and venous thrombosis,
ureteral anastomotic leakage, or necrosis, wound infection or dehiscence, and
lymphocoel [33-35]. Rarely, bowel may be traumatized or arterial insufficiency
of the ipselateral leg may be exacerbated. Postoperative ileus can occur with
refeeding and occasionally can lead to bowel perforation. Early surgical
complications that affect graft function can be misdiagnosed and mistreated
as rejection with detriment to allograft and patient. Monitoring with ultrasound
and nuclear medicine scans facilitates accurate diagnosis. Allograft needle
biopsy is a relatively benign procedure which very occasionally can be com-
 179

plicated by bleeding or arteriovenous fistulae [36]. Ultrasound guidance helps

avoid overlying bowel or extrarenal vessels. The possibility of bleeding should
be anticipated before each biopsy so that transfusion, angiography, and even
urgent surgery are facilitated.

Postoperative medical care

The chief goals in postoperative medical renal transplant care are (a) man-
agement of known preoperative risk factors, (b) safe and effective dialysis
support, (c) infection management, (d) management of immunosuppression,
and (e) patient psychological well-being. Preoperative risk factors can be
divided into general risks and patient-specific factors. General risks stem
from the prevalence of cardiac disease in the dialysis production [5, 6, 8-14],
the effect of age on post transplant mortality and morbidity [37-45], and the
increased incidence of fluid, electrolyte, and hemorrhagic complications.
Volume expansion immediately postoperatively in an effort to help allograft
function [46, 47], may increase left ventricular work and risk myocardial
ischemia [48], cardiogenic pulmonary edema [49], and noncardiogenic pul-
monary edema with OKT3 [50].
Preoperative patient-specific conditions that might lead to postoperative
therapeutic urgencies include subtle preoperative glucose intolerance, par-
oxysmal cardiac arrhythmia, chronic seizure disorder, diverticulosis, and drug
allergy. The need for postoperative dialysis support varies widely. Hemodialysis
associated hypotension, bleeding (with heparin), and sudden loss of access can
be minimized by planning and attention to detail. Post transplant peritoneal
dialysis poses special risks of infection with immunosuppression, dialysate
leakage through the wound, and hyper-glycemia. Patient-specific problems
include dialyzer allergies, CAPD equipment incompatibilities, and reduction
of peritoneal volume after transplantation in small CAPD patients. The
therapeutic goal in this area is communication of the details of each patient's
preoperative status and continual postoperative awareness of same by the
medical team.
Prophylaxis and treatment for infection that arises in the immediate post
transplant period involve several areas. Cytomegalovirus (CMV) infection is
particularly common and can be particularly severe in seronegative recipi-
ents who receive kidneys from seropositive donors and in patients who receive
increased immunosuppression [51-53]. CMV prophylaxis is indicated in
high risk groups [54, 55] and is employed by some centers in all patients in
the postoperative period. Acyclovir, gancyclovir, or immune globulin are
used for CMV prophylaxis [54]; Gancyclovir is indicated for invasive CMV
disease [52] but not necessarily for viremia alone, which is much more common
[56]. Acyclovir is effective prophylaxis and treatment for herpes simplex
infection [57, 58]. Other factors which influence the occurrence and severity
of infection are delays in diagnosis and in decreasing or discontinuing immuno-
180

suppression once infection is present. Significant - even fatal - viral infec-

tion post transplant is not always preventable with current prophylaxis and
treatment.
Bacterial infection arises from common postoperative sources: wound, lungs,
intravenous access, and urinary tract [59]. Many programs prophylax with daily
oral antibiotics for urinary tract infection either permanently or for several
months post transplant [60]. Perioperative pulmonary care, surgical tech-
nique, and intravenous and bladder catheterization technique and maintenance
are obvious risk factors for bacterial infections. Opportunistic infections
from nocardia, candida, other fungal agents (blastomycosis, aspergillis and
coccidiodomycosis), and PCP are strongly influenced by the total amount of
immunosuppression, especially with OKT3 or ALG [51, 53, 59], and
environmental variables [59]. PCP prophylaxis can be accomplished with daily
trimethoprim-sulfamethoxazole [61] for several months post transplant.
Post-transplant infection with tuberculosis, coccydiodomycosis and at times
other fungi can represent a failure of pretransplant surveillance and/or
treatment for preexisting infection. At present, infection control remains a
major area of concern in transplant quality assurance.

Immunosuppression

The overall goals of immunosuppressive management are to maximize graft

survival but to minimize complications of therapy. Immunosuppression pro-
tocols differ in many respects among renal transplant centers. The intensity
and timing of postoperative immunosuppressive therapy markedly influences
early graft survival, but the latter is also a function of the number of immuno-
logically high risk recipients as well as medically high risk recipients in
whom immunosuppression may be more likely to be curtailed when intercurrent
medical problems arise. Routine induction therapy with OKT3 or ALG may
[62], or may not [63] offer long term benefit. Induction therapy is commonly
utilized in immunologically high risk patients [64, 65] and in recipients of
kidneys with ATN, who may [66, 67] or may not [68, 69] benefit from early
reduction or withholding of cyclosporin. The published response rate of
rejection to OKT3 is variable [70]. A low rejection response rate to OKT3
or ALG may reflect a high threshold for use of these agents. Failure of OKT3
therapy for rejection may also be related to deficient monitoring techniques
or to the need to stop treatment due to a variety of side effects [50]. Serious
complications from OKT3 are sufficiently infrequent to limit detailed
comparison of the individual experiences of most programs. However, infec-
tion is least likely during induction therapy [62] and more frequent when
total immunosuppression dose is higher, e.g., when OKT3 is used late in the
postoperative period or when 2 course of antilymphocyte preparation are given
[51, 53]. The incidence of lymphoma is also increased with use of antilym-
phocyte preparations [71].
 181

CSA levels tend to fluctuate early in the post-transplant period and cannot
be predicted accurately for a given patient from a standard intravenous or
oral dose [72]. Patients who receive daily cyclosporine (CSA) as initial therapy
must have accurate, daily trough levels, which can be monitored by one
of several techniques [73]. Although the precise value of CSA levels is
unsettled, in the absence of antilymphocyte (induction) therapy, adequate CSA
and steroid dosage is critical in the early post transplant period, to minimize
early graft loss [74].

Long-term graft survival

Many renal programs return patients after transplant to a referring nephrolo-

gist who is not affiliated with the transplant center. Therefore there may be
difficulty in monitoring and affecting long term allograft survival, which
nevertheless is a concern of every renal transplant program. Unfortunately,
the lack of complete consensus in the transplant community can promote
suboptimal chronic immunosuppressive regimens in inexperienced hands.
Chronic and acute rejection occur when CSA maintenance is reduced [75,
76], although some centers are comfortable with overall lower CSA doses than
others [77, 78]. Average effective maintenance chronic CSA dosing is in the
range of 4-5 mg/kg/day [75, 79] but due to wide variation in individual
pharmacokinetics, therapy must be guided by CSA levels. The benefit of
"low dose" azathioprine (l mg/kg) which is added to CSA and prednisone
is controversial [76, 78, 80]. Substitution of full dose azathioprine for
cyclosporine for chronic immunosuppression is at best controversial [78,
81-83] and is not practiced at most centers. Steroids add to the effectiveness
of CSA immunosuppression both clinically and experimentally [84-87], but
selected patients may tolerate gradual prednisone withdrawal [88]. Chronic
cyclosporine toxicity will impair graft function, [89,90]. A reliable cyclosporin
blood assay with which referring physicians are familiar is probably essen-
tial for good long-term graft survival. Although trough CSA levels may not
be as relevant as earlier post-dose levels or the area under the concentration-
time curve [91], they are easier to obtain on a routine, outpatient basis. Long
term graft survival will be influenced by patient compliance both in taking
medication and in accurate timing of "trough" CSA levels. Education of the
referring nephrologists as to current immunosuppressive practices is a major
factor in long term graft success.
A less frequent cause of graft loss or impairment is medical renal disease
without rejection. Particular concern regarding aggressive early recurrent
disease [92] may arise in some transplant candidates with end stage renal
disease due to hemolytic-uremic syndrome [93, 94] or focal sclerosis [95]. Low
dose antibiotic prophylaxis for urinary tract infection [60] is indicated in many
female if not all post transplant patients, especially those who have had
allograft pyelonephritis, which can seriously impair graft function.
182

Long term medical care

Long-term medical care of the renal transplant recipient involves a unique

constellation of problems [96]. Cardiac disease is responsible for an increasing
percentage of long-term post-transplant mortality [7, 97]. As most patients with
preexisting coronary artery disease are transplanted uneventfully [8, 9],
initially asymptomatic coronary artery disease will be present in many trans-
plant patients and can be expected to progress. Periodic cardiac assessment
and treatment of hyperlipidemia [98] are essential. Diagnosis of renal artery
stenosis leading to progressive hypertension and azotemia can be difficult [99].
Calcium channel blockers offer some advantages for management of hyper-
tension and may improve allograft function [100, 101]. Renal transplant
patients are at increased risk for neoplasm, particularly cutaneous neoplasms,
lymphoma, and renal tumors [19]. The multiple nonimmunologic complica-
tions of steroid therapy (osteopenia, cataracts, hyperlipidemia) may warrant
limiting chronic dosage to 0.15 mg/kg (or 0.30 mg/kg q.o.d.) or less, which
should not produce adrenal suppression [102]. For many programs, the quality
of long-term medical care is approached only through dialogue with
referring nephrologists concerning these specific health care issues.

Patient education and psychological assessment

Patient education is performed in association with the intake process; prospec-

tive recipients should be made familiar with the possibilities of a prolonged,
complicated hospital course, early graft failure, and perioperative mortality.
Patients should know that lifelong physician followup and lifelong medica-
tion will be necessary even with the most successful transplant. They should
be familiar with the expected complications and side effects of immuno-
suppressive medicines. Other issues which may be discussed are likelihood
of recurrent disease [92] and exacerbation of chronic hepatitis by immuno-
suppression [15, 16].
Patient psychological well-being immediately post transplantation should
receive direct attention from the transplant program. Patient perception of
medical events and resultant stress may not correspond to the clinical reality.
Patient expectations are also frequently high due to pretransplant patient denial
and encouragement from may sources. Psychological reactions to steroids,
cyclosporin, and OKT3 should be anticipated and discussed prospectively.
At times a preoperative trial of prednisone or CSA is helpful to assess effects
on mental status. Psychological events postoperatively may also reflect
inadequacies in preoperative screening. Patient perception of hospital and staff
cleanliness, infection control protocols, and medical sophistication of all
involved personnel may produce postoperative stress. Some programs may
decide to use a patient questionnaire in the first two or three months post
transplantation.
 183

Organizing quality assurance activity in a renal transplant program

The amount of time and effort spent on quality assurance activities varies
widely among renal transplant programs in the United States. Financial
considerations or lack of overt program commitment to quality assurance
may limit quality assurance activities. The present standard of practice in
this area is therefore somewhat vague and broad. A proposed basic plan for
the organization and activity of a renal transplant quality assurance program
is outlined in Table II. The categories for data collection closely relate to
the eight areas outlined in the section "Scope of Program Evaluation" above.
The emphasis given by such a quality assurance committee to data from each
area would vary from program to program and would depend on individual
program strengths and weaknesses and designated priorities. Other data
categories such as patient death, readmissions, early graft failure, and reop-
erations, will fall into the data categories already proposed in Table I. Although
the process of continuous quality improvement postulates that all areas of a
transplant program are subject to review and improvement by the transplant
team, emphasis and priority depend on available resources and perceived needs
in these different areas.
The difficulty in setting out specific protocols or standards for adequate
quality care of the transplant patient is illustrated by the following exercise.
Eight renal transplant programs in the United States were surveyed by tele-
phone. Two of these programs performed 80 to 100 transplants per year, and
the rest between 100 and 200 transplants per year. A physician or a trans-
plant coordinator who was knowledgeable in program practices and procedures
was asked a series of standardized questions concerning preoperative evalu-
ation procedures for potential candidates for renal transplantation and
postoperative care. The results are displayed in Table III in the text. Some

Table II. Organizing a quality assurance program.

A. Committee members: medical and surgical staff members and transplant coordinators; floor
nursing or administration where appropriate.
B. Meet regularly (monthly or bimonthly).
C. Gather and analyze data on:
(l) medical complications (cardiac, infectious, and dialysis related, related to inadequate
pretransplant evaluation, physician or communication, or to postoperative care).
(2) surgical complications (arterial or venous thrombosis, ureteral breakdown or stenosis,
wound infection or dehiscence, bleeding, biopsy complications).
(3) early allograft dysfunction (diagnosis, cadaver donor selection).
(4) rejection (incidence, efficacy of baseline immunosuppression, diagnosis and treatment
for rejection, complications of treatment).
(5) patient psychological status.
(6) long term complications (neoplastic, cardiac, infections, patient compliance).
(7) long term allograft function.
D. Institute changes where appropriate.
E. Evaluate effect of changes.
184

Table III. Patient evaluation and management: a survey of 8 large renal transplant programs.

Pretransplant patient evaluation protocols written and detailed: Yes 2; No 6

Patients seen and evaluated by surgeon: Yes 6; No 2
Seen by nephrologist: Yes 4; No 4
Most protocols understood but not in writing: Yes 4; No 4
Some protocols written (eg. cardiac evaluation): Yes 2; No 6

Infection prophylaxis:
Transplant CMV positive kidneys into CMV negative recipients?: Yes 7; No 1
Routinely prophylax with acyclovir: Yes 4; No 1; High risk patient only 1
Immunoglobulin routinely used for CMV prophylaxis?: No 7; Occasionally 1
Trimethoprim-sulfamethoxazole prophylaxis for PCP or UTI: Yes 6; No 2

Screening for neoplasm:

Native Kidney ultrasound: No 3; Occasionally 1; Polycystic kidneys 2; Often 2
Colonoscopy or barium enema: No 7; Yes 1 (over age 50)
Prostate specific antigen: No 7; Yes 1 (but results hard to interpret)
(All programs require chest x-ray, and pap smear and mammograms in most females.)

Cardiac evaluation:
Cardiac stress testing in a hypothetical 45-year-old male on hemodialysis for two years without
risk factors: Usually 2; Occasionally I; No 5
Diabetic patients: noninvasive "stress" testing (exercise tolerance test or persantine thallium scan):
Yes 8; No 0
Estimated number of patients in whom pretransplant cardiac bypass grafting or angioplasty would
be recommended: 5% or less 6; 5-10% 1
(All programs require baseline EKG and would pursue additional cardiac testing if EKG abnormal
or if patient had suggestive symptomatology.)

Other risk factors:

Cholecystectomy: Yes 1 (Screening ultrasound over age 30, cholecystectomy if multiple stones);
No 4
Screening VCUG: Usually 1; No 6
Weight limit and issue in obese patients: Rarely 3; Occasionally 3

programs did not respond in certain categories, so that in some categories

less than 8 responses are tabulated.
The survey revealed variability in may aspects of preoperative evaluation.
Many respondents cited clinical experiences as reasons for requiring certain
tests or evaluation protocols. However, some respondents felt that aside from
very basic screening, most problems were best dealt with after transplanta-
tion as the need arose. Among these several large and successful programs,
there was not uniformity of procedure in most areas, suggesting that each
program felt it had had a satisfactory experience with its own approach to
the pretransplant evaluation.
Likewise, prophylaxis for infection post transplant was not uniform, with
variable use of antibiotics and antiviral agents. Immunosuppressive proto-
 185

cols were not tabulated, but given the present day differences of opinion in
the literature, one would expect to see differences in immunosuppression
protocols if a similar survey were conducted. These results reinforce the
conclusion that acceptable practice may vary considerably in renal transplant
programs at the present time.

Acknowledgments

Nicholas Halasz, M.D. reviewed this manuscript and made helpful suggestions.
Dorothea Rehman diligently aided in manuscript preparation.

Note

1. The data in the UNOS 1988 & 1989 Annual Report on the Scientific Registry and the
Organ Procurement and Transplantation Network have been supplied by transplant
facilities and Scientific Registry subcontractors who continually upgrade and supply new
information to UNOS. The data reflect the database on April 7, 1990 and are subject to
change. The interpreting and reporting of data are the sole responsibility of the authors.

References

1. Berwick DM: Sounding board. Continuous improvement as an ideal in health care. New
Engl J Med 320(1): 53-56, Jan 1989.
2. Renal Transplants and Renal Transplant Centers: Annual Report on the U.S. Scientific
Registry for Organ Transplantation and the Organ Procurement and Transplantation Network,
1988 and 1989. UNOS, Health Resources and Services Administration, and the Division
of Organ Transplantation, p 40 Aug 1990.
3. Accreditation Manual for Hospitals: The Joint Commission on Accreditation of Healthcare
Organizations. Chicago, 1990.
4. Transplant Outcomes: Annual Report on the U.S. Scientific Registry for Organ
Transplantation and the Organ Procurement and Transplantation Network 1988 and 1989.
UNOS, Health Resources and Services Administration, and the Division of Organ
Transplantation, pp 69-74, Aug 1990.
5. Parfrey PS, Harnett JD, Barre PE: The natural history of myocardial disease in dialysis
patients. J Amer Soc Neph 2(1): 2-12, 1991.
6. Rostand SG, Gretes JC, Kirk KA et al.: Ischemic heart disease in patients with uremia
undergoing maintenance hemodialysis. Kidney Int 16: 600-611, 1979.
7. Hill MN, Grossman RA, Feldman HI et al.: Original investigations. Changes in causes of
death after renal transplantation, 1966 to 1987. Amer J Kidney Dis XVII(5): 512-518,
May 1991.
8. Philipson JD, Carpenter BJ, Itzkoff J et al.: Evaluation of cardiovascular risk for renal
transplantation in diabetic patients. Am J Med 81: 630-634, 1986
9. Holley JL, Fenton RA, Arthur RS et al.: Thallium stress testing does not predict cardio-
vascular risk in diabetic patients with end-state renal disease undergoing cadaveric renal
transplantation. Am J Med 90: 563-570, 1991.
10. Brown KA, Rimmer J, Haisch C: Noninvasive cardiac risk stratification of diabetic and
nondiabetic uremic renal allograft candidates using dipyridamole-thallium-201 imaging
and radionuclide ventriculography. Amer J Card 64: 1017-1021, Nov 1989.
186

11. Kletter K, Derfler K, Balcke P et al.: Thallium-20 1 dipyridamole (DPM) stress and
redistribution scintigraphy in hemodialysis (HD) patients and renal transplant recipients
(RTRs): 6 year follow-up study, J Am Soc Neph 1 (Abstr): 292, 1990.
12. Marwick TH, Steinmuller DR, Underwood DA et al.: Ineffectiveness of dipyridamole
spect thallium imaging as a screening technique for coronary artery disease in patients
with end-stage renal failure. Transplantation 49(1): 100-103 Jan 1990.
13. DeVault GA Jr, Wynn SW, Swanson S et al.: The impact of routine coronary screening
in diabetic candidates for renal transplantation. (Abstract) 10th Annual Meeting, American
Society of Transplant Physicians, Chicago, 1991.
14. DeVault GA Jr, Wynn SW, Holman JM Jr et al.: Coronary risk assessment limits the need
for routine coronary screening in diabetic candidates for renal transportation. (Abstract) 10th
Annual Meeting, American Society of Transplant Physicians, Chicago, 1991.
15. Parfrey PS, Forbes RDC, Hutchinson TA et al.: The clinical and pathological course of
hepatitis B liver disease in renal transplant recipients. Transplantation 37(5): 461-466
May 1984.
16. Huang C, Lai M, Fong M: Hepatitis B liver disease in cyclosporine-treated renal allograft
recipients. Transplantation 49(3): 540-544, 1990.
17. Port FK, Ragheb NE, Schwartz AG et al.: Neoplasms in dialysis patients: A population-
based study. Amer J Kidney Dis XIV(2): 119-123, Aug 1989.
18. Ishikawa I, Saito Y, Shikura N et al.: Ten-year prospective study on the development of
renal cell carcinoma in dialysis patients. Amer J Kidney Dis XVI(5): 452-458 Nov 1990.
19. Penn I: The changing pattern of posttransplant malignancies. Transplantation Proceedings
23(1): 1101-1103 Feb 1991.
20. Iwaki Y, Iguro T, Terasaki PI: Effect of sensitization on kidney allografts, in Terasaki PI
(ed) Clinical Kidney Transplants 1985 (Los Angeles, CA: UCLA Tissue Typing Laboratory),
pp 139-145, 1985.
21. Perdue ST: Risk factors for second transplants, in Terasaki PI (ed) Clinical Kidney
Transplants 1985 (Los Angeles, CA: UCLA Tissue Typing Laboratory), pp 191-203,
1985.
22. Almond PS, Matas AJ, Gillingham K et al.: Risk factors for second renal allografts
immunosuppressed with cyclosporine, Transplantation 52(2): 253-258, Aug 1991.
23. Holley JL, Shapiro R, Lopatin WB et al.: Obesity as a risk factor following cadaveric
renal transplantation. Transplantation 49(2): 387-389, 1990.
24. Garvin PJ, Castadeda M, Linderer R et al.: Management of hypercalcemic hyperpara-
thyroidism after renal Ttransplantation. Arch Surg 120: 578-583, 1985.
25. Dunn J, Golden D, Van Buren CT et al.: Causes of graft loss beyond two years in the
cyclosporine era. Transplantation 49(2): 349-353 Feb 1990.
26. Bay WH, Hebert LA: The living donor in kidney transplantation. Ann Int Med 106(5):
719-727, 1987.
27. Bertolatus JA: Renal transplantation for the nephrologist: Living donor kidney transplan-
tation: What did we learn during the 1980s? What should we learn during the 1990s? Am
J Kid Dis XVII: 596-599, 1991.
28. Mohr DN, Offord KP, Owen RA et al.: Asymptomatic microhematuria and urologic disease.
A population-based study. JAMA 256(2): 224-229, July 1986.
29. Rosenthal JT, Miserantino DP, Mendez R et al.: Extending the criteria for cadaver kidney
donors. Transplant Proc 22(2): 338-339, 1990.
30. Rao KV, Kasiske BL, Odlund MD et al.: Influence of cadaver donor age on post-
transplant renal function and graft outcome. Transplantation 49(1): 91-95, 1990.
31. Leunissen KML, Bosman FR, Nieman FHM et al.: Amplification of the nephrotoxic effect
of cyclosporine by preexistent chronic histological lesions in the kidney. Transplantation
48(4): 590-593, 1989.
32. Finn WF: Prevention of ischemic injury in renal transplantation. Kidney Int 37: 171-182,
1990.
33. Tilney NL, Kirkman RL: Surgical aspects of kidney transplantation. In Garavoy MR,
 187

Guttmann RD (eds) Renal Transplantation (Edinburgh: Churchill Livingstone), pp 93-123,

1986.
34. Belzer FO, Glass N, Sollinger H: Technical complications after renal transplantation. In
Morris PJ (ed.) Kidney Transplantation. Third Ed (Philadelphia: W.B. Saunders), pp
511-532, 1988.
35. Starzl TE, Broth CG, Putman CW et al.: Urological complications in 216 human
recipients of renal transplants. Ann Surg 172: 609, 1973.
36. Wilszek HE: Percutaneous needle biopsy of the renal allograft. Transplantation 50(5):
790-797, 1990.
37. Roza AM, Gallagher-Lapek S, Johnson CP et al.: Renal transplantation in patients more than
65 years old. Transplantation 48(4): 689-725, 1989.
38. Fehrman I, Brattstrom F, Duraj F et al.: Kidney transplantation in patients between 65
and 75 years of age. Transplantation Proceedings 21(1): 2018-2019, 1989.
39. Pirsch JD, Stratta RJ, Armbrust MJ et al.: Cadaveric renal transplantation with cyclosporine
in patients more than 60 years of age. Transplantation 47(2): 259-261, 1989.
40. Schulak JA, Mayes JT, Johnston KH et al.: Kidney transplantation in patients aged sixty
years and older. Surgery 108(4): 726-733, 1990.
41. Korb S, Kolovich R, Blackburn S et al.: Renal transplantation for older patients.
Transplantation Proceedings XX (Suppl 3): 201-203, 1988.
42. Murie JA, Lauffer G, Gray D et al.: Renal transplantation in the older patient.
Transplantation Proceedings 21(1): 2024-2025, 1989.
43. Cardella CJ, Harding ME, Abraham G et al.: Renal transplantation in older patients on
peritoneal dialysis. Transplantation Proceedings 21(1): 2022-2023, 1989.
44. Shah B, First MR, Munda R et al.: Current experience with renal transplantation in older
patients. Am J Kidney Dis XII(6): 516-523, 1988.
45. Howard RJ, Pfaff WW, Salomon D et al.: Kidney transplantation in older patients.
Transplantation Proceedings 21(1): 2020-2021, 1989.
46. Cuciani J, Frantz P, Thibault P et al.: Early anuria prevention in human kidney transplan-
tation. Transplantation 28(4): 308-312, 1979.
47. Carlier M, Squifflet JP, Pirson Y et al.: Maximal hydration during anesthesia increase
pulmonary arterial pressures and improves early function of human renal transplants.
Transplantation 34(4): 201-204, 1982.
48. Wizemann V, Kramer W: Choice of ESRD treatment strategy according to cardiac status.
Kidney lnt 33 (Suppl 24): S-191-S-195, 1988.
49. Himelman RB, Landzberg JS, Simonson JS et al.: Cardiac consequences of renal trans-
plantation: Changes in left ventricular morphology and function. J Am Coli Cardiol 12(4):
915-923, 1988.
50. Himelman RB, Landzberg JS, Simonson JS et al.: Cardiac consequences of renal trans-
plantation: changes in left ventricular morphology and function. J Am Coli Cardiol 12(4):
915-923, 1988.
50. Thistlethwaite JR, Stuart JK, Mayes JT et al.: Monitoring and complications of mono-
clonal therapy. complications and monitoring of OKT3 therapy. Am J Kidney Dis
XI(2):112-119, 1988.
51. D' Alessandro AM, Pirsch JD, Stratta RJ et al.: OKT3 salvage therapy in a quadruple immuno
suppressive protocol in cadaveric renal transplantation. Transplantation 47(2): 297-300 Feb
1989.
52. Dunn DL, Mayoral JL, Gillingham KJ et al.: Treatment of invasive cytomegalovirus disease
in solid organ transplant patients with ganciclovir. Transplantation 51(1): 98-106 Jan
1991.
53. Munda R. Hutchins M, et al.: Infection in OKT3-treated patients receiving additional
antirejection therapy. Transplantation Proceedings 21(1): 1763-1765, 1989.
54. Davis CL: In-depth review. The prevention of cytomegalovirus disease in renal trans-
plantation. Amer J Kidney Dis XVI(3): 175-188, Sept 1990.
55. Snydman DR, Werner BG, Heinze-Lacey Bet al.: Use of cytomegalovirus immune globulin
188

to prevent cytomegalovirus disease in renal transplant recipients. New Engl J Med 317:
1049-1054, 1987
56. Van Den Berg AP, Van Der Bij W, Wan Son WJ et al.: Cytomegalovirus antigenemia as
a useful marker of symptomatic cytomegalovirus infection after renal transplantation - a
report of 130 consecutive patients. Transplantation 48(6): 991-995, 1989.
57. Wade JC, Newton B, Flournoy N et al.: Oral acyclovir for prevention of herpes simplex
virus reactivation after marrow transplant. Arner Intern Med 100: 823, 1984.
58. Tomlanovich SJ, Sabatte-Caspill, J, Melzer J et al.: The incidence and impact of herpes
simplex virus infections in the first month following renal transplantation. Transplantation
Proceedings 21(1): 2091-2092, Feb 1989.
59. Rubin RH, Tolkoff-Rubin NE: Infections: The new problems. Transplant Proc 21(1):
1440-1445, 1989.
60. Maddux MS, Veremis SA, Bauma WD et al.: Effective prophylaxis of early post-
transplant urinary tract infections (UTI) in the cyclosporine (CSA) era. Transplantation
Proceedings 21(1): 2108-2109, Feb 1989.
61. Higgins RM, Bloom SL, Hopkin JM et al.: The risks and benefits of low-dose cotrimoxa-
zole prophylaxis for pneumocystis pneumonia in renal transplantation Transplantation 47(3):
558-560, 1989.
62. Shield CF: Use of OKT3 as prophylaxis in cadaveric renal transplantation. Transplantation
Proceedings 21(6) (Suppl 2): 15-18, 1989.
63. Monaco AP: Renal prophylaxis with orthoclone OKT3 in the United States. Transplantation
Proceedings 21(6) (Suppl 2): 7-13, 1989.
64. Melzer JS, Husing RM, Feduska NJ et al.: How to improve results for second renal allo-
grafts. Transplantation Proceedings XX(3) (Suppl 3): 176-179, June 1988.
65. Schroeder TJ, First MR, Mansour ME et al.: Prophylactic use of OKT3 in immunologic
high-risk cadaver renal transplant recipients. Arner J Kidney Dis XIV(5) (Suppl 2): 14-18
Nov 1989.
66. Benvenisty AI, Cohen D, Stegall MD et al: Improved results using OKT3 as induction
immunosuppression in renal allograft recipients with delayed graft function. Transplantation
49(2): 321-327, 1990.
67. Cohen DJ, Benvenisty AI, Cianci J et al.: OKT3 prophylaxis in cadaveric kidney trans-
plant recipients with delayed graft function. Am J Kidney Dis XIV(5) (Suppl 2): 19-27,
Nov 1989.
68. Michael HJ, Francos GC, Burke JF et al.: A comparison of the effects of cyclosporine versus
antilymphocyte globulin on delayed graft function in cadaver renal transplant recipients.
Transplantation 48(5): 805-808, 1989.
69. Pearson RC, Johnson RW, Bakran A et al.: A prospective study of prophylactic ATG
versus cyclosporine in regrafted and highly sensitized renal allograft recipients. Trans-
plantation 50(6): 1061-1063, 1990.
70. Todd PA, Brogden, RN: Muromonab CD3. A review of its pharmacology and therapeutic
potential. Drugs 37: 871-899, 1989.
71. Cockfield SM, Preiksaitis J et al.: Is sequential use of ALG and OKT3 in renal trans-
plants association with an increased incidence of fulminant posttransplant
Iymphoproliferative disorder? Transplantation Proceedings 23: 1:11 06-11 07, 1991.
72. Kahan BD: Individualization of cyclosporine therapy using pharmacokinetic and pharma-
codynamic parameters. Transplantation 5: 457, 1985.
73. Kahan BD, Shaw LM, Holt D et al.: Consensus document: Hawk's Cay meeting on
therapeutic drug monitoring of cyclosporine. Clin Chern 36(8): 1510-1516, 1990.
74. Kahan BD, Mickey R, Flechner SM et al.: Multivariate analysis of risk factors impacting
on immediate and eventual cadaver allograft survival in cyclosporine-treated recipients.
Transplantation 43(1): 65-70, 1987.
75. Salomon D, Brunson M, Vans ickier J et al.: A retrospective analysis of late renal graft
function: Correlation with mean cyclosporine levels and lack of evidence for chronic
cyclosporine toxicity. Transplantation Proceedings 23(1): 1018-1019, 1991.
 189

76. Wrenshall LE, Matas AJ, Canafax DM et al.: An increased incidence of late acute rejec-
tion episodes in cadaver renal allograft recipients given azathioprine, cyclosporine, and
prednisone. Transplantation 50(2): 233-237, 1990.
77. Vanderwerf BA, Serota AI: Low-dose cyclosporine for cadaveric renal transplantation.
Transplantation 45(2): 320-323, 1988.
78. Delmonico FL, Conti D, Auchincloss H Jr et al.: Long-term, low-dose cyclosporine treat-
ment of renal allograft recipients. Transplantation 49: 899-904, 1990.
79. Lewis RM, Van Buren CT, Radovancevie B et al.: Impact of long-term cyclosporine
immunosuppressive therapy on native kidneys versus renal allografts: Serial renal function
in heart and kidney transplant recipients. J Heart and Lung Transplant 10(1): 63-70, Jan/Feb
1991.
80. Brinker KR, Dickerman RM, Gonwa TA et al.: A randomized trial comparing double-
drug and triple-drug therapy in primary cadaveric renal transplants. Transplantation 50(1):
43-49, 1990.
81. Kramer NC, Peters TG, Rohr MS et al.: Benefician effect of cyclosporine on renal trans-
plantation. Transplantation 49(2): 343-348, 1990.
82. Hall BM, Tiller DJ, Hardie I et al.: Comparison of three immunosuppressive regimens in
cadaver renal transplantation: Long-term cyclosporine, short-term cyclosporine followed
by azathioprine and prednisolone, and azathioprine and prednisolone without cyclosporine.
New Engl J Med 313(23): 1499-1507, 1988.
83. Helling TS, Wood BC, Nelson PW et al.: Elective conversion from cyclosporine to aza-
thioprine in sensitized patients following cadaveric renal transplantation. Am J Kidney Dis
XV(2): 137-140, 1990.
84. Schulak JA, Mayes JT, Moritz CE et al.: A prospective randomized trial of prednisone versus
no prednisone maintenance therapy in cyclosporine-treated and azathioprine-treated renal
transplant patients. Transplantation 49(2): 327-332, 1990.
85. Reisman L, Lieberman KV, Burrows L et al.: Follow-up of cyclosporine-treated pediatric
renal allograft recipients after cessation of prednisone. Transplantation 49(1): 76-80, 1990.
86. Zanker B, Walz G, Wieder KJ et al.: Evidence that glucocorticosteroids block expression
of the human interleukin-6 gene by accessory cells. Transplantation 49(1): 183-185, Jan
1990.
87. Furue M, Ishibashi Y: Differential regulation by dexamethasone and cyclosporine of human
T cells activated by various stimuli. Transplantation 52(3): 522-526, 1991.
88. Griffin PJA, Gomes Da Costa CA, Salaman JR: A controlled trial of steroids in cyclosporine-
treated renal transplant recipients. Transplantation 43(4): 505-508, 1987.
89. Meyers BD, Sibley R, Newton L et al.: The long-term course of cyclosporine-associated
chronic nephropathy. Kidney Int 33: 590-600, 1988.
90. Mihatsch MJ, Thiel G, Ryffel B: Histopathology of cyclosporine nephrotoxicity. Transplant
Proc 20 (Suppl 3): 759-771, 1988.
91. Kasiske BL, Heim-Duthoy K, Venkateswara R et al.: The relationship between cyclosporine
pharmacokinetic parameters and subsequent acute rejection in renal transplant recipients.
Transplantation 46(5): 716-722, 1988.
92. Mathew TH: In-depth review. Recurrence of disease following renal transplantation. Am J
Kidney Dis XII(2): 85-96, 1988.
93. Stevenson JA, Dumke A, Glassock RJ et al.: Thrombotic microangiopathy: recurrence
following renal transplant and response to plasma infusion. Am J Nephrol 2: 227-231,
1982.
94. Hebert D, Sibley RK, Mauer SM. Recurrence of hemolytic uremic syndrome in renal
transplant recipients. Kidney International 30: S-51-5-58, 1986.
95. Cameron JS, Senguttuvan P, Hartley B et al.: Focal segmental glomerulosclerosis in fifty-
nine renal allografts from a single centre: Analysis of risk factors for recurrence. Transplant
Proc 21(1): 2117-2118, 1989.
96. Bruan WE: Long-term complications of renal transplantation. Kidney Int 37: 1363-1378,
1990.
190

97. Mahony JF, Caterson RJ, Pollock CA et al.: Coronary artery disease is the major late
complication of successful cadaveric renal transplantation. Clin Transplantation 4: 129-132,
1990.
98. Kasiske BL, Tortorice KL, Heim-Duthoy KL et al.: Lovastatin treatment of hypercholes-
terolemia in renal transplant recipients. Transplantation 49(1): 95-100, 1990.
99. Glicklich DG, Tellis VA, Quinn T et al.: Comparison of captopril scan and doppler ultra-
sonography as screening tests for transplant renal artery stenosis. Transplantation 49(1):
217-219, 1990.
100. Palmer BF, Dawidson I, Sagalowsky A et al.: Improved outcome of cadaveric renal trans-
plantation due to calcium channel blockers. Transplantation. 52(4): 640-645, 1991.
101. Kirk AJB, Omar I, Dark JH: Long-term improvement in renal function using nifedipine
in cyclosporine-associated hypertension. Transplantation 50(6): 1061, Dec 1990.
102. Shapiro R, Carroll PB, Andreas GT et al.: Adrenal reserve in renal transplant recipients with
cyclosporine, azathioprine, and prednisone immunosuppression. Transplantation 49(5):
1011-1013, 1990.
CHAPTER 12

Quality assurance In dialysis product manufacturing

RICHARD S. THUMA

Introduction

Quality Assurance has evolved from an essentially narrow, manufacturing-

based discipline to one with much broader implications for management.
Outside the confines of the manufacturing facility, quality has tremendous
impact on the strategic direction, operations, and, profitability of whole
industries. Quality assurance, as applied to the manufacture of products, focuses
on three phases of the manufacturing activity: (1) control of raw materials
and components; (2) process control; and (3) evaluation of outcomes. It is
popular to think of the entire manufacturing and control process as a system.
But in these days of increasing competitive pressure, focusing on control
of the three manufacturing stages in the manufacturing process is not enough.
An operation that seeks to develop competitive advantage must think in terms
of continuously improving its processes. This concept of Continuous Quality
Improvement (CQI) is one of the most powerful in quality assurance and is
symbolized by the Plan-Do-Check-Act cycle shown as Figure 1. In a manu-
facturing setting, CQI elements might employed as follows:

CQI

Fig. 1. Continuous quality improvement elements.

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 191-211.

© 1994 Kluwer Academic Publishers.
192

1. Use of statistical process control techniques (SPC) to monitor the

manufacturing processes. If necessary, the processes are broken into sub-
processes. SPC is used to identify process variation.
2. When unacceptable levels of process variation are detected using SPC, their
root causes are identified (see Fig. 2).
3. Once the cause has been identified, possible improvements can be deter-
mined.
4. The acceptability of the improvements must be determined using appro-
priate statistical rigor.
5. Finally, when improvements have been validated, they are implemented
and the process monitoring resumes.
In terms of general system theory, a manufacturing system may be viewed
as composed of three steps:
Inputs ::::} Process ::::} Outputs
Inputs to processes are both raw materials and information. Processes consume
the raw materials and information and produce outputs; the outputs can be
thought of as a mixture of two types: useful output (e.g., defect-free product)
and unusable output (scrap, energy loss, defective product). Juran [1], in an
interesting parallel, made the observation that, in order to achieve any level
of quality, the costs could be divided into avoidable and unavoidable.
Unavoidable costs were associated with prevention - inspections, sampling,
sorting, and other quality control initiatives. Avoidable costs were those related
to defects and product failures - scrapped materials, labor hours required to
rework and repair, complaint processing and financial losses resulting from
unhappy customers.

Engineering
Redesign

Resolve
wlvendor

Train Operstors 1--------.. .

Fig. 2. Generic problem-solving process applied to a manufacturing environment.
 193

Quality Assurance attempts to introduce systems for controlling processes

and setting checkpoints at strategic locations in the process. Its objective is
to increase the amount of useful output from the process. It does this by
focusing on prevention, i.e., controlling the quality of the inputs (both raw
materials and information must be controlled), monitoring the process in
real-time, and checking the outputs.
The purpose of this chapter is to give a brief overview of the manufac-
turing processes used to produce Hemodialysis and Peritoneal dialysis products.
Its emphasis will be on hemodialyzer and PD solution manufacturing. It is
not intended to be an exhaustive treatise on the mechanics of quality control
and quality assurance as applied to the manufacture of dialysis products, but
rather to point out the factors that determine quality control and assurance
process development.
Within the drug and medical device industry, some aspects of manufacturing
control systems are driven by the regulatory requirements applied to the
products produced. Peritoneal Dialysis solutions, for instance, are classified
as drug products, and are treated, by FDA and other regulatory agencies and
Ministries of Health worldwide, quite differently than hemodialyzers, which
are classified as devices.
The complexity of the manufacturing process and the quality assurance
strategies applied to it will vary, depending on the degree of vertical inte-
gration present within the manufacturing company. Highly vertically integrated
companies manufacture virtually all the components used in assembly of the
finished product; less vertically integrated operations source some or all of
their components from other manufacturers; still others purchase the finished
products complete with their own labeling.
A simplified manufacturing and control process is shown as Figure 3. This

Raw Mafl
Component
Inventory

Manufacturing SPC or Mfg In-process QA Final Finished

Process Inspections ~ Inspection ~ Goods

+ +
glllllllX B!!~Ul M!!!!tlOgs
• Field Problems Review
• Revlaw of returned product analysis results
• Review of rework Issues
• Revlaw of QC Data

Fig. 3. Simplified manufacturing and control process diagram.

194

generic process can be applied to the production of all dialysis products,

including hardware. The strategy the manufacturer uses to continually improve
the quality of its manufacturing processes is by monitoring each stage of the
process, use of statistical tools (e.g., Statistical process control or statistical
sampling procedures), in-process and finished goods inspections, and review
of quality data.
Raw Material/Component Inventory - Manufacturer's generally adopt one
or the other of two alternative strategies: (1) either they maintain capacity
and resources for testing and validating that incoming raw materials and
components meet specifications, and or (2) they "certify" some or all of their
vendors. Vendor certification is a process whereby the manufacturer lever-
ages the vendor to perform the component and material qualification studies
and certify the results meet the manufacturers specification. The focus is on
the critical parameters. Certification shortens the feedback loop to supplier
since he has responsibility for testing to specifications.
Manufacturing/Assembly Process - Generally, renal therapy products are
either manufactured in "lot" quantities or assembled one at a time (e.g.,
instruments). Process control strategies differ for each. For instance, depending
on the product, daily production status meetings might be held to review issues
identified during each day's production cycle or each work shift. Such meetings
would probably include inspection results, raw material/component problems,
and other issues. Problems that are identified would be assigned to correc-
tive action teams for root-cause determination and resolution.
Measurements reported by shift, daily, weekly, and monthly vary from
manufacturer to manufacturer and from product to product, but commonly
include such quantities as Process Capability (CpK), Defects per Million
(DPM), Defects per instrument, Final line rejects, and so on. Pareto charts,
process capability charts, fishbone analysis and other analytical and problem-
solving techniques are common quality and statistical tools used to focus on
the issues and determine root causes. Manufacturers train employees in the use
of such tools at regular intervals.
SPC/In-process Inspections - Statistical Process Control (SPC) X-bar and
R-charts are commonly used measurement methods in manufacturing plants.
These charts plot the mean (X-bar) and Ranges (R-charts) of attributes found
in a sample taken from the manufacturing process. The resultant points are
used to determine if the process is running within control limits. The control
limits are determined over time. Many manufacturing plants use SPC to
monitor every product manufacturing operation and have set process capability
goals. Other plants may use SPC to a lesser degree, dependent on the product
line, for instance, in instrument manufacturing, there is a greater reliance on
in-process inspection since each instrument is assembled one-at-a-time. Results
from SPC and inspections are used immediately, if an out of control condi-
tion is detected.
Final Inspection - all lots of products, including each instrument, receive
a final inspection by quality assurance personnel. These inspections are usually
 195

based on statistical sampling plans. Examples of measurements at this

manufacturing stage include final line rejects, conversion loss (scrap), release
times, yield, and product performance data (with trends).

Production of peritoneal dialysis solutions

PD products are made up of two basic components: a solution of dextrose,

electrolytes and water and a container - most often a bag. The bag is a clear
plastic container, usually of PVC although some suppliers use semi-rigid
polyethylene or other materials. The semi-ridgid containers make visual
checking of the solution difficult, since they are not transparent. Most
manufacturers make PD solutions available in a wide range of fill volumes
(1.5, 2.0, 2.5, or 3.0 liters) and concentrations of dextrose (1.5%, 2.5% or
4.25%). This wide variety of fill volumes and concentrations gives the
physician substantial flexibility in designing the treatment regimen. Figure 4
illustrates the manufacturing process for PD solutions.
Within each manufacturing stage, there are a number of manufacturing
and quality considerations as well as issues with potential impact on the therapy
(Fig. 5). A number of these critical factors are discussed below.

A. Raw Material and Components

Quality factors at this stage focus on a number of complex elements:

(1) Plastic variations - caused by minor variation in the component resins
used in the extrusion process. These rarely cause significant issues beyond
minor mixing and extrusion control issues. However, absence or con-
tamination of any of the minor ingredients can cause very significant
changes in the physical and chemical properties of the resulting film so
control of component additions is a key manufacturing process step.
(2) Extrusion defects - are generally only an issue during the process of
forming the bag or semi-rigid container. Some factors that cause extru-
sion defects include extrusion temperature, amount of regrind and
temperature of extrusion.
Manufacturers often use the scrap plastic collected during the container
forming stage and regrind it. This regrind is used in the extrusion of
new container material. The maximum percentage of reclaimed material
used in the extrusion process is controlled by the manufacturer and based
on appropriate analytical and toxicological testing. One of the problems
with regrind is related to the potential mixing of differing materials in
the reground materials. Mixing of materials will often cause tiny bits
of plastic to burn during the extrusion process, resulting in "pepper"
particulate matter embedded in the plastic sheeting.
(3) Seal integrity - in the formation of the bag or container, usually two halves
 .....
1.0
0\

Fig. 4. Simplified process diagram for manufacture of peritoneal dialysis solutions.

 Raw Mat'l & FIlling Sealing & Package and
Overpouch Sterilization
Components Process ACCllas Shipping

o Particulates oWater Quality o Solvent Variability o Material Choice o Orientation on o Materials

o Plastic Variations oFiltration o Seal Integrity o Sealing Process Sterilization Truck o Load Patterns
o Extrustion Def eets o Sanitization • Solvent Type o Loading Process o Carton Sealing
o Seal Integrity o Personnel o Dimensional o Cycle Variations o Labeling
o Solvent Constraints o Mat'l Variations Tolerances o Mix to Sterilization Time o Pest Control
o Solvent Variations o Mixing Process o Fill to Sterilization Time o Distribution Method
o Labeling o Fill Vol. Variations o Cooling Cycle o Storage Effects
Manufacturing o Vendor Approvals o MIg Environment o Sterilization to Pack o Traceability
laaues o Incoming Inspeetior oAnalytical Methods Time o Overall GMP
oFilter Testing Compliance
oStorage Conditions

o Leachables ~ Pyrogens o Sterility Barrier o Bioburden Levels o Package Integrity

o Clarity
o PVC Blend o Leachables • Color Coding o Moisture o Pyrogen Testing
o Pyrogens o Degradation o Microorganisms o Dextrose Degredation
Therapy o Antigenic Sub· Products o Ease of Opening Products
l8auea stances '" Surlace-to-
o Trace Elements Volume Ratio
o Bag Clarity ~ Stability
'" PM Tolerances

Fig. 4. A tabulation of factors that influence the quality strategies used in the manufacture of peritoneal dialysis solutions.
......
10
--.J
198

of the container are joined, using a sealing process. The integrity of the
seal is a function of the method chosen for forming the seal as well as
the composition and thickness of the container material. Some semi-
rigid containers are blow molded in one piece so seal integrity is an
issue only when access ports are added to the container.
(4) Solvent constraints - solvents are used routinely to seal medication and
access ports to the container. Most manufacturers use cyclohexanone, a
compound whose general use is coming under some scrutiny in some parts
of the world; cyclohexanone is prohibited in Japan, so suppliers in Japan
use Ultrasonic (RF) sealing strategies - which are usually more expen-
sive and difficult to control than solvent systems.
(5) Leachables - because the PD solutions are in contact with the container
for potentially long periods of time (most manufacturers put two or more
years of dating on their products), the plastic formulation must be shown
to be non-toxic and the long-term leachability of plastic components and
plasticizers must be taken into account in the design of the container
and choice of materials.
PVC has been used for many years in the production of medical devices
and as containers for solutions. This is due partly to its excellent bio-
compatibility, ease of processing, and several other characteristics. Most
of the medical grade plasticized PVC is consumed in the manufacture
of containers, gloves, and tubing. In the manufacture of flexible bags
for solutions, DEHP is the only plasticizer used routinely.
DEHP is the most widely used plasticizer in flexible PVC for medical
applications and its toxicological properties have been examined most
extensively. The acute toxicity of DEHP is very low.
When receiving fluids stored in PVC containers, patients may be
exposed to plasticizers; PVC tubing used either for administration of fluids
or for dialysis may also contribute to the exposure.
Data on DEHP exposure in patients undergoing medical treatment show
a wide variation. One contributor to this variation is explained by the
fact that the reliability of the analytical methods differ. However, recent
evidence estimates exposure by Hemodialysis patients at 50-370 mgt
treatment [2, 3] and Peritoneal Dialysis patients at 0.1-20 mg/day
[4, 5]. It should be noted that even the highest exposures of DEHP in
dialysis patients provides a significant margin of safety when compared
with levels of exposure which produced no effect in animal studies [6].
(6) Pyrogens - both the solutions and components must be non-pyrogenic
so pyrogen testing is a part of every manufacturers process. The pyrogen
test most frequently employed is the Limulous Amebocyte Lysate Test
(LAL test), an in-vitro test for bacterial endotoxins. The test detects the
presence of bacterial endotoxins by the formation of protein clots.
Extensions of the test include use of enzyme-linked reagents that cleave
chromophores in the presence of the endotoxins and allow a direct
colorimetric measure of the presence of the pyrogenic substances.
 199

From a quality assurance point of view, the LAL test requires appro-
priate reference materials as positive controls, proper training of personnel,
and, perhaps most important of all, pyrogen-free glassware and equipment
used in the assay.
The LAL test will detect only bacterial endotoxins and will therefore
be completely ineffective at detecting chemical pyrogens, therefore, in
some areas of the world, the rabbit pyrogen test is still performed.
(7) Trace elements - can be a concern, especially aluminum, and must be
minimized. This is usually accomplished through control of water used
to make the solutions and very tight control of the analyte raw mate-
rials.
(8) Bag clarity - the end user must be able to determine acceptability of
the solution (cloudy solutions may signal bacterial contamination or
presence of particulate matter). Since this is facilitated by the clarity of
the bag, the formulation of the container is important. As mentioned
earlier, most blow-molded and semi-rigid containers are not transparent,
thus making it difficult for the end user to determine solution acceptability.

B. Filling Process

(1) Water quality - pH, trace element content, and pyrogenicity all affect the
final product. Water processing and testing is a critical step in the
manufacturing process. Location of the manufacturing plant is determined
by a number of economic and logistical factors, including the avail-
ability of good water.
In the manufacturing process, the pH of the solution is lowered to about
5.5 to prevent caramelization of the glucose during sterilization.
(2) Filter and filter testing - filter testing is required before use and is accom-
plished by using a pressure test (bubble test). This assures filter pack
integrity. The objective of filtration is not sterilization but rather, by
passing the product through a series of filters of decreasing pore sizes,
elimination of particulate matter.
(3) Fill volume variations - the filling process uses automated filling machines
that may be fed manually or automatically. Checks on fill volume are
carried out by weighing samples at set intervals. Process control charts
(X-bar and R-charts) are common methods used to track fill variations
during the filling process. Filled container weight variations that fall
outside the control limits for the process trigger immediate corrective
action.
(4) Manufacturing environment - filling generally takes place in a controlled
access, class 100,000 or class 10,000 clean room to control bioburden
and particulate matter. Access to the room is usually restricted and may
require donning special lint-free uniforms, shoe covers, head covers, and
gloves.
200

(5) Degradation products - these byproducts of dextrose degradation are a

direct result of the formulation. The magnitude of their formation is
influenced by the pH of the solution and the mixing, filling and steril-
ization processes. The primary degradation product is usually 5-Hydroxy
Methyl Furfurol (5-HMF). Excessive amounts of 5-HMF may produce
pain on infusion in some patients.
(6) Suiface-to-volume ratio - the amount of solution placed in the container
directly affects both the sterilization cycle time required to attain sterility
and the stability of the final solution. Composition of the container as well
as surface to volume ratios affect the water vapor transmission rate of
the container and thus the solution stability.
(7) Particulate matter - standards regarding particulate matter (PM) have been
set by USP and their counterparts around the world. In the U.S., PD
solutions must meet USP standards as applied to injectables.

C. Overpouch

(1) Choice of material - material composition determines (1) if the overpouch

can function as a sterile barrier; (2) whether the end-user can see the
solution and read the labeling; and (3) ease of opening the overpouch -
all are important manufacturing and end-user issues.
(2) Sealing process - since the overpouch must be sealed, many of the same
sealing process issues confronting container manufacturing will arise
during this process phase. Some strategies to minimize the physical
problem of inserting the filled bag into the overpouch include inserting
the bag into the overpouch before filling.
(3) Moisture - usually condenses between the overpouch and the container
after the sterilization process. This makes detection of leaking bags more
difficult, however, appearance of excessive moisture may be an indication
of a leaky bag.

D. Sterilization

(1) Mix-to-sterilization times affect opportunities for bacterial growth. Manu-

facturers limit the amount of time filled units wait prior to sterilization.
This issue is aggravated where manufacturing takes place in hot climates
and units are staged in hot warehouse areas for long periods of time
prior to sterilization.
(2) Physical/Handling issues such as
• Loading process
• Orientation on the sterilization truck
have potential impact on the final product. Some manufacturers have
automated the loading process in order to avoid possibility of overlap-
 201

ping containers. Overlapped containers can result in units ineffectively

sterilized or in prolonged cooling cycles which affect the color of the final
solution as well as the formation of 5-HMF.
(3) Sterilization cycle variations affect the amount of degradation products
present in the solution; degree of caramelization of dextrose is also
affected. Extended cycles or inefficient cooling cycles may cause dis-
coloration of the solution and a buildup of 5-HMF.

E. Package and Shipping

Here, many of the same issues covered in previous sections apply: choice of
packaging materials, load patterns on pallets, sealing of packing materials,
labeling, storage effects and pest control.

Hemodialyzer manufacturing

For the purposes of this discussion, only hollow fiber (capillary) dialyzers
will be considered. Conceptually, the dialyzer is a tubular container
separated into two compartments: a dialysate compartment and a blood com-
partment. A semipermeable membrane, made up of a bundle of thousands of
small, hollow fibers, separates the two. Access to the two compartments is
through two pairs of ports.
In hemodialysis, the blood flows from the patient into a port at one end
of the dialyzer and is channeled through the hollow fibers and back to the
patient. The dialysate solution flows around the outside of the fibers.
A number of advantages and disadvantages of hollow-fiber dialyzers can be
articulated [7] today with most attention focused on the membrane materials.
In general, membrane materials are classified into one of three types: cellu-
loses (cuprammonium cellulose, cuprammonium rayon), modified celluloses
(cellulose acetate and triacetate), and synthetics (polyacrylonitrile, PAN,
poly sulfone, polymethylmethacrylate). The various membrane types differ in
their physiologic effect. For purposes of this discussion, it is only important
to recognize that the membrane type determines many of the manufacturing
and quality control strategies.
Similar to the manufacture of PD solutions, there exist many factors that
influence the quality and production strategies used to produce dialyzers. A
number of these factors are tabulated in Figure 6 and described in more detail
below.
The manufacturing process flow for dialyzers can be generalized into the
steps tabulated in Figure 7. A number of these process steps are only relevant
for highly vertically integrated producers.
 1:5
N

A••emble Flnalln.pactlon
Raw Mat'l. Fiber Membrane Potting, Cutting Packaging. Storage and
and Sterilization
Components Winding Treatment Leak Te.tlng and Cleaning Shipping

o Component o Fiber Type o Lumen Ruld o Header Attach- o Type of o Sterile Barrier o Sampling Plan
Variation o Breaks o Solvent Type ment Method Compound Choice o Pyrogen
o Defects o Winding Method o Method o Tip Protector o Potting Method o Sterilization Testing
oVendor o Number of o Bundle Insertion Material o Polymerization Method o Leak Testing
Approvals spools Method o Stress Cracking Time o Effect on Mat'l o PM Testing
o Incoming oTension Control o Potting Cap - o Tip Protector • Cure Time o Safety o Performance
Manufacturing Inspections • Process Speed Plastic Type Fit • Cutting Method oLoad Pattems Testing
Isaue. o Pilot Lots • Bundle Wrap o Dryvs. Wet o Cutting Blade o Cycle Variations • Storage
o Injection molding o Pressure Parameters o Lumen Fluid Conditions
Decay vs. Gas o Durometer of o Degassing Time • Carton Sealing
Composition. Cured Potting o Pest Control
Compound • Distribution
Method
oTraceability
oOverallGMP
Compliance
o Leachables oHeader Leaks o Toxicology o Sterility o Pyrogens
Therapy o Plastic Blends o Tip Protector Fit o Leachables o Brittle Fibers • Performance
Issu•• o Pyrogens o Sterility of Blood • EtO Absorbtion o EtO Residuals
o Trace Path • Blood Leaks oBioburden
Elements o Fiber Leaks • Rattened Fibers
o PM Tolerances

Fig. 6. Tabulation of factors that influence the quality strategies used in the manufacture of dialyzers.
 203

Fig. 7. Hemodialyzer manufacturing process diagram.

A. Raw materials

Quality systems within manufacturing sites generally identify procurement

of raw materials and components as the first critical step in the manu-
facturing process. Most manufacturers are not totally vertically integrated
204

and source at least some raw materials and components from outside
vendors.
A dialyzer is made up of several basic components (see Fig. 8), including
a plastic case, headers or end caps, a hollow fiber bundle, potting compound,
and tip protectors. The heart of the dialyzer is the fiber bundle. Since the
performance of the dialyzer must meet certain AAMI guidelines, manufac-
turers must carefully qualify each lot of fiber used in the manufacturing
process.

Fig. 8. Schematic diagram of a typical holow fiber dialyzer.

When manufacturers use vendors to supply fiber, often the receiving and
inspection process includes physiologic testing and actual production of a pilot
lot of dialyzers using various lots of fiber furnished by the vendor.
Manufacturer's have a choice with respect to the fiber raw material, if it
is to be "wound" at the dialyzer manufacturing location, it is received as spools,
in which case the process of inspection includes identification, certification,
and permeability testing. On the other hand, if winding capability is missing
at the manufacturing site, it will be received as bundles of fibers already wound
and cut into the proper size for insertion into the dialyzer case. When fiber
bundles are received, inspection includes checking length and diameter of
the bundle as well as permeability or porosity.
As far as the rest of the dialyzer componentry is concerned, materials that
make up the case and headers (the rigid parts) are critical and must be of
appropriate materials to withstand the stresses of the manufacturing process
including the sterilization method (i.e., EtO, gamma, and steam). Also, the
chemicals used by dialysis clinics to process dialyzers for reuse may affect
some case and other component materials dramatically. Polycarbonate is the
most common case material but some manufacturers use acrylic. The choice
of materials is somewhat limited by the potential toxicity of the materials
and the potential for leachables to affect the dialysate and blood.
Other critical aspects of the case include the port connectors - both dialysate
and blood; both need to conform to international standards. There are two types
 205

of dialysate connectors: Hanson and Walter; the difference between the two
being the Walter connector incorporates an internal locking system with a
check-valve (most often found in European dialysis clinics).
The quality strategy in manufacturing revolves around the choice of
Statistical Process Control (SPC) vs. inspection by attributes after manufac-
ture.

B. Winding process

Manufacturers generally use either the "wheel" or "beam" as a strategy for

winding the fibers and forming the fiber bundle (see Fig. 9).
With either method, the winding apparatus is rotated about a central axis
while fiber is wound from multiple spools. The use of multiple spools and
the winding process causes the fibers to be laid down in a "woven pattern."
The more spools used, the less variability between dialyzers from a fiber
performance point of view, and the better the woven pattern. From an manu-
facturing and economics point of view, the use of the wheel winding method
forces a trade off between increased manufacturing speed and scrap levels
while the beam winding method allows winding of more fragile membranes
and generally produces less fiber scrap.

Wheel Winding Method

Fber
opools

Beam Winding Method

Fig. 9. Diagram of two fiber bundle winding methods.

206

Tension control as the fiber is wound is critical and can be accomplished

in two ways (1) resistance at the spool, or (2) use of a "dancer" arm (a
method borrowed from the textile industry). Of the two, the dancer arm applies
a more consistent pressure to the fiber as it is wound, however, while the
dancer arm may cause more consistent pressure application, it does have a
significant disadvantage: when more fragile fibers are wound, this method
brings the fiber into contact with an additional surface and may raise the
probability of fiber abrasions and subsequent weakening.
Generally, the type of fiber dictates the winding and tension control strate-
gies. Cell uloses are tougher fibers than synthetics and are readily adapted to
the wheel winding method while the more fragile synthetics do best using
the beam method.
The winding process may break some fibers that would lead to blood
leaks. Manufacturers rely on two methods to detect leaks: mechanical or
electronic. In either case, when a break is detected, the ends must be found
and tied off within the bundle.
Immediately after winding, a wrap is placed over the wound fibers. This
provides some degree of configuration control, protection and facilitates further
processing. The wrap is usually a sheet of polyethylene or polycarbonate which
is taped in place. Once winding is completed and the wrap applied, the bundles
are cut from the wound fiber.

C. Membrane treatment

The need for further processing of the membrane depends on whether or

not the fiber contains lumen fluid With some types of membranes, lumen
fluid is put in place during the membrane manufacturing process and serves
to keep the fibers from collapsing. If the membrane contains lumen fluid, it
must be removed at some point in the dialyzer manufacturing process. Typically
this is done by centrifugation. Some manufacturers remove the lumen fluid
after winding and before assembly and some don't. One of the reasons for
not removing the lumen fluid prior to potting is that treatment is relatively
inefficient because of the difficulty of forcing solvent through the unpotted
fibers. In any case, a solvent system, usually consisting of alcohol, alcohol and
glycerol, freon, or a freon-alcohol azeotrope, is employed to wash the fluid out
of the fibers.
Most synthetic membranes are spun with water and have no lumen fluid.

D. Membrane insertion

Once any treatment of the fiber bundles has been completed, they are inserted
into the dialyzer case, either manually or by some automated process. After
insertion, plastic potting caps are put in place. The purpose of the cap is to
 207

create a barrier at each end of the dialyzer to contain the potting compound
during the potting process. These caps are either partially retained or removed
after potting.

E. Potting

Potting is the process of encapsulating the fibers by imbedding the ends in a

compound that is polymerized after it is applied to the ends of the assem-
bled dialyzer. All hollow fiber dialyzers contain potting compound.
There are two fundamental potting compound types: aliphatic and aromatic
polyurethanes. Their application is a matter of preference for the manufacturer.
Potting compound selection must be done with consideration given to toxi-
cological matters, including the property of the potting compound to "trap"
or absorb materials used in the manufacturing and sterilization process (e.g.,
Eta residuals). For instance, there is some evidence that ethylene oxide
residues are trapped in higher concentrations in aromatic-based than aliphatic-
based compounds. However, this problem can be easily solved by more
intensive degassing procedures.
The actual potting process may be either static (i.e., dipping the ends of
the dialyzer into the polymerizing solution), or based on centrifugal force.
In the centrifugal process, the potting mixture is poured into the assembled
dialyzer which is then centrifuged at a rate sufficient to keep the material at
the ends of the case until polymerization occurs. Centrifugation is the most
popular method for potting because it is easier to control compound distrib-
ution. Polymerization time is a critical quality factor and relates to extractables.
A key quality factor of the potting compound, the potting procedure and
in regulating the polymer mixture is the accuracy of the ratios of the two
compounds used in the mixture. Poor quality potting leads to blood leaks at
the header and toxicological problems from residual unpolymerized material.
The polymer ratio also affects potting compound adherence to the dialyzer
case. Other quality problems that occur during the potting cycle include blocked
fibers and fiber bundles that pull away from the edge of the case.
Cure time after potting is critical. It determines how soon the next step in
the process can be intiated and varies by urethane type from a few hours to
a few days.

F. Cutting

After encapsulation and curing, a cutting process is used to reopen the end
of the membrane - essentially by slicing off the ends of the potting cap. This
is generally done using knives which either perform shear or radial cuts. In
radial cutting, blades are spun at a high rate of speed, like a radial saw blade,
and applied to the dialyzer end. In shear cutting, a blade is pressed down on
208

the dialyzer potting cap ends, slicing it off. There is no question that the
spinning radial blade method for cutting is a more rapid process but the
blades tend to generate additional particulate matter which must be removed
in later processing steps.
In the cutting process, factors affecting the quality of the cut and, conse-
quently, the fiber, are:
• number of passes made by the cutting blades
• blade geometry
• blade rotation speed
• penetration speed
• blade angle
All of these parameters, are affected by the durometer (or flexibility) of the
cured urethane potting compound, type of membrane used, and the size of
the bundle. Common defects that influence quality of the final product include
flatness and membrane adherence to the urethane. If cutting results in severe
flattening of the fibers, it can influence the fit of the header. Also, dull or
high speed cuts cause fibers to pull away from the urethane.
The choice of potting compounds also affects the cutting process. There
is some evidence that aromatic urethanes cut better.

G. Cleaning

Once cut, the dialyzer must be cleaned before the caps (headers) are applied
to the ends. The purpose of cleaning is both to remove particulate matter
and other by-products of the manufacturing process and remove lumen fluid.
Similar to the cleaning process described previously, this step is accom-
plished using solvents, usually one of the following: Freon, Freon-alcohol
azeotrope, alcohol + glycerol, water (for synthetics), or inert gas (if no lumen
fluid is present).
The actual cleaning process uses either a bath or a flushlflow through
method. Cleaning the blood path is the primary objective. Both methods are
multi-stage. No matter what solvent system is used, critical to the process is
the reclaim procedure (for both economic and environmental reasons).
At present most manufacturers use Freon, which is scheduled for phase-
out by 1995. This already presents substantial problems for dialyzer
manufacturers because there is no "drop-in" replacement solvent for the freon
washing system.

H. Assembly

Once cleaned the headers and tip protectors are assembled to the case con-
taining the potted fiber bundles.
Headers are attached using three differing methods: torqued screw-on, ultra-
 209

sonic sealing, and adhesive. It is not uncommon for manufacturers to use a

combination of two of these methods. The purpose of the sealing is to prevent
blood leaks from the headers. In those geographical locations where there is
a trend toward re-use, selection of dialyzers has narrowed to those whose
headers are torqued on, since this facilitates cleaning the dialyzers.
The tip protectors, assembled on the ends of the blood ports, are gas
permeable but used to assure sterility of the blood path. Material selection is
critical, since sterilization method choice may expose the tips to EtO, gamma
radiation, or steam.
Most quality issues with headers are related to development of stress
cracking. Stress cracking is an injection molding issue related to tempera-
ture of injection and molding cycle time.

1. Leak testing

All manufacturers use a 100% leak testing process. The effectiveness of the
testing depends on the type of leak testing selected. There are two basic
types: (1) pressure decay testing, where about 20 PSI filtered air is applied
to the dialyzer and a change in pressure with time determined; and (2) gas
composition testing. The gas composition method is considered to be signif-
icantly more sensitive in leak detection.
Pressure decay testing options include (1) pressure blood side; (2) pressure
dialysate side; (3) negative pressure the blood side; (4) negative pressure the
dialysate side. Whatever leak testing strategy is chosen by the manufacturer,
it should be understood that leak testing does not totally prevent leakers.

J. Packaging

There are two strategies for the packaging of the finished dialyzer: either
(1) the package provides sterile barrier; or (2) the package is only a dust
cover - the tip protectors provide sterility assurance.
The type of packaging selected by the manufacturer is determined by both
the sterilization method and whether the package will represent a sterile barrier.

K. Sterilization

There are three common method for sterilization. Each has its own set of
advantages and disadvantages.
Gamma irradiation is fairly common and is very effective. However, irra-
diation has a negative effect on most components of the dialyzer - it tends
to make some fibers more brittle and can cause discoloration of the plastic case
and other components. This effect is related to the total sterilizing radiation
210

dose received by the dialyzer. Manufacturers use dosimeters to measure the

average exposure and assure that a minimum exposure is achieved (usually
about 1.8 mrad). The manufacturer must develop data to support the minimum
exposure dose.
Ethylene oxide (EtO) has been a popular sterilizing agent for many years
and is considered very effective but has some disadvantages, including the fact
that it is a toxic, mutagenic, and the sterilizing gas mixture is explosive.
Furthermore, emission of EtO residues into the environment pose a real
disposal problem for the manufacturing facility and a potential health hazard
for the environment in the vicinity of the manufacturing plant.
Another consideration is the degassing time required for product after
sterilization. It can take several weeks for product to degas enough to reach
what is considered a safe EtO level. Additionally, FDA and other regulatory
agencies worldwide have begun to focus on and develop guidelines limiting
EtO residuals which remain in the products and which then expose the patient
to these materials. These residuals are ethylene glycol, ethylene chlorhydrin,
and ethylene oxide. Manufacturers must measure EtO residuals in the products
they distribute and assure quantities of residuals fall below minimum estab-
lished levels. In the U.S., the current limits of these compounds is 250, 25, and
25 ppm respectively. Presently, work is progressing in the development of
ISO standards and a Tripartate Agreement that will base residual limits on
the gram weight of the device. These regulatory standards cause manufacturers
to rely on forced aeration strategies for reducing residuals in their products.
Steam sterilization is restricted to only certain membrane types.
Ultimately, however, it is the choice of membrane materials that is the
primary determinant of sterilization method. Membranes with high glycerin
concentrations in the lumen fluid usually cannot be gamma or steam sterilized.

L. Final testing

Once the dialyzer has been sterilized, the manufacturer, using a statistical
sampling plan, performs some testing on the finished product. This testing
usually includes pyrogen testing using the LAL method, leak testing, blood
path monitoring (wherein the dialyzer is flushed with a known volume of
fluid and the effluent examined for particulate matter), performance testing,
including UFR and clearance measurements.

Conclusion

Manufacturers are faced with a great number of environmental, material-

related, process-related, regulatory-driven, and people-dependent factors that
influence the manufacturability of dialysis products. Some of these factors
impact the therapy alone while others are restricted to manufacturing control
 211

issues. These factors can be generally classified as falling into one or more
of the following categories: biology, pyrogenicity, toxicity, chemistry, and
general manufacturing. It is only by applying quality and manufacturing
technology that the manufacturer can provide a truly quality product to the
patient.

References

1. Juran JM, Gyrna FM Jr, Bingham RS Jr: Quality Control Handbook, fourth ed. (Manchester,
MO: McGraw-Hill), 1988.
2. Pollack GM et al.: Circulating concentrations of di(2-ethylhexyl)pthalate and its
deestrified pthalic acid products following plasticizer exposure in patients receiving
haemodialysis. Toxicol Appl Pharmacol 79: 257-267, 1985.
3. Ganning AE et al.: Regular exposure from plasticizer can have detrimental effects on health.
Liikartidningen 81: 4389-4392, 1984.
4. Henderson IS et al.: Factors affecting plasticizer content of unused CAPD fluid. Peritoneal
Dialysis Bulletin Suppl 7(2): 539, 1987.
5. Nassberger Let al.: Exposure of patients to phthalates from polyvinyl chloride tubes and
bags during dialysis. Nephron 45: 286-290, 1987.
6. European Chemical Council (CEFIC) draft position paper on Medical Applications of
Plasticised PVC, 1993.
7. Van Stone JC: Hemodialysis Apparatus. In Daugirdas JT, Ing TS (eds) Handbook of Dialysis
(Brown: Little, Brown and Co), 1988.
CHAPTER 13

Human resource issues in quality management

GAIL S. WICK

While Deming and Juran [3, 7] focus heavily on the use of statistical tech-
niques in total quality management, they also emphasize management's
leadership role as a key factor in success. State of the art equipment, abundant
supplies and pleasant environment facilitate provision of quality products
and services, but ultimately the humans making and using products and
providing services determine the success or failure of all quality manage-
ment efforts. As Hoesing and Roey [2] point out, employees' behavior or
performance determine patient clinical and satisfaction outcomes. These
outcomes drive organizational profit (market share and margin). Clearly then,
quality management (quality assurance, control and improvement) is an
employee driven concept.
Although employees ultimately determine success or failure of quality
efforts, the fact that "workers are responsible for only 15% of quality problems
and the system for the other 85%" illustrates the challenging task that man-
agement faces. It must create a system in which quality management activities
can flourish and succeed rather than flounder and die [7]. The purpose of
this chapter is to focus on the responsibility and role of management in leading
personnel to and through quality management in the health care setting.
Corporate culture must embrace and nurture the concept of quality manage-
ment and do this through judicious implementation, evaluation and continuous
improvement. While quality is a total company commitment, it must come
from the top down.

Quality management

Quality is built into the process by developing and promoting standards that
reflect the organization's definition of quality services. Standards are supported
by policies while procedures promote positive outcome. In other words,
standards state what the organization believes quality is. Standards can also
be used to monitor, evaluate and for improvement [2]. Policy states clear
expectations of what can and is to be done; and procedures define the exact

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 213-223.

© 1994 Kluwer Academic Publishers.
214

steps or process which will achieve the desired outcome in an efficient manner.
The role of a company's employees is to assist by (1) defining quality through
the development of unit specific standards, (2) identifying indicators to measure
quality, (3) determining procedures to achieve it and (4) developing a futur-
istic plan of continual improvement based on statistical analysis and problem
solving.
Management is responsible for the education of the entire organization on
how to "do it right" and then leading it into a commitment. Mission, values
and guiding principles must be established early, because they form the basis
for all other activities. An organization's mission statement communicates
the broad vision and purpose of the organization, and gives employees
direction and focus to pursue their tasks to achieve this mission. The mission,
along with guiding principles of operation, clarifies the important roles that
all employees of the organization have in accomplishing this mission. From
this mission statement, goals which clearly communicate how to get, maintain
and measure quality can be collaboratively established [7]. When an organi-
zation clearly states and consistently acts on its commitment to quality, then
pride, security, loyalty and commitment will increase.

Quality management in the dialysis setting

As with any concept, there are barriers to implementation and success. Deming
[2] outlines certain barriers that result in quality failure. These barriers include:
and emphasis on numbers, not quality; turning out products (services) quickly
rather than properly; turning a deaf ear to suggestions; too much time spent
on rework; using poor or improper tools and difficulties with incoming
materials. These barriers can be found within the dialysis setting and must
be identified in order to guard against them. Are staff mix and staff/patient
ratios set regardless of patient acuity? Do caregivers rush to get patients on
and off dialysis to meet "quotas" in lieu of proper assessment, following proper
procedure and clinical trouble shooting? Are preventive care and patient
satisfaction conscious goals? Have staff meetings and inservices been cut from
the schedule for perceived lack of time thus hindering important communi-
cation and educational opportunities? Are unnecessary complications and
deficiencies occurring due to the implied need to focus on mechanics and
quotas rather than the total health care picture? Is the staff apathetic and
dispirited? Are tools for quality management and care such as statistical
tools, state of the art equipment, adequate supplies and qualified personnel
available to accomplish quality goals? Management must identify barriers
and eliminate them through effective leadership and provision of adequate
resources.
 215

The human resource issues in quality management

Juran [3] states that the goal of leadership activity is to (1) improve the
system through input from the team and (2) to facilitate consistency in
performance so that over time, employees perform with fewer variations from
the established process. In other words, the goal is for everyone to do what
he is supposed to and for the team to continually evaluate and improve the
process in order to get the desired results. This can only occur if manage-
ment selects the right team, prepares team members for their roles within
the organization and provides them with a work environment that is con-
ducive to the provision of quality services.
Key human resource issues in quality management are therefore (1) the
selection process, (2) training and education activities and (3) team develop-
ment. Provision of quality services must be built into the system. Human
resources must be maximized and self leadership capabilities must be
promoted.

The selection process

Central to all quality activities within a company is the selection of employees

with the necessary skills, knowledge and experience to do the job right, and
who have the values and work ethics to enhance quality management goals.
As Leeds [4] points out "nowhere do you have more at stake than in hiring
a new employee". Quality staffing is based on credentials, references, work
ethic, attitude and compatible values and goals. Organizational standards will
not influence an employee's behavior, actions or attitude unless he accepts and
values those standards. For this reason, identification of potential employees
with similar values and work ethics is as important, if not more important, than
experience and educational preparation. Knowledge and skills can be taught;
values and work ethics are not easily changed. The goal, then, of the selec-
tion process is to select the person who brings all the necessary requirements
to the position. If quality management is important to the organization, then
commitment to quality must also be a value that a potential employee
possesses. Interview questions which focus on this include:
"What constitutes quality care?"
"What did you do in your last job to improve care and the work
environment?"
"What do you think your specific role will be in quality assurance
and improvement in this position?"
"What connection do you see between process and patient outcome?
For example, what relationship does technique have to infection? How
do you prevent or minimize complications?"
216

Acceptable answers to these questions reinforce the feeling that to enhance

a quality management program, a person must have self discipline and assume
responsibility for hislher own actions. The prospective employee must also
clearly understand what will be expected in terms of performance and attitude.
Showing the appraisal format that will be used for evaluations clarifies the
relationship of standards, employee performance and quality management in
written form.
To begin the selection process one must define what constitutes the right
people for the job and team. A role description is not enough for this task.
Specific job requirements, standards which indicate quality, characteristics
of the person most likely to meet these standards and specific values and
work ethics that match or complement those of the company must be identi-
fied. Identification of beliefs is particularly important since beliefs are
fundamental to thinking [5] and, ultimately, to behavior. Questions [4] which
help clarify these issues include:
What specific work or tasks need to be done?
What knowledge and skills are necessary to successfully accomplish
the desired work?
What are the toughest aspects of the job? What characteristics are
necessary to cope with them?
What constitutes quality in this particular job?
What talents, knowledge, skills and attitudes are necessary to accom-
plish it?
What is the company's mission? What are its values? What values will
complement the company's values?
What type of person will complement the current team?
Once this job analysis is finalized, the interviewing process, if conducted
properly, will identify those individuals who will enhance the team's pursuit
of quality. The interviewing process also identifies education and training needs
as well as self leadership tendencies. Information to assist the applicant to make
an informed decision about the position should also given. When a manager
hires an employee, he assumes the responsibility for that employee's success
or failure, as well as hislher contribution to the organization whether it be
positive or negative. Great care must therefore be taken to hire the right
person for the position and once hires, to set them up for success in the position.
Thus, the selection process is tremendously important and should not be under-
estimated.
 217

Setting the stage for success

Once the manager has made a commitment to an employee through the hiring
process, the challenge is to provide the employee with the necessary knowl-
edge, skills and resources to meet, and hopefully exceed, the quality standards
established by the institution. Education and training lay the groundwork for
positive patient outcome and practitioner performance, both of which directly
affect bottom line as well as quality. An additional benefit is that a clear
understanding of expectations and indicators of acceptable performance
provides focus and direction for personnel. Managers and subordinates can
collaboratively measure and monitor performance as well as outcome.
Initial training and education for a direct care giver in the dialysis setting
should include, but not be limited to the following:
• overview of chronic and acute renal failure with indications for therapy,
including the principles and practice of dialysis
• nutritional and psychosocial considerations
• manifestations of renal failure
• medical complications of dialytic therapy
• patient teaching needs and techniques
• specific procedures and rationale for the steps
• the interrelatedness of technique, assessment and preventative measures
to quality care
Regardless of exact content, nationally recognized standards of practice such
as those of the American Nephrology Nurses Association and the American
Association of Medical Instrumentation, should serve as the basis for educa-
tional programs. These standards can also be useful guidelines to coach,
counsel and appraise personnel involved in direct patient care as well as the
technical activities of dialysis.
Quality management requires a cooperative atmosphere where departmental
goals enhance the overall efforts of the organization. For this reason, team
members must understand the role and relationship of all departments to the
quality management effort. Orientation of support and non-direct care per-
sonnel should include a basic introduction to ESRD, its treatment, financial
aspects of the program as well as challenges to quality management. Direct
care givers, in tum, must understand the specific role and importance of other
departments in the quality effort.
Because quality management is so important to overall patient outcome
and program success, employees must be taught additional techniques and
skills. These should include problem solving skills, creative brain storming,
communication skills, time management and data collection and analysis.
All are important factors in a successful employee driven quality manage-
ment program. The principles and practice of care identify (1) what aspects
of care are high volume, high risk or problematic and (2) reasonable indica-
tors for monitoring quality. Data collection and statistical analysis come next,
followed by activities which require the ability to identify problems, imple-
218

ment solutions and evaluate results in a team environment. Viable solutions

come from a team environment which promotes creative brain storming without
fear of secondary agendas. Clear communication promotes successful outcome
of the entire quality process. Time management skills insure that adequate time
is available even in a busy facility.
Initial and continuing education sessions as well as staff meetings are ideal
times to reinforce the organization's mission, values, and commitment to
quality. Standards of work as well as patient care, and the specific role that
teamwork plays in quality management are crucial to successfully meeting
organizational commitments. Expectations, responsibilities and areas of
accountability must be clearly communicated. Well informed, knowledge-
able employees whose values match those of the employer are able to perform
in a safer, more efficient and effective manner. In other words, they can provide
quality care and services. Shortcutting initial training and education is a
costly mistake both in management time and productivity as well as quality.
Formal training and education are necessary for quality management [7].
Part of the educational process for all employees of the dialysis facility
includes understanding quality management and its relationship to patient care.
Acceptance of the pursuit of quality on a continuous basis as part of one's work
ethic is an essential requirement in employees. Employees who are taught
the principles of quality and application of the principles in a cost and time
efficient manner are prepared to design better ways to do things, and thus,
improve quality. Quality assurance and quality improvement are often thought
of as difficult, perhaps unpleasant, but necessary tasks. This negative perception
embraces the obstacle thinking concept described by Manz and Sims [7]. For
quality management to succeed, the manager's challenge and responsibility are
to get subordinates to change their behaviors in areas that adversely affect
process and outcome.
For professional growth, a highly rated motivational factor, and commit-
ment to quality to be sustained, opportunities for continued education and input
into quality issues must be provided. This can take the form of regularly
scheduled inhouse in services and seminars, time off to attend outside offer-
ings, and financial assistance for education. Suggested topics are theoretical
advances, current and future issues in nephrology, principles and practice of
dialysis related to new technology, and problem solving skills in the clinical
setting. These help arrest the tendency to focus only on the technical aspects
of dialysis versus the team oriented care of the total patient and disease entity.
Employees who have lost focus of the total picture or who have stopped
learning, do not strive to continuously improve care nor do they necessarily
stay with the organization. Many of those who do stay lack true commit-
ment to quality service and improvement.
 219

Environments that foster quality

Work environments that foster quality services and products demand the highest
standard of care and work. They are consistent in promoting quality man-
agement and strive to keep employees satisfied. They provide for clear and
frequent two way communication and have a committed team that is
dedicated to the concept of quality.
In corporation of quality management programs has an added bonus of
helping create a success oriented work environment. Three steps to incor-
poration of quality management programs include (1) careful and honest
assessment of the current situation, (2) analysis of the answers and gearing
up for action, and (3) development of programs which give employees
something to work toward [8]. Quality management carried out by self-
managed teams can provide most, if not all of the conditions employees feel
are most important in a work environment.

Standards and goals

Because people like to work for and with the best, they expect certain stan-
dards and rules which promote the best service or product. Helping to establish
the standards fosters ownership. Much of the time, the goals and standards
set by employees are as high if not higher then those of management. In
addition, people forced to carry out standards and goals that they had no
hand in establishing, do so half heartedly and without uniformity [7]. This can
be very damaging to quality efforts because quality requires total commit-
ment throughout the organization as well as uniformity in quality performance.
Self imposed goals are an important ingredient for successful self leader-
ship. They help employees establish self direction and priorities for immediate
work tasks and long term career achievements [5]. If employees have been
carefully selected and adequately educated and trained, they are likely to choose
goals which promote quality as well as support the principles of the organi-
zation. When employees establish goals they are more likely to strive to
make the process and outcome successful. Active involvement in the direc-
tion of the organization also self guards against boredom and loss of focus
on the ultimate goal - quality. Good questions to ask in assessing status and
setting new goals include:
• What did we do right this week?
• What do we want to accomplish within the next 90 days?
• How can we do it?
• What can we do to make things better?
Once agreed upon, goals must be clearly communicated and followup
monitored.
In addition to establishing standards and goals, employees can and should
be involved in regularly answering the following:
220

• What are the outcomes of our efforts?

• Are we achieving the desired patient outcomes?
• What barriers have we encountered to quality? Why?
• How can we overcome them?
• Are we using the best and most efficient process?
• What do we want to happen this week? Next week? Next month? Next year?
• Has patient satisfaction increased? If not, why? If so, why?
• Is performance improving or status quo? Why?
• Are we staying within our cost restraints?
• Are the cost restraints reasonable?
Answering these simple questions is the foundation of the complex issue
of quality assurance and improvement.
While establishing goals is important, assuring that the goals of the various
departments complement rather than conflict with each other must also be
considered and carefully monitored. Teamwork and effective communication
are instrumental in achieving this.

Teamwork

Teamwork in the quality management environment is critical. Not only must

goals and activities not conflict, but frequent communication must occur to
prevent overlap and redundant activities. All departments and personnel
categories must recognize that they share in the vision or mission of the
organization. Although different levels of responsibilities may exist, all must
have the same commitment and direction. Duplication of effort must be avoided
and all activities should compliment the effort rather than conflict with or
take away from efforts. Clear, frequent communication and coordination of
activities are necessary before, during and after action has been taken to
avoid duplication of effort and conflict as well as to facilitate organization wide
problem solving [2]. A good model for achieving this is the self-managing
team concept proposed by Manz and Sims [5]. This concept places a high
degree of decision making and problem solving control within the jurisdic-
tion of teams that have been carefully trained and educated in their specialty.
In addition, they should be well trained in the theory and practice of group
dynamics, quality management and effective communication and problem
solving. Although models vary, one which should prove effective in the dialysis
setting with respect to quality management is the three hierarchial level model.
Upper management comprises the first level called the support team. This
team handles many of the routine management responsibilities such as strategic
planning, contract negotiation and broad operational activities. The support
team supervises and provides support rather than direction for the next level
called work-team coordinators. Coordinators are responsible for the overall
activity of one or more teams. The third level is called the self-managing
team and may have anywhere from three to twenty or more members. A leader
 221

is elected for each team and has the responsibility for facilitating meetings,
communication and problem solving as well for his/her normal work respon-
sibilities. A pay differential accompanies the added leadership responsibilities.
The purpose of the self-managed team is give control of performance and
outcome of efforts to employees who have been prepared philosophically
and educationally for such responsibilities. The responsibilities include such
areas as budget preparation, scheduling, data collection and analysis and
problem solving. Communication within and between teams is facilitated
through specific channels. Short, weekly team meetings provide the formal
arena for communication and group problem solving, while daily communi-
cation among team members is possible because of close work proximity.
Coordinators become involved in team discussions only if necessary for
specific problems or situations. Communication between teams and the three
team levels is maintained through posted written agendas and minutes of all
meetings along with regularly scheduled leadership level meetings. Information
sharing is free and open with the exception of those matters which are truly
personal. Motivation and discipline come from within the teams, thus the
term self-leadership. They do not come from management. Problems and
quality issues are identified and solved by the teams rather than by manage-
ment. Solutions are therefore employee owned, which increases motivation
to make the plan of correction and subsequent improvement succeed. In
actuality, when practiced successfully, the self-managed team concept addresses
almost all the qualities that people want most from their job [1] which are
to:
• work for efficient managers
• think for themselves
• see the end result of their work efforts
• be assigned interesting work
• be informed
• be listened to
• be respected
• be recognized for the efforts
• be challenged
• have opportunities for increased skill development
Quality is the end product of the coordinated efforts of all departments
within an organization. Self-managed transdepartmental teams with well estab-
lished channels for communication is a mechanism for serious consideration.
The role of management is to influence the team members to provide quality
and to create a positive work environment themselves rather than for man-
agement to issue directives or to do it themselves [5]. In other words, control
lies with employees rather than management. Since quality control is an
employee driven activity, the organizing, directing and monitoring functions
of self-leadership and self-managed teams should promote successful quality
programs if adequately planned for and implemented. The enthusiasm and
motivation necessary to assure quality within a dialysis unit can come from
222

a fully developed team which works together to solve problems and to do

the job well [5]. Responsibility and the rewards of a job well done are shared
by all.
Whether assessing the work environment for motivational purposes or
quality management purposes, the resulting plan of action involves input and
action on the part of personnel. Therefore, it is critical to utilize the princi-
ples of communication and feedback. Personnel, if involved in the process,
become owners of the program or goals, thereby increasing the determina-
tion to succeed. Additionally, loyalty to the organization and its program are
increased.

Job satisfaction

A sense of achievement increases job satisfaction and thus employee retention.

With an active quality management program, dialysis is no longer just a task.
The overall focus shifts to care for the total person. The entire organization
becomes involved. The focus shifts to methods to achieve better results. An
added feature is increased efficiency, therefore decreased energy expenditure
as well as financial outlay. The work itself has the potential to become much
more rewarding with increased opportunities for growth through learning
and experience. The roles of all persons from the receptionist to the billing
department to the direct care staff become apparent and the interrelatedness
more appreciated.

Summary

Care of the person with renal insufficiency is a complex and challenging

task for all persons on the health care team. A renal health care team consists
of both direct care providers as well as support personnel. The finest equip-
ment is only as good as those producing and using it.
Quality management in any health care setting is no easy task and dialysis
is no exception. Dialysis is a complex process which demands total commit-
ment, adequate resources and time to plan, organize and evaluate. Departing
from traditional human resource management may be necessary to insure
provision of quality care as well as successful quality management programs.
Quality, whether from the manufacturing or health care provider perspec-
tives, requires commitment to and implementation by individuals. For this
reason, leadership is one of the major factors controlling quality. Fortunately,
the leadership skills and attitude~ necessary to promote quality are well
established and proven. Leadership, particularly leadership in the form of
helping others to lead themselves and to take control of their performance
and outcome, is required from the initial selection process to the final analysis
of outcome.
 223

References

1. Dell T: An Honest Day's Work: Motivating Employees to Give Their Best (Los Altos,
California: Crisp Publications), 1989.
2. Hoesing H, Kirk R: Common sense quality management. JONA 20, Number 10, October
1990.
3. Juran JM: Juran on Leadership for Quality, An Executive Handbook (New York: The Free
Press), 1989.
4. Leeds D: Smart Questions - A New Strategy for Successful Managers (New York: McGraw
Hill), 1987.
5. Manz C, Sims H: Super-Leadership (New York: Prentice Hall Press), 1989.
6. Peters T, Austin N: A Passion for Excellence (New York: Random House), 1985.
7. Walton M: The Deming Management Method (New York: Putnam Publishing Company),
1986.
8. Wick G: Stimulating peak performance in the dialysis setting. NRAA Journal 11, Number
I, Spring/Summer, 1990.
CHAPTER 14

Quality criteria for the clinical record

CINDY JO PING, SUSAN GROSS and CORRINE E. ALGRIM

The clinical records in dialysis vary from the official medical record of the
patient to a multitude of unofficial, yet vital, forms and tracking systems.
All of these allow the clinician to monitor the many variables associated
with management of acute and chronic dialysis patients. Quality manage-
ment improvement in the clinical dialysis setting requires the tracking and
trending of data which can be used to identify and analyze areas for poten-
tial change which will bring about a new and hopefully a better level of care
for the renal patient.

Documentation systems

In order to implement and maintain data systems which will provide mean-
ingful information for continuous quality improvement, the clinical records
must be evaluated for their relevancy to the services being delivered and for
the monitoring of those services through the quality improvement process. This
evaluation process is ongoing and action-oriented. Data collection systems
should support the evaluation process and be designed to facilitate easy
identification of the trends, actions, and outcomes. Finally, the data collec-
tion systems should support activities which will lead to future improvements
in service.
Continuous improvement of health care services is being promoted by
regulatory and advisory agencies in health care today. One such entity is the
Joint Commission of Accreditation for Health Care Organizations (JCAHO),
which is implementing a plan of change over the next five years [1]. This
plan will require hospitals, and therefore hospital-based dialysis units, to
demonstrate that components of the Continuous Quality Improvement CCQI)
[2] model are integrated into the quality monitoring process in use in the
facility. This process was developed by W. Edwards Deming and Walter
Shewhart [3-5]. The essential elements of their theories are: (1) Establish a
constancy of purpose, (2) Utilize statistical analysis for evaluation of the quality
of services, (3) Examine the variations found in the statistical analysis, (4) Use

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 225-238.

© 1994 Kluwer Academic Publishers.
226

the knowledge gained from the analysis to correct the systems and bring
about improvement in the quality of the service.
In the future, it is reasonable to expect mandatory guidelines will be
developed based on national standards, that will require both hospital-based
and free-standing dialysis units to meet criteria which will measure the quality
of patient outcomes. It is also reasonable to expect that the process of
measuring these standards will be based on the principles of continuous quality
improvement.

Measurement

Searching for better ways to provide dialysis treatment is not new to dialysis
professionals who have been involved with the research and application of
the technology. What is new, however, is the concept of measuring the rela-
tionship of clinical response of the patient to technological advances.
Conceptually, both in industry and in the healthcare setting, measurement of
quality has focused on the methods and the products used in the treatment
of patients as well as the quality of the environment where patient care was
given. These methods for measuring quality focused on the process and
structure of the organization. Now the emphasis of measuring quality is focused
on the patients' response to the health care intervention. The emphasis on
outcomes has added a new dimension to health care quality measurement [6-7].
Previously, quality assurance plans for health care concentrated on counting
the presence or absence of specific criteria. Outcome based quality monitors
encourage the use of tracking information for the purpose of identifying trends
on which decisions regarding the quality of the care being delivered can be
made. The additional concept of continuous improvement then becomes the
driving force for elevating the level of care beyond established industry
standards.
Continuous Quality Improvement acts as a catalyst. The catalyst is the
challenge Berwick refers to as a continuous reach for opportunities to improve
all aspects of health care [8]. This dimension stimulates action to intervene
and improve patient care and decrease risk through system changes and
improvement in practice.
A search for opportunities to improve care should not require intensive
detective work to uncover meaningful trends in the care being given. Trends
should be easily identified from the data collection process. In addition, the
people who collect the data are usually the patient care staff who are invested
in improving the outcome of what they do for patients. Since time allotted
for patient care is the priority in the clinical setting, data collection systems
should not detract from the caregiver's primary responsibility. The data systems
should support the care given both from a legal perspective, as well as from
a treatment perspective. Therefore, the data that is recorded should be concise,
accurate, timely, legible, factual and clinically pertinent [9]. The data
 227

collection method developed for the clinical dialysis setting should promote
these criteria. Additionally, it should be streamlined to achieve efficiency in
the daily work of all patient care providers.

Analysis of documentation

Continuous quality improvement is a model for managers to use as a basic

foundation for all systems which support the end product, in this case,
documentation of the dialysis patient's treatment and the results of those
treatments. Application of the CQI model requires that the results of treatments
be reviewed for opportunities to improve. This analysis is derived from many
sources. The clinical records in the dialysis setting have evolved to meet the
requirements of policies, procedures, guidelines, protocols, regulations and
laws. These internal and external factors have impacted not only the amount
of data, but the kind of data kept by the dialysis centers. This proliferation
of data has resulted in numerous forms required to initiate and manage patient
care over an extended period of time. Data overload among staff, therefore
is not uncommon. Due to the volumes of data requested or required on a
daily basis, the staff may become less concerned about the accuracy and
legibility of what they are recording. When faced with too much data, staff
may not refer to previous documentation to assist them in assessments of
current problems, let alone, use the data to search for opportunities to improve.
In some situations tests are repeated unnecessarily because the results could
not be found. This has a negative impact on the efficiency and effectiveness
of the care that can be given to the patient and increases the cost of
providing the dialysis treatment.
One solution to the documentation issue is computerization of the clinical
record [10-12]. What must occur before computerization can take place,
however, is the same process which needs to take place in a system without
computer support. A complete evaluation of all documentation sources and
current practice is the first step before any plan for revision can take place.
The following five questions should be asked by every dialysis manager in
order to identify existing areas of integration and overlap:
1. Who is collecting and reporting the data?
2. What data should be collected?
3. When is the data collected and recorded - concurrently or retrospectively?
4. Where is the data collected and where is the data analyzed and reported?
5. Why is the data collected and reported? Is it useful to do so?
6. Is it a regulatory requirement?
By carefully analyzing the answers to each of these questions, areas of
duplication and overlap should begin to become apparent. Further analysis can
then identify each factor necessary for the achievement of a goal. In the clinical
dialysis setting, cause and effect diagraming would assist the manager in
organizing the clinical data record based on the individuals who record the
228

patient treatment information. Flow diagraming is an excellent tool to use when

determining the effectiveness or efficiency of a system [13]. By plotting each
step required to complete the dialysis treatment, the points of interdependence
are identified before the pro.::ess can move on to the next step and it becomes
clearer which points cause the process to stall. The manager can then analyze
what changes in the system will make the process flow more smoothly. This
applies to the clinical record as well. If information is not recorded so that
it can be readily found and used efficiently, cause and effect diagraming and
flow diagraming can help depict the relationships between the steps required
to produce a complete record or to identify points where the process of charting
becomes cumbersome and breaks down.

Influences of the clinical setting

The Clinical setting also influences the type of data necessary for a quality
recording system. Location and acuity level are influencing factors in the
planning for documentation systems. The patient who is acutely ill and
receiving treatment in an intensive care unit will have many core data elements
which are similar to the stabilized patient and the home dialysis patient. On
the other hand, each of these groups of patients have a set of data which is
unique to their setting and care requirements. In addition to the location and
acuity of the patient population, the need to interface with other health care
providers also influences the design of the documentation system. In the
hospital setting, the dialysis record must be part of the centralized medical
record system. In the free-standing dialysis clinic, mechanisms must be in place
for the communication of clinical data between the center and other health
systems which provide support to the nephrology team for the care of the
patient. The home dialysis record must support an even broader network of
health care providers such as the patient's local physician, community health
agencies and home equipment and supply companies.

Criteria for design

Finally, the data should be reviewed for its accuracy and validity. The facts
relating to the patient's care should be recorded in a timely manner and in
an organized format so that there is a clear chronological picture of what
has happened to the patient during the course of the treatment by each provider.
The people who use the documentation system are the best individuals to
ask for input when designing a new form or an entire system. If the system
is simple, user-friendly, flexible and designed with future improvements in
mind such as computerization, the system will be used. The forms them-
selves should be organized and printed so as to guide the user in a sequential
and efficient manner from the assessment of patient care through planning,
 229

implementation and evaluation. This process then sets the data collection in
a logical sequence for monitoring continuous quality improvement oppor-
tunities.

The medical record

The most familiar document in the clinical setting is the patient's medical
record. It is a tool for communication between professionals about the care
of the patient. It serves as a diary of chronological facts pertaining to the
diagnosis and course of the treatment and serves as a reservoir of data for
evaluation of the comprehensive treatment of the patient. A good medical
record is a map of the what, when, and how a patient received care [14]. The
medical record also acts as the one permanent record of care and treatment
of the patient from admission to discharge. A well kept record can assist the
provider in substantiating standards of practice as well as standards of care
based on recorded outcomes. On the other hand, a poorly kept record can make
anyone appear to have given poor care [9].
Good medical records provide an accurate account of what has happened
to the patient with evidence of interventions, rationale by the provider and
response to the intervention by the patient. The medical record also serves
as a tool from which useful data for research and development of new
approaches to patient management can be obtained. Given the potential uses
for the medical record, another requirement is protection of the information
to maintain patient confidentiality [15]. Medical records should be kept in
an area which can be secured from the general public and only individuals with
official business and approved releases should be allowed to use the records
for any purpose.
The contents of the dialysis medical record are influenced by many legal,
regulatory, accrediting and policy directives. The following list is a com-
posite of some of these directives. They have been divided into recommended
and optional; requirements may differ from unit to unit depending on the
location of the unit.

MEDICAL RECORD
Recommended Optional
Intake Admissions Data Chart Checklist
Dialysis Consent Pre Dialysis Checklist
Medication Record Problem List
Long Term Care Plan Health Maintenance Record
Modality Selection Transfusion Record
Pre and Post Dialytic Assessment Research Protocols
Modality Selection
Intradialytic Flow Sheet
Dialysis Adequacy
Patient Care Plan by Dialysis Team Individualized
230

Recommended (Continued)
Laboratory and Diagnostics
Medical Directives (Living Wills)
Narrative Records
Patient Bill of Rights
Patient Grievance Procedure
Viral Surveillance
Infection Rates
History and Physical
Discharge Summary

Facility records

In addition to the patient medical record many unofficial records must be

maintained by the facility in order to meet license, certification, accredita-
tion and legal requirements. Once again, these lists have been divided into
recommended and optional. Many of these records support the systems
necessary to run the units, maintain equipment or staff the patient care area
and each, once again, may vary in their requirements depending on the location
of the facility.

FACILITY RECORDS
Recommended Optional
Water Treatment Records Mission Statement
Reuse Records Research Protocols
Machine Maintenance Records Patient Classification System
Credentialing Records Career Ladder Records
Performance Appraisals Continuing Education
Morbidity Mortality Data Annual Report
Patient Standards of Care Operational Directives
Policy and Procedures Organizational Chart
Infection Control
Staffing Patterns
Required Individual Orientation and Inservice
Records
Facility Mandatory Inservice Records: Fire/Safety,
CPR, Infection Control
Quality Program
Occurrence Reports
Staff Meeting Records
Hazardous Waste Program
Medical ByLaws
Governing Board Minutes
On Call Records
Home Supply Records
Billing Records
 231

Both of these lists of records found in the dialysis setting have many
common data elements. Because the number of reporting requirements have
grown over time, much of the information is repeated from form to form.
Reducing the amount of recording by staff through an organized and effi-
cient information system is the first step in ensuring quality of care in the
dialysis unit.
Information systems which streamline the process of recording, interfacing
and integrating the data will lead to better and more efficient use of the records
by staff in their care of the patient. Without a streamlined information system,
quality programs become unmanageable, discouraging and eventually ignored.
Once the information systems for recording and maintaining patient data
has been established, these records should be key tools or sources for quality
monitoring data [16].

The clinical record and quality monitoring

The most common format for monitoring quality in the clinical setting today
is based on the JeARO 10 Step model [17]. Adaptions of this model to
incorporate more continuous improvement principles has already occurred and
over time the model will evolve as the concept becomes refined in the health
care setting. The relevance of this model and of the Deming industrial model
or the adaption of Deming for health care by Batalden and Vorlicky is that
there is a definitive evaluation plan [4]. This plan will guide an organization
to establish standards of quality, monitor the organization's ability to uphold
the standards and ultimately to improve upon the standards.
The goals of quality monitoring are:
1. To assure identification and correction of system wide and unit based
problems.
2. To promote care that is cost effective and clinically effective.
3. To promote optimal patient care through the ongoing assessment of all
aspects of care, together with the correction of identified problems.
4. To identify and reduce risk factors while at the same time managing known
risks.
5. To meet the requirements of external agencies.
In the future new models will be developed to approach quality improve-
ment, but the basic structure of the JCARO model is simple and easy to
implement in a relatively short period of time.
The JCARO 10 Step model can be compared to other models already in
place in the health care field. The nursing process is one example where a
systematic approach of problem analysis can be used to determine the care a
patient will need and the patient's response to the care. In research, the problem
solving process is orderly and disciplined to acquire reliable information and
to analyze that information in an orderly manner.
All of these models or processes focus on improving quality of care and
232

should be a dynamic and therefore, a continuous process. Documentation

should support the outcome of improvement through recognition of varia-
tions from established standards as a result of data analyzed from the clinical
record [18].
Quality improvement is an analytical process and requires critical thinking
on the part of the individuals involved. Critical thinking by staff should lead
to actions which promote successful experiences. A team working together can
be more creative and innovative than one individual working alone toward a
solution. The analytical process for the team should include brain-storming
which is free from professional barriers or the stigma of organizational hier-
archy. This type of team process will lead to better problem solving and
efficiencies will increase.
Quality improvement ideally produces standards which surpass minimum
expectations [19]. It begins with assessment of state, federal, and institu-
tional requirements. Once the minimum requirements are identified, one can
explore the need for expansion, modification or reduction of current docu-
mentation or record keeping systems. The goal should be to reduce unnecessary
paperwork and unnecessary recording. A good documentation system should
support proactive planning for quality improvement.
Most dialysis centers have some type of quality monitoring program in
place, but may not know how to direct the documentation system toward
quality improvement [20-21]. Once again this must begin with a thorough
assessment of the program's scope of service by involving the personnel who
use the documentation process. Since the medical record is the central
document in all dialysis facilities, monitoring the quality of care through
documentation should begin with the medical record.
The most common standards used in monitoring of the medical record is
the standard of clinical pertinence. Clinical pertinence of the medical record
can be determined when a set of criteria are met which demonstrate that the
care rendered addressed the individualized medical needs of the patient. One
method of monitoring clinical pertinence of the dialysis medical record is to
establish a standard set of criteria and the expected rate of completion. A review
of the medical records of the dialysis population then can be assessed for trends
where those standards are not being met. Examination of the variance from the
expected rate through statistical analysis will determine if the variance is
significant [22]. If there is indication that the variance should be explored,
use of cause and effect or flow diagram tools should be implemented to work
toward a solution and to point out opportunities for improvement. The
following example demonstrates a standard set of criteria by which the clinical
record can be measured for quality in the dialysis setting.

CLINICAL PERTINENCE STANDARD

The clinical record documentation will reflect individualized appropriate medical
treatment.
 233

CLINICAL PERTINENCE CRITERIA

1. The History and Physical provides adequate information to assess the condition
of the patient and begin the process of diagnosis and treatment.
2. Diagnostic and Therapeutic orders are documented.
3. Clinical observations and results of treatment are documented.
4. Reports of procedures and tests are documented.
5. Diagnostic impressions are documented.
6. Progress notes adequately reflect the patient's condition and course of treat-
ment.
7. Plans are developed for follow-up care and discharge instructions pertinent to
the care of the patient.
8. Conclusion-final disposition of the patient's condition is documented.
9. Instructions and patient education are documented.
10. Signature of the physician is present.
11. Handwriting is legible.
12. Consultations and referrals are documented.
13. Documentation is appropriate for the medical situation.
14. Abnormal study results are addressed.

A practical approach to begin the process of monitoring the documenta-

tion of criteria which support a quality clinical record is to return to the JeARO
10 STEP model. This tool will act as a guide to formulate a quality improve-
ment model for the specific dialysis setting being monitored.
The most knowledgeable group of individuals to work on establishing a
quality improvement model for the dialysis unit are the team members who
must daily use the forms and meet the needs of the patient. Input from all
disciplines is imperative so as to fully address every issue pertinent to the
quality standards desired. The patient's perception of the quality of the services
they receive should also be included [23]. The most common method of
obtaining patient data is from patient satisfaction surveys [24]. These surveys
must be administered carefully in the chronic dialysis population in order to
protect anonymity and promote honest feedback. If done correctly, the
information derived from the survey will be invaluable to the quality moni-
toring program.
Once the team has been identified who will develop the quality moni-
toring program, the next step is to define the framework in which the program
will function. The following example will demonstrate how the 10 STEP model
can be applied to refining a clinical record monitoring program. The example
will also show how this framework can be applied to other clinical issues in
the dialysis facility which rely on dialysis records as a source of data.

Dialysis quality improvement model

1. Assign responsibility among the Multidisciplinary Team

2. Delineate scope of care - Identify the functions/services provided to what
type of patients in what type of setting.
234

3. Identify important aspects of care - Select the functions/services most

important to the quality and success of the program. One approach to
this selection process is to identify which services represent the high
volume, high risk, or problem prone areas [22]. These three factors when
applied to specific programs such as acute services, chronic services
and home services, will pinpoint the same similarities and differences
of the aspects of care for each service. Identification of the aspects of care
which are critical to the quality of the care received by the patient must
be accomplished before proceeding to the next step.
4. Identify monitoring indicators - Selection of an indicator of quality and
the criteria which support the indicator must be specific to the dialysis
setting. A clinical indicator as defined by JCAHO is a "quantitative
measure that can be used as a guide to monitor and evaluate the quality
of important patient care and support service activities [25-26]." The
indicator should be clearly defined and comprehensive.
A selected group of quality indicators representing structure, process
and patient outcome evaluation of services follows. Each indicator should
be supported by criteria derived from the clinical record.

INDICATORS
STRUCTURE INDICATOR
The medical record organization reflects the chronological factual occurrences
of the patient from admission to discharge.
PROCESS INDICATOR
Documentation in the medical record is meaningful, significant issues are iden-
tified and management of the patient is consistent with the findings.
OUTCOME INDICATOR
The quality of care provided to the patient is clearly documented.
DIALYSIS CLINICAL RECORD INDICATOR
The clinical record documentation will reflect individualized and appropriate
medical treatment.

CRITERIA FOR MONITORING

ACUTE CRITERIA
1. The medical record reflects adequate information to assess the condition
of the patient and begin the process of diagnosis and/or treatment.
2. The progress notes adequately reflect the patient's condition and hospital
course.
3. The tests documented in the medical record are consistent with the patient's
condition.
4. The medical record reflects that abnormal study results are addressed.
5. The medical record reflects planning for chronic ESRD treatment with
the patient and family.
6. The medical record reflects quality of care provided.
a. The discharge summary reflects an accurate reason for admission.
 235

b. The final diagnosis coincides with the findings and treatments docu-
mented.
c. Plans for follow-up care and discharge instruction are pertinent to
the care of the patient.
CHRONIC CRITERIA
1. The medical record reflects that appropriate ESRD treatment options
have been discussed with the patient and family and that they have
participated in the choice of treatment.
2. The medical record reflects assessment of the patient prior to initiation
of dialysis treatment in the following areas:
a. Access
b. Volume - target weight
c. Potential infection
d. Laboratory values
3. The medical record reflects ongoing assessment of the patient receiving
chronic dialysis therapy in the following areas:
a. Understanding of treatment modality
b. Understanding of present health problems
c. Coping mechanisms
d. Rehabilitation
4. The medical record reflects participation by the patient and his ability to
make decisions in the planning of his own ESRD care.
HOME DIALYSIS CRITERIA
1. The medical record reflects ongoing clinical assessment of the appropriate-
ness of the ESRD modality of choice.
2. The medical record reflects ongoing assessment of the patient's and/or
family's satisfaction with the treatment modality.
3. The medical record reflects appropriate referrals to community resources.

5. Establish statistical measurements for assuring quality of the medical

record [22] - Statistical expectations are usually derived from expected
norms as established in the literature or through historical practice.
Caution should be taken however, when using a predetermined norm as
a measurement of quality; an artificial limit of quality can be created.
If in fact, the current acceptable level of meeting a particular monitoring
criteria is less than 100% of the time, a study of the factors which
prevent that from occurring is indicated. Deming believes that the majority
of factors which impede quality are system problems and therefore can
be corrected either by correcting the system or by educating the staff [27].
In the case of providing medical care one must objectively assess all
data related to the measurement of quality services to patients and deter-
mine which of the variances from the standards set are a result of systems
failures and which are due to uncontrolled independent variations in
patient response to therapy.
6. Collect and organize data - Once the criteria for clinical record review
has been established, the quality monitoring team should decide how to
collect data from a representative sample of records in a systematic fashion
236

so as to obtain valid data. This is the most important and demanding

step as it requires careful planning and an understanding of statistical
approaches to problem solving. If this step is done correctly, the remaining
process will be simple to complete and rewarding to the individuals
participating in the quality improvement program. Depending on the
size of the population served, either all of the records or a percent of
randomly selected records should be reviewed [20]. A common practice
is to review 25 records or 5% of the total population which ever repre-
sents the most valid representation for statistical purposes.
The monitoring program should be designed to collect data on each
of the criteria periodically. The list of criteria can be prioritized to
determine the appropriate monitoring frequency.
7. Evaluate the data - When the preceding step has been carefully executed
the results of the data can be analyzed for meaningful trends. Two
approaches for review of the data should be performed at this point.
First, does the variance from the standard set a trend which needs
correction? Second, if a variance does not exist and the standard has
been set at less than 100%, should the standard be reviewed for the
potential to improve the quality of the criteria monitored? These critical
questions should be asked when reviewing the results of all criteria studies
and in particular of criteria dealing with system-wide issues.
8. Plan action to improve care - The quality monitoring team is the most
useful for planning the way in which to improve patient care utilizing
the results of the analyzed data. Brainstorming by a group which includes
members who represent all of the departments which are interdependent
in the achievement of improving system quality, can produce innova-
tive ideas to better the overall system.
9. Take action to improve care - A plan of action should be implemented
to bring about permanent improvement in the quality of the service
through both correction of the system and through education of the staff.
A permanent change in practice by staff cannot be guaranteed without the
proper know ledge on their part as to the importance of the change, how
it will benefit the work they do and how it will benefit the patient.
10. Evaluation-Re-evaluation - Once a change has been made, ongoing
evaluation of the criteria should determine if the intervention made any
difference and if it is permanent. Re-evaluation is necessary to deter-
mine if the highest level of quality has been achieved or if the process
can be improved [28].
In summary, the model proposed here is a continuous review of the criteria
which support the established standards a unit has determined as representing
the quality delivered to the patients served. Whether these standards repre-
sent clinical pertinence or represent patient care delivery as reflected in the
clinical record, monitoring of the practice of care is the only method for
objectively measuring the actual quality being delivered.
Once the documentation of the clinical record has been validated, further
 237

quality monitoring can be achieved with efficiency because the data will be
readily retrievable. Quality monitoring can take the route of examining the
structure of the organization to determine if the organization itself promotes
quality outcomes for patients. Another route of quality monitoring is to look
at the process of the organization to determine if the systems are in fact
supportive of quality care. The emphasis, however, should be focused toward
the outcome of the patient. Patient outcome results should demonstrate that the
interventions produced desired results.

Conclusion

The importance of the clinical records in the dialysis setting cannot be empha-
sized enough. As demonstrated through the various examples in this chapter
of how to establish a quality monitoring program the core element is a valid
clinical record. Successful implementation of the continuous improvement
model requires an understanding of the theory of quality management and how
these concepts integrate with medical practice and the influence of the
environment on that practice. Continuous quality improvement is a concurrent,
dynamic, action oriented process based on objective analysis of valid data. This
model simultaneously evaluates quality, assesses, prevents and reduces risk,
implements cost-effective resource management measures, and identifies and
combats potential problems. A well organized, streamlined, interactive clinical
record system is the key to a successful continuous quality improvement
program.

References

1. The Joint Commission Guide to Quality Assurance, 1988: The Joint Commission on
Accreditation of Healthcare Organizations, Chicago.
2. Conway-Welch C: Entering a new era of quality care. ANNA Journal 16(7): 469-471,
Dec 1989.
3. Deming WE: Out of crisis. Cambridge, MA: Massachusetts Institute of Technology, Center
for Advanced Engineering Studies, 1989.
4. Walton M: The Deming management method (New York: Putnum Publishing Group), 1986.
5. Vlchek DL, Day LM: A quality improvement management model for renal care. Nephrology
News & Issues Special Supplement.
6. Milton 0: Challenges of quality assurance program evaluation in a practice setting. J Nurs
Qual Assur 2(4): 25-34, 1988.
7. McCormick KA: Future data needs for quality of care monitoring, DRG consideration, reim-
bursement and outcome measurement, Image: Journal of Nursing Scholarship 23(1): 29-32,
Spring 1991.
8. Berwick OM: Continuous improvement as an ideal in health care. American Journal of
Medicine 320(1): 53-56, Jan 5, 1989.
9. Professional Liability Program, Farmers Insurance Group of Companies, Keeping the Record
Straight: Guidelines for Charting, QRC Advisor 6(3): 7-9, Jan 1990.
10. Pollak VE, Peterson OW, Flynn J: The computer in quality control of hemodialysis patient
care. QRB, 202-210, June 1986.
238

11. West E: Designing information systems to increase quality care. Computers in Healthcare,
Sept 1990.
12. Kaiser LR: Anticipating your high tech tomorrow. Healthcare Forum, 12-20, NovlDec 1986.
13. McDonald IC, Newton GA: The patient flow management model: A process for quality
assurance. The Health Rec Manage 10(3): 32-43, 1990.
14. Mogli GD: Role of medical records in quality assurance program. Am Ro 30(4): 11-15,
Nov 1989.
15. Chasteen IE: For the record. Dental Assisting, 23-26, Sept/Oct. 1987.
16. Donabedian A: The quality of care: How can it be assessed? JAMA 260(12): 1743-1748
Sept 23/30, 1988.
17. Joint Commission on Accreditation of Healthcare Organizations, 1988: Accreditation Manual
for Hospitals, Chicago.
18. Hexum 1M: Monitoring standards instead of problems. J Nurs Qual Assur 1(3): 8-13,
1987.
19. Laffel G, Blumenthal D: The case for using industrial quality management science in
health care organizations. JAMA 262Z(20): 2869-2873, Nov 24, 1989.
20. Bednar B, Neff M: Preparing for inspection: A tool to maximize quality and minimize
risk. ANNA Journal 17: 159-164, April 1990.
21. Rajki KL, Feltman BA, Smeltzer CH: Assessing the quality of nursing care in a dialysis
Unit. ANNA Journal 12(1), Feb 1985.
22. Schyve PM, Prevost IA: From quality assurance to quality improvement. Psychiatric Clinics
of North America 13(1): 61-71, March 1990.
23. Louden TL: Customer perception counts in quality assurance, Hospitals 84, Jan 20, 1989.
24. ANNA, Quality Assurance for Nephrology Nursing, First ed, 1989.
25. Characteristics of Clinical Indicators. QRB 330-339, Nov 1989.
26. Parker 1: Reduction in ESRD reimbursement rate: Identifying research priorities and quality
indicators. ANNA Journal 17(2), April 1990.
27. Getting Ready for Continuous Quality Improvement: Fundamentals for Department
Managers. American Hospital Association, 20 June 1991.
28. Van Valkenburgh D: Continuous quality improvement (CQI) of current quality - assur-
ance programs. Dialysis & Transplantation 20(9), Sept 1991.
CHAPTER 15

Continuous quality improvement and the

best demonstrated practices program

ARTHUR HOLDEN and ROSALIE VILLANO

Background

The commitment to quality improvement has been widely accepted in product

and service industries, with quality programs in place in many industries
worldwide. In dialysis, quality assurance (QA) programs, although relatively
late in developing, have become increasingly integrated into dialysis units as
their importance is recognized. The need for QA is accentuated by: (1) greater
awareness of quality problems, (2) increased patient care demands that
accompany an aging population, (3) shortages of qualified nurses, (4) increased
cost pressures; (5) heightened regulatory involvement and (6) more com-
petitive dialysis environments. The extensive focus on dialysis prescription
and delivery in both HD and PO has also accelerated the acceptance of
quantitative measurement and prescription modification.

Dialysis quality programs

Comprehensive dialysis quality assurance programs begin with collection

and evaluation of clinical, therapeutic and patient satisfaction measures. In this
chapter we will address a support program enhancing existing QA programs
for peritoneal dialysis (PO), the Best Demonstrated Practices program (BOP).
Through BOP, data collected from an individual dialysis unit is compared to
national and regional norms and to historical center performance. Through this
introspective process improvement, issues are identified and evaluated. Causes
of changes are explored, and solutions sought.

The Best Demonstrated Practices (BDP) program, a quality program

In 1986, when the focus on quality assurance was emerging in dialysis, the
Renal Division of Baxter Healthcare Corporation initiated a quality improve-
ment-support program for peritoneal dialysis, the Best Demonstrated Practices

L. W. Henderson and R.S. Thuma (eds.), Quality Assurance in Dialysis, 239-251.

© 1994 Kluwer Academic Publishers.
240

(BDP) program. The objective of BDP is to assist professionals in analyzing

their PD performance. The two aspects of the program include (1) measure-
ment of individual center performance and (2) determination of 'best' practices
from excellent centers.
The two components of BDP, analysis of patient data and practice improve-
ment are integrated into a process depicted in Figure 1.

1. Data collection; Patient

Retention Analysis
evaluation (or re-
evaluation

4. Hypothesis testing; 2. Data analysis:

Practice changes and identification of program
evaluation strengths & weaknesses

3. Program assessment &

evaluation;: Comparison to
'best' practices; Education

Fig. 1. Integration of BDP components into a continuous improvement process.

The program facilitates understanding of the factors related to patient

retention - which practices enhance success and which accentuate problems.
By analyzing treatment outcomes and modality transfer patterns, professionals
have quantitatively identified strengths, challenged weaknesses, and modified
practices to affect outcomes.
Today, BDP is used in 6 countries as a continuous quality assessment tool.
Over 400 dialysis centers have participated in the data collection and analysis,
with many more utilizing the practice guidelines that have evolved from this
program. Then and now, BDP has helped professionals evaluate and upgrade
treatment practices.

Outcome measurement

Patient retention and therapy transfer

Lifetable analysis is an important technique to quantify patient survival and

retention. It offers dynamic insight into trends, (Fig. 2) something a "snap-
shot", at one point in time cannot do (Fig. 3). However, both provide insight
when used in conjunction with each other. In BDP, lifetables show the prob-
ability over time that (1) a patient survives or (2) will remain on dialysis. They
are commonly used indicators of patient success with a technique.
Patient retention is measured and causes of patient and technique success
 241

100 .-;:~
90 ~_ _

.- ~~
.!-
c •
,! ~ .~~ ------ .......

.-. .,.
"'-,s
'0 • 60 •
#.= 50 - - - - ._ _ _ _
:.,s
40 ---.
.~
'!i "'- 30
EE
o:::I u0 20
10
O~--------------------------------~
o 6 12 18 24 30 36
InlelV,1 numb.,

Fig. 2. CAPD survival rate.

Transfer facility Tran:splanl

2.4% 17.0%
Transfer 10 Hemo
22.0%
Died
18.0%

Still on CAPD
40.3%

Fig. 3. Current patient status.

identified. Analysis is followed by exploring most likely causes of problems

and hypothesis testing. The process requires data collection, analysis, and
re-evaluation.

Morbidity measures

Morbidity measures provide important insights into dialysis quality. Infection,

a major complication in both PD and RD, can be reduced by understanding
242

and systematic evaluation. Reducing infection is a clear example of a QA goal.

For example, identifying the number and type of infections may serve as
a baseline measurement. Improvement begins with identifying the causes,
comparing practices with standards, challenging existing practices which
influence infection, testing hypotheses, re-measuring outcomes, and integrating
practice upgrades. Other complications can be reduced by attention to the same
process.

The BDP process

To participate in BOP, 3 years of patient data are collected for a Patient

Retention Analysis (PRA), the outcome report. Data collection and defini-
tions are standardized to facilitate evaluation, comparison and repeat
measurement. The PRA provides information about specific outcome measures.
From this report, improvement opportunities are reviewed, priorities estab-
lished, and objectives formulated. Team discussion leads to selection of the
most probable root causes of a problem, and leads to generation of potential
solutions. Where possible, outcomes are tied to practices. Potential solutions
are reviewed and a "best" solution adopted. An action plan is formulated
and solution implemented. Repeat measurement provides a basis for evaluating
the impact of changes.

Analytical feedback through the patient retention analysis

The PRA reports treatment outcomes based on retrospective patient data. It

profiles historical patient retention performance and reasons for leaving PO
over time. For each center, the report contains:
- PO lifetables: patient survival and technique survival (Fig. 1),
- Factors related to therapy change (Fig. 2),
- Patient population risk and demographic profiles (Table I); and peritonitis
rates.
Factors related to therapy change are also presented by 6 month intervals, to
facilitate shifts over time. These measures are used to evaluate and improve
PO technique success.

Table [. Patient demographic risk factors.

Year 1 Year 2 Year 3 Total

Average age 58.0 55.4 58.8 57.5

% diabetic 40% 42% 40% 41%
% cardiovascular disease 53% 45% 43% 49%

Caption: Shifts in the patient demographics may influence outcomes.

 243

The PRA forms the basis for problem evaluation, prevention and practice
improvement. Quantification of treatment outcomes facilitates comparison
within and to other dialysis units. The PRA thus provides a basis to assess
program strengths and weaknesses, to target areas for improvement, and assess
the impact of practice changes through re-analysis.

Practice improvement through professional recommendations

The second goal of BDP is integration of 'best' practices. Professionals can

evaluate and compare their practices to those of experts, to help correct
weaknesses. Through BDP, professionals share information in meetings and
print. Both formats help improve clinical practices.

Print Materials

Several educational materials have been created which document successful

techniques from programs with superior patient retention [1]. Treatment guide-
lines and best practices are created by (1) physician advisory committees
and (2) documented practices of excellent centers. These practices have been
published in journal articles, practice checklists, and monographs.
A collection of best practices from BDP "Centers of Excellence" was
originally compiled by Baxter and physician advisory committees in 1986-
1987. Thirty two highly successful centers with superior overall patient
retention and extraordinary results in one or more practice areas and physi-
cian advisory panels have served at the basis for practice information. Resource
materials, articles, educational materials were published on "best" practices:
- Peritonitis management and antibiotic therapy practices recommendations
on the diagnosis and treatment of peritonitis and exit site infection [2]
- Catheter and exit site practices an analysis of catheter implantation tech-
niques, break-in, exit site infection and ongoing catheter care [3]
- Patient education and training practices, a staff resource manual outlining
successful education and training practices [4]
- Modality selection practices criteria used to evaluate and select patients
for HD and PD [5]
Several recommendations have been updated through advisory committees:
Peritonitis [6, 7], Catheter [8, 9], and Modality Selection [10].

BDP Meetings

Meetings based upon the BDP process have been held in several countries.
Following presentations on clinical topics, meetings focus on workshop
sessions. Emphasis is placed on exploring practical management techniques
244

and potential solutions to key issues. Participants reexamine current issues and
practices, and provide the basis for individual quality assurance linkages.
Suggestions for future research are outlined, as well as practical ideas to
evaluate.
Specific practice improvements are generated for individual problems by
professionals who have managed the complication area well, based on demon-
strated low drop out from PD in specific practice areas in the PRA. Centers
with low dropout due to the complication use their own experience as a catalyst
for practice brainstorming with other professionals. Specific practices are
shared, in search of "best" practices. The professionals (1) clearly define PD
complications for their group (2) identify potential solutions/new ideas for
PD patient management and (3) evaluate and support implementation of
practice changes.
At a Canadian BDP meeting in December 1992, BDP program partici-
pants met to understand Canadian issues in patient and center management
practices. Meeting objectives included:
1. Share Canadian BDP Program results.
2. Share current PD issues and Practices.
3. Understand Canadian issues in PD patient and center management prac-
tices.
4. Share information from Canadian Centers of Excellence in each Practice
area.
S. Evaluate the Canadian BDP program effectiveness as a quality assur-
ance/assessment tool [11].
The attendees consisted of medical and nursing professionals from 12 Canadian
centers. The centers represent 7 provinces, 16% of the total Canadian dialysis
centers, 39% of all Canadian PD patients. The average PD usage is 49%.
Outcome/results are discussed later.
Professionals compared their individual center results to Canadian averages
and identified individual strengths and weaknesses in 4 clinical areas:
Peritonitis, Catheter issues, Adequacy, and Psychosocial/Non compliance.
Workshop sessions were held to discuss clinical practices and develop
recommendations for 'best' practices in these areas. Each group then presented
their recommendations to the entire audience. Through this process, program
participants established Canadian best demonstrated practices. Recommenda-
tions assist the QA teams in each center in improving practices and evalu-
ating quality on an on-going basis. A summary publication was developed
and distributed to all Canadian dialysis centers. Thus, Canadian practices
of excellent centers was disseminated to all Canadian centers as a reference
guide and forum for sharing "Best Demonstrated Practices".
At one U.S. meeting, peritonitis was explored as a reason for therapy
transfer. Of note, 29% of patients transferring to HD due to peritonitis did
so because of an unwillingness to continue CAPD, in the absence of medical
contra-indications. This finding highlighted the need to understand the
psycho-social aspect of therapy transfer. A similar finding in the Dec '92
 245

meeting highlights the need to better understand the psychosocial aspect of

therapy transfer. Practical management and prevention of peritonitis focused
on training, staff support, techniques used, and management of exit site
infection. This same process was used in evaluating the role of inadequate
dialysis, catheter complications, and psychosocial issues [12].

Measurement and change

Once identified, practice changes can be implemented, tracked, and evalu-

ated. Annual use of the PRA provides a measurable way to document and
evaluate changes. This commitment to a quality process requires ongoing
measurement as a catalyst for improvement. BDP provides an important
analytical approach for PD program improvement. It facilitates this process
of seeking, learning, implementing and measuring.
In evaluating the impact of therapy changes, quantitative measures are
important. As an example, setting a goal of reducing a complication becomes
a basis for action. In the case of catheter complications, an identified action
may include a change in catheter break-in, insertion policies, or treatment
regimen.

A success model

In an earlier report, ninety seven centers prior to participating in BDP

experienced a 2.4% per month or 29% per year dropout rate to hemodialysis.
With BDP, in six months, the rate dropped fifty percent or 1.2% per month
or 14.4% per year [1]. The BDP meeting also provides another successful
model for use of the PRA and practice improvement tools. These materials
have provided a basis for and problem solving and practice improvement
discussions in several regions of the world. The data provides a basis for
evaluating root causes; the professionals and potential solutions, which can
then be implemented.

Case study briefs

A large university based CAPD program identified their dropout issues through
data analysis in the PRA. Catheter problems were identified as a primary cause
of early dropout. The data provoked investigation of catheter practices; it
was determined that variation existed in catheter implantation procedures.
Following a meeting with surgeons, nephrologists and CAPD nurses to review
the latest in catheter insertion practices, a common protocol was developed
for catheter insertion which improved early catheter success.
A small, rural, free-standing CAPD program improved their training
246

materials by evaluating and incorporating information from the Patient

Education and Training Module. An increase in patient compliance was noted.
A large, established CAPD program with a small nursing staff did not
have the time to evaluate program statistics on an ongoing basis. BDP provided
the framework for analysis.
A large hospital-based center with many referring physicians did not have
a standard regimen for the treatment of peritonitis. Through BDP, specific
antibiotic dosing protocols were adopted, thus reducing dropout to peritonitis.
A mid sized PD program has used the PRA on an annual basis for the
past several years to identify shifts in the pd program. As a result of these
annual analyses, the pd team has evaluated practices affecting several outcome
measures, and have successfully modified and improved their outcome
measures. Many dialysis centers utilize the BDP data analysis program on
an annual basis to evaluate the impact of therapy changes.
An 11 center dialysis chain was concerned about their high modality transfer
rate. In evaluating the patient outcome in each center, it became apparent
that the patients leaving PD were lost to death. However, there was tremen-
dous variability within the individual dialysis units. This heightened the
awareness of (1) unique issues in several individual units and (2) issues that
impacted across the units. The physicians are probing specific information from
the analysis such as causes for death, management of higher risk patients,
and understanding when PD is used as a 'last choice' modality and its impact
on modality transfer. Guidelines are being created for policies across centers.

Patient retention database

The collection of data from individual dialysis units has permitted a macro-
analysis of a large number of dialysis facilities in several countries. The U.S.
database-analysis and a Canadian BDP meeting data are contained below:
U.S.: As of June 1991, 356 U.S. centers had submitted data for a patient
retention analysis of 21,497 patients over a 10 year period. These data
represent unique views into PD use due to its longevity and large number of
patients. Figure 5 compares the aggregate patient and technique survival for
the 21,997 patient. Several points are noteworthy. First, according to this
data, the CAPD patient has a 50% chance of technique success after 5.5 patient
years. Secondly, in this data base, 362 patients had been successfully managed
on the therapy for over 6 years.
Canada: As a part of a multi-center BDP meeting in December 1992, 13
centers submitted data for a PRA for 1,990 patients spanning a 10 year period.
Figure 1 shows the aggregate patient and technique survival. At 3 years, 80%
of patients were maintained on PD excluding death and transplant. The
Canadian patient has a 50% chance of technique success after 6.2 patient years,
and 40% were still on the therapy at ten years. Transfer to HDS occurred
most frequently in the first six months of PD treatment (30%). Transfer
 247

100

90 - - Patient Survival

80 -0-- Technique Survivial

70
iij
> 60

..
~
::J

C
50

..
8....
Q.
40

30

20

10

0
.. ., ..,
0 <D CII
-
CI)
CII
<D

Mon1hs on PO
o
<D
<D
'"

Fig. 4. Patient and technique success, lifetable, U.S. Database. The PD patient has a 50%
probability of remaining on PD at 5.5 years.

10%

21%

32%

23% o Catheter
• PsychologlcaVNon-compllance

Other Medical

Peritonitis

Inadequate Dialysis

Fig. 5. Reasons for modality transfer, U.S. database. Peritonitis is the prmary factor for transfer
to hemodialysis over the entire data base.
248

steadily declined over the scope of the analysis, with death and transplant
increasing:

Table II. Canadian patient status, 13 centres.

Patient status by time to dropout -

Percent of Reasons at Selected Intervals:

0-6 months 19-24 months 31-36 months Current status

Still on CAPD 31.4% 41.5% 44.3% 40.3%

Transferred to HD 30.4% 19.8% 12.3% 22.0%
Moved 4.7% 2.8% 1.0% 2.4%
Received transplant 15.8% 11.5% 13.4% 17%
Died 17.5% 24.1% 28.8% 18%

Peritonitis (30%) accounts for the major reason for therapy change over
the ten year period (Fig. 6). Other medical problems account for the second
largest reason for leaving (23%). In the first six months of use, other medical
factors accounts for the largest reasons for changing therapy (30%). This

18%

23%

o Catheter
• PsychologicaVNon-compliance

• Other Medical

D Peritonitis

Inadequate Dialysis

Fig. 6. Reasons for drop-out to hemodialysis - Canadian data. Peritonitis accounts for 30% of
the reasons for modality change over the entire period.
 249

data supplemented individual center data and served as a basis for therapy
practice analysis in the December BDP meeting.

Conclusion

Baxter developed the BDP program with the goal of increasing PD technique
success by improving the clinical and training practices associated with the
modality. Results support the merits of this type of clinical management process
in PD. The BDP database did not show significant outcome differences in
diabetes, cardiovascular disease, and age, as having significant influence on
PD technique success. Rather, more subjective elements, such as professional
commitment to PD, quality practices, ongoing commitment to improvement,
and a full line of dialysis therapeutic options appear to positively affect program
success.
The use of BDP has been varied. In those centers with strong focus on
quality, BDP has supported outcome measurement and change. There are
successful reports of its use in individual centers as a means of providing
annual analytical feedback. The large database analysis has also successfully
provided practitioners a basis for discussion of outcomes and practices in BDP
meetings.
BDP's success lies in its two part approach to problem discovery and
change. It has increased the confidence of physicians and nurses in their clinical
practices and programs. The process has provided access to successful prac-
tices and has enhanced communication among professionals. The challenge
is to understand the dynamics behind success and failure and apply the proper
means to correct problems. Professionals using the program report that BDP
supplies the process; however, it is the individual's commitment that is the
underlying success factor.
Embedded in the BDP program are several quality improvement concepts
and tools, including: Personal involvement, goal setting, education, commu-
nication, striving for excellence, team-work, specific quality improvement
projects, and an on-going improvement process. Results are never instanta-
neous; ongoing attention, review and commitment to the process are required.
BDP provides a framework for quality improvement, assisting professionals
with their quality assurance efforts.

Integrating dialysis quality

It is well recognized that renal replacement therapy is dynamic, with patients

often using more than one therapy over the course of treatment. While this
change is part of the reality of renal patient care, therapy change has both
human and financial costs. While transplant is considered a positive outcome
and death unavoidable, unnecessary changes in modality can create a burden
250

on patient and staff. Modality transfer/patient retention provide clear, quan-

tifiable and repeatable measures which impact patient well-being, dialysis
unit functioning and financial performance. Regardless of the length of time
a patient is treated or method used, it is necessary to understand the dynamics
of therapy transfer. Ultimately, this enhances patient care and reinforces
organizational excellence.

The impact of therapy transfer at a practical level

Increased patient retention yields psychological benefits to patients, who may

be able to continue home care instead of radically changing lifestyle to
accommodate in center care. Increasing the average length of time a patient
is treated at home also leverages the dialysis unit's orientation and training
costs associated with establishing a home patient.
Effectively integrating the QA process into the daily life in a dialysis unit
as ongoing process is a significant challenge. Simple methods and techniques
are required, such as simple, consistent methods of data collection, utilizing
standardized definitions. A quantative baseline is required. Trial and integra-
tion of key practices consistently across the dialysis team is required. For a
QA program to be successful, staff education and cooperation are essential.
Only then can practice enhancements be adopted. To be effective, this process
must be integrated into the very fabric of the dialysis unit as a continuous,
supported activity, involving all staff members. The process tends to empower
the people using it, and is in of itself, rewarding. The impact on dialysis quality
becomes evident, and is the truest reward.

References

1. Holden A, Graumer G: Best demonstrated practices program promoting CAPO patient reten-
tion in the United States. Advances in Continuous Ambulatory Peritoneal Dialysis, aI86-191,
1987.
2. Best Demonstrated Practices, Peritonitis Management and Antibiotic Therapy Practices
Module. Baxter Healthcare Corporation, 1987.
3. Best Demonstrated Practices, Catheter and Exit Site Practices. Baxter Healthcare
Corporation, 1987.
4. Best Demonstrated Practices, Patient Education and Training Module. Baxter Healthcare
Corporation. 1987.
5. Best Demonstrated Practices, Modality Selection Practices Module. Baxter Healthcare
Corporation. 1987.
6. Keane WF et al.: Continuous ambulatory peritoneal dialysis (CAPO) peritonitis treatment
recommendation: 1989 update. Peritoneal Dialysis International 9: 247-256, 1989.
7. Keane W et al.: Peritoneal dialysis-related peritonitis treatment recommendations. Peritoneal
Dialysis International 13: 1, 1993.
8. Oreopoulos DG et al.: Peritoneal catheters and exit site practices: Current recommendations.
Peritoneal Dialysis Bulletin 7: 130-137, 1987.
 251

9. Gokal R et al.: Peritoneal catheters and exit-site practices: towards- optimum peritoneal
access. Peritoneal Dialysis International 13: 1, 1993.
10. Hamburger RJ et al.: A dialysis modality decision guide based on the experience of six
dialysis centers. Dialysis & Transplantation 19: 65-69, 84, 1990.
11. Suart Chatterley: J. Best Demonstrated Practices - The Canadian Results, Dec. 1992.
12. Finkelstein F et al.: Initiatives in Peritoneal Dialysis: Where do we go from here? Peritoneal
Dialysis International 11: 274-278, 1991.
Subject index

AAMI Water Standards 89 in transplantation 169

adequacy JCAHO requirements 225
in peritoneal dialysis 38, 61 operations and controls 47
NCD Study 110 outcomes 48, 51
nutritional status as an index 114 root cause determination 192
aluminum toxicity 85 customer satisfaction
autonomy 30 in PD 41

Basic Personality Inventory 141 degradation products 199

best demonstrated practices 239 DEHP 198
Demographic Risk Factors 14
cadaver kidney quality 178 Dialysis Somatic Symptoms Distress Scale 141
cellulose membranes 201 Dialyzer-Reuse lSI
clearance 127 HCFA Study 152
Clinical Efficacy Assessment Project 159 NIH study 153
clinical indicators 234 Disease Stress Scale 140
clinical Intervention documentation systems 225
criteria for evaluation of 165
clinical pertinence standard 232 education assessments 33
clinical record entrepreneurship 30
clinical setting influences 228 EPA Regulations 87
clinical studies ESRD Regional Networks 171
peritonitis 162 exit site care 33
clinical trials
guidelines 156 facility records 230
hierarchy 157 Fatigue Rating Scale 141
multicenter 159 fractional clearance 128
Standards Established 159
co-morbid conditions 15 HCFA
CQI case review criteria screens 75
as a management model 227 HCFA requirement in transplant quality 171
as an objective 1 Health Care Finance Administration 171
best demonstrated practices 239 heavy metal toxicity 86
characteristics 48 hemodialysis
clinical inputs 50 AAMI water standards 89
elements 169 membrane selection 109
elements 191 quality of life 133
in Multicenter Systems 47 water treatment 85

253
254
hemodialyzer manufacturing 201
high flux membranes 118
home health services 78
home hemodialysis 63
human resource issues
selection of employees 215
human resources
key issues 215
immunosuppression 180
indicator 5
defining 226
development 51
home dialysis 73
types 2, 3, 4
ion exchangers 98
JCAHO
accreditation manual 65
indicators 234
model 231
job enrichment 28
job satisfaction 222
long-term graft survival 181
long-term medical care 182
management of equality
education and training 216
education requirements 26
measurement
reliability 142
subjectivity 143
validity 143
measurement approaches
disease-specific 139
global 139
medical record 225
contents 229
home care dialysis 71
mission statement
quality management 214
monitoring
water quality 104
monitoring criteria 234
multicenter communication 54
NCOS 53,110
nutritional Status 114
outcomes of dialysis 13
patient care 67
patient care plan 68
patient education 33
patient retention analysis 240, 246
patient rights 66
PO program requirements 31
PO Solution Quality
leachables 198
pyrogen testing 198
peritoneal clearance 122
peritoneal membrane 123
peritonitis
clinical studies 162
diagnosis 39
prevention 39
Physiological Index 140
population norms 144
productivity 26
professional enhancement 28
Psychosocial Stress Scale 140
pyrogens 198
quality
and patient involvement 24
clinical decisions ISS
improvement model 233
in the dialysis center 23
organizational issues 23
patient treatment related factors 19
philosophy of self-care 23
program integration 23
role of GMPs 50
technical factors 18
quality assurance
definition I, 11, ISS
elements 9, 64
human aspects 5
in Clinical Trials ISS
in home dialysis 64
in manufacturing 191
in transplantation 167
in water treatment 85
medical records 229
principles 128
quality of life measures 134
regulatory aspects 50, 65
Treatment Strategies lIS
quality assurance programs
home dialysis 73
management 77
quality commitment 26
quality data
HCFA2
quality environment 219
quality goals 219

quality management
communication issues 40
CQI as a model 227
education and training 217
human resource issues 213
implementation 6
in a dialysis setting 215
indicators 40, 226
job satisfaction 222
leadership goals 219
measurement 226
role of employee 214
role of management 214
role of standards 215
team approach 28
teamwork 26, 220, 233
quality Measures 194
quality model 52, 193
quality monitoring
JCAHO 10 step model 231
quality of care
definition 64
education of patient 69
home care records 71
impact of BDP 250
infection control 70
safety management 69
quality of dialysis
impact of selection 16
quality of life 109
definition 135
domains and components 136
patient survival 73
physical functioning 136
psychological functioning 137
social functioning 138
quality of life measurement
norms 44
subjectivity 143
quality of therapy
impact of reuse 151
membrane effects 128
quality programs
outcome measurement 240
policy and procedures manual 31
quality Systems
administrati ve support 25
255
Peritoneal Dialysis 23
regulatory aspects
in water quality 85
of Quality Assurance 50, 65
of transplantation 171
reuse 151
reverse osmosis 94
risk factors 14
Safe Drinking Water Act 85
selection
effect on survival rates 15
for home dialysis 68
Sleep Adequacy 140
Social Risk Factors 14
social worker's role 27
solution manufacturing 195
standards 155
clinical pertinence 232
clinical practice 31
standards of practice 217
sterilization 200
survival factors 81
synthetic membranes 201
teamwork 26
total clearances 38
transplant quality
management of living donors 177
transplant QA
organizing a program 183
patient education and psychological 182
postoperative medical care 179
preoperative screening 172
regulatory requirements 171
surgical care 178
treatment related factors 19
urea kinetic modeling 110, 161,38
water quality
microbiological considerations 87
water system design 101
water treatment quality 103
winding process 205
 Developments in Nephrology

1. J.S. Cheigh, KH. Stenzel and A.L. Rubin (eds.): Manual of Clinical Nephrology o/the
Rogosin Kidney Center. 1981 ISBN 90-247-2397-3
2. KD. Nolph (ed.): Peritoneal Dialysis. 1981 ed.: out of print
3rd revised and enlarged ed. 1988 (not in this series) ISBN 0-89838-406-0
3. AB. Gruskin and M.E. Norman (eds.): Pediatric Nephrology. 1981
ISBN 90-247-2514-3
4. O. Schuck: Examination o/the Kidney Function. 1981 ISBN 0-89838-565-2
5. J. Strauss (ed.): Hypertension, Fluid-electrolytes and Tubulopathies in Pediatric
Nephrology. 1982 ISBN 90-247-2633-6
6. J. Strauss (cd.): Neonatal Kidney and Fluid-electrolytes. 1983 ISBN 0-89838-575-X
7. J. Strauss (ed.): Acute Renal Disorders and Renal Emergencies. 1984
ISBN 0-89838-663-2
8. LJ.A. Didio and P.M. Motta (eds.): Basic, Clinical, and Surgical Nephrology. 1985
ISBN 0-89838-698-5
9. E.A. Friedman and C.M. Peterson (eds.): Diabetic Nephropathy. Strategy for Therapy.
1985 ISBN 0-89838-735-3
10. R. Dzurik, B. Lichardus and W. Guder: Kidney Metabolism and Function. 1985
ISBN 0-89838-749-3
11. J. Strauss (ed.): Homeostasis, Nephrotoxicy, and Renal Anomalies in the Newborn.
1986 ISBN 0-89838-766-3
12. D.G. Oreopoulos (ed.): Geriatric Nephrology. 1986 ISBN 0-89838-781-7
13. E.P. Paganini (ed.): Acute Continuous Renal Replacement Therapy. 1986
ISBN 0-89838-793-0
14. J.S. Cheigh, KH. Stenzel and AL. Rubin (eds.): Hypertension in Kidney Disease. 1986
ISBN 0-89838-797-3
15. N. Deane, R.J. Wineman and G.A. Benis (eds.): Guide to Reprocessing of
Hemodialyzers. 1986 ISBN 0-89838-798-1
16. C. Ponticelli, L. Minetti and G. D'Amico (eds.): Antiglobulins, Cryoglobulins and
Glomerulonephritis. 1986 ISBN 0-89838-810-4
17. J. Strauss (ed.) with the assistence of L. Strauss: Persistent Renalgenitourinary
Disorders. 1987 ISBN 0-89838-845-7
18. V.E. Andreucci and A Dal Canton (eds.): Diuretics. Basic, Pharmacological, and
Clinical Aspects. 1987 ISBN 0-89838-885-6
19. P.H. Bach and E.H. Lock (eds.): Nephrotoxicity in the Experimental and Clinical
Situation, Part 1. 1987 ISBN 0-89838-997-1
20. P.H. Bach and E.H. Lock (eds.): Nephrotoxicity in the Experimental and Clinical
Situation, Part 2. 1987 ISBN 0-89838-980-2
21. S.M. Gore and B.A Bradley (eds.): Renal Transplantation. Sense and Sensitization.
1988 ISBN 0-89838-370-6
22. L. Minetti, G. D' Amico and C. Ponticelli: The Kidney in Plasma Cell Dyscrasias. 1988
ISBN 0-89838-385-4
23. AS. Lindblad, J.W. Novak and KD. Nolph (eds.): Continuous Ambulatory Peritoneal
Dialysis in the USA. Final Report of the National CAPD Registry 1981-1988. 1989
ISBN 0-7923-0179-X
 Developments in Nephrology

24. V.E. Andreucci and A. Dal Canton (eds.): Current Therapy in Nephrology. 1989
ISBN 0-7923-0206-0
25. L. Kovacs and B. Lichardus: Vasopressin. Disturbed Secretion and its Effects. 1989
ISBN 0-7923-0249-4
26. M.E. de Broe and l.W. Coburn (eds.): Aluminum and Renal Failure. 1990
ISBN 0-7923-0347-4
27. K.D. Gardner Jr. and 1. Bernstein (eds.): The Cystic Kidney. 1990
ISBN 0-7923-0392-X
28. M.E. de Broe and G.A. Verpooten (eds.): Prevention in Nephrology. 1991
ISBN 0-7923-0951-0
29. T.A. Depner (ed.): Prescribing Hemodialysis. A Guide to Urea Modeling. 1991
ISBN 0-7923-0833-6
30. V.E. Andreucci and A. Dal Canton (eds.): New Therapeutic Strategies in Nephrology.
Proceedings of the 3rd International Meeting on Current Therapy in Nephropology
(Sorrento, Italy, 1990). 1991 ISBN 0-7923-1199-X
31. A. Amerio, P. Coratelli and S.G. Massry (eds.): Turbulo-Interstitial Nephropathies.
Proceedings of the 4th Bari Seminar in Nephrology (April 1990). 1991
ISBN 0-7923-1200-7
32. M.G. McGeown (ed.): Clinical Management of Renal Transplantation. 1992
ISBN 0-7923-1604-5
33. C.M. Kjellstrand and J.B. Dossetor (eds.): Ethical Problems in Dialysis and Transplan-
tation. 1992 ISBN 0-7923-1625-8
34. D.G. Oreopoulos, M.F. Michelis and S. Herschorn (eds.): Nephrology and Urology in
the Aged Patient. 1993 ISBN 0-7923-2019-0
35. E.A. Friedman (ed.): Death on Hemodialysis: Preventable or Inevitable? 1994
ISBN 0-7923-2652-0
36. L.W. Henderson and R.S. Thuma (eds.): Quality Assurance in Dialysis. 1994
ISBN 0-7923-2723-3

Kluwer Academic Publishers - Dordrecht / Boston / London