See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/8551344

Clinical Informatics in Critical Care

Article  in  Journal of Intensive Care Medicine · June 2004

DOI: 10.1177/0885066604264016 · Source: PubMed

CITATIONS READS

63 100

3 authors, including:

George Daniel Martich Michael Imhoff

University of Pittsburgh Ruhr-Universität Bochum
25 PUBLICATIONS   1,089 CITATIONS    134 PUBLICATIONS   1,538 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The creation of a new A-Fib detection and monitroing technology. View project

What MCG Technology is designed and engineered to deliver. View project

All content following this page was uploaded by Michael Imhoff on 03 November 2016.

The user has requested enhancement of the downloaded file.

Clinical Informatics in Critical Care
G. Daniel Martich, MD*
Carl S. Waldmann, MA, MB, BChir, FRCA, EDIC†
Michael Imhoff, MD, PhD‡
Health care information systems have the potential to
enable better care of patients in much the same manner as
the widespread use of the automobile and telephone did
in the early 20th century. The car and phone were rapidly
accepted and embraced throughout the world when these
breakthroughs occurred. However, the automation of
health care with use of computerized information systems
has not been as widely accepted and implemented as com-
puter technology use in all other sectors of the global
economy. In this article, the authors examine the need,
risks, and rewards of clinical informatics in health care as
well as its specific relationship to critical care medicine.
Key words: electronic medical record, electronic health record,
clinical information systems, clinical decision support, comput-
erized physician order entry, health care informatics, critical
care information systems, leapfrog, Institute of Medicine, med-
ical errors, clinical decision support, return on investment
The New York Stock Exchange developed its first
automated quotation system in 1953, electronic
ticker display boards in 1966, and by 1996, it had a
fully integrated technology network with wireless
data system. Stock markets throughout the world
have similarly relied on electronic transaction sys-
tems for years. Banks and mutual fund companies
do not balance their books at the end of the day
using paper and pen, but rather through electronic
spreadsheets with linked databases. The airline
industry, too often used as a benchmark for the
health care industry, could not fly the thousands of
flights with disparate equipment to all corners of
the globe without electronic tracking and reserva-
tion systems. Each of these industries can rightfully
From the *Department of Critical Care Medicine, University of
Pittsburgh School of Medicine; †Intensive Care Unit, Royal
Berkshire Hospital, Reading, UK; and the ‡Surgical Intensive
Care Unit, Klinikum Dortmund gGmbH, Dortmund, Germany.
Received Jul 14, 2003, and in revised form Jan 28, 2004.
Accepted for publication Jan 31, 2004.
Address correspondence to Dr G. Daniel Martich, Department of
Critical Care Medicine, 3550 Terrace Street, Room 602B,
Pittsburgh, PA 15261.
Martich GD, Waldmann CS, Imhoff M. Clinical informatics in
critical care. J Intensive Care Med. 2004;19:154-163.
DOI: 10.1177/0885066604264016
154
Downloaded from jic.sagepub.com by guest on November 3, 2016
claim that efficiencies, greater productivity,
increased revenues, and improved customer satis-
faction have been a direct result of automation.
Still, hospitals are at least as complex as banks, air-
lines, or steel plants. Even without an electronic
patient record, a 300-bed hospital in the United
Kingdom will generate 500,000 transactions every
day; about the same number of transactions as a
180-branch bank network in the same time. In the
United Kingdom, intensive care costs about £1.5
billion ($2.35 billion) annually. Can you name
another multimillion-pound business that runs
without computerization?
We in health care have fallen woefully behind
other industries in automation of many processes.
• Is there a need for clinical information systems
(CIS) in critical care?
• What should a CIS do for critical care?
• Why is there a reluctance to embrace information
technology in health care?
• Can clinical decision support improved care?
• What is the return on investment for CIS?
In this article, we will address the above ques-
tions and discuss the state of the art of health care
informatics.
Is There a Need for CIS in Critical Care?
Perhaps the most gripping reason for development
of CIS comes from the reports of the Institute of
Medicine (IOM). The IOM, part of the National
Academy of Sciences, is an adviser on scientific and
technological matters to the US government. In
June 1998, the IOM Quality of Health Care in
America Committee was formed and ultimately
concluded in their report “To Err Is Human” that as
many as 98,000 Americans die each year from pre-
ventable medical mistakes they experience during
hospitalizations [1]. The report goes on to indicate
that there are more deaths in hospitals each year
from preventable medical mistakes than there are
from motor vehicle accidents, breast cancer, or AIDS.
Copyright © 2004 Sage Publications

The single leading type of error is the medication
error. Estimates range from 4% to 20% of all hos-
pitalized patients encountering medication errors
[2,3]. These errors result in increased charges of
$2900 to $9000 per error, save litigation costs [4-6].
One of the oft-repeated recommendations from this
IOM report is that information systems could
reduce these mistakes considerably.
Information Overload
at the Point of Care
An intensive care unit (ICU) is a data-rich environ-
ment. When so many data elements need to be
turned into information and knowledge, errors
occur because of sheer volume. A physician may
be confronted with more than 200 variables during
typical morning rounds [7]. However, even an
experienced physician is often not able to develop
a systematic response to any problem involving
more than 7 variables [8]. Moreover, humans are
limited in their ability to estimate the degree of
relatedness between only 2 variables [9]. This prob-
lem is most pronounced in the evaluation of the
measurable effect of a therapeutic intervention.
Personal bias, experience, and a certain expecta-
tion toward the respective intervention may distort
an objective judgment [10].
While the first IOM report did not focus on care
in ICUs, the report does suggest that critically ill
patients are at a particularly high risk for adverse
events. That ICU risk, as high as 17.7% for death or
disability and nearly 46% for any type of adverse
event, was highlighted in “To Err Is Human” [11].
The abundance of information generated during
the process of critical care can now be captured
and stored using CIS, which provide for complete
medical documentation at the bedside. The clinical
usefulness and efficiency of CIS have been shown
repeatedly [12-15].
On the level of the hospital enterprise, informa-
tion overload is an issue too. In addition to the
administrative aspects of such a complex enter-
prise, new and much more complicated reimburse-
ment and accounting policies lead to an un-
matched complexity of data and information [16].
The volume of scientific literature is growing
exponentially. As of 1993, the number of systemat-
ic reviews in medicine increased 500-fold during
a 10-year period [17]. It is impossible for the indi-
vidual health care professional to keep track of all
relevant medical knowledge even in very narrow
subspecialities.
Journal of Intensive Care Medicine 19(3); 2004
Downloaded from jic.sagepub.com by guest on November 3, 2016
Clinical Informatics in Critical Care
Compliance With Accreditation Rules
The every-3-year assessment of hospitals in the
United States by the Joint Committee on Accredi-
tation of Healthcare Organizations (JCAHO) man-
dates compliance with certain standards. This year,
the JCAHO has developed 6 new core competen-
cies for all ICUs. These include the following [18]:
• ICU-1, ventilator-associated pneumonia (VAP; VAP
prevention: patient positioning): numerator = num-
ber of ventilator days in which the patient’s head of
bed is elevated equal to or greater than 30°;
denominator = total number of ventilator days.
• ICU-2, stress ulcer disease (SUD) prophylaxis:
numerator = number of ventilator days in which
patients received SUD prophylaxis; denominator =
total number of ventilator days.
• ICU-3, deep vein thrombosis (DVT) prophylaxis:
numerator = number of ventilator days in which
patients received DVT prophylaxis; denominator =
total number of ventilator days.
• ICU-4, central line associated primary blood stream
infection (BSI): numerator = number of central
line–associated BSI by type of ICU; denominator =
number of central line days by type of ICU.
• ICU-5, risk adjusted ICU length of stay by type of
ICU: continuous variable statement = mean length
of stay for ICU patients by type of ICU.
• ICU-6, risk-adjusted hospital mortality for ICU
patients: numerator = total number of adult patients
having had an ICU stay and whose hospital out-
come is death; denominator = total number of adult
patients having had a qualified ICU stay.
Each of these measures makes a strong case for
capturing data electronically at the point of care.
For instance, for the VAP core competency, the
measurement needed by the JCAHO requires
reporting the number of days that the head of the
bed is tilted at least 30° upward in a mechanically
ventilated patient (numerator) over the total num-
ber of ventilated days (denominator) for the
patient. This would need to be done for all venti-
lated patients and would be a daunting task with-
out electronic documentation by the bedside nurse,
data capture, and database analysis tools.
In Germany, in conjunction with disease-related
groups (DRGs) as the sole method for inpatient
reimbursement, health care authorities will require
hospitals to publish quality benchmarks and also
accumulate accreditation data after 2005. This can
realistically be captured only by means of an elec-
tronic patient record.
In the United Kingdom, a recent government ini-
tiative has set up the Modernisation Agency for
Critical Care to try to ensure standardization of pro-
tocols for transfers and the introduction of care
bundles for stress ulcer prophylaxis and DVT pro-
155
Martich et al
phylaxis with hopes to address other issues such as
the management of head-injured patients not
requiring neurosurgery in district general hospi-
tals. Now that more than two-thirds of UK ICUs
subscribe to the Intensive Care National Audit and
Research Centre, it has been possible for ICUs to
measure themselves on case-mix adjusted outcome
against other units. Each of these initiatives, includ-
ing data collection and standardization, lends itself
perfectly to the use of CIS for the ICUs.
Shortage of Intensivists
It is well known, at least among intensivists, that in
addition to national and global shortages of nurses,
there are too few intensivists to staff ICUs across
the United States and the world. While intensivist-
led care teams demonstrate improved mortality, it
would be impossible to physically provide each
ICU with enough intensivists on site to demonstrate
the better outcomes. Therefore, 1 possible solution
using information technology (IT) in the ICU is to
create an e-ICU. One commercial model for such a
virtual ICU care team is that established by VISICU
[19]. The VISICU model is dependent on 2 essential
factors: technology at the bedside capable of cap-
turing vital signs, medication administration, docu-
mentation, and laboratory information and hospi-
tal/physician practice change behavior to enable an
off-site intensivist to lead the care team.
What Should a CIS Do
for Critical Care?
The Leapfrog Group
In addition to the reasons mentioned already, there
are many other compelling initiatives that push for
the development of CIS. The first of these ideas
should be well known to intensivists around the
world despite being a proposal by American com-
panies to improve health care in the United States.
This initiative, called Leapfrog, was presented on
January 26, 2000, to the Senate committee of
Health, Education, Labor and Pensions by Dr Arnie
Milstein, the medical director of the Pacific Group
on Health. A group of large employers, such as
General Electric, International Business Machines,
General Motors, and Boeing, who have also
focused on the issue of medical errors from the
156
Downloaded from jic.sagepub.com by guest on November 3, 2016
payer’s perspective, has explicitly defined 3 issues
that will influence their choice of health care organ-
izations for their respective companies’ health
plans. These 3 major goals would result in leaps
forward for health care, with a 105% to 50% reduc-
tion in patient mortality. Attendant to these 3 aims
was the explicit cost savings associated with better
care to their employees both directly by decreasing
the cost of the health insurance and indirectly by
reducing lost days of work by the employee. The
medical directors of many Fortune 500 companies
determined these goals. The goals were chosen on
the basis of likely impact on health care quality,
attainability, and ease of explanation to large seg-
ments of the population. The initiatives can be
found in Table 1, with the translation to the sixth
grader next to the original goal.
Data Security and Confidentiality
Yet another requirement for pursuing CIS imple-
mentations in hospitals includes the Health
Insurance Portability and Accountability Act
(HIPAA) in the United States. This act demands that
hospitals know which individuals have accessed
patient-specific data. This daunting task cannot be
done with paper charts. Only with electronic time,
date, and person stamping will this degree of iden-
tification be in compliance with the law. Compara-
ble rules and legislation apply in many other
developed countries, where federal or state super-
vision of confidentiality of private data, including
specifically patient data, has a long tradition.
Current State
Documentation in Critical Care
The patient record remains the principal instrument
for ensuring continuity of care. The numerous
drugs and devices for diagnostics, monitoring, and
therapy combined with the steadily increasing acu-
ity of our patients prompt a demand for an
improvement of integration and acquisition of data
[20,21]. The current methods for documentation are
fraught with inefficiencies. The paper chart can be
used by only one individual at a time. The paper
chart can be in only one location at a time. The
paper chart is fragmented (Fig 1) and does not
allow for extraction of data across a defined popu-
Journal of Intensive Care Medicine 19(3); 2004
 Clinical Informatics in Critical Care

Table 1. Leapfrog Group Standards for Improving Health Care

Translation From Mother
Leapfrog Initiative Goal to Her Sixth Grader
1. Computer physician order entry (CPOE): physician order entry in hospitals Doctors have bad penmanship so
should be computerized. Adverse drug treatments are the leading cause of they should type directly what
avoidable death and disability in American hospitals. they really want for the patient
into a computer.

2. Evidence-based hospital referral: patients requiring selected complex Practice makes perfect.
treatments should routinely be referred to hospitals that offer the best odds
of survival, based on risk-adjusted comparisons or based on a hospital’s
annual volume of such treatment.

3. Intensive care unit physician staffing standard: physicians with credentials Really sick patients need doctors
in critical care medicine should at least, during the day, actively monitor who specialize in care of really
patients in intensive care units. sick patients.

These items represent in numerical order up to a 50%, 30%, and 10% reduction in mortality according to the literature. The common-
language “translation” of each of these initiatives is stated in the right-hand column.

lation of patients. The discordance between the

mass of data and the capacity of paper-based doc-
umentation leads to major defects in information
processing. As many as 30% to 50% of all interven-
tions and observations are not recorded, especially
in emergency situations [16,20,22].
Although many vendors offer CIS software sys-
tems for critical care, only a very small minority of
ICUs inside and outside the United States are using
paperless documentation. Despite its very small
market penetration, CIS for critical care can be con-
sidered a mature product. In many instances, sys-
tems have been on the market today for more than
20 years.
Some ICUs have replaced the paper record at the Fig 1. A partial medical record. Three of 7 total paper
bedside by a computerized patient record. Early portions of a patient’s current inpatient paper chart that
studies reported optimistic figures for achievable has been debrided in an intensive care unit because of
time savings with electronic ICU documentation. the size of the chart and inability to close the binder.
Two of the other parts of the current admission’s chart
Estimates of 30 to 90 minutes saved per shift per could be found on other medical/surgical units. Two
nurse with a CIS compared to manual documenta- parts of this patient’s chart were never found.
tion were reported [12,23,24]. In recent years, a
more realistic evaluation of the potential of time
saving from electronic documentation alone has at the point of care is the indispensable prerequi-
prevailed. More recent studies suggest that only site for any workflow automation and for any effec-
minor time savings are feasible but that adoption of tive implementation of computerized physician
an electronic flow sheet (Fig 2) significantly order entry (CPOE).
improves the quality of documentation and care
[20,25].
It must be assumed that documentation alone, Why Is There a Reluctance
while definitely feasible with CIS, is not an end in to Embrace IT in Health Care?
and of itself. Electronic documentation at the bed-
side becomes extremely productive when the data Most health care institutions in the Western world
are used for further analysis and processing (eg, have already implemented IT systems for business
charge capture, quality control) and for medical administration, billing, and accounting. Although
process control (eg, online decision support, elec- these systems provide some of the IT infrastructure
tronic protocols). Electronic data capture with a CIS needed to support clinical systems, most of these

Journal of Intensive Care Medicine 19(3); 2004 157

Downloaded from jic.sagepub.com by guest on November 3, 2016

Martich et al

Fig 2. Electronic data capture. A sample computerized spreadsheet with the patient’s noninvasive blood pressure high-
lighted with the detail of who, what, and when the data were entered on this test patient, automatically captured dur-
ing the documentation process. Similar data capture occurs each time a patient’s record is accessed electronically.

installations cannot handle the kind of medical
information that is necessary to manage the process
of care. Today there are only very few examples of
robust, integrated hospital-based clinical informa-
tion systems, notably most in the United States [26].
Only about 20 full CPOE systems are installed in US
hospitals [27].
Moreover, investments into IT systems to im-
prove the quality of care have to compete for fund-
ing with legally required changes, such as HIPAA in
the United States or the introduction of DRGs in
Germany. It is also noteworthy that about 70% of
all medical IT projects do not fully succeed [27].
(Lack of) Process Control
Point-of-care CIS allows the broadest control over
the process of care. Outstanding examples for the
management of process control are computerized
protocols, clinical pathways, and guidelines. Order
entry and decision support systems offer a unique
opportunity to dramatically improve the quality of
care while reducing overall costs [28,29]. In the ter-
minology of total quality management, an explicit
method (eg, a computerized protocol) is part of the
stabilization of the process necessary to improve
quality [30,31]. With the above background infor-
mation, the major focus in health care informatics
in the United States has been CPOE with or with-
out electronic decision support systems. While
endorsed by the Leapfrog Group and recommend-
ed by the IOM, many hospitals and vendors are
struggling to implement these extremely complex
systems. Physician order entry can be considered
the key to medical process control. It is probably
the most complex functionality in any CIS.
The major factor in the recent and remote failures
of CPOE has been the lack of process control in
hospitals [32-34]. Physicians have long been taught
to “trust no one but yourself” and are in many ways
independent practitioners of their trade despite
being members of private group practices or aca-
demic departments. Therefore, when CIS and
CPOE “force” the individual physician down a
pathway that is not to his or her liking, he or she
will rebel. That rebellion has stopped the imple-
mentation of CPOE efforts despite all of the hereto-

158 Journal of Intensive Care Medicine 19(3); 2004

Downloaded from jic.sagepub.com by guest on November 3, 2016


fore mentioned need at Cedars-Sinai Medical
Center in Los Angeles [35].
This example shows how important it is that
complex medical information systems fit into the
clinician’s workflow. The importance of under-
standing medical workflows on a very detailed
level before implementing medical information sys-
tems cannot be overstated [34,36].
Time Is Money
Before process control manifests itself as an issue,
however, physicians most often disdain new infor-
mation systems because it costs them extra time.
Unlike attorneys who bill quarter hours for even
thinking about a case, intensivists in the United
States can charge a 99291 code beginning at only
the 31-minute threshold and up to 75 minutes for
the time spent actually caring for the patient. The
extra time required logging in to CIS or waiting for
the processing of lab data, orders, and so forth
results in frustration from simply waiting for com-
puter screen flips. While realtors worldwide con-
stantly speak the mantra of “location, location,
location” when it comes to properties they wish to
sell, physicians who are selling CIS deployments
stress to our technology colleagues the need for
“performance, performance, performance.” In
other words, lack of speed kills when it comes to
physician adoption of CIS. Early analysis of the
extra time spent by medical residents at Regenstreif
Institute found that it took an extra 5.5 minutes per
patient per day to enter orders using CPOE versus
handwriting orders [37]. When that is multiplied by
an average ICU census of 12 critically ill patients,
more than an hour will be added to the intensivist’s
day in performing a function formerly done by unit
clerks. Several investigators have found and sup-
ported what most of us feel when working with
computers: speed is everything [38,39]. It has been
shown that fast response times are what medical
users value most in their IT systems. For CPOE and
clinical decision support (CDS) systems, typical
response times should be subsecond [36].
Don’t Surf My Electronic Records!
All of the advantages of sharing information via CIS
can be washed down the drain with one example
of inappropriate sharing or access to electronic
patient records. Not exclusive to physicians or
other clinicians, the concern over widespread dis-
Journal of Intensive Care Medicine 19(3); 2004
Downloaded from jic.sagepub.com by guest on November 3, 2016
Clinical Informatics in Critical Care
tribution of confidential medical information has
been mentioned numerous times by the lay press
with a multitude of high-profile cases. The physi-
cians who begin with skepticism over adoption of
CIS can rightly point to the lack of data security
stewards scrutinizing access to every piece of a
person’s record.
Expensive
Experience shows that the largest investments are
needed to provide the information infrastructure
such as electronic patient records and physician or-
der entry systems, while the cost of implementing and
maintaining online CDS in these systems is moder-
ate [39-41]. Depending on the kind of decision sup-
port and the level of standardization, even expen-
sive systems have the potential to generate signifi-
cant financial return on investment [16,42].
The implementation costs for CIS can be stagger-
ing, in the range of $5 to $20 million for a single
hospital. Therefore, it comes as no surprise that
CPOE is not widely implemented in US hospitals or
around the rest of the world. Hospital operating
margins in the United States average less than 3%;
indeed, many hospitals are hemorrhaging red ink.
In the United Kingdom, the National Health
Service has experienced embarrassment following
massive investment in computer systems. The most
recent fiasco concerned the information system
introduced for dealing with calls by the London
Ambulance Service. The system reportedly resulted
in ambulances being sent to wrong addresses and
delay in care resulting in several deaths [43]. It is no
wonder, then, that hospital administrators and
boards of directors are often fearful of implement-
ing an unfamiliar technology such as CIS. With no
clear-cut vendor-provided electronic health record
system leading the marketplace and the bursting of
the dotcom balloon, hospital’s capital-improvement
dollars are more often spent on bricks and mortar
than on picks and clicks on a computer screen.
Further Problems
Also, significant technological hurdles have still to
be overcome. Most decision support applications
and any kind of qualified data analysis require con-
sistent and standardized medical vocabularies,
which are still lacking in all developed countries
despite the formidable efforts of international and
national organizations [44].
159
Martich et al
But there may be even more daring challenges to
the implementation of IT to reduce medical errors
[3]:
• The successful implementation of medical informa-
tion technology often requires changes in the way
health care professionals think and act. These may
include, among other issues, the standardization of
care, the abandonment of personal style, and
development of trust in computer systems.
• The small number of fully evidence-based guide-
lines and pathways indicates the need for sub-
stantiation and validation of medical guidelines and
common practices in a timely and cost-efficient
manner.
Can CDS Improve Care?
Assuming that physicians and hospital organiza-
tions buy into the notion of process and change
control as well as acceptance of evidence-based
medicine in their daily practice, then the imple-
mentation of CIS, CPOE, and, perhaps most impor-
tant, CDS is possible. The goal of CDS is to supply
the best recommendation under all circumstances
[45]. In its 2001 report, the IOM strongly recom-
mended the use of sophisticated electronic CDS
systems for radically improving safety and quality
of care [12]. Clinical information systems that gen-
erate CDS-driven electronic reminders and CDS-
assisted CPOE have been shown to improve both
quality and cost-effectiveness of care [4,20].
Electronic reminders can significantly improve
physicians’ compliance with guidelines, reduce the
rate of human errors, and make physicians more
responsive to specific clinical events [6,46].
The most detailed and explicit algorithms in clin-
ical decision making use rule-based computer sys-
tems [27,47]. Using data from one of the most com-
prehensive clinical data repositories in the world
(LDS Hospital, Salt Lake City, Utah), the group of
Morris developed a rule-based CDS for the
mechanical ventilation of the critically ill [48]. Their
CDS generates, on the basis of actual patient data,
explicit, executable, and reproducible instructions
or recommendations for the next therapeutic step.
Their study found that it was possible to control
more than 95% of the ventilation times with these
protocols, while intermediate and final clinical out-
comes showed a beneficial effect [30].
Another tangible benefit of a CDS system is evi-
dent in the example of a large institution in which
the CDS checked all doctors’ orders for drug inter-
actions and approximately 400 of 15,000 orders were
changed daily. Most of these CDS-recommended
160
Downloaded from jic.sagepub.com by guest on November 3, 2016
changes were to avert potential adverse drug
events [49].
What Is the Return on
Investment for CIS?
The increasing pressure on health care expendi-
tures and cost-effectiveness of care has made
return on investment (ROI) analysis more impor-
tant for the health care professional [13,29,45]. In
other industries, the scope of ROI most often is
very clear. It is typically analyzed from only the
perspective of the investor, which in most cases is
also the company benefiting from this investment.
Normally, this perspective is the only applicable
one.
In health care, there may be many different per-
spectives to assess ROI. We need also to keep in
mind that investor and beneficiary are not always
the same. Therefore, we need to define which per-
spective we address before we engage in an ROI
analysis. Different perspectives in health care may
include
• the individual care provider (eg, the primary care
provider, surgeon, nurse, dietician),
• the health care organization (eg, a hospital, clinic,
nursing home),
• a health maintenance organization,
• an entire health care system (ie, the public purse,
Medicare),
• the employer of the patient, or
• the individual patient.
There are significant differences between health
care and other industries in how ROI can be gen-
erated [21,29]. Of course, from a monetary per-
spective, there are still only 2 basic approaches to
improve ROI: increase revenue or reduce cost. The
differences are how likely each of these approach-
es can be followed in health care compared to
other industries.
In health care, the emphasis has been on cost
reduction rather than on increasing revenue. The
highest cost in health care is also the most limited
resource: personnel. With that in mind, the greatest
potential for cost reduction is associated with
changing the process of care in a way that allows
a reduction in personnel [17,50]. As alluded to
above, CIS, CPOE, and CDS are possible (and
potentially profitable) only if associated with large-
scale acceptance of standardized process of care.
However, the more complex the change in the
processes of care, the more difficult it is to carry
out an appropriate ROI analysis. Many analyses are
Journal of Intensive Care Medicine 19(3); 2004

based on assumptions derived from studies in the
literature. Whether these assumptions are actually
justified is often questionable [51]. Differences in
health care systems and the practice of medicine
may render it impossible to extrapolate results from
one country to another. For example, CPOE sys-
tems and CDS systems have been shown in US
studies to reduce the rate of medical errors [4,20].
As medical errors are a problem in all developed
countries and their underlying causes are compara-
ble, it is probable that findings from these studies
also apply to other countries. Nevertheless, the
absolute ROI may be different because of different
cost structures, ancillary workflows, and litigation
practices.
ROI analysis is feasible even in the highly com-
plex setting of a health care institution. It is neces-
sary to justify the significant investments into med-
ical information technology that will be needed in
the upcoming years to cope with the problems of
cost reduction, medical errors, and quality of care [21].
But even on a limited scope, ROI analysis is a for-
midable task that carries a high level of uncertainty.
Finally, ROI and cost-related issues may become
moot as initiatives such as Leapfrog, the IOM
reports, and especially the medical malpractice cri-
sis in many US states mandate that CIS and CPOE
be implemented. The only ROI justification needed
may be that these information technologies are
essential to fulfill requirements and permit hospi-
tals to stay in business. Therefore, CIS and CPOE
could become just part of the cost of doing (health
care) business [52].
Discussion
Medical errors are without any doubt one issue of
major public interest and concern. The IOM report
“To Err Is Human” describes a situation that is in-
tolerable in the United States. Although the report
covers only the US health care situation, its findings
also give a broad idea of how the situation is in
other countries with advanced health care systems.
There are studies that make it very likely that the
problem of medical errors is as grave in Europe as
in the United States. Although the debate on med-
ical errors has not fully started in Europe yet, the
question is not whether it will become a high pri-
ority issue, but only when this will happen.
Information management, given its broadest
meaning to include communication among care-
givers, process control, and standardization to
achieve efficiencies, is at the core of many efforts
Journal of Intensive Care Medicine 19(3); 2004
Downloaded from jic.sagepub.com by guest on November 3, 2016
Clinical Informatics in Critical Care
to reduce medical errors. Therefore, medical infor-
mation management, with or without the help of
computers, will play a decisive role in any solution
posited for the growing concerns of quality in
health care. The application of technology to the
solution is a certainty given the complexity of the
problem with the caveats of cost, timing, and
acceptance.
Specific recommendations from many groups
focus on point-of-care CIS with CPOE and CDS sys-
tems. For the development of medical IT systems,
the implications are the following:
• The development of medical IT must be in point-
of-care information systems integrated into the
health care enterprise-wide information manage-
ment approach.
• The integration of information and knowledge
must be seamless along the continuum of care and
across the entire health care enterprise.
• Clinical information systems, electronic order entry
systems, CDS systems, medical knowledge bases,
and ultimately the electronic patient record provide
the most powerful solutions to the problem of
medical errors. These systems may even become in
part legally compulsory or may be enforced by
major payer groups.
It is also clear that IT must reach a critical mass
to be effective. For instance, payback from a physi-
cian order entry system can be expected only if it
covers most of the continuum of care. This means
that massive up-front investments are needed.
Moreover, most advanced online CDS systems
require comprehensive patient information in elec-
tronic format, which can be provided only by clin-
ical information systems.
There is no doubt that medical information man-
agement is one answer to the problem of medical
errors. Standardization of care, process control,
physician order entry systems with online CDS, and
the electronic patient record have the potential to
make health care safer, improve quality of care,
and reduce cost in the interest of all parties
involved.
We do not know how long it will take nor how
expensive it will be. But we know that action, in-
vestments, and commitment are needed now. And
we know that the way we practice medicine will
never be the same again.
References
1. Karlsson G, Johannesson M. Cost-effectiveness analysis
and capital costs. Soc Sci Med. 1998;46:1183-1191.
2. Leape LL. Error in medicine. JAMA. 1994;272:1851-1857.
161
Martich et al
3. Imhoff M. Health informatics. In: Sibbald W, Bion J, eds.
Evaluating Critical Care: Using Health Services Research to
Improve Quality. Update in Intensive Care and Emergency
Medicine 35. Heidelberg, Germany: Springer; 2000:255-
269.
4. Bates DW, Spell N, Cullen DJ, et al. The costs of adverse
drug events in hospitalized patients. Adverse Drug Events
Prevention Study Group. JAMA. 1997;277:307-311.
5. Classen DC, Pestotnik SL, Evans RS, et al. Adverse drug
events in hospitalized patients. Excess length of stay, extra
costs, and attributable mortality. JAMA. 1997;277:301-306.
6. Raschke RA, Gollihare B, Wunderlich TA, et al. A comput-
er alert system to prevent injury from adverse drug
events—development and evaluation in a community
teaching hospital. JAMA. 1998;280:1317-1320.
7. Morris A. Algorithm-based decision-making. In: Tobin JA,
ed. Principles and Practice of Intensive Care Monitoring.
New York, NY: McGraw-Hill; 1998:1355-1381.
8. Miller G. The magical number seven, plus or minus two:
some limits to our capacity for processing information.
Psychol Rev. 1956;63:81-97.
9. Jennings D, Amabile T, Ross L. Informal covariation assess-
ments: data-based versus theory-based judgements. In:
Kahnemann D, ed. Judgement Under Uncertainty:
Heuristics and Biases. Cambridge, UK: Cambridge
University Press; 1982:211-230.
10. Guyatt G, Drummond M, Feeny D, et al. Guidelines for the
clinical and economic evaluation of health care technolo-
gies. Soc Sci Med. 1986;22:393-408.
11. Andrews LB, Stocking C, Krizek T, et al. An alternative
strategy for studying adverse events in medical care.
Lancet. 1997;349:309-319.
12. Bosman RJ, Rood E, Oudemans-Van Straaten HM, et al.
Intensive care information system reduces documentation
time of the nurses after cardiothoracic surgery. Intensive
Care Med. 2003;29:83-90.
13. Imhoff M. A clinical information system on the intensive
care unit: dream or nightmare? In: Rubi JAG, ed. Medicina
Intensiva 1995, XXX. Murcia, Spain: Congreso SEMIUC,
Murcia Pictographia; 1995:17-22.
14. Imhoff M. Three years clinical use of the Siemens EMTEK
System 2000: efforts and benefits. Clin Intensive Care.
1996;7(suppl):43-47.
15. Imhoff M, Lehner JH, Lohlein D. Two year clinical experi-
ence with a clinical information system on a surgical ICU.
In: Monduzzi, ed. 7th European Congress on Intensive Care
Medicine. Bologna, Italy: European Congress; 1994:163-
166.
16. Haux R. Aims and tasks of medical informatics. Int J Med
Inf. 1997;44:9-20.
17. Chalmers TC, Laus J. Meta-analytic stimulus for changes in
clinical trial. Stat Methods Med Res. 1993;2:161-172.
18. JCAHO. Six core ICU measures. Available at:
http://www.jcaho.org/pms/core+measures/candidate+core
+measure+sets.htm.
19. Celi LA, Hassan E, Marquardt C, et al. The eICU: it’s not just
telemedicine. Crit Care Med. 2001;29:N183-N189.
20. Balas EA, Weingarten S, Garb CT, et al. Improving preven-
tive care by prompting physicians. Arch Intern Med.
2000;160:301-308.
21. Imhoff M, Gather U, Morik K. Development of decision
support algorithms for intensive care medicine: a new
approach combining time series analysis and a knowledge
base system with learning and revision capabilities. In:
Brugard W, ed. KI-99: Advances in Artificial Intelligence.
Bonn, Germany: Springer; 1999:219-230.
22. Gardner RM. Computer technology and its application to
cardiovascular nursing. J Cardiovasc Nurs. 1987;1:69-71.
23. Leyerle BJ, LoBue M, Shabot M. Integrated computerized
databases for medical data management beyond the bed-
side. Int J Clin Monit Comput. 1990;7:83-89.
162
Downloaded from jic.sagepub.com by guest on November 3, 2016
24. Shabot MM, LoBue M, Leyerle BJ, et al. Decision support
alerts for clinical laboratory and blood gas data. J Clin
Monit Comput. 1990;7:27-31.
25. Fraenkel DJ, Cowie M, Daley P. Quality benefits of an
intensive care clinical information system. Crit Care Clin.
2003;31:120-125.
26. Committee on Quality of Health Care in America, Institute
of Medicine. Crossing the Quality Chasm: A New Health
System for the 21st Century. Washington, DC: National
Academy Press; 2001.
27. Blum MS, Cole C, Leviss J, Shabot MM, Weisenthal A.
Advanced EMR implementations: practical experience from
those who have been there. Presented at: AMAIA
Symposium 2003; November 8-12, 2003; Washington, DC.
28. Classen D, Evans R, Pestotnik S, et al. The timing of pro-
phylactic administration of antibiotics and the risk of
surgical-wound infection. N Engl J Med. 1992;326:281-286.
29. Pestotnik SL, Classen DC, Evans RS, et al. Implementing
antibiotic practice guidelines through computer-assisted
decision support: clinical and financial outcomes. Ann
Intern Med. 1996;124:884-890.
30. Morris AH. Computerized protocols and beside decision
support. Crit Care Clin. 1999;15:523-545.
31. Walton M. The Deming Management Method. New York,
NY: Putnam; 1986.
32. Massaro TA. Introducing physician order entry at a major
academic medical-center: 1. Impact on organizational cul-
ture and behavior. Acad Med. 1993;68:20-25.
33. Massaro TA. Introducing physician order entry at a major
academic medical-center: 2. Impact on medical-education.
Acad Med. 1993;68:25-30.
34. Maviglia SM, Zielstorff RD, Paterno M, Teich JM, Bates DW,
Kuperman GJ. Automating complex guidelines for chronic
disease: lessons learned. J Am Med Inform Assoc.
2003;10(2):154-165.
35. Ornstein C. Hospital heeds doctors: suspends use of soft-
ware. Los Angeles Times. January 22, 2003:B1.
36. Bates DW, Kuperman GJ, Wang S, et al. Ten command-
ments for effective clinical decision support: making the
practice of evidence-based medicine a reality. J Am Med
Inform Assoc. 2003;10(6):523-530.
37. Tierney WM, Miller ME, Overhage JM, McDonald CJ.
Physician inpatient order writing on microcomputer work-
stations—effects on resource utilization. JAMA.
1993;269:379-383.
38. Lee F, Teich JM, Spurr CD, Bates DW. Implementation of
physician order entry: user satisfaction and self-reported
usage patterns. J Am Med Inform Assoc. 1996;3(1):42-55.
39. Shojania KG, Yokoe D, Platt R, et al. Reducing vancomycin
use utilizing a computer guideline: results of a randomized
controlled trial. J Am Med Inform Assoc. 1998;5:554-562.
40. Overhage JM, Tierney WM, Zhou XH, et al. A randomized
trial of “corollary orders” to prevent errors of omission. J
Am Med Inform Assoc. 1997;4:364-375.
41. Teich JM, Glaser JP, Beckley RF, et al. Toward cost-effec-
tive, quality care: the Brigham Integrated Computing
System. In: Proceedings of the 2nd Nicholas E. Davies CPR
Recognition Symposium. Washington, DC: Computerized
Patient Record Institute; 1996:3-34.
42. Teich JM, Glaser JP, Beckley RF, et al. The Brigham inte-
grated computing system (BICS): advanced clinical systems
in an academic hospital environment. Int J Med Inform.
1999;54:197-208.
43. Cocks RA, Glucksman E. What does London need from its
Ambulance Service? BMJ. 1993;306:1428-1429.
44. Work Group on Computerization of Patient Records.
Toward a National Health Information Infrastructure:
Report of the Work Group on Computerization of Patient
Records. Washington, DC: US Department of Health and
Human Services; 2000.
45. McDonald CJ, Wilson FJ Jr. Physician response to comput-
er reminders. JAMA. 1980;244:1579-1581.
Journal of Intensive Care Medicine 19(3); 2004

46. McDonald CJ. Protocol-based computer reminders, the
quality of care and the non-perfectability of man. N Engl J
Med. 1976;295(24):1351-1355.
47. Organisation for Economic Co-operation and Develop-
ment. A Comparative Analysis of 29 Countries. OECD
Health Data. Paris, France: Organisation for Economic Co-
operation and Development; 1999.
48. Thomsen GE, Pope D, East T, et al. Clinical performance
of a rule-based decision support system for mechanical
ventilation of ARDS patients. Proc Annu Symp Comput
Appl Med Care. 1993:339-343.
Journal of Intensive Care Medicine 19(3); 2004
Downloaded from jic.sagepub.com by guest on November 3, 2016
View publication stats
Clinical Informatics in Critical Care
49. Sharpe V, Faden A. Medical Harm. Cambridge, UK:
Cambridge University Press; 1998.
50. Leape LL, Bates DW, Cullen DJ, et al. Systems analysis of
adverse drug events. JAMA. 1995;274:35-43.
51. Eckhart JG. Costing out nursing services: Examining the
research. Nurs Econ. 1993;11:91-98.
52. Bauer JC. Return on investment: going beyond traditional
analysis. J Health Inf Manag. 2003;17(4):4-5.
163