Task Overlap Among Primary Care

Team Members: An Opportunity

for System Redesign?
Richard G. Best, Ph.D., health services investigator, Veterans Evidence-Based Research
Dissemination and Implementation Center, San Antonio, Texas; Sylvia J. Hysong,
Ph.D., health services investigator, Houston Center for Quality of Care and Utilization
Studies, and instructor of medicine, Baylor College of Medicine, Houston, Texas;
Jacqueline A. Pugh, M.D., professor of internal medicine, University of Texas Health [First Page]
Science Center, and staff physician, South Texas Veterans Health Care System, [295], (1)
San Antonio; Suvro Ghosh, database analyst, Veterans Evidence-Based Research
Dissemination and Implementation Center; and Frank I. Moore, Ph.D., professor,
University of Texas School of Public Health, San Antonio Regional Campus Lines: 0 to 35
———
* 31.06001pt PgVar
———
................................................................................................ Normal Page
E X E C U T I V E S U M M A R Y * PgEnds: Eject
This article presents the results of research1 on a systematic approach to the as-
signment of primary care work in the Veterans Health Administration. Based on [295], (1)
a functional job analysis protocol, the study identified overlap in the performance
of primary care tasks among multiple occupational groups as prima facie evidence
of opportunities to reallocate work responsibilities. Results show that registered
nurses, physicians, advanced practitioners, and licensed vocational nurses reported
performing 60 percent to 97 percent of the same tasks, while clerks and health
technicians appeared to be underutilized. The frequency and duration with which
occupational groups performed each task were also examined, providing additional
evidence to be used in improving clinic efficiency.
The management of healthcare personnel can be improved through systematic
analysis of the work, the worker, and the work organization and through more in-
formed decisions about the appropriateness of task assignment (or reassignment).
This article presents an evidence-based approach to personnel management with
important implications for clinic efficiency. The approach can be used to guide
strategic planning and staffing decisions by identifying not only who currently does
the work but, more importantly, who should be doing the work given the full array
of data.

For more information on the concepts in this article, please contact Dr. Best
at rbest@satx.rr.com.

295
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 295
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
 J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
T he fundamental change deemed
necessary to bridge the chasm
between current healthcare and
the ideal is thought to be systemic
in nature (Institute of Medicine
2001). According to the Institute for
Healthcare Improvement (2002), “We
firmly believe that performance at
this level cannot be achieved by the
simple accumulation of many small
process changes without the guiding
architecture of a new care system as a
whole. In other words, these dramatic
improvements are entirely out of reach
for healthcare systems under current
operating designs, without radical and
sustained system-level change.” The
Veterans Health Administration (VHA)
has embarked on systemwide change
to transform its mode of care delivery
from acute care with an inpatient focus
to one featuring primary care and
outpatient care as the principal points
of access. The VHA has implemented
a sophisticated computerized patient-
record system with clinical reminders
for guideline-concordant actions and
pharmacy alerts for unsafe prescribing
practices (Pugh 2001). The VHA also
regularly monitors the quality of
outpatient care performance. However,
what has not yet been addressed by the
VHA is the allocation of work within
the practice team, or delivery system
redesign (Von Korff et al. 1997). The
objective of this article is to examine
current work assignments among VHA
primary care personnel to identify
delivery system redesign opportunities.
RELEVANCE OF DELIVERY
SYSTEM REDESIGN
Delivery system redesign is empha-
sized by VHA managers at multiple
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
levels because staffing mix and the
assignment of task responsibility
are high priorities for VHA primary
care. Indeed, research in the VHA has
indicated considerable variation in
the assignment of task responsibility
for clinical practice guidelines (Pugh
2002). Reallocation of work assign-
ments is also important to initiatives
to improve timely access to care. One
key principle of improving access is
managing constraints or bottlenecks [296], (2)
that restrict patient flow through the
primary care clinic. Constraints are
Lines: 35 to 51
“rate-limiting steps” that determine
———
clinic throughput; they often include
0.0pt PgVar
primary care providers who perform ———
work that uniquely adds value to the Long Page
system (Murray and Berwick 2003). PgEnds: TEX
Accordingly, work assigned to primary
care providers should maximize their
[296], (2)
skills and expertise.
Although the importance of task
reassignment for optimizing clinic
efficiency is clear, little guidance exists
on how to systematically identify
work tasks that could be reassigned.
Research on a primary care system
in Colombia found that registered
nurses (RNs) and licensed vocational
nurses (LVNs) shared 95 percent of
their work tasks (Moore 2000). More
recently, a systematic review found
that nurse practitioners performed
work comparable to that of physicians,
with equivalent or superior short-
term outcomes (Horrocks, Anderson,
and Salisbury 2002). These studies
point to an overlap in the performance
of primary care work as prima facie
evidence that multiple occupational
groups can do and are doing the same
work. The identification of overlap
in work performance suggests that
296
the permission
of 296
/ Vol. 51, No. 5/ Page Health
1st proof
 Tas k Ov er l ap Among Pr imar y C ar e Team M ember s
opportunities exist to reassign task
responsibilities. Therefore, systematic
examination of primary care work
for the purpose of identifying overlap
is the foundation of delivery system
redesign.
WORK-DOING SYSTEMS
THEORY AND FUNCTIONAL
J O B A N A LY S I S
We used Fine’s work-doing systems
theory to guide our examination of
current patterns in primary care work
in the VHA (Fine and Cronshaw 1999).
This framework posits a dynamic
interaction of three components
of organizational systems: (1) the
work organization (its purpose,
goals, and objectives); (2) the worker
(capacities, experiences, education, and
training); and (3) the work content
(the functions, subfunctions, activities,
tasks, and associated performance
standards). The advantages of this
conceptualization are that it contributes
to the organization and marshalling
of resources (technologies, materials,
workers) for optimum efficiency and
that it serves the master purpose of
the organization (Fine and Cronshaw
1999). The work-doing systems theory
provided a comprehensive architecture
for the examination of current work
patterns within the VHA.
Functional Job Analysis
Functional job analysis (FJA) is a job
analytic methodology based on the
work-doing systems theory. FJA has
a venerable history as the original
foundation on which the Dictionary of
Occupational Titles (a comprehensive
classification system developed for
the Office of Personnel Management)
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
was based. This is important for two
reasons: (1) job analytic methods are
required for legally validating person-
nel decisions, and (2) they provide
an empirical foundation for evidence-
based management (Pfeffer and Sutton
2006; Walshe and Rundall 2001).
FJA analyzes task statements of
work as the basic building blocks of
human resource management and
organizational strategic planning. Task
statements explicitly implicate the three [297], (3)
components of work-doing systems—
work content, worker characteristics,
Lines: 51 to 80
and the work organization—using the
———
following elements:
0.0pt PgVar
———
• who performs (the subject matter Long Page
expert) PgEnds: TEX
• what action
• with what tools, materials, or work [297], (3)
aids
• upon what instructions (including
the requisite knowledge, skills,
abilities, and performance standards
for task performance)
• to accomplish what organizational
outcome or result?
Tasks are also rated according to
functional skills requirements that
define the complexity of performance
across cognitive, interpersonal, and
physical dimensions (see Fine and
Cronshaw 1999 for the full description
of the scales). Given the current
climate for patient safety, we also
used a dimension that evaluates the
potential consequence given an error
in performance (Institute of Medicine
2000; Gilpatrick 1972). These ratings
help describe the relative simplicity or
complexity of the work content (Fine
297
the permission
of 297
/ Vol. 51, No. 5/ Page Health
1st proof
 J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
and Getkate 1995). Thus, functional
complexity ratings provide additional
guidance for decisions about task
assignment. Tasks may be assigned to
maximize the unique skills and exper-
tise of workers (promoting employee
growth and satisfaction) as well as
to ensure that competent personnel
perform the work (enhancing quality
of care and patient safety). Indeed,
the rich array of information at the
task level highlights the utility and
flexibility that result from aligning the
work with the worker.
Our research used FJA to system-
atically investigate VHA primary care
work. We hypothesized that consid-
erable overlap in task performance
occurred among (1) physicians and
advanced practitioners (i.e., nurse
practitioners and physician assistants),
(2) physicians and RNs, (3) RNs and
LVNs, and (4) nurses and clerks or
health technicians.
Although we focused on overlap
in task performance as prima facie
evidence that multiple occupational
groups perform many of the same
tasks, the identification of overlap
by itself is insufficient for actual task
reassignment and delivery system
redesign. Assignment of primary care
task responsibility could be informed
by additional factors such as (1) the
frequency and duration by which
different occupational groups perform
the same task; (2) the functional skills
required to perform the task, here
represented as the level of complexity
rating on data, people, reasoning, and
human error consequence of the tasks
being considered for reassignment;
(3) the scope of practice based on
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
licensure and certifications; and
(4) local contextual issues regarding
number of staff and/or process issues.
The frequency and duration by
which different occupational groups
perform the same tasks circumscribe
the extent to which reassignment is
both feasible and beneficial. For exam-
ple, efficiency is less likely when task
responsibility is reassigned to an occu-
pational group that performs that work
less frequently and for longer duration. [298], (4)
Although overlap in task performance
indicates who currently is doing the Lines: 80 to 97
work, functional and content skill
———
ratings offer guidance as to who should 0.0pt PgVar
perform tasks given the complexity of ———
task performance. If, for example, an Normal Page
error in task performance is consid- PgEnds: TEX
ered consequential, allocation to an
occupational group with more train- [298], (4)
ing and experience is preferred in the
interest of patient safety. Whereas the
primary objective of this article is to
demonstrate the use of a job analytic
database to identify overlap in task
performance among multiple primary
care occupational groups, we will also
illustrate the frequency and duration
of task performance by occupational
groups and the level of task complexity
in simulated reallocation decisions.
METHODS
Study Design
The FJA protocol is well documented
and uses a mixture of qualitative and
quantitative methods to develop task
databases (see Fine and Cronshaw
1999; Moore 1999; Fine and Getkate
1995). Thus, our research design used
298
the permission
of 298
/ Vol. 51, No. 5/ Page Health
1st proof
 Tas k Ov er l ap Among Pr imar y C ar e Team M ember s
qualitative focus groups and quantita-
tive survey data collection to develop
and verify a database of task statements
describing VHA primary care work.
Site Selection
We convened an expert advisory panel
(EAP), consisting of VHA managers at
multiple levels, to guide our research
design, including nominating and re-
cruiting participating facilities. Our EAP
identified organizational characteristics
believed to influence assignment of
task responsibility (i.e, academic affil-
iation, size, and the degree of primary
care service line implementation) and
nominated facilities for inclusion based
on these characteristics as well as the
EAP members’ experience with the
facilities and the likelihood of facility
interest in participation. We engaged
six facilities that met these inclusion
criteria.
Subjects
We invited between six and eight sub-
ject matter experts (SMEs) or job in-
cumbents to participate in separate
focus groups of each of six primary
care jobs (physician [MD], nurse practi-
tioner [NP] or physician assistant [PA],
RN, LVN, health technician, and clerk).
The SMEs were incumbents with the
requisite experience of the job being
analyzed. We invited SMEs with varying
lengths of tenure both in the job and
in the VHA. We invited all primary care
personnel from participating sites to
complete the survey.
Procedure
Development and verification of the task
database. We conducted two-day focus
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
groups with each of six core primary
care occupations across the six partici-
pating sites. To minimize the burden to
each facility, we restricted the number
of occupational groups to three per
facility (i.e., no more than three job-
specific focus groups per site). Across
the six participating sites, however, we
conducted two to three focus groups
for each occupational title. In total,
we conducted 15 focus groups among
77 healthcare personnel to compose [299], (5)
primary care task statements for the
database. Per FJA protocol, task state-
Lines: 97 to 115
ments generated in the focus groups
———
were evaluated and edited by three
0.0pt PgVar
certified job analysts and reviewed by ———
the focus group participants to ensure Normal Page
linguistic fidelity. As the final step in PgEnds: TEX
developing task statements, two job
analysts independently rated each task
[299], (5)
according to the functional complexity
dimensions. Rating discrepancies were
resolved through a consensus process
and mediated by a third job analyst
as necessary. All task statements were
merged into a final database consisting
of 243 unique task statements of VHA
primary care work.
The final set of task statements was
compiled into a machine-readable
survey format and sent to the local
principal investigators at the six par-
ticipating facilities. The principal in-
vestigators distributed the survey to all
primary care personnel. For each task
statement, participants were asked to
indicate whether or not they performed
the task in question (task endorsement),
how frequently they performed the task
per week (frequency), and how many
minutes it took them to complete
one instance of the task (duration).
299
the permission
of 299
/ Vol. 51, No. 5/ Page Health
1st proof
 J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
These frequency and duration measures
reflect actual time-spent metrics with
demonstrated reliability (Albert et al.
1997). We also asked respondents
to report the number of years they
had been employed, the number of
years they had been in the current
position, and the highest degree they
had earned.
To test our hypothesis of sub-
stantial overlap in task performance
among relevant occupational groups,
we examined the survey data using
relational database utilities. We first
created occupation-specific tables fea-
turing the subset of task statements
endorsed by each occupational group.
Then we analyzed task endorsement
(e.g., responses to the question, “Do
you perform this task?”) to identify
tasks performed by each of the core
occupational groups of interest.
We also computed the average
frequency (per week) and duration (in
minutes) of performance for each task
statement in each occupation-specific
table. The cross-products of these data
were used as estimates of the number
of minutes per week spent on each task
endorsed by each occupational group.
Time estimates were then used in con-
junction with cost information from
the VHA’s Personnel Automated Infor-
mation Database and from the Office
of Personnel Management to calculate
average cost per week for each task.
To examine overlap in task per-
formance, we queried the occupation-
specific tables and identified task state-
ments that were common across oc-
cupational dyads (e.g., MD/advanced
practitioner, MD/RN, MD/LVN, MD/
clerk, and MD/health technician, and
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
so on). The queries generated separate
tables featuring task statements mutu-
ally endorsed by occupational group
dyads. Next we calculated the percent-
age of overlap from the perspectives of
both occupational groups in each dyad.
Specifically, both ratios featured the
number of overlapping tasks (i.e., mu-
tually endorsed) in the numerator and
the number of endorsed occupation-
specific tasks in the denominator. To
illustrate, the two perspectives featured [300], (6)
in the MD/advanced practitioner dyad
are represented analytically as follows
Lines: 115 to 142
(see Table 1):
———
0.0pt PgVar
• MD perspective = the number of ———
overlapping tasks divided by the Normal Page
number of MD tasks, or 121/141 * PgEnds: Eject
• Advanced practitioner perspective =
the number of overlapping tasks [300], (6)
divided by the number of advanced
practitioner tasks, or 121/134
R E S U LT S
We received 231 responses to the sur-
vey from a distribution of 619 for an
overall response rate of 37.32 percent.
Among the survey participants, 51
percent were MDs, 39 percent were
RNs, 35 percent were NPs or PAs, 24
percent were LVNs, 12 percent were
health technicians, and 18 percent were
clerks. Slightly more than half of the
respondents indicated that they had
been in their current position five years
or less and the average number of years
of education reported was 15.3.
Tasks Performed by Occupational Groups
We defined “work performed by tar-
geted occupational groups” as the set
of task statements, from among the
300
the permission
of 300
/ Vol. 51, No. 5/ Page Health
1st proof
 Tas k Ov er l ap Among Pr imar y C ar e Team M ember s

•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••

TA B L E 1
Number of Tasks Endorsed, by Occupational Titles and Dyads

MD NP/PA RN LVN MAS/PSA Health Technician

MD 141
NP/PA 121 134
RN 109 114 173
LVN 85 87 129 133
MAS/PSA 20 18 41 34 44
Health technician 33 31 46 42 24 48

Note: The numbers in the diagonal represent the total number of tasks endorsed by each occupational title (i.e., of [301], (7)
the 243 full set). The numbers in the off-diagonal reflect the number of tasks mutually endorsed by the associated
dyad.
••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• Lines: 142 to 198
———
0.0pt PgVar
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
———
Normal Page
TA B L E 2
Percentage of Primary Care Tasks Endorsed by Each Occupational Title * PgEnds: Eject

Percentage of the
[301], (7)
Full Set of 243 Tasks

MD 58.02
NP/PA 55.14
RN 71.19
LVN 54.73
MAS/PSA 18.11
Health technician 19.75
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••

full set of 243 tasks in our database, the fewest number (i.e., 44/243, or
that were endorsed by at least two 18.11 percent of the full set).
job incumbents from two different
VHA facilities. As seen in Tables 1 and Task Overlap Among
Occupational Groups
2, the MDs, NPs and PAs, RNs, and
The percentages listed in Table 3 reflect
LVNs reported performing the greatest
overlap from the perspectives of each
number of tasks and endorsed the occupational group in each dyad com-
highest percentages of the 243 tasks. parison. Results revealed substantial
RNs endorsed the most tasks (173/243, overlap in task endorsement among
or 71.19 percent of the full set), while MDs, NPs and PAs, RNs, and LVNs.
clerks (MAS/PSA) reported performing Overlap among these occupational

301
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 301
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
 J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006

•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••

TA B L E 3
Percentage of Task Overlap Among Occupational Titles

MD NP/PA RN LVN MAS/PSA Health Technician

MD 90.30 63.01 63.91 45.45 68.75

NP/PA 85.82 65.90 65.41 40.91 64.58
RN 77.30 85.07 96.99 93.18 95.83
LVN 60.28 64.93 74.57 77.27 87.50
MAS/PSA 14.18 13.43 23.70 25.56 50.00
Health technician 23.40 23.13 26.59 31.58 54.55

Note: Percentages reflect the perspective of one occupational group (listed in the top row) compared to the other
group (listed in the left-hand column) in each dyad. The percentages are derived from a ratio in which the numera-
tor is the number of overlapping tasks and the denominator is the number of occupation-specific tasks.
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
groups ranged from a low of 60.28
percent (LVNs endorsed 85 of the 141
tasks endorsed by physicians) to a high
of 96.99 percent (RNs endorsed 129
of the 133 tasks endorsed by LVNs).
Conversely, clerks endorsed the small-
est percentage of work also endorsed
by the other occupational groups (i.e.,
ranging from 13.43 percent of the tasks
also endorsed by advanced practition-
ers to 50 percent of the tasks endorsed
by health technicians). The percentage
of task overlap among nursing person-
nel (RNs and LVNs) and clerks and
health technicians ranged from 77.27
percent (LVNs endorsed 34 of the 44
clerk task statements) to 95.83 percent
(RNs endorsed 46 of the 48 health
technician statements). These findings
support our hypotheses of overlap in
task endorsement among the targeted
occupational groups.
Frequency and Duration of
Task Performance
To illustrate a potential application of
frequency and duration information
in task reassignment analyses, we ex-
[302], (8)
Lines: 198 to 230
———
amined the reported frequencies per 0.0pt PgVar
week for each of the 129 overlapping ———
RN/LVN tasks. There was substantial Normal Page
variation in the frequency of the tasks PgEnds: TEX
performed by the two groups. We
created three categories of tasks from [302], (8)
this distribution: those tasks that LVNs
reported performing five or more times
per week more than RNs (n=44); those
that RNs and LVNs reported perform-
ing approximately the same number
of times per week (n=41); and those
that RNs reported performing five or
more times per week more than LVNs
(n=14). For purposes of illustration,
we targeted the tasks performed more
frequently by RNs for opportunities
to assign work to optimize time and
cost savings. Specifically, we identified
two tasks from the set of 14 that were
performed more frequently by RNs but
more efficiently (i.e., less duration) by
LVNs. We programmed optimization
models in Excel to estimate time and
cost savings that would result from
reassigning RN responsibility to LVNs
for the two tasks. These reallocation

302
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 302
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
 Tas k Ov er l ap Among Pr imar y C ar e Team M ember s
scenarios resulted in an estimated
weekly time savings of 308.35 minutes
and 143.03 minutes and projected cost
savings of $212.65 and $161.67 for the
two tasks, respectively.
DISCUSSION AND
I M P L I C AT I O N S
Job analytic methods for delivery sys-
tem redesign represent an innovative,
systems approach to evidence-based
management of human resources.
Our research illustrates how system-
atic identification of overlap in task
performance can be used as the first
step in delivery system redesign. When
multiple occupational groups perform
the same work, as we found in our
analysis of VHA primary care, opportu-
nities exist for delivery system redesign.
Our finding that the bulk of tasks in
our database were performed by RNs,
MDs, NPs and PAs, and LVNs raises the
question of whether health technicians
and clerks are being used effectively in
primary care. Perhaps MDs, advanced
practitioners, and RNs and LVNs could
be relieved of the responsibility for
tasks already being performed by clerks
and health technicians. In our study,
RNs reported performing all but three
tasks endorsed by clerks and all but
two tasks performed by health techni-
cians. This signals a tremendous oppor-
tunity to redesign the delivery system
by offloading work, if the functional
complexity of task performance permits
the work to be done by others. In
theory, relieving RNs of responsibility
for performing low complexity tasks
would allow more time for tasks that
are higher in functional complexity and
that make better use of RNs’ skills and
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
training. Optimizing the fit between
worker competencies and task complex-
ity should also yield greater comfort
and satisfaction among healthcare
workers by minimizing consequences
such as boredom (underutilization of
worker skills) and burnout (overutiliza-
tion of worker skills). Actual reassign-
ment decisions, however, must consider
contextual factors such as process flow,
local skill levels, and clinic layout.
Evidence-based decisions regarding [303], (9)
the appropriateness of task assignment
may also assist the implementation of
Lines: 230 to 241
clinical practice guidelines. Systematic
———
assignment of task responsibility for
0.0pt PgVar
practice guidelines may help ensure ———
greater adherence to the guidelines Normal Page
and avoid the default of allowing PgEnds: TEX
“whoever has time” to perform the
work (Pugh 2002). This approach also
[303], (9)
affects quality of care: task assign-
ments are not based exclusively on
who can do the work, but rather on
who is best suited given the full array
of information.
Using a task database also helps to
enhance timely access to care. Theories
of advanced clinic access advocate
redistribution of work among other
qualified members of the care team
to minimize constraints in clinic flow.
Our results offer evidence to support
such redistribution efforts. In addition,
the use of a primary care task database
also affects patient safety. For example,
task ratings that evaluate human error
highlight the potential ramifications
of errors in task performance. This
evidence may be used to guide task
allocation such that the most conse-
quential tasks are delegated to workers
with the greatest skill, training, and
303
the permission
of 303
/ Vol. 51, No. 5/ Page Health
1st proof
 J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
experience. Although it may be argued
that licensure and certification already
circumscribe work responsibilities for
patient safety insurance, a job analytic
database offers greater precision by
drilling down to task-level specificity.
Whereas licensure and certification
focus on scopes of practice in general,
our approach guides the determination
of who should perform individual
tasks, the smallest identifiable unit of
work content. In addition, task state-
ments afford maximum flexibility in
analyses of work processes (Fine and
Getkate 1995).
Limitations
Our study was inherently descriptive
and intended solely to illustrate the
potential utility of task analysis toward
redesigning current patterns of work as-
signment. Our goal was to construct a
database describing primary care service
delivery. Other functional areas such
as administrative and logistic (supplies
and materials) responsibilities require
further development. The 243 task
statements should not be construed
as the entire universe of primary care
work in the VHA. We used survey in-
formation to profile task endorsement,
including the frequency and duration
of performance. Accordingly, our data
are subject to the psychometric concern
for accuracy that is common in survey
research. Even though we examined
the distribution of responses and sta-
tistically trimmed data points beyond
two standard deviations of the mean,
the usual caveats regarding survey data
apply here (i.e., sample dependence,
cross-sectional data collection, and self-
report bias). However, while we cannot
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
claim complete reliability with regard
to our task performance data (e.g.,
frequency and duration), our research
clearly illustrates the potential utility of
a primary care task database methodol-
ogy in evidence-based decision making
regarding personnel.
Caution should be taken in inter-
preting overlap among the occupa-
tional dyads. Our finding that tasks
endorsed by RNs overlapped most
often with other tasks performed by [304], (10)
the other occupational titles is not
unexpected, given that RNs reported
Lines: 241 to 252
performing the greatest number of
———
tasks. However, to the extent that these
0.0pt PgVar
data accurately reflect current patterns ———
of work responsibility, these percent- Normal Page
ages signal extensive opportunities for PgEnds: TEX
delivery system redesign.
[304], (10)
CONCLUSION
A primary care task database has great
potential for calibrating human re-
sources and reengineering work respon-
sibilities. In contrast to subjective ap-
proaches to personnel decision making,
FJA explicitly uses subject matter exper-
tise to compile a rich array of evidence
to better align the work content with
requisite worker characteristics. Based
on the high levels of overlap among
primary care team members, there
appears to be tremendous opportunity
to reallocate work more efficiently and
effectively. Identification of overlap in
task performance, however, is but one
step in the process of delivery system
redesign. The functional complexity
ratings must also be used to guide the
assignment of task responsibility to
personnel with the required compe-
tencies to perform the task. Likewise,
304
the permission
of 304
/ Vol. 51, No. 5/ Page Health
1st proof
 Tas k Ov er l ap Among Pr imar y C ar e Team M ember s
delivery system redesign should not
focus on any one dimension (e.g., cost)
and should consider the possibility
of unintended consequences of the
reallocation of tasks (e.g., worker dis-
satisfaction).
Healthcare managers will find
the validity of job analytic methods
useful in personnel decision making,
both in terms of legal defensibility
and in terms of quality of care and
patient safety (U.S. Department of
Labor 1978). The proposition that job
analytic databases provide evidence to
help guide delivery system redesign is
philosophically analogous to imple-
menting evidence-based medicine to
guide healthcare quality. In a sense, the
primary care database may be likened
to a toolkit for guiding evidence-based
management of human resources.
This methodology is not a standard
pathway to evidence-based personnel
management. Rather, a primary care
task database is a tool that must be
adapted to the local context (including
clinic layout, staffing configurations,
and skill levels) in a collaborative effort
among SMEs, management, and labor
(Bray et al. 2000; Kowalski et al. 2003;
Yorks et al. 1996). Indeed, this method-
ology has been used in many different
industries and has a time-honored
history (Fine and Cronshaw 1999; Fine
and Getkate 1995; Moore 1999 and
2000; Gilpatrick 1972). Thus, while the
methodology is generalizable across
different settings, it must be adapted for
implementation in the local context.
According to the Institute of Med-
icine (2000), “tens of thousands of
Americans die each year from errors in
their care, and hundreds of thousands
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout
of Healthcare Management
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
suffer or barely escape from nonfatal
injuries that a truly high-quality care
system would largely prevent.” To help
bridge this quality chasm, systematic
analysis of the work, worker, and work
organization provides evidence to
guide one aspect of the “radical and
sustained system-level change” recom-
mended by the Institute for Healthcare
Improvement.
[305], (11)
Acknowledgment
The authors thank Dr. Judith Patter-
son, director of the Biometry Core at
Lines: 252 to 284
VERDICT, for her assistance in database
analyses. ———
0.0pt PgVar
———
Note Normal Page
1. The research reported here was sup- PgEnds: TEX
ported by the Department of Veterans
Affairs, Veterans Health Administration,
Health Services Research and Develop- [305], (11)
ment Service (IIR-01–185). The views ex-
pressed in this article are solely those of
the authors and do not necessarily reflect
the position or policy of the Department
of Veterans Affairs or the University of
Texas School of Public Health.
References
Albert, W. G., W. Bennett, Jr., K. Pemberton,
C. N. Holt, and P. Waldroop. 1997.
“Evaluation of Automated Technology in
Basic Military Training Using Computer-
Based Surveys.” Presentation at the 38th
annual conference of the International
Military Testing Association (IMTA),
Sydney, Australia.
Bray, J., J. Lee, L. Smith, and L. Yorks. 2000.
Collaborative Inquiry in Practice: Action,
Reflection, and Making Meaning. Thousand
Oaks, CA: Sage.
Fine, S. A., and M. Getkate. 1995. Benchmark
Tasks for Job Analysis: A Guide for Func-
tional Job Analysis (FJA) Scales. Mahwah,
NJ: Lawrence Erlbaum.
Fine, S. A., and S. F. Cronshaw. 1999. Func-
tional Job Analysis: A Foundation for Hu-
305
the permission
of 305
/ Vol. 51, No. 5/ Page Health
1st proof
 J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006

man Resource Management. Mahwah, NJ:
Lawrence Erlbaum.
Gilpatrick, E. 1972. Health Services Mobility
Study: Final Report for the Period October
1967 Through March 1972. Technical
Report no. 11 [ERIC Reports no. ED 066
613]. New York: City University of New
York Research Foundation.
Horrocks, S., E. Anderson, and C. Salisbury.
2002. “Systematic Review of Whether
Nurse Practitioners Working in Primary
Care Can Provide Equivalent Care to
Doctors.” British Medical Journal 324:
819–23.
Institute for Healthcare Improvement. 2002.
“Idealized Design of Clinical Office Prac-
tices.” [Online information; retrieved
10/12/02.] http://www.ihi.org/idealized/
idcop/ background.asp#Model)-a.
Institute of Medicine. 2000. To Err is Human:
Building a Safer Health System, edited by
L. T. Kohn, J. M. Corrigan, and M. S.
Donaldson. Washington, DC: National
Academies Press.
. 2001. Crossing the Quality Chasm:
A New Health Care System for the 21st
Century. Washington, DC: National
Academies Press.
Kowalski, R., J. Harmon, L. Yorks, and D. Ko-
walski. 2003. “Reducing Workplace Stress
and Aggression: An Action Research
Project at the U.S. Department of Vet-
erans Affairs.” Human Resources Planning
Journal 26 (2): 39–53.
Moore, F. I. 1999. Functional Job Analysis:
Guidelines for Task Analysis and Job Design.
Prepared for the World Health Organi-
zation. [Online information; retrieved
7/25/06.] http://www.who.int/hrh/tools/
job analysis.pdf.
. 2000. Guidelines for the Application of
Functional Job Analysis. Chapter 13,
Tool Kit for HRH. Geneva: World Health
Organization, Division of Development
of Human Resources for Health.
Murray, M., and D. M. Berwick. 2003. “Ad-
vanced Access Reducing Waiting and
Delays in Primary Care.” JAMA 289 (8):
1035–40.
Pfeffer, J., and R. A. Sutton. 2006. “Evidence-
Based Management.” Harvard Business
Review 84 (1): 63–74.
Pugh, J. A. 2001. “Knowledge Management
and Clinical Practice Guideline Imple-
mentation.” Proposal no. CPI 99129–1.
San Antonio, TX: Veterans Evidence-Based
Implementation Dissemination Center. [306], (12)
. 2002. “Knowledge Management and
Clinical Practice Guideline Implementa-
tion.” Final Report for VA Health Services Lines: 284 to 349
Research & Development Project No. CPI
99129–1, San Antonio, TX.
———
U.S. Department of Labor. 1978. “Uniform -2.75499pt PgV
Guidelines on Employee Selection Pro- ———
cedures.” In Chapter 60: Office of Federal Normal Page
Contract Compliance Programs, Equal Em- PgEnds: TEX
ployment Opportunity. [Online informa-
tion; retrieved 10/31/02.] http://www.
dol.gov/dol/allcfr/Title 41/ Part 60– [306], (12)
3/toc.htm.
Von Korff, M., J. Gruman, J. Schaefer, S. J.
Curry, and E. H. Wagner. 1997. “Collab-
orative Management of Chronic Illness.”
Annals of Internal Medicine 127: 1097–
102.
Walshe, K., and T. G. Rundall. 2001.
“Evidence-Based Management: From
Theory to Practice in Health Care.” The
Milbank Quarterly 79 (3): 429–57.
Yorks, L., J. O’Neil, V. J. Marsick, G. E. Nil-
son, and R. Kolodny. 1996. “Bound-
ary Management in Action Reflection
Learning™ Research: Taking the Role of a
Sophisticated Barbarian.” Human Resource
Development Quarterly 7: 313–29.

P R A C T I T I O N E R A P P L I C A T I O N
Robert P. McDivitt, FACHE, deputy network director, U.S. Department of Veterans
Affairs, VA Midwest Health Care Network (VISN 23), Minneapolis, Minnesota

T he primary question posed by Dr. Best and his colleagues is whether the use
of functional job analysis and other traditional tools for selecting and training

306
Photocopying and distributing this PDF of the Journal of Healthcare
Management is prohibited
BookComp/ Health Administration Press/ Journalwithout the permission
of Healthcare Management of 306
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
 Tas k Ov er l ap Among Pr imar y C ar e Team M ember s

healthcare workers can be expanded to improve task assignment in a primary care

setting. Based on the data provided, the answer clearly is yes.
For years primary care managers have struggled to determine an appropriate
mix of physicians, mid-level providers, RNs, LVNs, health technicians, and clerical
support. We have used guides such as panel size, workload projections, and a
subjective staffing “feel” developed through years of experience. The authors assert
that staffing assignments can be based largely on objective, measurable criteria.
Best and colleagues propose that evidence-based validation takes much of the
guesswork out of the staffing process.
An unsurprising conclusion of the study is that physicians and other inde-
pendent practitioners (physician assistants, nurse practitioners, and clinical nurse
specialists) perform many of the same duties. We must strive to eliminate these [307], (13)
duplications. However, the large percentage of a physician’s time that is spent on
duties more appropriate to less complex positions is disturbing. Appropriate redis-
Lines: 349 to 360
tribution of these duties should allow healthcare managers to significantly improve
———
primary care efficiency. The study suggests that there is clear opportunity to move
* 25.53252pt PgVar
more routine duties from our most expensive staff (physicians) to others. This is a ———
finding that must be further explored and developed in the practical laboratories of Normal Page
our hospitals and clinics. * PgEnds: PageBreak
The primary analytical tool used in the study—functional job analysis (FJA)—is
a technique developed in the 1950s by Sidney Fine and others seeking a databased,
[307], (13)
equitable method to select employees for promotion or training opportunities.
Best and his team have transferred this tool to the arena of healthcare. The 243-
question survey, administered by professional job analysts and supported by clin-
ical and administrative subject matter experts (SMEs), provided a supportable
scientific foundation for the kind of decision making that is currently lacking in
most healthcare staffing models.
Implementation of FJA in a primary care setting is fully compatible with cur-
rent methods of flow analysis such as advanced clinic access (ACA), which, like
FJA, breaks our workflow processes into their component parts. As a national
leader in ACA, the U.S. Department of Veterans Affairs is well suited to comple-
ment ACA initiatives with FJA and is appropriately viewed as a potential pilot im-
plementation site. Variables in the implementation process that may substantially
affect task assignment are clinic layout and the existing skill set of providers and
support staff.
The authors have identified significant duplication and potential misassignment
of duties among providers and staff from the six primary care sites participating
in the study. FJA techniques, combined with other process improvement activities
such as ACA, show great promise in injecting evidenced-based decision making
into our primary care staffing plans and improving efficiency by assigning the right
work to the right staff.

307
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 307
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.