|
Items in cart [0] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 144
THE COST OF GRADUATE MEDICAL EDUCATION IN OUTPATIENT SETTINGS*
Judith R. Lave, Ph.D.
Professor of Health Economics
University of Pittsburgh
Introduction
Until recently, medical students and residents received most of their clinical
training in hospital inpatient settings. Educators and others who are interested in
training physicians for practice in primary care have been critical of the focus on
inpatient training for some time (MacLeod, 1970; Schroeder et al, 1986; Karpf and
L.evey, 19861; they are now joined by those who realize that, because of changes in
medical practice, there is a need to conduct more physician training outside the
inpatient setting for all physicians.
However, as training is shifted from the inpatient to the outpatient sector, a
whole set of issues are raised. How should training in the outpatient sector be
organized? How much does it cost to train residents in the outpatient setting? Is
it more costly to train medical students in the outpatient setting than in the
inpatient setting? What will be the effect of developing a program in outpatient
training on the net revenues of an institution? Do we need different ways of
paying for services if we change the locus or training?
The questions above are very general. The term "outpatient setting"
embodies a range of practice settings. Formal outpatient clinics, Health
Maintenance Organizations, prepaid group practices, surg~-centers and even doe in
the box, are all outpatient settings. Some of these have been training sites for
medical students and residents since medical schools were first established.
However, in the past these settings have been of very limited importance, and the
type of training provided in those sites was not geared towards developing
· . · -
pr~ma~ care p. Scans.
* ~ would like to thank Lester Lave, Gordon MacLeod and Jessica Townsend for
helpful comments on an earlier draft; Lawrence CIare for bibliographic help and
Bruce Block, Marian Block and Michael Karpf for describing their training
programs.
~44
OCR for page 145
In this paper I address some of the issues related to the cost of training
medical students and residents in primary care. I begin by reviewing briefly the
literature on approaches to estimating the cost of graduate medical education in
the inpatient setting. Next I develop an analytical framework for examining the
cost of training in the outpatient sector and the cost of introducing medical
students and residents into outpatient clinical practice is explored at some depth.
I then examine the relevant empirical literature related to this issue. I go on to
examine cursorily the "financial" impact of the implementation of training
programs in primary care. I then address a number of other selected issues
related to training in primary care and conclude with some recommendations for
future research in this area.
The Cost of Medical Training in Hospitals
Teaching hospitals are multiproduct firms; they produce medical training,
patient care, research and community services. For the many services that are
joint products, it is impossible to allocate the costs among them in a nonarbitrary
fashion. Conceptually, the cost to society of medical education and training in the
hospital setting is equal to the difference in total costs (both physician and
hospital costs) between a hospital which provides these educational services and
one that does not assuming they differ only in the educational programs. It is
necessary to look at both physician and hospital costs because the trainees provide
patient care services in place of those ordinarily provided by trained trained
. · ~
P nys~c~ans.
With some exceptions, analysts have not tried to estimate the cost of
medical education as defined above because of the difficulty in determining the
costs associated with providing physician services in hospital settings. In general,
hospital accounting records include information on all of the expenditures incurred
by hospitals (nursing salaries, pharmacy drug costs, lab equipment and so forth).
These expenditures include the direct costs of the hospital training programs
(salaries of teaching physicians, resident stipends and fringe benefits as well as
hospital overhead costs that are allocated to the training programs). Data on
payments for services rendered to patients by physicians in hospital settings are
kept by patients and in the claims Blles of the third party payers.
In general analysts have focused on
gross costs of graduate medical education.
determining what ~ will refer to as the
The gross costs of graduate medical
education have two components: direct costs of the hospital residency programs
(defined above) and indirect costs which are the increase in patient care costs
incurred by the hospital because it is involved in graduate medical education.
Patient care costs are higher because there are increased space needs, additional
145
OCR for page 146
record keeping requirements, decreased productivity of other staff members, and
excess test ordering by residents because of their inexperience.
Estimates of the direct costs of graduate medical education are obtained
from hospital accounting records. Since the indirect costs cannot be measured
directly, they are determined statistically (Lave, 19851. In most cases, analysts
estimate "cost functions" in which hospital inpatient costs (excluding the direct
costs of graduate medical education) are regressed against a series of variables
known to influence hospital costs such as case-mix, hospital wages, hospital bed
size and some indicator of the hospital's teaching status. (Anderson and Lave,
1986; Sloan, Feldman and Steinwald, 1983; Thorpe, 19881. The specification of
the estimating equation varies across studies as does the measure of hospital
teaching activity. (The most common measures are the number of residents per
_
~ ~ · ~ · ~ ~ ~ ~ ~ · 1 ~ ~ 1 ~ · "01 · ~ · 1 ~ ~ ~ ~ ·1
bed or an indicator of the hospital s teaching Eaten - member of the Bounce
of teaching hospitals; teaching hospital but not a COTH hospital, or nonteaching).
Analysts thus estimate the incremental effect of teaching on hospital costs (minus
the direct costs of teaching) once the effect of other factors known to influence
these costs have been taken into consideration.
The results of these analyses indicate that the indirect costs of teaching are
positive; i.e., that hospital inpatient costs increase with the extent of the hospital's
involvement in medical education. However, the estimated size of the indirect
costs depends upon the group of hospitals studied as well as measures used to
control for other factors known to influence costs across hospitals. Nationally, a
one percentage point increase in the number of residents per bed is associated
with about a 5 percent increase in inpatient care costs per case.
Analysts are aware that these are "gross" costs of graduate medical
education. Some investigators have tried to estimate "net costs" by estimating
physician costs associated with a hospitalization either by obtaining estimates of
physician services from medical records (Arthur Young, 1986) or by merging
information on physician claims with cost information from hospital accounting
records (Cameron 19851. The results of these studies indicate that residents are
partial substitutes for fully trained physicians. Cameron for example, found that
in California payments to physicians for services provided to Medicaid inpatients
were lower in major teaching hospitals by Medicaid's share of the residents
salaries in those hospitals.
These studies shed no light on whether graduate medical education
programs lead to increased hospital profits (or reduced losses), increase physician
incomes, convey prestige on a institution, or are the source of other unmeasured
benefits. The profitability of graduate medical education programs will depend on
how these programs affect hospital admissions and on how hospitals and other
:46
OCR for page 147
providers of services are paid for the services that they provide. Under the
financing arrangements that prevailed in 1989, available information indicates that
graduate medical education programs in hospitals were profitable.
Cost of Training in The Outpatient Sector:
An Analytical Approach
In considering the cost of training programs in the outpatient sector, the
appropriate question to be raised is: what additional resources are needed to
accomplish this task or what is the increased monetary cost of adding this activity
(Delbanco and Calkins, 1988; Gavett and Mushlin, 1988)? (If all the resources are
costed out, then these questions will have identical answers.) However, these
questions are different from an equally important one: will the implementation of
this program generate net positive revenues? To the sponsoring institution or the
manager of the training program, the impact on net revenues is the more
important question.
The cost of the training program will depend on its nature. Training
programs can range from the development of a residency program in family
practice in a nontraditional setting, the enhancement of an outpatient focused
primary care component in an internal medicine residency program, or the
implementation of a clerkship for medical students in an HMO or family
practitioners' office for 4 weeks during the summer. Since the programs differ, so
too will their costs.
Consider the cost of adding residents to a primacy care practice. (Medical
students could easily be added to this example.) This case which addresses the
following question - is it more costly to provide medical care in the ambulatory
setting in which residents are both being trained and are providing services than
it is to provide care by fully trained physicians - is explored in depth because it
has generated the most interest. In the discussion that follows, the term FT
(Fully Trained) physician is used to describe physicians who have completed their
residency training.
In the ambulatory setting, residents acquire skills while "learning by doing".
However, while they are practicing, they receive faculty input through direct
supervision, case conferences, chart review and consultations. The nonphysician
component of the cost of care may vary with the level of training of the resident.
For example, residents may order more tests, may require more nursing time per
visit or may use more examining room time than more experienced physicians;
that is, there may be some indirect costs of graduate medical education (as the
term is used in the inpatient setting) in the outpatient sector.
i47
OCR for page 148
In order to focus on the most important factors influencing the cost of
training, assume that only physician services are needed to produce patient visits.
Figure One indicates both how productivity varies with the amount of training
that a physician has as well as the amount of faculty input provided to the
resident. The figure indicates that a FT physician working full time in the
clinical setting provides O] visits per unit time while the number of visits that a
resident provides varies with the year of training as indicated by the curve CD.
As suggested by the curve EF, the amount of faculty input needed per average
visit provided by the resident decreases as the residents become more experienced.
In Figure Two, OS indicates the average cost of a visit provided by a FT
physicians. RT indicates the average cost of a visit provided by the resident. At
first, the cost per visit provided by a resident is higher than that provided by a
FT physician, however it begins to fall as faculty input decreases and resident
productivity increases. The curve suggests that the cost may begin to increase
towards the end of training because the residents' salaries increase and the
number of hours worked decrease leading to an increase in the cost per hour.
Although the general shapes of Figures One and Two should be true for all
settings, the details of the curves will vary from site to site. For example, the
position of CD in Figure One will depend upon whether there is an adequate flow
of patients through the clinic, the efficiency with which the practice is operated,
and the pace at which residents practice. (The pace could influence both the
quality of patient care provided and the quality of the clinic experience from the
training perspective). The position of EF will vary with the amount of training
and feedback the resident actually receives.
· ~ ,, ~ , ~
Finally, the relative positions of OS
and RT will depend upon the hourly cost of residents and the faculty.
Thus, the net cost of training will in large part depend upon the amount of
faculty input into the training process, the relative income of the faculty and
residents and the flow of patient visits. It will also depend on the mix of
residents: in general the net cost will be positive for medical students and for
residents early in their training and negative for more senior residents. Because
the net cost of training depends upon so many factors, it is not surprising that
empirical studies of the cost of training, which usually considers single sites, often
reach very different conclusions about both its size and direction.
i48
OCR for page 149
Patients
Per Hour
Cost Per
Visit
S
o
J
C
O
~7 Output
-
/ Resident
t - Output
Faculty
lo) Input
0 12 24 36
Month of Traintug
FIGORE ONE
\
Bcsiten~
Fully Trained Ml)
Year of "atn
FIGURE "0
149
Faculty HD
Input Per
Resident Vis
T
OCR for page 150
Empirical Studies of The "Net Cost" of Training In The Outpatient Setting
Several studies have attempted to estimate the cost of training medical
students and residents in the outpatient setting. Some of these studies have
examined only the gross costs of education; but many have investigated the net
costs of education. In this respect, this research is different from that on the cost
of graduate medical education in the inpatient setting. The work on the net cost
of graduate medical education can be classified into four groups: time and motion
studies, replacement cost studies, relative cost studies and marginal productivity
studies.
Time and Motion Studies
In time and motion studies, the investigators directly observe the various
providers of care to determine how they spend their time. However these studies
often do not provide useful information into the incremental cost of providing
training. This is a particular problem when the faculty members are training
residents at the same time that they, the faculty, are providing care or are
training a number of people simultaneously. However, time and motion studies
can provide some information on the amount of faculty time that the residents use
when they, the residents, are the primary providers.
In a recent study of the cost and efficiency of prodding services in 15
general internal medicine ambulatory practices in teaching hospitals, Kosecoff and
associates (1987) used time and motion studies to determine how much time
patients spent with their primary providers) (faculty member, resident or nurse)
as well as how much time they spent with the attending physician and nurse
during their visit to the primaly providers. The investigators found that for the
patients' first visits, PGYls
In these practices patients were assigned to one of several types of
practitioners, residents or fellows, faculty members or nurses who served as
the primary provider during that visit. However, during a visit to a
primary provider, other providers would be involved in the care. A nurse
could be present in the room or provide distinct services; the attending
physician could see the patient being treated by the resident or consult with
the resident on the case.
i50
OCR for page 151
(residents in their first post graduate year) spent a statistically significantly longer
period with the patient than did the other classes of primary providers (MDs and
nurses); while for follow up visits, faculty physicians spent a significantly shorter
time. They also found that the attending physicians spent a significantly longer
period of time with PGYl's for both first and follow up visits (when the PGY1
was the primary provider) than they did with did PGY2s and PGY3s. There was
no significant difference in the amount of time the patient spent with the nursing
staff by type of primaly provider.
Kosecoff et al. estimated the direct physician and nursing cost per visit first
by estimating the average hourly wage for each type of provider and then by
multiplying that by the proportion of time the physicians spent with a patient by
type of primary provider. In the institutions studied, the average PGY1 earned
$23,000 a year and worked 70 hours a week while the average associate professor
earned $85,000 and worked 55 hours a week. The estimated average cost for a
visit by type of primary provider is shown in Table 1. The main factor
contributing to the difference in the costs across visits by type of primary provider
is the amount of faculty time per visits.
On average, the faculty members spent very little time interacting with
PGY2s and PGY3s. However, there was considerable variation in the amount of
time the faculty spent with residents across the 15 practices. For example, for a
follow up visit, when the PGY2 was the primary provider, the attending spent on
average less than a minute with the patient in 9 of the 15 practices, between 1
and 6 minutes in 4 of the practices and over 6 minutes in 2 of the practices.
(The amount of the time the attendings spent discussing the case with the
residents outside of the examining room was negligible.)
2
For example, the cost of the first visit is highest when the faculty member
is the primary provider because most of the care is provided by the mostly
costly provider. The cost of a visit to a PGYl is high because, although the
hourly cost of the PGY! is low, the attending physicians spends a
significant amount of time with the patient and resident.
i5i
OCR for page 152
Table ~
Average Nurse and Physician Costs of Visits
To Fifteen Internal Medicine Group Practices
(in 1983 Dollars)*
Type of Primacy Prowder
First Visit Follow-Up
Mean
visits cost visits
Mean
Cost
n
n
up
Faculty 87 21.40+ 233 10.43#
Postgraduate Year ~32 18.03 80 6.33
Postgraduate Year II 35 9.24 145 5.44
Postgraduate Year Ill 42 9.55 120 5.47
Nurse Practitioner, 10 17.85 91 S.20
Physician Assistant
* Based on actual time spent with patient (time-and-motion study).
+ No significant difference with postgraduate year I; p < 0.01 with postgraduate
years I] and ITI.
# p ~ 0.01 with all three postgraduate years and nurse practitioner.
~52
OCR for page 153
Replacement Cost Studies
In a replacement cost study, the following question is posed: if outpatient
clinics continued to serve the same number and mix of patients, what would be
the cost of closing the teaching programs and using full time practicing physicians
to provide the services that were previously provided by the residents.
Researchers who have used this approach have carefully studied the clinic practice
(i.e. examined the number of patients seen per hour by faculty and by residents,
asked questions about the use of other personnel, noted the use of examining
rooms) and estimated how changing the physician mix would change the costs in
that clinic. me '' ' ' ' ' -' ~ ~ ~
data.
-sneer estimates are cased on ~uagements not statistical analyses of
Two recent replacement cost studies have been reported in the literature.
Delbanco and Calkins (1988) studied the effect of replacing residents with full
time physicians at Health Care Associates, a primary care practice at the Beth
Israel Hospital in Boston; while Gavett and Mushlin (1988) studied the same
problem at a primary care practice at the Strong Memorial Hospital in Rochester.
Residents in all three post graduate years provided patient care in both sites.
Table 2 presents estimates of how replacing residents with FT physicians
would change the average cost per visit and its component costs at the two sites.
As indicated by these numbers, Gavett and Mushlin assume that residents use
space less efficiently than FT physicians; while Delbanco and Calkins assume that
residents use relatively more of other staff time than do the FT physicians.
Gavett and Mushlin do not indicate that replacing the residents would reduce the
time of attending physicians, whereas Delbanco and Calkins indicate that such
costs would be decreased by $3.59 per "visit". Neither team of researchers
estimated the effect of the change on the use of ancillary services.
Gavett and Mushlin conclude that replacing residents with FT physicians
would decrease the cost of care, whereas Delbanco and Calkins conclude that such
a replacement would increase costs. The main factor leading to these different
conclusions is not the different assumptions about how residents and FT
physicians differ in their demands on other resources (space, nurses, etc.) but
rather the assumptions about the relative costs of residents and FT physicians in
the two sites. Delbanco and Calkins allocated about 5 percent of a resident's
salary for a half day clinic session (this was based on a resident's work week of
~53
OCR for page 154
Table 2
Change in Cost Per Visit If Trained Physicians Replace Residents
Other Net
Faculty Resident Training Fringe Space Staff Change
GMi $ 9.09 $-7.92 $-2.85 $ -1.68
DC2 12.67 -1.68 $-3.59 $ 1.34 $-1.34 8.40
iCalculated from data presented in Gavett and Mushlin, 1988.
2Calculated from data presented in Delbanco and Calkins, 1988.
~54
OCR for page 162
the income the faculty receive as attendings when family practice patients are
hospitalized as well as income from foundation grants, state allocations, federal
grants as well as contributions from the sponsoring institutions.
In 1982 Ciriacy et al. surveyed a national sample of family practice
residency programs asking about program costs and revenues. They found that,
on average, patient income covered 31 percent of the total costs of the programs.
In 1984, Ramsay conducted a national survey of family practice residency
programs. He too found that that patient revenues accounted for 31 percent of
the costs of the program. (In both surveys there usable response rate was about
50 percent).
The proportion of program costs covered by revenues from the family
practice clinics appears to be low. Three explanations have been offered: (~) The
volume of patients seen is small. Ramsay believes that the number of patients
could be increased by about 30 percent without impairing either patient care or
residency training (in fact it could even enhance training). If the patient volume
met his suggested targets (see Table 3), then practice revenues would cover about
45 percent of the program costs. (2) Collection rates are lower than necessary.
(3) Fees may be set too low. However, the fees charged at these clinics have to
be competitive with fees charged for visits in the clinic's market area. Since
insurance coverage for outpatient services is much less generous than it is for
inpatient coverage, clinic prices cannot be much higher than those at private
physician offices.
General Internal Medicine Ambulatory Practices in Teaching Hospitals
In 1980 the Robert Wood Johnson Foundation sponsored a program to
develop general internal medicine primacy care practices in 15 teaching hospitals.
These practices were established to deliver primary care and to train residents. In
theo~, they were to replace the old clinic practices and to emphasize continuity in
care, and to be a setting in which residents were trained in the behavioral and
social aspects of being a primary provider. These practices were established in
low income areas.
Kosecoff et al. (1987) and Brook et al. (1987) evaluated the demonstration
sites. They looked at the cost of teaching, the efficiency with which care was
provided and the financial impact of the ambulatory practices on the sponsoring
hospitals. (They were not not interested in the financial impact of establishing a
training program in primacy care but rather on the financial impact of
establishing ambulatory practices which served low income populations and in
which residents would be trained in primary care).
162
OCR for page 163
Table 3
Visit Targets for the Residency Sessions
Level Session/wk Pts/Session Pts/wk # wks Total
1 1 5 5 48 240
2 3 10 30 48 1,440
3 4 12 48 48 2,304
TOTAL 8 27 83 48 3,984
Average visit/resident .R fi4R = 1,328
4
Expected income per resident 1,328 x $27.06* = $35,936
Actual income per resident 1,015 x $27.06 = $97 4fi4
Difference per resident $ 8,472
*Income per visit derived from study data.
~63
OCR for page 164
The investigators found that the revenues associated with these practices
exceeded the costs the hospitals incurred in operating them. Defined revenues
include not only revenues received for the visits but also revenues received from
ancillary tests as well as any resulting hospital admissions. However, depending
upon the assumptions that were made about staffing patterns, Practice revenues
(revenues from the sernces directly provided by the residents and faculty
members) did not always cover practice costs.
~ , _ _ · · ~,,~
In general, the researchers found that the practices were managed veer
inefficiently; that staff time was not well utilized, that patient scheduling was poor
and that the use of space was inefficient.
Summary
The implementation of a training program in primaly care is accompanied
with a complicated flow of costs and revenues throughout the sponsoring
institutions. No study has tried to trace these flows in their entirety. but each
~.,
A,
study nas focused on a few or teem. In general, patient care revenues which are
received as a result of the provision of services in the primal care practices will
cover the costs of those practices, if the practices are allocated only a proportion
of the residents and faculty members costs - a proportion that reflects the time
they spend in the clinics. Clinic revenues, however, are not sufficient to cover the
full cost of the residents' salary and the salary of the faculty and administrative
staff who are directly the responsibility of the program. The practices associated
with the training program may be a new source of patients to the sponsoring
hospitals. In this case, the hospital may make profits off the ancillary services
and admissions that result from the program. Other benefits and costs of which
are a consequences of these training programs have not been measured.
Discussion: Other Issues
Issues touched on too briefly above include: the efficiency of the training
process; some of the differences between training in the inpatient and the
outpatient setting; the level of efficiency in the production of ambulatory services
in teaching settings, and some of the differences between training in an HMO and
a fee for service setting.
i64
OCR for page 165
The Efficiency of the Training Process
There were only a few studies which systematically examined the way that
physicians were being trained. Most of these studies described the structure of
the training programs, rather than indicate what actually happened in the
residency settings (Goodson et al., 19861.
The quality of the training received in the primary care clinic will depend in
part on the amount of faculty input as well as the volume of patients that are
seen. Figures Three and Four below display some hypothetical relationships
between resident "learning' and faculty input and patient flow. Since faculty input
and patient flow are the two variables that influence the cost of training, it would
be useful to know something about actual quantitative relationships depicted in
these figures. The importance of this knowledge is underlined by the fact that
faculty member input is the most important factor influencing the cost of care
provided by residents (and medical students) and because the studies discussed
above indicate that there is ode variation in the amount of faculty time actually
given to training in the outpatient setting.
There is also very little known about the level of training required of the
physician "trainer". In the inpatient setting, the resident is both a student and a
teacher. Residents spent some proportion of their time training medical students
(Institute of Medicine, 19761. However, at least as reported in the published
literature, the teaching role of the resident seems to be nonexistent in the
ambulatory setting. This points to the question: What is the role of the resident,
particularly the third year resident, as teacher in the ambulatory setting.
It is important to determine ways of decreasing the costs of training -
particularly for medical students and junior residents. Educators have suggested
the use of professional patients, computers, and simulations. Shueser et al (1985)
describe an interesting use of the wdeo in training medical students in the
emergency room.
The Patient's Role
The patient, a necessary ingredient in the training process, is different in the
inpatient than in the outpatient setting. In the inpatient setting the patient is
usually in bed. Consequently, rounds can be organized solely for the convenience
of the physicians. House staff can interact with the patient when they want to.
Patients rarely refuse to answer yet another resident or student asking the same
set of questions or being examined more than once. In the outpatient setting,
~ 65
OCR for page 166
Residents
Residents
Learning
/
a
a
~ _
FIGURE _
166
AD Input
P6r V1sic
Pat1~nt#s
Su./Hour
OCR for page 167
however, patients may be less willing to donate their time to the educational
process. This setting requires coordination of the activities of the residents and
the attendings. It also means that the number of residents (and medical students)
who can learn from a given patient is limited.
Secondly, patients in the outpatient setting are more likely to want a doctor
who is fully trained rather than one who is still being trained. Although in the
inpatient setting people of all social and economic backgrounds are used to
interacting with the house staff, many patients have "their own doctor". However,
in the outpatient setting, one physician (either resident or FT physician) becomes
the patient's primary doctor. Many people, particularly in the higher social
economic groups, may be unwilling to be assigned to the resident's panel of
patients. This problem will be exacerbated in establishing a panel of patients for
the first year residents.
Thirdly, in the outpatient setting patients are more mobile. With the
increasing supply of physicians, people have more choices. This means that the
University based practices have to be competitive with community physicians in
order to get patients. However, this may exacerbate town and gown relationships
(Medical News, 19891.
These observations lead to the conclusion that the socio-economic background
of people in the residency practices will be different from those seen by the faculty
in their private practices. This difference means that that patients in the
residency clinics are less likely to have health insurance coverage. If the patients
are insured, they are more likely to be covered by less generous insurance
programs such as Medicaid.
The Efficiency with Which the Practice is Operated
Almost all studies which compared the cost of treating people in the
outpatient clinics as opposed to private physician offices have found the costs in
the former to be higher (Lion et al., 19851. Kosecoff et al. in their evaluation of
the ambulatory practices supported by the Robert Wood Johnson Foundation
found that the practices were very inefficiently run. They found that there were
problems with patient scheduling and scheduling of physicians. Once in the office,
patients waited longer to see the physician than they do in the offices of
community physicians. Some of these problems may be due to the differences
between the types of patients seen in these practices compared to those in private
physicians offices. For example, Kosecoff found that the no show rate was much
higher than that in private physician offices. A high no-show rate makes it
particularly difficult to schedule appointments. It may be possible to reduce the
~ 67
OCR for page 168
no-show rate but it is possible that it is related to the socio-economic
characteristics of the patient population.
These observations lead to the conclusion that it will be very important to
pay attention to the efficiency with which the primary practice clinics are
managed.
Training in the Health Maintenance Organization
There is increasing interest (Isaacs and Madoff, 1984) in using the health
maintenance organization as training sites. In both HMOs and other settings the
training of medical students and junior residents will add to the cost of care.
There is, however, one difference between the implications of training in the two
sites. We noted above that there were two sources of "indirect costs of graduate
medical education" in the ambulatory setting: one is the higher (nonphysician
costs) of residents visits (we noted that there were data suggesting that nursing
costs and space may be higher ); the second is increased ancillary testing and
perhaps increased admissions. In both the fee for service sector and the HMO
there is no increased revenue to offset the first type of increased costs; however,
in the fee for service setting there is increased revenue to offset those indirect
costs whereas in the health maintenance organization there is not.
Conclusions
There are many different ways of looking at the question of the cost of
training medical students and residents in the outpatient setting. A number of
analysts have been interested in the net cost of education in the ambulatory
setting where net cost is defined as the difference in the cost of producing a given
number of patient visits by residents who are also being trained and by full time
physicians. The net cost was found to depend upon a number of factors including:
the flow of patients through the clinics, the faculty members' input into the
resident's practice and the relative salaries received by the resident and the FT
physician. In general, we found that if the outpatient settings are allocated a
proportion of the resident's salary (that proportion that they spend at the clinic),
then the net cost of graduate education is negative for second and third years
residents, and positive for some of the first year students. This finding, however,
was not universal. This conclusion needs to be accompanied by a caveat. While
the net cost of training in the primary care clinic may be negative for a second
and third year resident, it does not follow that the net cost of training second and
~68
OCR for page 169
third year residents is negative. These residents spend time in subspecialty clinics
such as orthopedics, dermatology and gynecology where their productivity level is
not as high as it is in the primary care clinic.
Most, but again not all, studies found that there were indirect costs of
graduate medical education in the outpatient setting. Most of this work has
concentrated on the indirect costs associated with the provision of patient visits
(space, time and ancillary personnel). There needs to be more work done on
examining resident practice as it relates to the ordering or tests and to hospital
admission decisions.
The income generated by primary care practices is not sufficient to cover the
cost of the primary care training program where the training programs is
responsible for the resident's full salary. The main reason for this discrepancy is
that the residents spend significantly less than 50 percent of their time in the
. · -
prlma~y care c. .lnlcs.
There is some evidence that the primary care clinics are not efficiently run.
There also is some evidence that a number of clinics - particularly those that are
part of residency programs in family practice programs do not have as many
patient visits as they would like. It is not clear whether the shortage of patients
is due general competition from private physicians or because of certain attributes
of the clinics per se.
It is reasonable to conjecture that the insurance coverage of patients being
seen at the primary care clinics will be less complete that that of patients seen at
the faculty physicians private practice of by community physicians.
More needs to be known about the training process. Training in the
outpatient setting is probably more costly than it is in the inpatient setting. Not
only is more training done by attendings rather than residents, but also there
appears to be a higher attending/resident ratio. In the studies that were examined,
there was considerable variation in the amount of faculty time that was provided
to the resident in the clinic setting. Since faculty time is very expensive, research
on the training process per se would seem to be called for.
i69
OCR for page 170
BIBLIOGRAPY
Anderson OF and Lave JR, "Financing Graduate Medical Education Using
Multiple Regression to Set Payment Rates," Inquiry 23: 191-199, 1986.
Arthur Young, Study of the Financing of Graduate Medical Education: Report Il.
Hospital Cost Analysis, October 1986.
Bankhead CD, "Town and Gown," Medical World News, pp. 44-52, Manual 23,
1989.
BerkeThamer ATE, "Charges by Residents and Faculty Physicians in a University
Hospital Pediatric Practice," Journal of Medical Education 61: 303-307, April
1986.
Brook RH, Fink A., Kosecoff d., et al, "Educating Physicians and Treating Patients
in the Ambulatory Setting: Where Are We Going and How Will We Know
We Arrived?" Annals of Internal Medicine 107~3~: 392-398, 1987.
Cameron dM, "The Indirect Costs of Graduate Medical Education," New England
Journal of Medicine 312~19), May 9, 1985.
Ciriacy EW, Liang FZ, Godes dR, et al, "The Cost and Funding of Family Practice
Graduate Medical Education in the U.S.," Journal of Family Practice 20~3~:
285-295, 1985.
Colwill dM and Glenn ~K, "Patient Care Income and the Financing of Resident
Education in Family Medicine," Journal of Family Practice 13~41: 529-536,
1981.
Delbanco TE and Calkins OR, "The Costs and Financing of Ambulatory Medical
Education," Journal of General Internal Medicine 3 (Supplement): S34-S43,
1988.
Gavett JW and Mushlin AI, "Calculating the Costs of Training in Primary Care,"
Medical Care 24~4~: 301-312, April 1986.
Goodson ~D, Goroll AH, Barsly AJ, Treadway KK, Thibault GE and Stoeckle JD,
"The Training of Physicians Outside the Hospital," Archives of Internal
Medicine 146: 1805-~809, September 1986.
.
170
OCR for page 171
Page 2
Hadley J. Medical Education Financing: Policy Analyses and Options for the
l980s, PRODIST (New York, NY) 1980.
Institute of Medicine, National Academy of Sciences, Medicare-Medicaid
Reimbursement Policies: Social Security Studies. Final Report, March 1976.
Isaacs JO and Madoff MA, "Undergraduate Medical Education in Prepaid Health
Care Plan Settings," Journal of Medical Education 59, August 1984.
Kahn L, Wirth P and Perkoff G. "The Cost of a Primary Care Teaching Program
in a Prepaid Group Practice," Medical Care (XVI): 61-71, 1978.
Karpf M. and Levey GS, "Training Internists for the Changing Medical Scene,"
Annual of Internal Medicine 104(4), April 1986.
Kirz HE and Larsen C, "Costs and Benefits of Medical Student Training to a
Health Maintenance Organization," Journal of the American Medical
Association 256~6~: 734-739, August 8, 1986.
Kosecoff J. Brook RH, Fink A., et al, "Providing Primary General Medical Care in
University Hospitals: Efficiency and Cost," Annals of Internal Medicine 107:
399-405, 1987.
Lave JR, "The Medicare Adjustment for the Indirect Costs of Medical Education:
Historical Development and Current Status," Association of American Medical
Colleges, ~Janua~g 1985.
Lion J. Malbon A, Henderson MG and Friedman RH, "A Comparison of Hospital
Outpatient Departments and Private Practice,' Health Care Financing Review
6~4~: 69-81, September 1985.
Panton DM, Mushlin Al and Gavett dW, "Marginal Ambulatory Teaching Costs
Under Varying Levels of Service Utilization," Medical Care, (XVITI: 61:
668-674, June 1980.
Pawlson EG and Watkins R., "The Costs of a Family Practice Residency
Ambulatory Care Program," Journal of Family Practice 9~61: 1059-1061,
1979.
171
OCR for page 172
Page 3
PawIson KG, Schroeder SA and Donaldson M, "Medical Student Instructional
Costs in a Primary Care Clerkship," Journal of Medical Education 54~71:
551-555, 1979.
Rabkin, MT, "Reducing the Costs of Medical Education," Health Affairs, Fall 1986.
Schroeder SA, "Group Practice Recommendations of the Committee on the Costs
of Medical Care," Milbank Memorial Fund Quarterly 56~2~: 169-~86, 1978.
Schroeder SA, Showstack, JA and Gerbert B., "Residency Training in Internal
Medicine: Time For a Change?," Annals of Internal Medicine 104~41:
554-561, April 1986.
Shanke] SW and Mazzaferri EL, "Teaching the Resident in Internal Medicine:
Present Practices and Suggestions for the Future," Journal of the American
Medical Association 256~6), August 8, 1986.
Shesser R. Smith M, Kline P. Rosenthal R. Turbiak T and Chen H. "A Cost-
Effective Emergency Medicine Clerkship, Journal of Medical Education 60:
288-292, April 1985.
Sloan FA, Feldman RD and Steinwald AB, "Effects of Teaching on Hospital Costs,"
Journal of Health Economics 2: 1-2S, 1983.
Stern RS, et al, "Graduate Education in Primary Care: An Economic Analysis,"
New England Journal of Medicine 297(12): 683-643, 1977.
Thorpe KE, "The Use of Regression Analysis to Determine Hospital Payment:
The Cost of Medicare's Indirect Teaching Adjustment," Inquiry 25~2~:
219-231, 1988.
172
Representative terms from entire chapter:
medical students
