592
September 2002
Family Medicine
Residency Education
Does Community- or University-based Residency
Sponsorship Affect Future Practice Profiles?
Peter J. Carek, MD, MS; Dana E. King, MD; Stoney Abercrombie, MD
Background and Objectives: The educational sponsor of a residency program (ie, community based versus university based, opposed versus unopposed) is one of the most influential factors in medical students’
choice of a training program. This study examined the practice profiles of South Carolina family practice
residency graduates to determine whether there were significant differences based on program sponsorship. Methods: A survey was mailed to the 1,335 graduates of South Carolina Area Health Education
Consortium (SC AHEC)-affiliated residency programs. Data were summarized and analyzed using the
Epi-Info® statistical program. Significance was preset at the P<.01 level. Results: A total of 720 (56.0%)
surveys were returned. Community-based program graduates were more likely to practice in the state, in
a rural area, and closer to their residency site. University-based program graduates were more likely to
practice in academic settings (18.9% versus 6.3%). Several minor differences in procedures performed
were noted, with some procedures performed more commonly by physicians trained in university-based
programs and others by physicians trained in community programs. No differences in salaries or personal
or professional satisfaction between groups was found. Conclusions: The current study found some differences in location of practice but found few differences in procedures performed, salaries, or satisfaction
based on sponsorship of residency.
(Fam Med 2002;34(8):592-7.)
The educational sponsor of a residency or program
“subtype” (ie, community based versus university
based, opposed versus unopposed) is one of the most
influential factors in medical students’ selection of a
training program in family practice.1 Fourth-year medical students often express a preference for community
hospital-sponsored training programs, especially programs located in hospitals without other residency trainees (ie, an unopposed program). Whether consistent
differences in the qualifications of faculty, training received, or hospital privileges received by graduates of
community- or university-based programs affect selection is the subject of debate among students, residents,
and educators.1-3 Despite a lack of clear evidence that
training differs between community- and universitybased residency programs, recent trends in the National
Resident Matching Program (NRMP) indicate that uni-
From the Department of Family Medicine, Medical University of South
Carolina.
versity-based programs are having a more difficult time
recruiting residents.4
Indeed, university affiliation is second only to the
regional location of a program as a consistent factor in
predicting fill rates.5 The viability of family practice
residency programs at medical schools could be threatened if the trend of students preferring communitybased programs continues or increases.6 Further, programs located in larger hospitals serving as the primary
sites of training for multiple residencies (ie, opposed
programs) could also be threatened. Without prosperous residency programs in the medical schools or larger
teaching hospitals, family medicine may lose academic
credibility, and contact between family physicians and
medical students could diminish.
If students are concerned about their future ability to
practice in a community setting, is the strong student
preference for community-based, unopposed residency
training justified? Does the presence of other residency
or training programs at a given hospital correlate with
later practice patterns? Are graduates of communitybased programs different from graduates of university-
Residency Education
based programs in their practice patterns, hospital privileges, or procedural skills?
This study examined the practice profiles of the family practice residency graduates from programs sponsored by the South Carolina Area Health Education
Consortium (SC AHEC). Specifically, the aim was to
determine whether there are significant differences in
practice patterns of physicians based on the program
subtype of the residency in which they trained.
Methods
Subjects
Subjects for this study were all graduates of the seven
family practice residency programs located in South
Carolina. Of the seven SC AHEC-affiliated programs,
two are considered university-based programs, and five
are community-based programs. The two universitybased programs have been in operation for 25 and 30
years, respectively. The age of the five community programs ranged from 20 to 30 years.
Of the five community-based programs, three are unopposed, and two are opposed. An opposed community-based program was defined as a family practice
residency program whose primary hospital was also a
primary hospital for other specialty training programs.
An unopposed community-based family practice residency program was defined as one that is the only graduate medical education training program sponsored by
the primary hospital.
Instrument
A questionnaire was designed specifically for this
project (available from authors on request). The questionnaire was modeled after previous surveys. The questionnaire was reviewed by the authors and several other
family physicians for readability and understanding, and
it was reviewed and approved by all of the program
directors of the SC AHEC Program Directors Council.
Procedures
A complete list of graduates was obtained from each
SC AHEC-affiliated family practice residency program.
The questionnaire was mailed to each graduate with a
cover letter explaining the purpose of the study and
inviting the graduate to participate in the study. Completed questionnaires were returned in an accompanying self-addressed, stamped envelope. A second mailing with the same cover letter was sent to nonrespondents 6 weeks after the initial mailing. The responses on
completed questionnaires were recorded in a computerized database.
Vol. 34, No. 8
continuous variables and t tests for continuous variables.
To adjust for multiple comparisons, significance was
defined as P<.01 level of confidence.
For analysis, residency programs were divided into
groups. The initial analysis compared graduates of the
community programs with those from the universitybased programs. In further analyses, the communitybased programs were divided into two groups: opposed
and unopposed.
For selected outcomes that were found to be significantly different between the graduates of the community- and university-based programs, a model using logistic regression that included age, gender, and race was
developed. This further analysis ensured that the differences found were not due to differences in demographic factors. Race was subsequently removed from
the model due to the small number of non-white respondents.
Results
Based on the information obtained from the individual family practice residency programs about their
graduates, a tota1 of 1,233 surveys were mailed; 720
(58.4%) surveys were returned and used for analysis.
The response rates of graduates by program type are
presented in Table 1.
Of the respondents, 564 (79.6%) were males, and
145 (20.4%) were females. The average age of the subjects was 42.7 (+7.5) years, and a majority (95.3%) identified themselves as white, non-Hispanic. The gender
and ethnicity of the graduates were not significantly
different when comparing community- versus university-based programs (P=.85 and P=.27) and opposed
versus unopposed programs (P=.03 and P=.25).
Community- Versus University-based Programs
Practice Site. When compared with graduates of the
university-based programs, the graduates of community-based programs were more likely to practice in
South Carolina (52.1% versus 35.4%, P<.01). Similarly,
a higher percentage of graduates from university-based
programs practice more than 120 miles from the site of
training (74.6% versus 60.8%, P<.01).
Table 1
Response Rates of Graduates by Program Type
Surveys mailed (#)
Data Analysis
Descriptive statistics were used to characterize and
summarize the data obtained. The information was then
analyzed using the Epi-Info® statistical program. Analysis was completed using a chi-square statistic for non-
593
Surveys returned (#)
Response rate (%)
Community
Based
811
426
52.5
University
Based
Opposed Unopposed
422
453
358
294
69.7
200
44.2
226
63.1
594
September 2002
Family Medicine
Practice Setting. The practice arrangement, practice
setting (respondent self-classified by city, town, rural
community, etc) and setting size (ie, population range)
were significantly different between graduates of community- and university-based programs (Table 2). While
the highest percentage of graduates of both types of
programs practice in a small family practice group (ie,
< five members), a higher percentage of graduates from
university-based programs (18.9%) practice in an academic setting, compared with graduates of communitybased programs (6.3%) (P<.01). Further, a higher percentage of graduates from community-based programs
are practicing in towns and rural communities while
graduates of university-based programs predominantly
practice in cities and suburban areas. On further analysis, the graduates of community-based programs were
less than one third as likely (P<.01) to practice in an
academic setting when compared to graduates of university-based programs (odds ratio [OR]=.3, 95% confidence interval [CI]=.16–.52). In contrast, graduates
of community-based programs are 2.5 times as likely
(P<.01) to practice in a town or rural community
(OR=2.5, 95% CI=1.77–3.41).
Salary. The percentage of graduates of community programs with salaries within specific ranges (ie, <$50,000,
$50,000–$99,999, $100,000–$124,999, $125,000–
$149,000, $150,000–$174,999, $175,000–$199,999,
>$200,000) did not significantly differ from those physicians who graduated from university-based programs
(P=.39).
Satisfaction. The satisfaction with specialty choice or
personal life did not differ significantly between the
two groups (P=.57 and P=.19, respectively) (Table 3).
Specifically, 95.8% of the community-based program
graduates and 97.8% of the university-based program
graduates were either extremely or somewhat satisfied
with their specialty choice. Further, 77.1% of the community-based program graduates and 81.3% of the university-based program graduates were either extremely
or somewhat satisfied with their personal life.
Procedures. A comparison of procedures performed at
least once per month by individual physicians found
that nasolaryngoscopy is performed more often by
university-based program graduates, while respirator
management is performed significantly more often by
Table 2
Practice Demographics of Graduates
Community Based (%)
University Based (%)
Practice arrangement
Academic
HMO (staff model)
HPO
Large family practice*
Multispecialty group
Small family practice**
Solo
Other
6.3
.2
3.7
12.2
8.3
39.5
19.0
10.7
18.9
1.5
5.5
14.5
12.7
26.5
10.5
9.8
Practice setting
City
Indian reservation
Rural community
Suburban community
Town
Other
30.1
.5
29.9
12.0
25.5
1.9
40.4
1.1
14.4
21.3
20.2
2.5
Setting size
<5,000
5,000–10,000
10,000–50,000
50,000–100,000
100,000–500,000
>500,000
13.4
3.0
17.4
36.8
16.1
13.4
5.5
7.9
16.1
22.0
20.5
28.0
P Value
<.01
Opposed (%)
Unopposed (%)
4.7
.0
3.7
11.0
8.4
46.6
17.3
8.4
8.0
.5
3.8
13.7
9.0
31.6
20.8
12.7
29.8
.0
27.7
13.6
26.2
2.6
30.6
.9
31.5
11.1
24.5
1.4
11.2
3.5
20.0
32.9
17.1
15.3
15.6
2.6
15.1
39.6
15.1
12.0
<.01
.60
<.01
HMO—health maintenance organization
HPO—
* Large family practice > five members
** Small family practice < five members
P Value
.11
.43
Residency Education
Vol. 34, No. 8
Table 3
Satisfaction With Specialty Choice and Personal
Life of the Graduates by Program Type
Community
Based (%)
University
Based (%)*
Opposed (%)
Unopposed (%)**
Satisfied with specialty choice
Extremely satisfied
Somewhat satisfied
Somewhat dissatisfied
Extremely dissatisfied
69.4
26.5
3.4
.7
69.9
27.9
1.8
.4
67.6
29.3
2.7
.5
70.3
24.5
4.2
.9
Satisfied with personal life
Extremely satisfied
Somewhat satisfied
Somewhat dissatisfied
Extremely dissatisfied
24.8
52.3
17.4
5.4
30.8
50.5
15.8
2.9
21.8
52.7
18.1
7.9
27.4
51.9
17.0
3.8
595
Board Certification. Overall, graduates of university-based programs
were significantly more likely to be
certified by the American Board of
Family Practice (ABFP) than were
graduates of community-based programs (96.7% versus 90.5%, P<.01).
While the graduates of communitybased programs were found to be 2.5
times less likely (P=.02) to be board
certified, after controlling for age and
gender in the logistic regression
analysis, the difference was no longer
significant.
Opposed Versus Unopposed
The likelihood of practicing in
* P=.57 for community- versus university-based program graduates extremely or somewhat satisfied
South
Carolina and the distance of
with specialty choice; P=.19 for same groups extremely or somewhat satisfied with personal life.
practice from individuals site of resi** P=.60 for opposed versus unopposed program graduates extremely or somewhat satisfied with
dency training did not differ signifispecialty choice; P=.29 for same groups extremely or somewhat satisfied with personal life.
cantly between graduates of opposed
versus unopposed programs (P=.87
and P=.24, respectively). The practice arrangement, practice setting,
graduates of community-based programs (Table 4). Othand setting size also did not significantly differ between
erwise, no significant differences were noted between
graduates of opposed and unopposed family practice
procedures performed by physicians in the two groups.
residency programs (Table 2). The highest percentage
Additionally, no significant difference was noted beof graduates from each group are practicing in small
tween the type of residency program and whether the
family or solo practices. Further, a majority of gradugraduate provided prenatal care with or without delivates from both types of programs are practicing in citery: 11.7% and 10.9% of graduates of university-based
ies, towns, and rural communities. Similarly, the salary
programs provide prenatal care with or without delivranges (ie, <$50,000, $50,000–$99,999, $100,000–
ery, respectively, and 13.4% and 9.1% of community$124,999, $125,000–$149,000, $150,000–$174,999,
based program graduates do so (P=.63 and P=.58).
$175,000–$199,999, >$200,000) of the graduates of
Table 4
Procedures Performed At Least Once Per Month by Program Graduates
Community Based (%)
Central line placement
14.6
Chest tube placement
13.4
Colonoscopy
4.9
Endoscopic gastroduoenoscopy
3.6
Exercise stress testing
23.7
Flexible sigmoidos copy
34.7
Joint aspiration/injection
90.8
Nasopharyngoscopy
6.7
Paracentesis
18.4
Respirator management
43.9
Skin biopsy
87.4
Swan-Gantz catheter
3.4
Thoracentesis
19.8
Vasectomy
7.2
Short arm or leg casts
50.9
Surgical assistant
10.9
University Based (%)
11.5
8.7
8.2
4.5
19.5
44.0
87.8
18.0
20.5
31.2
85.5
3.0
18.6
9.1
45.2
10.7
P Value
.26
.06
.08
.55
.21
.02
.22
<.01
.49
<.01
.47
.78
.71
.40
.16
.92
Opposed (%)
16.4
13.8
6.7
3.9
22.1
42.1
92.2
7.8
18.1
41.9
86.2
3.4
18.5
10.0
60.8
10.6
Unopposed (%)
13.0
12.6
3.0
3.5
26.4
27.9
88.7
5.9
18.7
44.8
88.6
3.5
20.1
5.0
42.0
11.1
P Value
.35
.72
.09
.82
.33
<.01
.24
.48
.89
.57
.47
.95
.68
.06
<.01
.88
596
September 2002
two types of community programs, opposed and unopposed, did not significantly differ (P=.59).
The satisfaction with specialty choice or personal life
did not differ significantly between the two groups
(P=.60 and P=.29, respectively) (Table 3). Specifically,
96.8% of the opposed program graduates and 94.8%
of the unopposed program graduates were either extremely or somewhat satisfied with their specia lty
choice. Further, 74.5% of the opposed program graduates and 79.2% of the unopposed program graduates
were either extremely or somewhat satisfied with their
personal life. None of the differences were statistically
significant.
The procedures performed at least once per month
by individual physicians were compared for graduates
of opposed and unopposed community-based programs.
As noted in Table 4, flexible sigmoidoscopy and short
arm or leg casts are performed significantly more frequently by graduates of opposed residency programs
than by graduates of unopposed programs. Otherwise,
no significant differences were noted between groups
in procedures performed. Additionally, no significant
difference was noted between the type of residency
program and whether the graduate provided prenatal
care with or without delivery.
No difference was noted in ABFP certification status between graduates of opposed and unopposed residency programs (91.4% versus 89.3%, P=.47).
Discussion
The results of this study are consistent with previous
research showing that university-based program graduates are somewhat more likely to practice in larger cities and are more likely to pursue academic careers.1 In
addition, graduates of community-based residency programs are more likely to remain in the state in which
they did residency training and to work in a more-rural, single-specialty practice than graduates of university-based programs. No differences were noted, however, in salaries or professional or personal satisfaction.
One common view regarding sponsorship of residency training is that graduates trained in community
and unopposed programs may receive more direct training in common procedures due to less “competition”
for procedures from residents in other programs and
therefore will perform these procedures more frequently
in practice. This view is not supported by our data.
While community-based program graduates were more
likely to perform respirator management, universitybased program graduates were more likely to perform
nasolaryngoscopy. No differences were found for the
rates of performance of 20 other procedures, including
prenatal care and obstetrical delivery. Whether the differences in the two procedures performed were based
on the specific training of the graduate or other factors,
such as hospital policy, could not be determined.
Family Medicine
Previous research of family practice residency graduates has demonstrated little difference between physicians trained at university- and community-based programs. Davidson and Kahn 4 compared graduates of the
University of California-Davis programs and found no
difference in the ABFP examination scores or predominant type of practice. In general, graduates of both types
of programs felt well prepared in most content areas.
Similarly, a recent national survey of family practice
residency programs regarding procedural skills training showed no differences based on type of program or
university affiliation.7
Based on the results of our study and others, therefore, the differences in practice profiles between community- and university-based program graduates would
not appear to support the influence that this status has
on student preferences for site of residency training.
The overwhelming majority of graduates from all types
of programs enter full-time solo, partner, or group practice regardless of the educational affiliation of the residency at which they trained. Further, graduates have
similar salaries, similar professional satisfaction, and
perform similar procedures. Student preferences for
residency programs should thus be based on the quality of the educational programs, interpersonal issues,
and other characteristics of program quality and reputation, rather than academic affiliation status.5,8
Limitations
This study has several limitations. The overall response rate (56%) and the response rate of the graduates by program category may have affected the results.
In addition, the study included graduates from the relatively small number of programs and only programs
located in a single state. Generalization of the results
to other states and regions in the country should be made
with caution since regional and local differences may
overshadow the true differences between program types.
However, the limited differences we found between
graduates of community- and university-based programs is consistent with studies done in other states
and nationally.1,7
The study is also limited by self-report data, and it
does not include data from practice audits or claims
data to confirm practice patterns and performance of
procedures. Even if physicians overestimate their performance of procedures or hospital privileges, however,
little reason is present to suspect a consistent inaccuracy based on residency affiliation. Finally, many other
factors not studied may affect physician practice patterns and the student residency selection.
Conclusions
In summary, the current study found few differences
in type of practice or in the rate of performance of office or hospital procedures based on university affilia-
Residency Education
tion among graduates of South Carolina family practice residency programs. University-trained residents
are somewhat more likely to enter faculty practice,
teaching, or fellowship training, while communitytrained residents are slightly more likely to do rural
practice near where they trained. Other than that, most
residents enter full-time solo or group practice and perform numerous procedures regardless of residency type.
Community- and university-based, opposed and unopposed training programs produce physicians whose
practic e profile s are basica lly indistinguishable .
Whether community-based program training will remain a factor considered by residency applicants remains to be seen.
Vol. 34, No. 8
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
Corresponding Author: Address correspondence to Dr Carek, Medical University of South Carolina, Department of Family Medicine, 9298 Medical
Plaza Drive, Charleston, SC 29406. 843-824-9574. Fax: 843-818-2990.
carekpj@musc.edu.
597
Davidson RC, Kahn NB Jr. Acomparison of university-based and community-based family practice residency programs. J Fam Pract 1984;
18:581-6.
Hueston WJ. A comparison of university- and community-based family practice physician educators. Fam Med 1993;25(9):576-9.
Weiss BD. Family physicians in university hospital intensive care units.
J Fam Pract 1984;17(4):693-6.
Pugno PA, McPherson DS, Schmittling GT, Kahn NB Jr. Results of the
2000 National Resident Matching Program: family practice. Fam Med
2000;32(8):543-50.
Skinner BD, Newton WP. Long-term perspective on family practice
residency Match success: 1984–1998. Fam Med 1999;31(8):559-65.
Halvorsen JG. Family medicine’s failures: reflections on Keystone III.
Fam Med 2001;33(5):390-2.
Tenore JL, Sharp LK, Lipsky MS. A national survey of procedural skill
requirements in family practice residency programs. Fam Med 2001;
33(1):28-38.
Kikano GE, Galazka SS, Flocke SA, Saffran E, Zyzanski SJ. Markers
of successful recruitment of students to family practice residency programs. Fam Med 1994;26(8): 492-6.