Family Medicine Residency Programs and the Graduation of Rural Family Physicians

Robert C. Bowman, M.D. and Joan D. Penrod, Ph.D.     July 7, 1997  

See updated regression model 2002 adding Title VII and osteopathic variables

Background and Objectives: Family practice residency programs, the largest source of rural practitioners, graduate about 600 rural physicians each year. Increases in resident positions have not increased the numbers of residents choosing rural practice. This study examines the relationship between the characteristics of the program and the graduation rate of rural family physicians.

Methods: Program directors from 352 (96% response rate) provided data for the study. Weighted least squares regression was used to analyze the relationship between program factors and the percentage of family practice residency graduates in the program that choose practices in towns of less than 25,000 not adjacent to a larger metropolitan area in 1992, 1993, and 1994. survey form example

Results: Programs that graduated more rural physicians were in more rural states and had more required rural training months, a rural mission, required ambulatory clinics in rural locations, fewer residents who were minorities, and more training months in obstetrics. They also had a program director as the primary rural contact.

Conclusions: Rural training and other rural characteristics of family practice training programs are associated with the graduation of more rural family physicians. Interventions to encourage more rural training or a more rural location could result in increased numbers of rural family physicians.

Listwise Deletion of Missing Data


Multiple R .70808

R Square .50138

Adjusted R Square .48221

Standard Error 32.79559

Analysis of Variance:

DF Sum of Squares Mean Square

Regression 13 365552.90 28119.454

Residuals 338 363536.08 1075.551

F = 26.14424 Signif F = .0000

------------------ Variables in the Equation ------------------






Sig T

Number of rural months






% of resgrads who were minority






Program has full rural mission






Per cent of state pop that is rural






Director is the rural contact






Any type of procedural fellowship






Number of other residencies that were present












OB months






Partial rural mission






Not found significant


City population


% rural faculty


Public hospital


Log-likelihood Function =-1469.930450


Family practice residency graduates are the major source of rural physicians, particularly for the smallest towns (aafp).1 From 1985-1987, an average of 685 graduating residents chose rural practice each year. In 1994-1996 only 580 did so.2 The most significant decreases have occurred in the smallest rural category, graduates choosing towns of less than 2500 people not adjacent to a larger city. Recent increases in residents and residencies have not significantly increased the numbers of family practice residents choosing rural practice.

There have been few significant rural medical education studies in recent decades. At the predoctoral level, medical schools in more rural states, public medical schools, and medical schools with a mission for rural health produce more rural physicians.1 Studies have demonstrated that family practice residents tend to choose practice sites near their residency location (Dorner, Denton)3,4 Studies at the state level have also suggested that minority (Lin)5 and female physicians (West)6 do not choose rural locations as often. Hoping to influence the training and career choices of residents, faculty have initiated rural preceptorships (Norris),7 rural moonlighting (Glenn),8 and rural training tracks (Rosenthal),9 and they have emphasized "hands-on" training and fellowships where residents get to do procedures and make decisions on patients (Norris).10

Preceptorships are the most common rural intervention. A national study of family practice programs noted that 135 family practice programs offered rural rotations, but only 84 required them (Blondell).11 A retrospective national study of rural physicians noted that 31.5% had taken a required rural rotation in residency, 48.5% had taken an elective rural residency month, and 59.5% had taken a rural clerkship in medical school (Norris).12 Another study noted that twice as many rural-interested students took rural and international experiences and volunteered for local organizations when compared with their peers in medical school (AAMC GQ).13 Obviously these trainees had directed their own careers because the American Medical Association found in 1992 that only 20% of medical schools required training in rural locations. Rural experiences have been found to influence the location patterns of nurse practitioners and physician assistants, but these trainees are much closer to graduation than medical students (Fowkes).14 No similar regional or national study exists in family medicine.

The impact of preceptorships in medical education is difficult to determine because the studies have limited designs. The studies have been local, advocate-directed, or retrospective. The students select the rural experience and location. There are few prospective, randomized studies in all of medical education (Pathman).15 The main studies regarding preceptorships have limited inquiries regarding past participation. The curriculum, location, and goals of the preceptorship are unknown. Some note that the preceptor may be more influential than the preceptorship (Ackerman).16

The impact of preceptorships also goes beyond the influence on the trainee, although this is rarely examined. Some program directors note that communities working with residents during rural preceptorships often do better with their recruitment. Often rural experiences are implemented to augment training, not just influence career choice. Physicians participating in rural training may stay longer in practice and this may improve physician distribution (Fryer, Stratton).17,18     

Why a Preceptorship Is Better

Most successful rural interventions involve admissions. Rabinowitz demonstrated that special programs could select students with rural background and family practice interest that were likely to choose and stay in rural practice.19 This was still true in 1995 as twice as many rural-interested students as seniors noted that they knew their specialty and practice location prior to medical school when compared to their non-rural peers  (AAMC GQ).13 Unfortunately the numbers of rural-interested seniors and rural background students have declined steadily over the past fifteen years. Some note that there may not be enough rural-interested students to supply the number of rural physicians needed by the country (Kassenbaum).20

Admissions Package

Studies note that initial rural interest may need to be reinforced by rural training experiences (Foley).21 The effectiveness of programs in the final years of training may therefore depend upon earlier efforts. Studies have documented positive impacts of programs intervening at multiple levels and starting in the early clinical years (Boulger, Brazeau, Roberts) 22,23,24, but few medical schools take this approach (Bruce).25 The success of medical school interventions may depend on access to and selection of graduate experiences that accommodate or facilitate rural interest.

The final decision for rural practice is often made outside of the residency process. Some residencies have been able to reverse a trend of graduates to locate in larger cities or outside the state. In Mississippi, administrators and faculty connect the residencies to communities with rural training and facilitate the recruitment efforts of underserved communities. Few states facilitate the entire sequence of admissions, medical school experiences, residency selection, residency training, and the recruitment process.

Rural Curricula, Strategies, Guidelines

Family practice programs face difficulties when they attempt to implement rural training. A national study noted that only 8% of family practice programs utilized Area Health Education Center resources (Blondell).11 Rural hospitals rarely apply for graduate medical education dollars even when they have residents on a regular basis. Funding for rotations often had to come from the communities, hospitals, physicians, or the residents themselves. Financial incentives also come into play as programs lose clinical and graduate medical education dollars when residents are away for rotations.

Barriers to Graduate Medical Education in Rural Areas

Beyond financing, there are other barriers to the development of rural experiences.  Accreditation guidelines are a concern. The Residency Review Committee governing family practice restricts residents to two months away from the continuity ambulatory practice in any year of training. Accreditation restrictions also make it impossible for smaller locations to initiate graduate training. Faculty attempting to introduce new rural rotations into their program often face significant resident resistance regarding separation from family and friends. Continuity ambulatory practices in nearby rural locations require extra coordination and development, not to mention the barriers and risks of travel for trainees and faculty both.

Many programs are not certain that they would benefit from rural training. Family practice residency programs face real and challenging local needs and the longer term needs of the state and nation are far more distant. Studies that demonstrate positive education or workforce impacts could influence programs to develop rural emphasis. This study assesses the relationship between various family practice program characteristics and the programís graduation of rural physicians.

Facilitating More and Better Rural Docs

Rural Faculty Development: Facilitating Town Plus Gown


Data Collection

Researchers obtained information about family residency programs from three sources. First, family practice residency directors were surveyed about program characteristics, including opportunities for rural experiences, practice locations of graduates, numbers of faculty who had been rural physicians, and obstetrical training, Second, data regarding program characteristics was obtained from the 1992 and 1993 Directories of Family Practice Residency Programs published by the American Academy of Family Practice. Finally, data from the U.S. Census was obtained regarding the population of the city and state of the program site.

Measures and Variable Specification

Dependent variable. The dependent variable for the analysis was the percentage of family practice residency program graduates who graduated in 1992, 1993, or 1994 and chose rural practices as their initial location. A rural setting was defined in the program director survey as a town of less than 25,000 people that was not adjacent to a metropolitan area.

Independent variables. The following characteristics of the residency programs were included in the analysis:

The availability of rural training experiences was measured with four separate variables indicating whether the program had (1) required rural block rotations; (2) required rural ambulatory clinics; (3) a rural training track; or (4) a rural fellowship program. The required rural rotation was a continuous variable from 0 - 36 representing the number of months of training that occurred in a rural location. The rural ambulatory clinic, rural track, or rural fellowship variables were dichotomous variables each with a value of 1 if the residency had the rural clinic, track, or fellowship.

The percentage of rural faculty, i.e., number of faculty who had been in rural practice expressed as a percentage of the total M.D. and D.O. faculty, was included as an indirect measure of resident exposure to rural advisors or faculty who could shape the residency experience.

The survey asked directors to list a rural contact person for the program. This resulted in three additional dichotomous variables. One represented programs where the program director was the rural contact, another noted whether a faculty person was the rural contact, and the third was a reference variable for programs without a rural contact listing.

An explicit commitment to rural training was measured with two dichotomous variables indicating whether the program director considered the program to have a full or partial rural mission. In the analysis, programs with no rural mission served as the reference category.

An "other residency" continuous variable with a value of 0 - 6 was included to indicate the number of other types of graduate training programs (internal medicine, etc.) that the family practice residency used for rotations. This variable indicated a potential impact on the ability of the resident to do procedures or have direct responsibility for patients.

The study included the total months of required months of obstetrical training as a continuous variable from 2 to 6 as obtained from the program directors. Programs with an obstetric (OB) fellowship were coded with a 1 if the program offered this fellowship.

Characteristics of the residents themselves were compiled from the program directory. The first continuous variable was the percentage of residents in the program that were minorities and the second noted the percentage of residents who were female.

Two continuous variables were included to measure the rural nature of the program site. Their inclusion in the regression model acts as a control for the known relationship between the program location and tendency of graduates to locate nearby. First, the population of city or metropolitan area of the program was included as a continuous variable; however, if the city was adjacent to a metropolitan area, the population of the metropolitan area was used. The second variable was the percentage of the population of the state living in non-metropolitan areas.

Residency program size was measured by the number of (1) residents in the program in 1994 (continuous) and (2) the number of hospital beds at the program (continuous). Additionally, a dichotomous variable indicated whether the sponsoring hospital was a public institution.

Program type was indicated by dichotomous variables as follows: (1) Community-based hospital; (2) Community-based, university affiliated, (3) Community-based, university administered; (4) University-based. Community-based hospitals (Type 1) served as the reference category with the other three variables included in the final regression model.

Analytic Plan

The conceptual model relates characteristics of family medicine residency programs to the proportion of program graduates practicing in rural communities. Consequently, a weighted least squares regression analysis was used to examine the influence of program characteristics on the proportion of graduates practicing in rural areas. Weighted least squares regression estimates the coefficients in a linear regression when some observations are less precise than others (Weisberg).26 In particular, residency programs with more graduates are given greater weight in determining the regression coefficients relative to programs with fewer graduates.


Survey Response

There were 396 non-military Family Medicine residency programs in the United States in 1994. Each program received mailings and follow-up telephone communications regarding program characteristics and graduate locations. The final database did not include 29 programs that did not yet have graduates. Researchers received complete data from 352 of the 367 remaining programs. The survey response rate was 96%.

Choices in Graduate Programs

Characteristics of the Sample

Table 1. Characteristics of Family Medicine Residency Programs (N = 352)

Program Characteristics

Mean (SD)

Graduates practicing in rural settings

29% (23%)

Female residents in the program

37% (15%)

Minority residents in the program

22% (23%)

Programs with a required rural rotation

14% (35%)

Average months of required rural training

2 (6)

Programs with a rural training track

11% (31%)

Programs with a rural fellowship

5% (21%)

Faculty with rural practice experience

32% (26%)

Program director is rural contact person

Faculty member is rural contact person

No identified rural contact person

32% (47%)

22% (41%)

46% (12%)

Programs with full rural mission

Programs with a partial rural mission

No rural mission

15% (36%)

27% (45%)

58% (19%)

Programs with an OB fellowship

12% (32%)

Average months of OB training

3 (1)

Average population of the city or metro area at program site

584980 (1367402)

Percentage of state population living in rural areas

25% (17%)

Average number of beds at sponsoring hospital(s)

479 (283)

Sponsoring hospital is public

27% (44%)

Average number of residents in the program

19 (10)

Average number of other residency programs at the site

2 (2)

Community-based residency program (Type 1)

3% (18%)

Community-based with academic health center (AHC) affiliation (Type 2)

63% (48%)

Community-based and AHC administered (Type 3)

19% (39%)

AHC-based (Type 4)

15% (36%)


As indicated on Table 1, about 29% of family medicine residents in the responding programs chose rural practice settings during the three year period. Twenty programs graduated over 75% of their residents into rural practices and 49 programs graduated over 50% into rural practice. About 165 programs (48%) had some form of rural training. Family practice residents provided 3742 months of medical services to rural communities in 1994 on required rural rotations alone.

The 63 programs with a rural mission graduated an average of 50.9% of graduates into rural practice while those with no rural mission graduated an average of 21.3 % into rural practice. Programs with only 2 months of obstetrical rotations graduated an average of 23.8 % of graduates into rural practice and programs with 5 or more months of obstetrical training graduated 42.1% into rural practice.

The Effect of Program Characteristics on Production of Rural Graduates

The results of the weighted least squares regression of percentage of rural graduates on characteristics of residency programs are summarized in Table 2. The model accounts for 48% of the variance in the percentage of rural graduates.

Table 2. Weighted Least Squares  Estimates of the Effects of Program Characteristics on Rural Graduates: Unstandardized Coefficients and Significance Levels (n= 352) 



Effects on Proportion of Graduates

in Rural Practices

Female residents in program

B = -10.8

ns, p =.08

Minority residents in program

B = -18.8

t = -4.2, p < 0.0001

Required rural ambulatory clinic

B = 6.7

t = 2.3, p = 0.022

Months of required rural training

B = 0.7

t = 3.6, p = 0.0004

Rural training track in program

B = 5.1

ns, p =.087

Rural fellowship

B = -3.0

ns, p = .5

Faculty with rural experience

B = 0.05

ns, p = .26

Program director is rural contact

B = 6.5

t = 2.7, p = 0.007

Faculty member is rural contact

B = 2.8

ns, p = .26

Full rural mission at program

B = 11.1

t = 3.9, p = 0.0001

Partial rural mission at program

B = 5.3

t = 2.4, p = 0.016

OB fellowship at program

B = 6.0

t = 2.2, p = 0.028

Months of OB training

B = 2.5

t = 2.6, p = 0.011

Population of program site

B = -9.7E-7

ns, p = .16

Ruralness of the state

B = 0.2

t = 2.6, p = 0.009

Hospital beds at sponsoring hospital

B = 0.001

ns, p = .76

Sponsoring hospital is public

B = 3.6

ns, p = .1

Number of program grads 1992-94

B = -0.1

ns, p = .11

Number of other residency programs

B = -1.0

t = -2.0, p = 0.045

Type 2 (affiliated) program

B = 4.7

ns, p = .36

Type 3 (administered) program

B = 5.1

ns, p = .34

Type 4 (academic) program

B = 4.9

ns, p = .38


B = 15.0

t=2.1, p= 0.034

Adjusted R-squared


F = 15.99 Signif F = .0000


The results indicated that programs that require more months of rural training and programs that include rural clinic experiences had a higher percentage of graduates who began practice in rural communities. Programs with an obstetrical fellowship, a full or partial mission for rural health, and more months of obstetrics produced more residents who chose rural communities. Programs in more rural states graduated more rural physicians. Programs who listed the residency director as the rural contact person graduated more rural physicians. On the other hand, programs with more minority residents graduated fewer rural physicians.


Number of required rural months (Number of programs with rural months)













Graduates choosing rural practice








Number of obstetrical months taken

(Number of programs)











Graduates choosing rural





Number of other graduate programs

(Number of programs)









4 - 6


Graduates choosing rural (mean)








The results from this study indicate that Family Medicine Programs can increase the proportion of their graduates who practice in rural communities. First, required rural rotations make a difference. The influence of a few months of required rural training is all the more remarkable, considering the restrictions placed upon these months, the late timing of many rural rotations, and the lack of significant curricular support for rural medical education. It may be too early to see the full impact of rural experiences at the graduate level as many rural interventions are new. Programs with easy access to rural areas can benefit from rural ambulatory clinics. Rural ambulatory clinics involve more effort in scheduling, faculty time, and travel, but they do influence the program and its graduates.

Programs with a commitment to rural health and with a more rural location did produce graduates who practiced in rural areas. It seems likely that programs with a rural mission or location would attract students with a least some interest in rural medicine. However, those programs may additionally influence students without an initial interest through other mechanisms. Program location may be important to state or federal health planners who could intervene with incentive grants or through graduate medical education funding to encourage more rural locations for residency programs.

Which Medical Schools Produce Rural Physicians?

Programs with more emphasis on obstetrics graduated more to rural communities. Studies of individual residents and their actual obstetric volume may be even more revealing. Rural and procedural fellowships often have significant obstetric components. It may be that residents select programs with the potential for obstetrical volume or advanced training. The desire for hands-on training may also be a factor. Another possibility is that the presence of an obstetrical fellowship is a proxy for either the existence of family practice faculty who have advanced obstetric skills or else the presence of a working relationship between family practice and obstetric leaders at the site. Considering the decline in the numbers of rural physicians doing obstetrics and the need for access to prenatal care, it is a significant concern that it is getting harder to find residency programs with an emphasis on obstetrics or extramural rotations in obstetrics. Another controversy is setting a minimum number of required deliveries. In some ways this would increase the obstetric volume of all residents and force changes in many program, but this policy could also result in a decrease in family physicians doing deliveries if the policy results in fewer deliveries for residents who are more interested in obstetrics.

Previous studies have been mixed regarding female or minority trainees and the choice to practice in a rural community. This analysis indicates that as the proportion of residents who are minority increases, the proportion of rural graduates decreases. The American Association of Medical Colleges recently noted that minority residents are more likely to choose underserved locations. The results of this study demonstrate that these are not likely to be rural locations. Gender did not contribute to the final regression model, but a trend is apparent. With increases in both female and minority graduates, it is important to identify interventions that could be implemented so that more would choose rural practice.

Gender and Rural Practice Info and Links

Minorities, Admissions, and Underserved

Nebraska's Rural Family Practice Programs

The study has several limitations. First, the results are descriptive. However, prior studies have not included the number of program characteristics available in this unique data set. Second, as with most descriptive studies, it is not possible to determine whether the rural characteristics resulted in more graduates or whether the programs accommodated the desires of their rural-interested residents by increasing rural training. Nor does the study note whether the rural characteristics influenced a new decision for rural practice or just reinforced past influences at the medical school level. Future studies are needed to explore these issues.

It is encouraging to find program characteristics that could be modified to increase the number of rural physicians. However, the overall trend in family practice graduates is stable or decreasing. Future indicators are not good: the numbers of rural background and rural-interested medical students is near an all time low, many of the new resident positions are located in larger towns, programs created in later years have also tended to graduate fewer rural physicians, and adequate obstetrics volume is a problem in many programs. Finally, many of the rural training interventions are new and their effectiveness at preparing residents for rural practice is still unclear.

Family practice residency programs continue to be the hope for small towns across the nation. Further study of these programs and the development of effective interventions at the program level may help the nation address the chronic maldistribution of physicians.

Hope: Students From the Underserved, For the Underserved

The impact of leadership positions in directing programs is also upheld by this study. See also Leadership Factors in Developing RME with other references.


Bowman RC. Continuing family medicineís unique contribution to rural health. American Family Physician 1996;54:471-483.

Report on Survey of Graduating Family Practice Residents 1975, 1980, and 1985 to 1996. American Academy of Family Physicians, Kansas City, MO.

Dorner FH, Burr RM, Tucker SL. The geographic relationships between physicians residency sites and the locations of their first practices. Academic Medicine 1991;66:540-4.

Denton DR, Cobb JH, Webb WA. Practice locations of Texas family practice residency graduates, 1979-1987. Academic Medicine 1989;64:400-405.

Lin G, Rosenthal TC, Horwitz M. Primary physicians practice location choice: an analysis of factors in New York State. New York Rural Health Research Center, State University of New York, Buffalo NY. July 1994.

West PA, Norris TE, Gore EJ, Baldwin LM, Hart LG. The geographic and temporal patters of residency-trained family physicians: University of Washington Family Practice Residency Network. WAMI Rural Health Research Center, Seattle, Washington. February 1995.

Norris TE, Norris SB. The effect of a rural preceptorship during residency on practice site selection and interest in rural practice. J Fam Pract 1989;27(5) 541-4.

Glenn JK, Hofmeister RW. Rural training settings and practice location decisions. Journal of Family Practice 1981;13;3:377-382.

Rosenthal TC, Maudlin RK, Sitorius M, Florence JA, Markowski G, Cleveland PD, Schneeweiss R. Rural training tracks in four family practice residencies. Academic Medicine 1992;67:685-91.

Norris TE, Acosta DA. A fellowship in rural family medicine; program development and outcomes. Fam Med 1997;29(6):414-20.

Blondell RD, Smith IJ, Byrne ME, Higgins W. Rural health, family practice, and Area Health Education Centers: A national study, Fam Med 1989;21:183-186.

Norris TE, Coombs JB, Carline J. An educational needs assessment of rural family physicians. J Am Board Fam Pract. 1996;9:86-93.

Association of American Medical Colleges. Medical School Graduation Questionnaire; 1980,1983,1985,1990-1995. Washington, DC.

Fowkes VK, Gamel NN, Wilson SR, Garcia RD. Effectiveness of educational strategies preparing physician assistants, nurse practitioners, and certified nurse-midwives for underserved areas. Public Health Reports 1994;109:673-82.

Pathman DE. Medical Education and physician s career choices:Are we taking credit beyond our due? Academic Medicine 1996;71:963-968.

Ackerman RJ, Comeauz RW. Mercer University School of Medicine: a successful approach to primary care medical education. Fam Med 1996;28:395-402.

Fryer GE, Stine C, Krugman RD, Miyoshi TJ. Geographic benefit from decentralized medical education: student and preceptor practice patterns. Journal of Rural Health 1994:10:193-198.

Stratton TD, Geller JM, Ludtke RL, Fickenscher KM. Effects of an expanded medical curriculum on the number of graduates practicing in a rural state. Academic Medicine 1991; 66:101-105.

Rabinowitz HK. Recruitment, retention, and follow-up of graduates of a program to increase the number of family physicians in rural and underserved areas. N Engl J Med 1993;328:934-9.

Kassenbaum DG, Szenas PL. Rural sources of medical students and graduates choice of rural practice. Academic Medicine 1993;68;3:232-6.

Foley AE. A strategy to increase the number of urban family practice resident physicians who enter rural practice. The Journal of Rural Health, 1994;10:119-121.

Boulger JG. Family medicine education and rural health: a response to present and future needs. J Rural Health. 1991;7:105-115.

Brazeau NK, Potts MJ, Hickner JM. The Upper Peninsula Program: a successful model for increasing primary care physicians in rural areas. Fam Med 1990;65;12 Suppl:27-31.

Roberts A, Foster R, Dennis M, Davis L, Wells J, Bodemuller MF, Bailey C. An approach to training and retaining primary care physicians in rural Appalachia. Academic Medicine 1993;68:122-5.

Bruce TA. Physicians for the American homelands. Academic Medicine 1990;65 (12Suppl):S10-14.

Weisberg, S. (1985). Applied Linear Regression. New York: John Wiley and Sons, pp. 81-84.