This is a study of employment and education history of primary care graduates who graduated from medical school at the University of Nebraska Medical Center in 1978, 1979, and 1980. This group included who entered primary care practice with 46.5% of this group still at their initial location in 1993. Upon graduation from residency 35% entered rural practice (town of less than 25,000) and 46% of the primary care cohort entered rural practice at some point during the 12 years. Only 9 doctors switched from rural to urban while 11 switched from urban to rural. Of the eleven PC Physicians beginning practice in towns of less than 10,000, five stayed throughout the study, five left for towns of over 25,000 and one left for a town of 10-25,000. Those choosing rural practice were more likely to have rural backgrounds, a positive rural preceptorship, state scholarship obligations, a preference for rural practice during medical school and family practice training, particularly in the state and in the Lincoln Family Practice Program. Those staying longest in rural practice ...

The distribution of physicians continues to be a problem, especially for rural and underserved areas(frenzen ref 30 in wami(ref 2,3,33 in wami hsa, nhsc,tarlov).). Additionally, the predictions of excess physicians of the Council of Graduate Medical Education are being realized (COGME). As economic pressures mount in favor of a choice of primary care, the specialty distribution of graduates is improving. Whether this will redistribute physicians geographically is unknown.

Past studies have confirmed the associations of choice of rural practice and rural background (wigton, others), preference for rural lifestyle, the spouse's background and preferences, the availability of group practice, and the location of training (Cooper, Cordes, Grimes, Crawford, Teplin).

There are indications that fewer primary care graduates are choosing rural practice with each passing year. Family practice residency graduates have been by far the largest graduate medical education resource for rural practice, but even in family practice only a minority of programs have rural training missions and curriculum, although the numbers of rural training tracks are increasing (RCB). More troubling is a downward trend in the choice of rural practice by family practice graduates (see table from AAFP data). Other studies reveal the lost potential for rural workforce. Nearly half (44.7%) of the graduates of a rurally-focused family practice residency chose urban locations despite having a preference for a rural location (Hartlaub and Gordon 1993). Although initially disappointing, do these rurally interested graduates represent a pool of physicians who will eventually chose a rural location?

Some controversy surrounds the impact of a rural preceptorship, but studies in Nebraska support such types of training. Wigton Paulman

Few studies examine a cross section of primary care physicians during a period of time. Horner found that rural physicians made up 33% of the primary care physicians who started practices in 1981-1989 in North Carolina (Horner et al 1993). Rural doctors showed an average stay of 4.6 yrs as opposed to 4.4 for their urban counterparts. Half of the 33% who started rural were still in the same county by 1989 and another 10% were in a similar rural county. The authors of this study suggested further studies such as in the midwest where populations were less dense and they suggested programs to encourage retention.

Others have noted that young physicians are a mobile group with 33% changing jobs in first five years (Willke, 1991). In order to attract physicians to underserved areas, Willke suggests looking for public med school graduates that not currently located in their state of training and who are employees in their first three years of practice.

Longitudinal studies could help to answer questions about the career choices of primary care graduates as well as retention rates and factors important in choosing them and preparing them for rural practice. This study of all medical students choosing primary care careers who graduated from the University of Nebraska Medical Center in 1978, 1979, or 1980 attempts to examine these areas using the primary care output of a rural midwestern public medical school.

other intro poss

Only a few studies look beyond the first years to examine the location of physicians over a decade or more. Some studies begin with physicians who are already at a rural location. Parker and Sorensen, 1978; Hassinger, 1980 parker, hassinger results.... review ebb and flow article

METHODS

Data was collected prior to graduation in 1980 from of students, in 1985 on of primary care practitioners in the cohort, and in 1992 on 113 respondents. A longitudinal database maintained at the University of Nebraska Medical Center has tracked these students over time. This database has been updated regularly and has been used in other published research (Wigton). The survey in 1993 used a two page mailed questionnaire asking respondents to complete a chronology of their educational and employment experiences after graduation from medical school. By using this chronology, another 17? who would have been considered primary care were excluded. These included ER, spec, etc.

To determine the rurality of practices, researchers located each community where the respondent lived and coded it with the 1990 census figure. In nearly every rural community, this was not much different than the 1980 figure. If the community was adjacent to a larger metropolitan area, the population of the larger area was used. For example, one physician indicated that he currently practiced in North Kansas City, Missouri (population 4,130), which was clearly part of the Kansas City metroplex. A population of 435,000 was coded for this physician. To be comparable with existing categories of rural (AAFP, WAMI) and with physician perceptions of rurality (Hartlaub 1993), practices in towns with a population less than 25,000 were considered rural.

The data included information on "the location of first practice," whether physicians "ever practiced in a rural area," "the years practiced in a rural area," and "the final practice location." The questionaire included a chronological listing of practice sites and years at that location. Other variables measured included additional educational experiences and the spouse's or former spouse's hometown. The rural background and preferences of the graduates dated from the initial 1980 survey.

The most complete is for the variable "location of first practice" (n=__), while only __ are used in the analyses that include "ever practice in a rural area" and "final practice location."

Data were entered into a mainframe computer and analyzed using SAS. Preliminary descriptive statistics were run including percentages and means. For the analysis of correlated proportions, such as comparisons of first position with final position, McNemar's test was calculated by hand.¹

All 139? were sent a questionnaire with 113 returning on either the first or second mailing, for a response rate of 81%. Among non-respondents, ___ had no forwarding address. (__%) were female and ___ (__%) were male. The average time to a first job change among those who moved at least once was 3.99 years, but a remarkable 46.5% of the 113 primary care respondents stayed at one location for the entire 12 years of the study.

Upon graduation ____chose family practice ( ), ___ chose internal medicine ( ),and ____ chose pediatrics initially for a total primary care percentage of ____. _____ or % entered primary care practice with % of family physicians , % of internists, and % of pediatricians staying in primary care for at least six months. The true primary care percentage for UNMC for these three years was _____.

A total of 33 (35%) entered practice in a small town with 49% of all primary care graduates practiced in a town with a population less than 25,000 at some point during the first twelve years of their medical career. In 1993, 37.5% were practicing in towns of less than 25,000. Those in towns of less than 2500 numbered 16; in towns of 2500-4999, 29; 5000-9999, 58; and in towns of 10000-24999, 82. Only 9 doctors switched from rural to urban while 11 switched from urban to rural (McNemar p < .05). Of the eleven PC Physicians beginning practice in towns of less than 10,000, five stayed throughout the study, five left for towns of over 25,000 and one left for a town of 10-25,000.

During the twelve years the ?? graduates who chose rural locations provided a total of ____ yrs of service to rural communities.

Primary care graduates with a rural background chose rural and stayed in rural locations longer than other primary care graduates. When controlling for specialty choice by examining only family practice graduates, this difference in location and retention _____.

Primary care graduates who had a positive rural preceptorship experience chose rural practices ______ and had ______ retention.

The study also notes the association of these three factors in retaining??? rural physicians.

Spouse's background in rural was strongly associated with "ever practicing in a rural area."

Results leftovers

Graduates from a rural background ??? chose rural????? ( vs ?, stayed in rural areas longer ( vs urbans), stayed? in the state ?? more ( ) and chose FP more often???? ( ) . Family physicians chose rural ( ) and stayed? rural longer ( vs ) . Those who had a positive impact during their rural preceptorship chose rural ???more and stayed ????longer than other graduates who chose rural.

what about those who were positively impacted by rural preceptorship 62% +, 6% neg, 32% neutral

Graduates who had an urban background and did not have a preference for a rural location and were not positively influenced by a rural preceptorship ..... did or did not go to rural locations.

The most important factor in ever chosing a rural location was (preference, background, or response to preceptorship). Does this vary by specialty choice?

Family practice graduates chose .... the most rural locations, followed by im? and peds. Of those who chose a rural location, fp's stayed in a rural location for x years as compared with im stay of x years and peds stay of x years.

FP or IM or peds were most likely to choose a location in nebraska. FP or IM or peds were most likely to make a career change to specialize. The average time until that change after graduation from residency was x years.

One of the difficulties in evaluating interventions in medical education is the long wait from intervention to graduation to beginning practice. For small scale interventions of only a few students a year, more years are needed to allow program development and sufficient outcome numbers to evaluate. Mobility and retention measurements take even longer. Studies of graduates over many years are more difficult, but add important information about retention and the total impact of a rural medical education program.

The use of the initial practice location may be a sufficient initial outcome marker for institutions and policy makers to utilize for further interventions and workforce predictions. As noted in the study, this marker may underestimate the total contribution of a program over many years. The study also reminds communities and faculty that physicians continue to choose rural practice even after many years.

A common belief is that physicians migrate mostly from rural to urban practices over time. This study contests that belief. Physicians did tend to move from smaller rural to larger rural, but the numbers moving from rural to urban were actually less than the opposite.

The choice of rural practice seems to be a steady state with a fairly even exchange between physicians coming to rural locations and those leaving them. This may vary with the rurality of the state and other factors. In a state such as Nebraska with fairly equal urban and rural populations, a balance is likely. In a more urban state, the urban pool would be much larger. Relatively few urbans choosing rural would equilibrate with the fewer rurals moving urban.

2-3 paragraphs after data updated

background, spouse background, obligation, state school, preceptorship, fp program location impact discussed - Ties together suggestions by many separate studies over the years.

impact of gender small sample but more likely to choose rural than females and stay longer ???

impact of preceptorship - does it confirm the rural background and encourage the urban background to consider rural positive impact of rural, are these the ones who choose rural initially, or perhaps later, are these the urban backgrounds who chose rural?

what about retention of diff pc grads im, peds, fp

The loss of primary care physicians to subspecialization is a well known phenomenon. The loss to careers not involving primary care non-practicing specialization or other careers is not as well detailed.

Nebraska continues to produce rural physicians for the smallest locations just as Madison reported back in .......

The study has limitations. One of the difficulties of tracking physicians over many years is the existence of multiple databases and varying types of software. These studies also extend beyond the career lengths of many faculty, particularly those at one institution. The study also includes data on one particular state school. Nebraska is also one of the most rural states with a medical school, having only 1.6 million people mostly crowded in the eastern counties and along the Platte River. There are only 9 cities of over 25,000 and 7 are in the eastern metro counties.

- omaha, lincoln, grand island, bellevue, north platte part of urban - papillion, ralston, la vista, south sioux city

Another limitation is the relevance of a study of medical school students who graduated years before. It is perhaps significant that the late 1970s represented a time when family practice was receiving more credibility in the academic arena just as it is receiving more attention in the current health care scene. Further reflection on the data in this study may help with some difficult decisions that we face at the state and national level.

some of the subcategories have fairly small numbers and a biased response could

Physician mobility in this study was comparable with other studies by Bruce and Horner. In Improving Rural Health, Bruce included 700 family practice, general practice, and internal medicine physicians in practice in Arkansas from 1962-1974. The Bruce study used 16,000 as the cutoff for rural instead of 25,000, but Arkansas at that time had similar numbers of towns of over 16,000 as compared to Nebraska towns of over 25,000.

Better measures of the impact of medical education programs are needed and these include studies following graduates over several years. Research needs to focus on the natural mobility of physicians and the effect this has on the physician workforce

Studies of the physician workforce should involve evaluations of several years of practice rather than just the first location or initial preferences.

Physicians did have a great deal of mobility. The initial practice chosen is a good marker of location. It may be surprising to some that more moved to rural locations than left them over time, but this may be a function of an equilibrium between urban and rural locations with a larger urban pool. This distribution may vary by the rurality of the state. Other factors of selection, training, the state practice environment, and state efforts may also be important.

The location of the first practice site should not be the only outcome considered. The concept of number of years spent in rural locations may be a more useful outcome indicator than just the first practice type and location. It is certainly possible for an intervention to improve retention while diminishing recruitment. Since rural physicians have been found to impact local economies by $380,000 a year and 18 jobs, each year spent in a rural location is very important.

The turnover ratio in a practice may also contribute to the decline in rural market share. Keeping rural physicians at their current locations is very important.

?? Studies on rural physician retention have found .... but this neglects the possible confounding influence of the natural attrition of physicians who tend to move regardless of other factors. If the variables associated in past studies are also correlated with the natural movement of physicians, then the impact of some of these may overestimate the true impact.

State and National Health Policy

As the nation considers an excess of physicians, it is more important than ever to consider the needs of underserved areas, especially small and medium rural locations. Current recommendations to reduce the class size should not be at the expense of those medical schools producing rural physicians. This will be difficult to maintain because these schools are often facing the most difficult state budgetary problems. The expected declines in graduate medical education funding and the challenge of integrating managed care with the missions of academic centers will also make rural preparation difficult.

Even though there are over 700 family practice faculty who have practiced in rural areas, there are few rural faculty and fewer with significant amounts of supported time to create, manage, and direct the medical education efforts that would preserve and increase the rural production (RCB).

It is important to remember Doeksen’s studies of the economic impact of the rural physician in the equation. (doeksen) Each rural physician is worth $380,000 and 18 jobs to the rural community. Loss of rural physicians may also impact the local hospital and other health services. Health care is also a primary determinant of job recruitment and retention. Coupled with the above impact is the loss of dollars due to transportation costs and lost time when rural citizens must access services at more distant locations. Improvements in the selection, training, location, and retention of rural physicians may be worth additional expenditures.

Concentrating on an initial rural practice location may be a worthwhile effort. State and federal scholarship and loan repayment programs may be undervalued. Rosenblatt’s recent article notes the impact of NHSC on subsequent career choices, with several rural family physicians choosing other rural or underserved practices. When compared to NHSC family physicians with family physicians from a rural midwest state medical school regarding recruitment and retention, it would be better to train physicians in the rural state.

Because of primary care physician mobility, researchers must consider more detailed measures for the evaluation of medical education interventions. The location of the first practice site should not be the only outcome considered. The concept of number of years spent in rural locations may be a more useful outcome indicator than just the first practice type and location. It is certainly possible for an intervention to improve retention while diminishing recruitment.

Screening programs to examine rural and service motivations prior to medical school may help improve the production of rural physicians or increase their years in rural practice. Only a handful of medical schools have undertaken this effort and even fewer have enough experience to provide guidance to others considering these options. More rural physicians for less cost would be a good option for the next decade.

In many ways the results of the study are not surprising. Many rural workforce experts have noted that retention may be more important than recruitment. Physicians who stay longer may also encourage their colleagues to stay longer, providing a greater measure of primary care stability to an area. This may well be a key factor in retaining and increasing local market share. Physician stability would certainly help retain local health services and facilities.

The study indicates a benefit for recruiting rural physicians from urban practice sources. If good information sources such as a tracking database are available, it would be possible to recruit those currently in urban locations who are most likely to respond. Currently the recruitment situation favors an urban location as there is less information about rural practice opportunities. Rural communities are also less organized in their recruitment efforts.

Future study

Future study is needed on the factors related to why physicians tend to move early in their careers. Further work could identify factors relating to longer term stayers to examine retention. Interview studies of some of the physicians in the cohort are proceeding in this area. Other studies could identify why some choose rural at a later date.

As managed care and networking spreads across the nation, will certain types of primary care physicians redistribute to rural locations or will they just commute more rural to work? If they do relocate rural, will they stay rural or will they move about more frequently?

As we face increasing problems in many urban locations, will physicians choose more rural and presumably safer locations to raise their families? Other physicians may continue to move from rural areas for better educational or job opportunities for their families.

Cooper, J.K., Heald, K., Samuels, M., and Coleman, S. (1975). Rural or urban practice: Factors influencing the location decision of primary care physicians. Inquiry, 12,18-25.

Cooper, J., K. Heald, M. Samuels, and S. Coleman. "Rural or urban practice: Factors influencing the location decision of primary care physicians," Inquiry. 1978; 12:18-25.

Cordes, S. M. "Opinions of rural physicians about their practice, community medical needs, and rural medical care," Public Health Report. 1978; 362-368.

Crawford, R.L., McCormack, R.C. (1971) Reasons physicians leave primary practice. Journal of Medical Education, 46, 4, 263-268.

Cullison, S, Reid, C, Colwill, J Graduates of the University of Missouri-Columbia School of Medicine 1957-1973. (May, 1974). Missouri Medicine, 214-219

Frenzen PD the increasing supply of physicians in suburban and rural areas, 1975-1988 aam j pub hlth, 1991 81 1141-1147

Grimes, R.M., Lee, J.M., Lefko, L.A., & Hemphill, F.M. (1977). A study of the factors influencing the rural location of health professionals. Journal of Medical Education, 52, 9, 771-772.

Hassinger, E.W., Gill, L.S., Hobbs, D.J., & Hageman, R.L.(1980). Perceptions of rural and metropolitan physicians about rural practice and the rural community, Missiouri, 1975. Public Health Reports, 95, 69-79.

Health Resources Administration Supply and Distribution of physicians and physican Extenders Hyattsville, MD GMENAC Staff Papers Publica Health Service, US Dept of Health Eucation and Welfare 1978 DHEW publication HRA 78-11

Horner, R. D., G. P. Samsa, T. C. Ricketts III. (1993). Preliminary evidence on retention rates of primary care physicians in rural and urban areas," Medical Care, 31, 7,640-648.

Madison, D. L. (1973). Recruiting physicians for rural practice, Health Services Reports, 88, 8, 758-762.

National Health Service Corps Program Unable to Meet the Need for Physicians in Underserved Areas, Washington DC US General Accounting Office, 1990 GAO/HRD-90-128

Parker, R. C., Sorensen, A. A. (1978). The tides of rural physicians: the ebb and flow, or why physicians move out of and into small communities, Medical Care, 16, 2, 152-165.

Pathman, D. E., Konrad, T. R., and Agnew, C. R. (1994) Studying the retention of rural physicians," The Journal of Rural Health, 10, 3, 183-192.

Tarlov, AR The increasing dispersion of specialists, NEJM 1980 303 1058-1059

Teplin, S. (1993), Life and Practice in Underserved Communities, The Robert Wood Johnson Foundation, Practice Sights: State Primary Care Development Strategies, National Program Office Forum, Asheville, NC.

Wigton, R. S. and W. C. Steinmann. "Plans for rural practice of medical students and residents at the University of Nebraska College of Medicine," Nebraska Medical Journal. 1981; April: 77-80.

Figure 1. Data collection matrix to get information on ....

Table 1. Crosstabulation of....

Tables to be added:

Start Job Location x Final Job location

Current Job

Rural Urban

Rural 24 9 33(35%)

Start

Urban 11 50 61(65%)

35(37%) 59(63%)

Start Job Location x Ever worked rural

Ever Rural

No Yes

Rural 0 34 34 (35.8%)

Start

Urban 48 13 61 (64.2%)

48(50.5%) 47(49.5%)

Bruce study with just pc

rural rur/rur rur/urb urb urb/urb urb/rur

FP 217 100 50 95 15 25 GP 20 9 5 8 2 3

IM 14 1 14 93 23 6

total pc 251 110 69 196 40 34

700 36 16 10 28 6 5

Wigton data

rural rur/rur rur/urb urb urb/urb urb/rur

FP

GP

IM

Peds

total pc

change all to percentages and compare as follows in table

rural rur/rur rur/urb

ark NE ark NE ark NE

Changes over time such as rural definition, move to more group and less solo,

The nc average stay noted by was similar to the figures noted in this study.

wami - more fp, the more rural produced, more rural the state the more rural produced

ark - 635 docs over 12 years 1962-1974 in towns of less than 16,000

half left their practice setting with those in smallest leaving more than those in largest,

 

of the 304 that left, 75% did so in the first two years,

 

towns - stayed left for <6000 left for >6000 left state

<1000 42% 11% 20% 27%

1-6k 58 9 12 22

6-16k 63 3 12 22

16000 as cutoff rural, 1700 total docs, used all specialties

41% urban and never moved

19% rural and never moved

13% moved urban to urban

13% moved rural to urban

9% rural to rural

6% urban to rural

with just pc

rural rur/rur rur/urb urb urb/urb urb/rur

FP 217 100 50 95 15 25 GP 20 9 5 8 2 3

IM 14 1 14 93 23 6

251 110 69 196 40 34

700 36 16 10 28 6 5

2/3rds of fp/gp had end point in rural 3/4 initially chose rural

60% no moves, 22% one move, 12 % 2 move, 7% 3+ moves

answer not recruit or retain but continuous, study obstacles and

make the next "desired" step easier

rur 25 11

urb 32 53

rur/urb 14 10

rur rur 13 3

urb rur 6 5

urb urb 7 18

Cullison, 1974; : u of MO 57-73 7% less than 5000 pop, half of docs to towns of less than 5000 from communities of that size previously

What is the effect of gender on the above? literature says females go less rural, do they stay less rural?

Towns >25k

Towns 10-25k

Towns <10K

male female