Influence of places of birth, medical education, and residency training on the eventual practice locations of family physicians: recent experience in Virginia.

To the Editor: Several earlier studies have demonstrated that the state where generalist physicians complete their residency training is an important predictor of eventual practice location. (1,2) The purpose of this study was to analyze the relative contributions of the locations of birth, medical education, and residency training in determining a family physician's eventual practice location. Data were obtained from the American Medical Association Physician Masterfile and the American Academy of Family Physicians files at the Robert Graham Center. (3) The study sample was limited to family physicians who completed their training from 1997 to 2003. Individuals were included if they were born, attended medical school, or completed family medicine residency training in Virginia. Individuals were excluded if any of these 3 locations were unknown or the practice location was a military address. The likelihood of practicing in Virginia was calculated for each of seven possible combinations of birth, medical education, and/or residency training in Virginia.

A total of 1,119 physicians met the inclusion criteria but 313 (28%) were excluded because of missing data and/or military practice addresses. For the remaining 806, the probability of eventual practice in Virginia varied according to previous contacts with Virginia (Table). The lowest likelihood of Virginia practice (6%) occurred among individuals who had been born in Virginia but had attended medical school and completed residency training in another state. The highest likelihood of Virginia practice (82%) occurred among individuals who had both attended medical school and completed residency training in Virginia.

More recent contacts with Virginia had the greatest influence in determining eventual practice location. For those physicians with only one previous contact with Virginia, the likelihood of Virginia practice was only 6% for those born in Virginia and 17% for those attending medical school in Virginia, but 49% for those completing residency training in Virginia. The likelihood of Virginia practice was even higher for a Virginia residency graduate if he/she had also been born or attended medical school, or both, in Virginia (74%, 82%, 81% respectively). Attending medical school in Virginia was the next most important influence. Among those who were born in Virginia but completed residency training in another state, the likelihood of Virginia practice was much higher among those who attended medical school in Virginia compared with those who attended medical school in another state (37% versus 6%). Having been born in Virginia was the least important influence but was associated with a higher likelihood of Virginia practice among those individuals who attended medical school or completed residency training in Virginia.

Careful analysis of factors that influence students' selection of family medicine residency programs in Virginia may be useful in efforts to increase the number of Virginia medical school graduates who eventually practice in Virginia. Creating new family medicine residency programs (or increasing the size of existing programs) would appear to be a logical response to the study findings, but the national match for family medicine residency positions has been increasingly difficult over the past several years. In 2004, for example, fewer than half of the available residency positions in family medicine were filled with graduating U.S. seniors. (4) Selection of family medicine residency applicants who attended medical school in Virginia or were born in Virginia may result in a higher proportion of residency graduates establishing practices in Virginia several years later.

This study has several important limitations. First, these data refer only to family physicians, and more specifically, only to family physicians who were born, attended medical school or completed residency training in Virginia. Whether these data can be extrapolated to others types of physicians or to family physicians from other areas is uncertain. The study sample is relatively large but the numbers in certain subgroups are relatively small, so the estimated likelihood of practice in Virginia is not as precise for these smaller groups. Finally, the possible importance of many other factors (such as having family in the state) in determining practice location could not be assessed because these other factors were not available in the database.

Table. Likelihood of a Virginia practice location for 806 family
physicians according to the locations of their birth, medical school
education and family medicine residency training (a)

                                             Number
Number of   VA birth  VA medical  VA FM      of          % practice
VA factors  location  school      residency  physicians  in VA

            Yes       No          No         139          6%
1           No        Yes         No         190         17%
            No        No          Yes        215         49%
            Yes       Yes         No          67         37%
2           Yes       No          Yes         19         74%
            No        Yes         Yes        102         82%
3           Yes       Yes         Yes         74         81%

(a) VA, Virginia; FM, family medicine.

References

1. Henderson T, Carrie F, Szwarc S. Practice location of physician graduates: do states function as markets? National Conference of State Legislatures: The Physician Workforce. Washington, DC: Institute for Primary Care and Workforce Analysis, 2003.

2. Seifer SD, Vranizan K, Grumbach K. Graduate medical education and physician practice location. JAMA 1995;274:685-691.

3. Bowman, RC. Bowman physician and FP graduate studies 2005. Available at: http://www.unmc.edu/Community/ruralmeded/bowman_fp_grad_2004.htm. Accessed February 25, 2005.

4. American Academy of Family Physicians Match Results and Information. Available at: http://www.aafp.org/match/table01/ Accessed February 25, 2005.

John A. Owen, ED D

Gregory F. Hayden, MD

Center for the Advancement of Generalist Medicine

University of Virginia School of Medicine 

Charlottesville, VA

Robert C. Bowman, MD

Department of Family Medicine

University of Nebraska Medical Center 

Omaha, NE