Robert C. Bowman, M.D.
Introduction
The United States has 20% of the population in rural areas, but only 10% of the nation’s physicians have rural locations.
The nation continues to fail to distribute physicians even though much is known about distribution. States and nations have distributed physicians with
• Medical school admissions involving rural background students, students interested in family medicine, or both.1 Admitting those who go and stay rural and in primary care is important.
• Medical schools or medical education in rural areas or less urban areas. Again stability in rural practice and family medicine are important.2-5 Decentralized health education has been effective for physician assistant, nurse practitioner, and nurse midwife programs.6 Losses in distributional health professions education are likely to result in fewer rural health practitioners. A wider viewpoint of entire states or nations as training classrooms with a focus on the needs of communities is important.7-10
• Improvements in education, health career orientation, and college access for rural origin students. Extending admissions opportunities to earlier levels of education can be effective. Admissions tracks beginning with rural high schools and rural colleges have distributed all rural health professionals, including physicians.11
• Coordinated statewide (or nationwide) efforts involving the entire lifespan of the physician from birth to training to retirement have been effective for Arkansas, North Carolina, West Virginia, Australia, and states in the Midwest and West with superior distributions of education. Multistate cooperatives involving the Northwestern states (WWAMI) and North Dakota have also been successful.
• Health policy in the United States emphasizing primary care and support of physicians outside of major medical centers.12
The same principles have also been successful in distributing physicians to low income and underserved urban areas. Rural areas and underserved urban areas experience the same barriers involving education, income, parents, and “distance” although the distance measures are increasingly social distance rather than geographic distance. Cultural differences are common to rural and to underserved areas and populations.
A study of physician distribution is a study in contrasts in the populations that make up America. The Asian populations are 12 million people and only 4% of the population but Asians are 23% of medical students.13 White, Black, and Hispanic medical students with the same parent income, urban origins, professional parents, and closest proximity to medical schools have the same highest level of admission. These elite medical students all have the lowest distribution outside of major medical centers. Those at the shallow end of admissions in the United States are most different in culture, socioeconomics, and geographic origins. Native American rural born medical students face some of the greatest barriers. Students with various combinations involving older age, lower income origins, and differences in language and culture face the most challenges, but they also have the greatest level of distribution to rural and to underserved areas.14
The great dilemma in distribution is that those most likely to gain admission are the least likely to distribute and represent populations in least need of health care. Those least likely to gain admission are the most likely to distribute and are the most representative of the population of the United States of America.
The greatest physician needs involve primary care and psychiatry for populations most different and distant. Efforts to admit more distant and different physicians have been successful, but major success is impossible without basic changes in distributions of education, child development, economics, and opportunity.15
Studies of various medical school characteristics have outlined the importance of rural mission, rural background admissions, public support, and the role of family medicine.16, 17 Studies of types of medical schools can reveal differences in rural graduation rates. These studies can also help in the development of health policy that could improve distribution of health resources.
The distributional medical education studies begin with comparisons of medical school types.
Methods
The 2005 version of the American Medical Association Masterfile listed 294,456 graduates from all medical school sources for the 2005 version. These physicians enter the Masterfile as medical students, residents, fellows, or practicing physicians. There were 293,794 that had a practice location listed as a zip code. There were 462 that did not have a zip code or that had an invalid 2007 zip code as listed by the United States Postal Service (USPS). The bulk of the graduates were allopathic public and private graduates, followed by international medical school graduates, and finally by osteopathic private and public school graduates. International medical school graduates were divided into the Caribbean medical schools, North American medical school graduates, and Distant International medical school graduates.
Physician practice location was determined using the OfficeMax version of the 2005 Masterfile. The OfficeMax software selects the most likely office location from up to 5 zip code locations available for each physician. The author also developed an extensive zip code database with 43,842 codes current as of January 2007. The two major dimensions in the database involved concentrations of people (rural to urban) and concentrations of physicians (major medical center to underserved).
Urban and Rural
Each zip code was matched to a Rural Urban Commuting Area (RUCA) zip code type. Each type was converted to a rural or urban determination using Categorization A for RUCA 2.0 coding. In Categorization A the zip code locations with over 29% commuting to urban zip codes for work are considered urban focused and were grouped with the urban zip codes. The remaining RUCA codes were rural. Rural codes can be subdivided into large, small, and isolated rural locations. The newer term for large rural practice locations is micropolitan and includes codes in the 4 – 6 range that were not commuting (4.1, 5.1). The codes ranging from 7 to 9 were small rural codes except for 7.1 and 8.1. The remaining codes in the 10 – 10.6 range were isolated rural codes (except for 10.1, commuting). Rural zip codes were estimated for Puerto Rican locations using RUCA criteria, population levels, and 2003 continuum codes.
Major Medical Center and Underserved
Concentrations of physicians can be defined as major medical centers. These can be medical school concentrations or concentrations at a specific location such as a clinic or hospital. Using the active physicians in the entire Masterfile, a major medical center was defined as a zip code where the total number of physicians in professional or alternative professional locations exceeded 74 physicians. Lack of concentration of physicians is more likely in underserved locations. Underserved zip codes were either high poverty or designated codes. High poverty zip codes had 20% or more of the zip code population in poverty. Designated zip codes were whole county primary care shortage areas, zip codes of Community or Migrant Health Centers, or National Health Service Corps sites. Zip codes not military, major medical center, or underserved were considered “served.” These codes can still be in rural or in urban RUCA locations. The underserved category is more suitable for measurements of primary care distribution which is more locally dependent.
Other Locations
The author determined military zip code locations using published internet sources (military listings and USPS listings) and concentrations of military physicians listed in the Masterfile. The military location category superceded all other codes such as RUCA or physician concentration. The physicians leaving the United States for international locations were also considered a separate category. Many of these appear in the Masterfile due to residency or fellowship years and most did medical school in Canada. A final coding consideration involved zip codes with no physicians or those with 75% or more physicians listed in home locations. For the purposes of the basic initial studies, this coding was not implemented. Also practice and home locations overlapped greatly with rural physicians.
Multiple Class Year Studies
The use of multiple class years at a single point in time reflects a workforce much closer to equilibrium. These studies are different than first time practice location studies or career choice studies involving the match. With increasing time after medical school graduation, physicians move steadily toward major medical center locations and subspecialties. This includes all physician and non-physician forms of primary care with the exception of family medicine and generalist pediatrics. This 40% sample of the physician workforce is a fair representation of all physicians as noted in the last row of Table I.18 The differences involve the additional categorization of the military zip codes.
Results
Physicians concentrate in urban areas. Only about 10% were found in rural locations even though 20.5% of the population is associated with rural zip codes. Urban locations with 80.3% of the people claimed 89.6% of physicians. The micropolitan 6.4% share of physicians did not match the 9.8% of the population in micropolitan areas. The 4.3% of the population in small rural areas had only 2.4% of the physicians and the 4.3% in isolated rural areas had 1.1% of the recent physician graduates.
Table I Medical School Type and Rural Distribution for 13 Recent Class Years
|
|
Urban |
Micro-politan |
Small Rural |
Isolated Rural |
All Three Rural |
Inter-national |
Military |
Type Totals |
|
Allopathic Private |
69516 |
3251 |
1076 |
499 |
4826 |
97 |
769 |
75208 |
|
92.4% |
4.3% |
1.4% |
0.7% |
6.4% |
0.1% |
1.0% |
100.0% |
|
|
Allopathic Public |
111848 |
9717 |
3475 |
1432 |
14624 |
88 |
1643 |
128203 |
|
87.2% |
7.6% |
2.7% |
1.1% |
11.4% |
0.1% |
1.3% |
100.0% |
|
|
Osteopathic Private |
13450 |
1511 |
698 |
310 |
2519 |
5 |
210 |
16184 |
|
83.1% |
9.3% |
4.3% |
1.9% |
15.6% |
0.0% |
1.3% |
100.0% |
|
|
Osteopathic Public |
4918 |
795 |
414 |
139 |
1348 |
3 |
62 |
6331 |
|
77.7% |
12.6% |
6.5% |
2.2% |
21.3% |
0.0% |
1.0% |
100.0% |
|
|
North American |
7370 |
473 |
214 |
110 |
797 |
1500 |
5 |
9672 |
|
76.2% |
4.9% |
2.2% |
1.1% |
8.2% |
15.5% |
0.1% |
100.0% |
|
|
Distant International |
48489 |
2645 |
1147 |
556 |
4348 |
225 |
43 |
53105 |
|
91.3% |
5.0% |
2.2% |
1.0% |
8.2% |
0.4% |
0.1% |
100.0% |
|
|
Caribbean |
4561 |
317 |
132 |
62 |
511 |
7 |
12 |
5091 |
|
89.6% |
6.2% |
2.6% |
1.2% |
10.0% |
0.1% |
0.2% |
100.0% |
|
|
Totals |
260152 |
18709 |
7156 |
3108 |
28973 |
1925 |
2744 |
293794 |
|
Per Class Year |
20012 |
1439 |
550 |
239 |
2229 |
148 |
211 |
22600 |
|
Physician % |
88.5% |
6.4% |
2.4% |
1.1% |
9.9% |
0.7% |
0.9% |
100.0% |
|
Population % |
80.3% |
9.8% |
5.4% |
4.3% |
19.5% |
|
0.1% |
100.0% |
|
Pop (Millions) |
226.3 |
27.8 |
15.3 |
12.1 |
55.1 |
|
0.4 |
281.8 |
|
All Physicians18 |
89.0% |
7.3% |
2.6% |
1.1% |
|
|
|
|
Distribution at or above the average physician % for each location is shown in the shaded cells.
Allopathic public, osteopathic, and Caribbean schools have graduated a greater percentage of physicians that remain in rural practice. Allopathic private and international medical schools without significant levels of American graduates have concentrated physicians in urban locations. The American born graduates of Ross, American University, St. George’s, Spartan, SABA, and Grace began with 11% rural distribution for urban born and extended from 26 – 46% for the various rural origin types. The rural contributions outside of American born graduates were no different than other international schools.
Distribution can also be compared by percentages across medical school type of physicians in a type of school compared to percentages found in various locations. The most dramatic change in this comparison involves North American schools with 3.3% of graduates listed in the Masterfile but this group is responsible for 77.9% of those leaving the United States for other nations. These are mostly Canadian medical school graduates going back to Canada after one or more years of residency training.
|
|
Urban |
Micro-politan |
Small Rural |
Isolated Rural |
All Three Rural |
Inter-national |
Military |
Total |
|
Allopathic Private |
26.7% |
17.4% |
15.0% |
16.1% |
16.7% |
5.0% |
28.0% |
25.6% |
|
Allopathic Public |
43.0% |
51.9% |
48.6% |
46.1% |
50.5% |
4.6% |
59.9% |
43.6% |
|
Osteopathic Private |
5.2% |
8.1% |
9.8% |
10.0% |
8.7% |
0.3% |
7.7% |
5.5% |
|
Osteopathic Public |
1.9% |
4.2% |
5.8% |
4.5% |
4.7% |
0.2% |
2.3% |
2.2% |
|
North American |
2.8% |
2.5% |
3.0% |
3.5% |
2.8% |
77.9% |
0.2% |
3.3% |
|
Distant International |
18.6% |
14.1% |
16.0% |
17.9% |
15.0% |
11.7% |
1.6% |
18.1% |
|
Caribbean |
1.8% |
1.7% |
1.8% |
2.0% |
1.8% |
0.4% |
0.4% |
1.7% |
|
|
100.0% |
100.0% |
100.0% |
100.0% |
100.0% |
100.0% |
100.0% |
100.0% |
Distribution at or above the specific average for the medical school type (right column) is shown in the shaded cells.
Distributional Medical Schools: The Lost Lesson of Specific Forms of Government Support
Allopathic public, osteopathic, and Caribbean schools have graduated increased concentrations of rural physicians. Allopathic public schools have been responsible for half of the rural physicians found in all three types of rural locations. Osteopathic medical schools contributed nearly twice the concentration of rural physicians compared to their physician composition.
Discussion
Despite a number of interventions spanning the past 40 years, the nation struggles with the distribution of physicians to rural areas. Despite successful distribution by nations, states, medical schools, and programs, distribution seems to remain a puzzle. One reason for the confusion may be failure to understand the principles that help determine distribution. Although it makes common sense that physicians for rural areas have rural origins and physicians for underserved locations have underserved origins, this is not what guides medical education, admissions, or the policies that shape medical classes from birth to admissions.
The usual perspective regarding career and location choice is a major medical center perspective. Researchers, including physician workforce researchers, are intimately connected with major medical centers and view distribution with a major medical center lens. Research is also concentrated in elite medical schools that are very different from the distributional types of medical schools. These are the medical schools that have the lowest levels of distribution. The major medical center perspective appears to be a limited perspective. Career and location decisions are primarily made by individual medical students, not medical schools. However for research to relate to the existing literature, initial studies can begin by taking this perspective. Eventually the studies can and should proceed to individual students to expand understanding beyond school type to the factors more directly involved in career and location decisions.
The medical schools with the most public support graduate the most physicians for the country, especially the lower and middle income and rural parts of the nation. The medical schools that admit the broadest range of students, admit the most medical students born in the United States, and graduate the most family physicians also distribute the most rural physicians.
Basic categories fail to capture the full range of distribution. Major medical centers and developing rural and urban centers concentrate physicians and attract patients from all other locations. Physicians are not as likely to be found in more rural locations. The 6 million people in rural counties that are predominantly Black, Hispanic, or Native American present the greatest challenges of all. The nation’s underserved and rural areas are also the locations most dependent upon Medicare, Medicaid, and specialized grant programs. They have the most primary care physicians that have the lowest levels of reimbursement. They are also the locations with the lowest levels of income, education, economics, political influence, and social organization.
The nation has health policy involving clinical reimbursement, corporate investment, graduate medical education funding, and research support that favors medical schools with the lowest levels of physician distribution. So far there has been little interest in restructuring health policies to improve areas such as physician distribution or teaching.
More physicians from international medical schools will also not help with rural distribution. Any temporary gains are lost when workforce is measured over a number of class years. Only the American graduates of the Caribbean forms of medical schools contribute to improved rural distribution. The origins of international graduates reveal that 80 – 90% were born in cities with medical schools, the largest and most urban locations. Consistent rural distribution has not been found for physician groups born and raised in the most urban locations, regardless of birth nation.
The nation will have more and more difficulty with rural physician distribution with
• Admissions of fewer rural origin medical students
• Graduation of fewer family physicians (doubles distribution)
• Health policy that favors major medical center careers and locations
• Lack of state health policy support for public schools
• Education and admissions policies that eliminate the most distributional types of medical students or force them to pay the most expensive tuition, delay their choice of rural location with military obligations, or force them to leave the nation to become physicians.
Simple comparisons such as public and private made more sense when public schools had significant levels of public funding. In recent years states have been less and less willing to support medical education. The federal government also supports medical education, but again these efforts have grown more and more limited. Changes in Medicare and Medicaid also present increasing difficulties for medical schools. Those most dependent upon these components are facing the greatest challenges. These include newer schools and even established schools cut off from insurance and paying patients such as Galveston. Primary care, teaching, and administrative efforts such as selecting the best medical students tend to lose out in a hierarchy restructured by research funding and specialized hospital and physician services.
The medical education of the nation has not always been so dependent upon market forces. Expanding medical schools during the height of federal and state support was an important contributor to increases in primary care graduation rates. A doubling of primary care choice and a doubling of medical school positions quadrupled primary care workforce output. These efforts defined the 1970s in medical education, including maximal growth of Medicare, Medicaid, medical education funding, primary care, and family medicine. The new medical schools created during this era continue to graduate more primary care, family medicine, rural, and underserved physicians. A similar structured health policy environment existed in the 1990s with emphasis redirected away from hospital-based careers and toward primary care and locations outside of major medical centers. It is uncertain whether the current expansion of medical schools during the current “market forces health policy era” will address needs for primary care, family medicine, rural areas, or underserved locations. It is possible that expansion may only keep pace with declines in current primary care workforce. This is similar to physician assistants and nurse practitioners where health policy diverts greater percentages away from primary care and practice outside of major medical centers.
Understanding more about the types of medical schools that distribute physicians may help guide federal and state support to schools that are most able to distribute physicians.
Distributional Medical Schools: The Lost Lesson of Specific Forms of Government Support
Physician Careers By MCAT Groups
Family Medicine Central: National Comparisons of Workforce
1. Rabinowitz HK, Diamond JJ, Markham FW, Rabinowitz C. Long-term retention of graduates from a program to increase the supply of rural family physicians. Acad Med. Aug 2005;80(8):728-732.
2. Ackermann RJ, Comeau RW. Mercer University School of Medicine: a successful approach to primary care medical education. Fam Med. Jun 1996;28(6):395-402.
3. Boulger JG. Family medicine education and rural health: a response to present and future needs. J Rural Health. Spring 1991;7(2):105-115.
4. Boulger JG. Minnesota bound. Stability of practice location among UMD family physicians in Minnesota. Minn Med. Feb 2000;83(2):48-50.
5. Easterbrook M, Godwin M, Wilson R, et al. Rural background and clinical rural rotations during medical training: effect on practice location. Cmaj. Apr 20 1999;160(8):1159-1163.
6. 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 Rep. Sep-Oct 1994;109(5):673-682.
7. Bowman RC, Crouse BJ. Community-driven medical education: the rural component. J Rural Health. Summer 2003;19(3):214-217.
8. Verby JE. The Minnesota Rural Physician Redistribution Plan, 1971 to 1976. Jama. Aug 29 1977;238(9):960-964.
9. Verby JE, Jr. Physician redistribution in Minnesota and the 84,000 square mile classroom. Minn Med. Oct 1985;68(10):757-758.
10. Verby JE, Newell JP, Andresen SA, Swentko WM. Changing the medical school curriculum to improve patient access to primary care. Jama. Jul 3 1991;266(1):110-113.
11. Jokela R. Rural Health Opportunties Program. http://webmedia.unmc.edu/rhen//2004RHOPbrochure.pdf .
12. Bowman RC. Five Periods of Health Policy and Physician Career Choice; 2006.
Five Periods of Health Policy and Physician Career Choice
13. Association of American Medical Colleges. Matriculants and Graduates 1993-2004. 2005.
14. Bowman RC. The Distributional Student, the Distributional School, and the Distributional Specialty. pending acceptance to Rural and Remote Health. 2007.
15. Nickens HW, Ready TP, Petersdorf RG. Project 3000 by 2000. Racial and ethnic diversity in U.S. medical schools. N Engl J Med. Aug 18 1994;331(7):472-476.
16. Rosenblatt RA, Whitcomb ME, Cullen TJ, Lishner DM, Hart LG. Which medical schools produce rural physicians? Jama. Sep 23-30 1992;268(12):1559-1565.
17. Rourke J, Dewar D, Harris K, et al. Strategies to increase the enrollment of students of rural origin in medical school: recommendations from the Society of Rural Physicians of Canada. Cmaj. Jan 4 2005;172(1):62-65.
18. Hart G. RUCAs and Physician Location, AAMC Presentation, May 2006. http://www.aamc.org/workforce/pwrc06/hart.pdf .
For info on the databases, sharing research, or ideas - [email protected]