Robert C. Bowman, M.D.
Ultimately the distribution of physicians involves improvements in the distribution of education, broader admissions of the students most like the populations in need of physicians, improved specificity of training for family medicine, improved specificity of training for underserved areas, and better health policy to distribute physicians and health resources.
The current time period involves very few of these areas although the past health policy efforts and various models continue to reveal that distribution is possible.
1. Background on Distribution
Centralization, Urbanization, High Standards
Education and Distribution
Admissions Forces: Academic and Accreditation Rewards and Punishments
Managed Care and Distribution
2. Theory of Career Choice and Distribution
For 100 years or more the nation has experienced the consequences of maldistribution of physicians. Geographic, education, culture, and income barriers have impeded health care access for significant components of the population. The forces that have helped create maldistribution include the centralization of medical education, the urbanization of the nation, and the increased demand for higher standards for physicians.
The nation has finally demonstrated that it can resolve maldistribution with better distribution of education, improved admissions, rigorous training, and national accountability in the form of career choices of physicians. The nation finally knows what to do from workforce research. It has documented successes with a variety of distributional models. It has even tasted distributional success with during the brief era of managed care. These efforts remain largely ignored and unreplicated. The nation also lacks the willingness to make small sacrifices so that others may have basic health care and has failed to develop the leadership that would accomplish this goal.
The primary task of distribution is the prioritization of breadth over depth. Depth involves programs that tend to concentrate resources around the "gifted" and those likely to go to college as compared to earlier efforts that result in higher high school graduation rates. The primary benefits of breadth are better long term outcomes regarding education and steady declines (relative to states or nations not investing broadly) in health care costs, prison costs, social program costs, and a broad number of areas. Bright Future Rankings
The breadth vs depth argument works in many systems, must dramatically in the function of the internet and in public health. For example, morbidity and mortality and cost outcomes are better for a large population when all blood pressures drop a small amount than concentrating on the fewer who have greater degrees of hypertension.
A major problem in health care is concentrating on those with depth of income while those at most risk of health problems lack the breadth of coverage. Concentrating on depth in education and medical school admissions improves the MCAT scores of medical students but insures that the physician leaders have less and less breadth of background and are less likely to distribute as physicians and also may be less and less able to relate well to their patients who are less and less like them with each passing year.
Choice of Family Medicine Regression
Centralization and urbanization are very similar. Urbanization is largely out of the hands of educators. Centralization does involve education and leadership. Medical education has continually chosen to concentrate resources in a relatively few highly urban areas of the nation. Over 96 % of medical education and graduate medical education occurs in the most urban areas of the nation. What schools and programs exist outside of urban areas are small in number of graduates.
The primary impact of centralization of college, medical education, and graduate medical education is to force those most likely to distribute through a number of adaptations and changes with every step likely to influence a "centralization" of location choice. The impact upon rural vs urban areas is even greater. Centralization represents a health services, economic, education, and leadership inequity for those in rural areas. In contrast, the few medical schools in the smallest locations have made significant contributions in all of these areas and more. Even in towns of 400,000 an established medical school is often the largest employer. In a town 10 times smaller the impact is awesome. The counties that have benefited from the few distributed medical schools have some of the highest ratios of medical students admitted compared to other counties across the nation. These benefits are not limited to medical school admissions. In contrast, urban medical schools make less impact compared to the greater economic and educational resources already present.
One other concept to include involves the Major Medical Center designation. Zip codes with over 75 total physicians or those that are associated with medical schools can be designated as Major Medical Centers (MMC). MMC locations claim 71% of US physicians, 75% of specialists, 70% in primary care internal medicine and pediatrics, but only 43% of family physicians. Family physicians concentrate outside of MMC locations while all others concentrate within MMC over time. The impact of Family medicine is important in physician distribution not only with first practice, but subsequent choices.
Family Medicine Contributes Much More
Decentralization efforts have been successful strategies for health professions education. These include short and long term training experiences in rural areas, training programs in rural areas, and hybrid training shared between urban and rural areas such as rural training track programs. Some schools such as those in Iowa, Nebraska, Arkansas, North Carolina, and Washington state have managed to compensate for centralization by adopting statewide or regional efforts involving admissions, curricula, family medicine training, and practice support for distributional physicians.
More at Centralization and Regionalization
The concentration of people, economics, and education in urban areas has presented a unique challenge regarding distribution of physicians. Students from all over the nation pour into prestigious urban colleges. Over 50 % of allopathic medical students are born in counties of over 1 million. Another 10 % are foreign born and over 60 % of these students were raised in these most urban counties. These urban and foreign born students are the least likely to choose family medicine and rural practice. There are 50 medical schools that have over 75 % of graduates who were born or raised in these most urban origins. Over 10% of medical students have rural origins despite 23 % of the population in rural areas.
Urbanization in the United States increasingly involves insulation and isolation, especially from those in most need. This does not appear to be compatible with an American way of life that has valued small groups and involvement. Urban attitudes can pop out in unusual ways such as when medical education leaders observe that rural areas lack physicians in part do to a “paucity of culture” and a lack of recreational opportunities.
Medical education should resist the forces of urbanization and should redress the sometimes less than tolerant attitudes that may impact students who might choose distributional careers.
Urbanization continues for suburbs and exurbs, but populations are leaving the core counties at high levels. The higher cost of living and especially housing appear to be major contributors.
Inequities in income and education are key issues related to urbanization and centralization and involve income, education, wealth, political power, professional and managerial types, and the decisions made at all levels. In states with significant retired, professional, and highest income types that do not realize the value of government investment in infrastructure, there are likely to be significant problems with distribution, especially in education and health and state cost and efficiency.
One hundred years ago the education of physicians was generally poor quality, but physicians had much better distribution in terms of geography, gender, income, and race. Poor education and preparation of students was the major problem. Reforms were needed but also closed schools that were supplying rural, female, and black physicians. Poor physician distribution was inevitable as medical students grew ever more elite and urban.
For many decades medical education had to work consistently to raise the bar in academic standards. With a consistent level of expectation, parents, high schools, colleges, and advisors had no choice but to greatly improve preparation. This was a great benefit not only to medicine, but to leadership development and higher education in the United States. The Medical College Admissions Test was the most important contributor regarding raising the standards and assessing readiness for medical school. Attrition rates have not been a significant problem for decades. History of the MCAT
If all US education and college backgrounds were similar, the work of admissions would be simple. If the highest scoring students were the most likely to distribute, it would be the best of all worlds. However educational backgrounds vary greatly and those most likely to perform well on speeded intellect tests are the most urban and highest income students, the least likely to distribute. Schools with more high scoring students do not distribute physicians. The assessment of students other than those with high scores and typical education and college backgrounds is far more complex. The MCAT fails most as an evaluation tool for these students. Additional assessment takes much time and effort. It also requires years of data collection on outcomes in education and distribution. Only one medical school, Jefferson, has consistently matched up data from before, during, and after training. The Jefferson Longitudinal Studies constitute a whole body of critical information regarding students of different ethnicity, speeded intellect, education and income background, and career choice. Clearly those students likely to distribute well have lower scores, lower income levels, and higher choice of family medicine. This effect of birth origin students at the national level is to confirm these direct studies. Students who are older, born instate, and born in less urban locations all represent lower income students who choose family medicine, rural family medicine, and rural practice in greater percentages. It is likely that they also choose underserved practices in urban areas to a greater degree even beyond tuition obligations.
To reframe these observations and the theory behind them:
Success in physician distribution involves primarily two areas, characteristics of the students prior to medical school and career choice impacts outside of medical school. The impact during medical school is much more difficult to engineer. Exceptional impacts can make a difference in career choice, but not the routine experiences and exposures in today’s allopathic medical schools. Schools, groups of students, and students with similar characteristics behave much the same regarding distribution. These students/schools/groups are also influenced by state and national events and policies. This baseline for career choice in students involves characteristics and upbringing, for schools it is about the admissions of certain types of students.
One way to visualize this would be a along a socioeconomic scale. Those at the highest end of preparation, education, and parent influences (income, education, degrees, expectations) tend to choose FP rarely, down to 2.2 % levels at the lowest group yet. Those at the other end of education and income who have a bit less speeded intellect (likely a more global distribution of intellect rather than inferior) tend to choose FP and distribution, up to 50% choice of family medicine in certain groups. Exceptional experiences and state and national events move choice of FP up and down this socioeconomic scale.
Some medical schools have chopped off the socioeconomic lower end (elite eastern private) with very little potential for distribution. Others have fewer at the high end of speeded intellect and income (osteopathic, some public schools). When looking at the percentage of students that are the most urban, the most foreign born, over 20 medical schools have 85 % or more of the students who are FP unlikely admissions year after year. Half of all US medical schools have 70 % of more of their classes that are FP unlikely. Those admitted to medical school are already not a bell shaped curve in income and intellect, the curve of parent/student income levels is skewed to the right in favor of higher income and gaining higher income admissions with each passing year.
Distributional medical schools have classes of markedly different composition with as low as 22 % FP unlikely. There percentages of older students are higher. Their school MCAT averages are lower. Their percentages of FP choice are higher, but their numbers and class sizes are low in the United States. This is a poor reflection of support of those most likely to embrace distribution as a mission.
Some medical schools have a higher average MCAT compared to similar schools with the same distributional markers. This includes states such as Iowa, Minnesota, and the WWAMI states. These states have clearly invested in education and also have students that are more prepared and committed to education. This can be measured by teaching efficiency and similar measurements. These states graduate not only higher numbers of distributional physicians, they are also the birth origins of more physician researchers. With the exception of these states the picture of US education does not look good from the perspective of admissions by state origins. States with stagnant or declining admissions compared to other states have lower educational investments and lower teaching efficiency scores with a correlation in the 0.3 to 0.4. States graduating fewer family physicians also have poorer distribution of educational resources (Funding Gap 2004). These studies measure differences between school district funding with rich compared to poor and high minority compared to low. One interpretation of this would be that any state can prepare high income or high intellect students for admission to medical school. However states that do not attend to the lower income students will not be able to meet distributional needs regarding medical professionals and probably other professionals as well. Even though family physicians are important, teaching professionals are more critical to the needs of the nation. Research documenting a link between poor distribution of the nation’s education resources and lack of teachers would be perhaps the most important education research of the current time period.
Students who could become family physicians will begin to melt away perhaps as early as elementary school or even in parent failures to read to their children from age 3 – 5, again a variable that involves lower income levels to a much greater degree.
Distributional medical schools also include the traditional black medical schools, who do well in choice of family medicine and also in providing graduates to serve underserved urban populations, again a confirmation of lower income and disadvantaged origins and connections to underserved outcomes.
Society rewards and punishes medical school in ways that are almost certain to distribute physicians poorly without active intervention. The danger in admissions at the low end of the socioeconomic scale for distribution is attrition. Medical school failure rates are vigilantly monitored. They are a constant concern of faculty meetings, task forces, school reputation, and accreditation visits. Beyond the apparent external concerns, there are internal risks. When admissions committee members understand what is involved in distribution, they are more willing to “take chances” admitting students who are likely to make a difference in the future. When the student has delays or failure, institutions seem to have long memories and personal reputations of such faculty can suffer.
In more recent years, the legal risks of admission have intensified. Now there is a lawyer in every delivery room and in every deliberation regarding the students admitted. These legal pressures are based on statistical analyses of MCAT scores without regard to what it takes to become a physician in a more realistic sense. Use of the MCAT and GPA alone would eventually make Stepford Children into physicians. This would destroy patient communication and relationships and medical quality. Distribution would have been long gone prior to this time.
Closely related is the decreased faculty time in medical education. Poor finance of medical education, increasing reliance upon research and clinical revenues have led to less and less time for committees, teaching, and other efforts.
Admissions is by far the most time consuming effort. Admissions committee members spend long hours and receive little academic recognition or reward. If numbers go south, they can be recognized in not so flattering ways. This has changed admissions progressively over the past decades in a way that has been poorly recognized. At a time when we need to be able to distribute physicians better there is less time to evaluate students. The distributional students also take far more time to evaluate, given their diverse origins.
As mentioned previously, the danger of admissions at the high end is lack of people skills and people orientation. This goes to the core of what most people in the US think of when they contemplate what a doctor is. They want and desire a doctor who is able to relate to people and care for them. The direction of admissions is a great concern. We are admitting a more and more narrow and elite group that may not be able to relate to people across the US who have an incredible variety of walks of life and incomes and cultures and geographic areas. This is not well studied or recognized yet. This is also a very sensitive area that most medical educators hope would go away. It will not.
It is my feeling that failure to raise the people skills bar in admissions to the same degree as the intellect bar is contributing to cost, quality, and access concerns in the nation. There are indeed studies that relate cost and quality to higher ratios of family physicians. The access and distribution figures for family medicine are unimpeachable. Moving toward higher MCAT scores and earlier admissions is not a good idea. Having some medical schools change and not others is also not a good idea when the nation funds research and specialism so much more than distribution. The only way to change Stepford Children is to change the way that they were influenced throughout their early years. It is too late to influence them toward people during medical school or after. This has to occur at all medical schools and US medical education has to stay the course of public outcry, even as legal pressures mount and outcry grows to a deafening roar. We cannot cave in like we did as a nation regarding managed care.
This managed care example may help illustrate the points. Managed care changed the pattern of medical student choice in some significant ways. Before and after the managed care effort involving the classes graduating 1995 - 1999, the choice of family medicine by birth origins was stable across 25 years. During managed care, all groups from urban to most rural had increases in choice of family medicine. However the largest groups of students, the most urban (50 % of students admitted), had the greatest increase in family medicine and primary care choice during the accountability years of managed care. Urban and foreign born students (11% now) had a 50 % increase in family medicine choice compared to 20 - 30 % increase for older and less urban students. Instate urban born had a particularly marked increase in choice of family medicine.Managed Care Comparison Table
During managed care "matches," medical schools with more urban instate born students gained ground in FP ratings as compared to medical schools with fewer urban and foreign born. The effect since managed care has been a reversal or worse as the nation experienced
1. rebound against managed care and those connected with it (FP)
2. marked changes in physician job markets due to decreased choice (anesthesia, radiology) and also the realization that limitations on physician specialty choice were over
3. increasing reports highlighting the problems facing front line patient care physicians, particularly paperwork, massive overhead increases, and cuts in reimbursement that had previously been spared those that were making distributional career choices.
Managed Care and Choice of FP - increasing in FP choice with accountability
A. Socioeconomic theory The choice of FP and likely other front line specialties like psych is made more probable by lower income related indicators
1. Older age at graduation- a particularly good magnifier of other distributional tendencies and may indicate comfort with direct patient care, with increases in % choosing family medicine and psychiatry with older age at graduation
2. Education barriers and distribution - studies across the nation at all levels of education, particularly those involving distribution of scholarships and the impact of standardized testing
http://www.ashe.ws/fellowship/summary_Farrell.pdf more education research and fellow presentations regarding income, ethnicity, funding distribution, outcomes, merit scholarships http://www.unmc.edu/Community/ruralmeded/understanding_higher_education.htm
3. Lower income - Students of lower income choose family medicine in greater percentages
|
Choice of Family Medicine |
Urban Influence Code |
RUCA |
Cooter and Jefferson Longitudinal |
MCAT Averages By School |
Asian Ethnicity |
|
Urban/Highest Income |
Code 1 13.3 % |
Urban/Urban-focused 10 - 14 % |
Highest Quartile 13 % |
Top 30 7 % |
Indian-Pakistani 2.3 % |
|
Other/Lowest Income |
Code 2-9 18-28 % |
Large Rural 20% to Isolated Rural 28 % |
Lowest Quartile 22 % |
Bottom 30 17 % * |
Vietnamese 24 % |
* Excluded 15 lowest MCAT schools and osteopathic.
See also Birth Origins and Distribution Tables for other proxies of income such as older, rural born, instate born and other outcomes beyond just family medicine
Jefferson longitudinal studies, Cooter publication, birth origins, admissions and FP choice by ethnicity also related to income levels
4. Instate born and less urban student origins also indicate lower income origins - the poor, the rural, older students, and those scoring lower on the MCAT have fewer options outside of instate medical schools.
County Birth Origins, Incomes, Choice of FP Probability of admission tables Admissions Ratio By Birth Origin
Forces away from distribution
The choice away from family medicine and toward specialty care is made more probable by younger age, higher income, higher MCAT (in student or school, a reflection of income and education and parents). This area needs great investigation. Privileged origins may mean a graduate who is not as comfortable with people. Emphasis on academics and test taking and higher income upbringing and private schools and top colleges may interfere with people development skills. Who needs them if test taking is the predominant mode for college and medical school? Non-academics are merely a distraction..
B. Orientation and Competence Theories - primary care is complex and takes many months if not years to understand. Just like you need dozens of deliveries of babies to live thru enough complications to gain confidence and experience the positives, you need at least 6 months of intense positive primary care (studies by Verby in RPAP in MN). After 3 months the RPAP students were overwhelmed with primary care (most med school exposures here), at 6 months the RPAP students were neutral to primary care, at 9 months they did not want to return home Many RPAP students chafe at the restrictions in the 4th year back in urban MN as they finish training.
The environment of the United States is growing increasingly unfriendly toward continuity medical care. Students in the US now compared to years ago have a lower and lower probability of experiencing such care. Even with the same continuity provider and patient, the quality has diminished (in the Future of Primary Care). Declines in continuity physicians in relation to the growth of the US population and other physicians will only worsen this situation. It is unlikely that medical schools will compensate by more orientation and more curricular time in primary care, in fact we may be going the other way.
1. Exposures to continuity care during youth- no studies yet on this but patterns are clear that most urban, highest income, most foreign born least likely to have a continuity physician - suburbans and rurals far more likely to have one
2. Exceptional efforts - significant orientation and intensity, hands on experiences - significant role models and role model experiences can have impact. This usually takes more than a few months and is not likely to be a factor in limited exposures.
RPAP - rural preceptorship for 9 months, was attempting an inner city replication recently
Accelerated FP - entire M-4 year family medicine, 50 % rural graduation rate, maturation selected and encouraged, no selection bias
Accelerated also a great model for inner city, faculty development alas it is gone
4 - 6 month preceptorships in various locations seem to have some impact, but all are closely coupled to admissions, less time may not mean the same with students who are not as well oriented or motivated to primary care.
3. Medical education retardation is unfortunately the rule - Medical education retardation is based on observations that US medical education is moving more and more passive and away from the student ability to "become a physician." The primary aspect is less ability to participate in medical decision-making and to actually "do" things. The reasons for medical education retardation include liability concerns, complete disruptions of continuity of care by the US health system (less in rural areas), mobility of the population, Medicaid changes, risk of Medicare audits, less and less teachers and teaching time from faculty, and accumulating impacts of previous inability to perform as a physician.
As students, residents, and fellows do less (and faculty) they are more likely to want to do more procedures. The hierarchy of fellows taking from residents from students then insures that there is less and less actual patient contact until later and later in training.
The numbers of procedures are one area of concern. Obstetrics is a particular decline area. Two other areas "expose" this long slow deterioration. These involve the repeated contacts with internationally trained students and faculty. The common conversation involves concerns why they are not allowed to "do" things in the US like they have been doing in Canada or other countries for a long time. The other area involves significant changes in US curricula that "accelerate" the development of a medical framework, especially one involving primary care.
Accelerated family medicine programs and longer term preceptorships are key examples. Accelerated grads have early choice and maturity and focus. They have a curriculum with the most orientation to FP and the most accelerated learning (back to back M-3 and PGY-1 years) with more rapid gains in confidence and competence and comfort with FP and primary care. The 150 recent accelerated grads have distributed better than most of the nations programs. The accelerated programs (10 out of 11) had no rural training location, no rural birth bias, and no bias toward selecting rural practice. Female accelerated grads chose rural in the same number and distribution to the most rural locations. Accelerated FP is one of the only training programs of any type to take urban born and graduate rural. This may be a function of older candidates, but still the process works. The accelerated difference involves a 50 % graduation rate in 150 graduates in 11 programs and this number was double the rural graduation rate of other FP programs in their states. There were 7 accelerated programs in the top 50 distributional FP programs. Accelerated Family Medicine Training Programs
See ranking medical schools and residency programs at Ranking Medical Schools and FP Residency Programs
More on Medical Education Retardation
C. State and National Impacts
States vary greatly in practice support, loan and scholarships, and how much they invest in stealing other states graduates. North Carolina and Wisconsin have particularly effective state recruitment programs. High growth states also steal other FP grads from other states. Most states donate to those with better recruitment and high growth. Graphics at Patterns of workforce Patterns of Rural Workforce
FP choice just took another punch with the Two Tales Wall Street Journal articles that have rapidly moved through AAMC and FP circles in a matter of days.
Distribution is about income, and also results in restoration, via education, health access, leadership, services, and economics. Lack of distribution is all about decline and hopelessness and injustice. Terrorism is all about hopelessness and poverty and inclination toward violence as a solution rather than education and jobs and sacrifice.
Last modifiers
On the personal level - Practice location, rural or inner city is a bit more problematic for certain graduates compared to others. This impact is less in the early years and more in the graduate years. The factors can include spouse/family effect, cultural familiarity and comfort levels (gender, culture, ethnicity, income), language issues, and a growing sense of reality of practice vs previous idealism regarding concept of practice and making a difference. Debt and income issues may also come into play more. Entire areas of the nation involve great cultural variety, including various rural, Appalachia, border areas, regions of the south, reservations, and pocket populations.
Observations in support of education/medical education and socioeconomics:
I also find it interesting that rural males and black males are admitted in the lowest ratios in the nation. This is also true regarding college. The impacts of education, parent income, society, and role models is interesting to consider. Low income may fuel male tendencies to choose other directions when the probability of usual success in academics and careers seems too remote.
Native Americans have disappeared from FP, my theory is that we have lost the reservation natives in favor of higher income urbanized natives. Over time as we admit more and more students of higher income origins, this has also included traditional underrepresented minorities. Also failure of education in many areas may lead to only extremely high intellect students escaping such origins. The high intellect and high income student would distribute not much differently that others who are white or Asian. This also may be a factor in decreasing the known association between minority status and service to the underserved. The arguments for income based determinations rather than ethnicity based are numerous and growing. Rural students and black students have the lowest admission ratios in the nation.
Asian 1967 allopathic medical students admitted 1994 - 2000 per 100,000 age 18 - 24, Urban White
Vietnamese medical student parents are the poorest of the Asia group. Indian Pakistani are the highest income of any group, similar if not greater than whites. Mexican Americans are the most rural and lowest income group in Hispanics. Vietnamese and Mexican Americans choose FP at 20 - 25 %. Indian Pakistani grads 2.3 % aided greatly by the managed care years. This study involved comparisons of all US allopathic grads of 1998 - 2001 by each ethnicity as the denominator compared to ethnicity of US allopathic FP grads of 2001 - 2003. Graphs of this and other ethnicities by parent income and FP choice at my web site. Parent Income and Ethnicity and FP Choice more of this at Medicine, Education, and Social Status
Without admissions that considers income and origins and education background, it is impossible to distribute physicians well, short of government quotas.
Robert C. Bowman, M.D.
rbowman@unmc.edu
Chair of STFM Group on Admissions
See also Birth Origins and Distribution Tables for other proxies of income such as older, rural born, instate born and other outcomes beyond just family medicine