Divisions in Physician Career and Location Choice Related to Age at Graduation

Robert C. Bowman, M.D.

The more things change the more they stay the same. Physician workforce has been in a constant flux under the assault of changes in medicine, health policy, American education, and patterns of immigration and migration. Longer term trends have progressively changed areas such as student income levels, concentrations of income, and the availability of opportunity. Other areas such as subspecialization, practice patterns, training methods, Medicare, Medicaid, and dominant types of medical schools have remained relative constants. In recent years schools, students, and class sizes have stabilized to allow closer observation using indicators such as physician age at graduation. A consistent factor such as age can help simplify workforce analysis.

Researchers have studied some of the factors related to older age at graduation. Older graduates may develop interest in becoming a physician after exposure to a health experience or health career. Life experiences may shape a later decision for a medical career.(xu) Others may be delayed in a more consistent fashion that may be worthy of study.

Delays in medical school admission may be due to obstacles of education, income, culture, college finance, professional school finance, or other differences. Those raised in lower income areas or populations may not attempt admission, they may try and fail, or they may gain admission at a later age. Maldistributions of income and education may also delay admission for all except the top 30% who can afford private school, college preparation, college, medical school preparation, and medical school. Higher status students may skip grades, benefit from the steady career advice of parents who are professionals, or gain early admission. Lower income, inner city, and rural origin students can be left behind without steady and significant investments in education infrastructure. Maldistributions can move up the socioeconomic scale to middle class students if allowed to persist and increase.

Progressive changes are most apparent in recent medical school admissions. The 16,000 medical school students admitted each year now include 3000 more medical students from parents making over $100,000 compared to 10 years ago. Medical students admitted from families making less than $40,000 have decreased by 1500 and the middle group has decreased by 1500 as well. The replacement of 3000 medical students each year with the broadest range of origins by 3000 or more of the narrowest range of origins can certainly impact physician workforce.  

If delays in medical school admission correlate with socioeconomic differences and geographic origins, older age graduates my have different career and location choices shaped by more extensive and different life experiences. Those who have been ministers, teachers, and serving professionals favor generalist primary care careers. (Madison, Service Orientation) Two major divisions of medical students emerge with relevance to physician workforce. One group has a broad range of humble origins, different education experiences, and different medical schools for training. The narrow group concentrates in specific colleges, medical schools, and fellowships. This is a group defined by thirty years of life in the most socially organized families, the most urban locations, and the highest income areas of the nation. Even without government investments, this group faces few obstacles to medical school admission. Opportunity for the middle and lower income group is far more dependent upon distributions of education and opportunity organized at the community and state level.  

A few medical schools apparently prefer or recruit older graduates.  Older Graduate Preference Older graduate career choices may fit with the admissions policies of a school or they may fit the mission of the school regarding training or preferences in careers or locations. A few have established admissions tracks for older graduates. Osteopathic schools and the allopathic medical schools graduating the most family physicians have twice the number of older graduates compared to the majority of allopathic medical schools.

For decades those entering medicine at older ages were considered by economists to be less likely to consider a prolonged medical education training involving surgery or a fellowship. This has been based on the assumption that older graduates are aware of fewer potential years of practice remaining and the need to have shorter duration of training. (Ernst and Yett) Massive changes in costs of college and medical school tuition and the costs of living prior to finishing training may shape career choices in different ways.

A number of other changes are likely to impact physician workforce and age at graduation.

·        Changes in Admissions By Age – Older graduates have only recently gained acceptance for admission. During the 1970s expansion, those graduating at older than 29 years of age increased from 7% to 17% of US MD Grads. Changes in age during this time period complicate workforce analysis related to age.

·        Changes in Admission By Origins – Increasing parent income levels and replacements of lower income, middle income, and rural born medical students are a few of the changes. Asian and foreign born medical students have increased to 30% of US MD Grads. These are all changes in admission that mean replacement of the most distributional types of students by those least likely to leave major medical centers. Even with the addition of new sources of youngest age student types, about 22% of US MD Grads remain older than 29 at graduation. Medical students have become both younger and older.

·        New Medical Schools - The medical schools created in the 1970s admitted more diverse students in socioeconomic and geographic origins. Those most different were also older and more likely to be found in family medicine, rural, and underserved practices.

·        Generalist Focus – The general practitioners before 1970 were older and different in other ways. Studies in the 1970s and 1980s tended to focus on generalism and combined primary care studies with less specific emphasis on family medicine. Career and location outcomes can be difficult to predict with a mixed set of subjects.

·        Shifting Sands Over Time - A study of 2500 graduates found 1100 different career pathways. (Zuckerman 1960) Career and location outcomes may well be shaped by the timing after graduation. In past studies physicians tended to move away from primary care and rural locations to specialization and major medical center location. (AAMC longitudinal Study of Medical School Graduates of 1960) These studies involved medical schools that included few physicians in rural states, few if any family physicians, and physicians graduating in a period with little or no expectation of state or federal support. Newer studies involve different schools, different health policies, and a new specialty that does not fit this pattern.

·        Admissions of Black and Hispanic Students – Until Civil Rights, admissions of diverse types of medical students involved only two schools. Limited studies associated exclusive career and location choices for the few Black medical students that were admitted. These studies may represent an atypical sample or the selection of only the most exceptional students. The Black and Mexican American students gaining admission to all medical schools represented the broadest range of origins and demonstrated the greatest interest in locating in underserved areas. (AAMC MIM) Black and Hispanic family physicians are found in underserved locations at 20 – 25% or 4 to 5 times the national mean of 5.4%. (eth)  

·        The New Specialty that Distributes and Lasts - Family practice residency graduates from 1997 – 2003 remained in family medicine at 98% and continued in office based primary care at 90%. Family physicians are found in rural locations at a constant 24% and underserved locations at a constant 10% for each class year since 1975. Both figures are double national distribution levels for physicians. Earlier graduates were found in both locations at slightly higher values. Family medicine concentrations outside of major medical centers remain a defining characteristic that maintain distribution years or decades later. The family physician primary care retention contrasts with all current and past forms of primary care.

·        Primary Care Retention – At least 43% of the 1960 – 1977 class years of US MD Grads that trained in primary care residencies left primary care entirely. More recent graduates do much better and much worse. Office-based general pediatrics has remained a constant 70% of pediatric residency graduates, but levels of general internal medicine have plummeted below 20%.(gar) Medicine pediatrics careers deteriorate throughout the match, residency training, and the first decade of practice to less than 20% levels. Physician assistants across the nation have declined steadily by about 1 percentage point a year for the past ten years in primary care, family medicine (40% to 28.5%), and rural practices (25% to 16%). (aapa) Nurse practitioners cannot demonstrate greater levels and all primary care continues to be shaped by the same training locations and health policies, but clearly the outcomes are not the same.

·        Limited Studies - Literature limited to individual programs or schools greatly limits evaluation of any type of practitioner. Decisions on national health policies must consider career and location choices across the nation and across the entire span of a physician career. Studies of single programs, first career choices, and first location choices can be misleading.

·        New Wrinkles in Workforce Interpretation – The timing of data collection in a specialty such as family medicine may not be critical. For specialties that collapse back into major medical centers or those that change careers over time, the timing may determine the career and location outcomes. Studies that closely follow residency graduation may not demonstrate age effects if older age graduates remain while younger graduates do not. After a decade has passed, the effects of age may be more easily seen. Even in family medicine, those found in nearly 100% office based care in rural or underserved urban locations are different than the typical 90% in urban office based family medicine, and are also different from other family physicians in office-based, teaching, research, administrative, or hospital based careers in major medical centers. Age and other characteristics may also be different. 

Career choice studies have also divided physician specialties into categories. One grouping includes surgical, person oriented, or mixed. (Yufit 1969, Wasserman 1969) Age was not used in this categorization, but few graduates of this period were different in age, diversity, and other origins. Changes in the concentrations of students, income, education, professionals, population, and physicians may mean the ability to classify physicians into various types, specialties, or categories by age at graduation.

Previous studies regarding older medical students were not possible. Only in recent decades has the nation admitted significant numbers of older medical students. Only in the past 40 years has the nation had health policy supportive of primary care, lower income areas, and rural locations. Only in the past 36 years has the nation had a stable primary care career that retained graduates in primary care and outside of major medical center locations.

Studies involving age, career choice, and practice location may help to understand physician workforce in some of the most important areas.

Methods

This is a descriptive study comparing age, career choice, and medical school factors. The 2005 OfficeMax version of the Masterfile includes 203,627 graduates of United States allopathic medical schools for the 1987 – 1999 class years. These are the most recent graduates in the Masterfile that have had time to distribute to near equilibrium conditions after the completion of training. This group represents 40% of total United States workforce. It is a group with a steady 16,000 graduates a year. This group of recent graduates is health policy neutral for the entire period since it includes relatively equal proportions of graduates with lower primary care choice (1987 – 1990 class years) and higher primary care choice (1995 – 1997 class years). This is also a group that preceded the rapid increases in parent income level beginning with the 2001 US MD Grads (1997 – current matriculants).

The US MD Grads that were missing career, location, or birth origins; those who were born in military bases, Puerto Rico, Guam, the Virgin Islands; those attending military or Puerto Rican medical schools; and those found in military locations or outside of the 50 states were excluded. The remaining group of 93% of the US MD Grads for 1987 – 1999 includes 189,163.

The major physician specialties were compared regarding the percentage in each specialty by age group. In addition the age mean for each specialty was calculated. The average MCAT scores for a medical school were collected from medical school web sites for 2000 – 2003. This average was compared to the percentage in each specialty to generate Pearson correlations. This is not individual MCAT data but secondary data from the medical school attended by the US MD Grad. This reflects the selectivity of the school, not the exclusiveness of the graduate. Medical school mean age may also reflect a similar selectivity, but in an opposite direction.

The usual age range of US MD Grads at graduation year is 25 – 30 years. Medical students age 26 are the most common type of medical student with 31%. Age 27 adds another 20%. Those younger than age 26 are about 5% of graduates. Those over age 29 are 22%.

Comparisons across the usual class years can divide US MD Grads into 3 groups: specialty choices that increase with age, are neutral, or that decrease with age. The changes would be more dramatic using a wider age range, but this would fall outside the normal age range of students. Those younger than 26 at graduation are an atypical group with more foreign born, Asian, urban origin, and highest income origin graduates and fewer rural born and lower income graduates so the age 25 and 26 graduate choices were combined for a baseline comparison.

The percentages of those in a specific career who were age 25 or 26 at graduation were compared to the percentages age 29 or 30 years of age. The career choices are also compared to medical school class characteristics for mean age of the graduating class for each medical school for the 1987 – 1999 graduates of the school and for the average MCAT score for the school for the 2000 – 2003 matriculants. Although the MCAT score is not class year specific to this group, sets of MCAT scores correlate with each other, with the MCAT average, with MCAT ranking, and with individual bioscience components at 0.86 to 0.96 or above.

A positive correlation with the age variable confirms career choice increase at the medical school level. A negative MCAT correlation means that the percentage of students at a school choosing a career such as family medicine decreases as the MCAT average score of the school rises. Both age and MCAT score appear to be measures of the selectivity of the medical school regarding admission. Both relate to important measures of physician workforce.

The database of medical schools used for correlations was a restricted list. Medical schools with specific younger age admissions (Kansas City, Northeast Ohio, Puerto Rican schools), the military school, and schools focused on family medicine and rural location (Mercer and University of Minnesota schools) were excluded.

Finally the author prepared 3 rating scales, one for General, one for Direct Patient Care, and one for Procedural/Technical/Subspecialty. Each career was scored from + to +++ in both scales. The table was ranked and grouped by correlation with mean age of medical school graduates.

Results 

Characteristics of Medical Student Careers Related to Age and School MCAT

* Other than marked with an asterisk, the correlation is significant at p < .05. Most are significantly higher. Rural pediatrics choice is higher in the older age ranges. Caution is needed in interpreting the medicine pediatrics and emergency medicine data. Only 1580 remain of 3100 who matched into medicine pediatrics for 1987 - 1999. Those remaining in this group are different. The emergency room data predates formal emergency room residency training and is more of a reflection of those that left other careers for emergency care. Changes in the terminology for Pulmonary (PUD) and Pulmonary Critical Care (PCC) were present in this cohort so both were combined. Radiology types were combined but all remained in Group III.

Physicians do layer out in three groups by changes in age at graduation, by mean age, and by average MCAT scores. The careers most involving generalism, primary care, and people and least involving procedures and technology increase with age and life experiences prior to medical school.

Careers with little change with age are more mixed in career specifications.

Careers that decline with older age involve less emphasis on direct patient care and more emphasis on sciences, technology, and academics.

Discussion

The careers that increase with older age at graduation in Group I represent the front line careers of medicine from primary care to mental health to emergency care. The medical students with the most exposure to people and life experiences before medical school may well be the ones most likely to consider similar orientation. Themes of impacts of life experiences, older age graduation, lower MCAT, biopsychosocial complexity, and direct patient care careers are supported. In addition the careers most needed for health care access are the most likely to be related to increased choice with age, graduation from medical schools with increased age medical students, and medical schools with lower MCAT scores. Research choice also is noted to increase with age.

The limitations of these studies involve variations that are inescapably related to most other important dimensions of geography, socioeconomics, status, concentrations, and shortages. For example, the same ratings for general, direct patient care, behavioral, science, technical, and subspecialty also would fit with other scales. Correlations between age means for specialties and other selectivity factors are high including the specialties most emphasized at elite medical schools (-0.61), preferred specialty choice in the match (0.56), board scores (-0.66), AOA status (-0.61) associated with specialties, or the percentage of foreign born US MD Grads in the specialty (-0.61).

The alignment of status, socioeconomics, exclusive admissions, standardized test scores, career choice, and physician distribution is a most important concept regarding physician workforce and also of great concern to those hoping to improve the distribution of professionals with a minimalist approach.

One would have to question the admissions trends of the last decade for allopathic medical schools. Asian medical students have increased far beyond the 4% of the population that is Asian to 23% of US MD Grads and 30% if including the non-Asian foreign born US MD Grads. These are changes that reflects concentrations of admissions in those of the highest levels of education, professionals, urban origins, and proximity to medical school. Admissions of all of the highest income medical students of all ethnicities and races is just as great, although not as dramatic as the more recent rise in Asian composition.

Other changes are readily apparent in direct studies of the admitted medical students. The major changes have been increasing parent income and increasing MCAT scores. MCAT Changes since 1992 The bioscience component for matriculants has increased at nearly 0.1 unit per year since the MCAT was last standardized in 1993. The increasing science subscores have driven MCAT average scores to record levels. Those with even slightly lower scores are increasingly left out of allopathic medical schools. These are medical students that are clearly acceptable and related to no greater risk of failure or no lower probability of distinction. (AAMC, Academic Medicine, Julian) However those gaining admission have scores that are slightly better, parents of higher income levels, and career choices that are the narrowest of all.

The fears of past medical educators appear to be the reality of today. Graduating physicians have increasingly narrow socioeconomic origins and much narrower focus on science and academics. The important balance between admissions of potential physicians with people skills as compared to science/technology orientation appears to have been lost or at least is in critical condition.

Medical education shares no different blame than all of the other contributors. State and national policies once supported a broader focus of education, broader admissions in college and professional school, and broader training of physicians. Now child development is restricted to those of higher status (hart and risley), neighborhood security is related to income, early education and all of education is limited to those not born in the gap areas of America (Funding Gap 2004), the top 146 colleges that feed most professionals and national leaders are 74% filled with those from the top 25% in income, and medical schools have the same admissions patterns as the top colleges. Opportunity as measured by economics and other areas is increasingly related to fortunate birth as noted by John Kenneth Galbraith and others.

The nation once specifically focused on the most important professionals of all types for the most needed locations. There was at least some recognition that distributions of education and professionals were critical to distributions of economics, jobs, education, leadership, and opportunity. These were also considered the cornerstones of democracy. The service oriented professionals in education, public service, and health were a top priority. During the 1960s, 1970s, and 1990s, the nation focused on the most physician careers and locations that were needed most by lower and middle class Americans. The nation now has very different policies that concentrate resources and opportunity. Although some level of concentration is to be supported and rewarded, there is the potential that concentrations can go so far as to inhibit opportunity, health, and education for far too large a segment of society. In such situations fewer and fewer make the effort to become educated or professionals. At this point further concentrations may compromise costs, quality, and efficiency. Clearly the highest income populations in the United States have separated from middle and lower income groups in income, education institutions, and personal contact. Caring for those most different or leading those most different may be more and more difficult without improved distributions.

1.         Xu G, Veloski JJ, Barzansky B. Comparisons between older and usual-aged medical school graduates on the factors influencing their choices of primary care specialties. Acad Med. Nov 1997;72(11):1003-1007.

2.         Madison DL. Medical school admission and generalist physicians: a study of the class of 1985. Acad Med. Oct 1994;69(10):825-831.

3.         Ernst RL, Yett DE. Physician Location and Specialty Choice. Ann Arbor: Health Administration Press; 1985.

4.         Zuckerman. 1960.

5.         Swanson AG. AAMC longitudinal study of 1960 medical school graduates: a 20-year effort in 28 schools, 1956-1976. J Med Educ. Dec 1986;61(12):991-992.

6.         Association of American Medical Colleges. Minority Students in Medical Education: Facts and Figures XI Available at https://services.aamc.org/Publications/showfile.cfm?file=version12.pdf&prd_id=89&prvid=87  Accessed April, 2003. Washington DC 1998.

7.         Bowman RC. The Distributional Student, the Distributional School, and the Distributional Specialty. pending acceptance to Rural and Remote Health. 2006.

8.         Bowman RC. Primary Care Retention.  http://www.unmc.edu/Community/ruralmeded/primary_care_retention.htm . Accessed October 26, 2006.

9.         Garibaldi RA, Popkave C, Bylsma W. Career plans for trainees in internal medicine residency programs. Academic Medicine. May 2005;80(5):507-512.

10.       American Academy of Physician Assistants. Data and Statistics.  http://www.aapa.org/research/index.html . Accessed October 26, 2006, 2006.

11.       Wasserman E, Yufit RI, Pollack GH. Medical specialty choice and personality. II. Outcome and postgraduate follow-up results. Arch Gen Psychiatry. Nov 1969;21(5):529-535.

12.       Yufit RI, Pollock GH, Wasserman E. Medical specialty choice and personality. I. Initial results and predictions. Arch Gen Psychiatry. Jan 1969;20(1):89-99.

13.       Bowman RC. Five Periods of Health Policy and Physician Career Choice; 2006.

14.       Julian ER. Validity of the Medical College Admission Test for predicting medical school performance. Acad Med. Oct 2005;80(10):910-917.

15.       Hart B, and Risley, T. Meaningful Differences in the Everyday Experience of Young Children. Baltimore: Paul H. Brookes; 1995.

16.       Carey K. Funding Gap 2004: The Education Trust; http://www2.edtrust.org/NR/rdonlyres/30B3C1B3-3DA6-4809-AFB9-2DAACF11CF88/0/funding2004.pdf  accessed April. 2005.

17.       Carnevale A,  Rose, SJ. Left Behind: Unequal Opportunity in Higher Education, Reality Check Series. In: Kahlenburg R, ed. New York: The Century Foundation Press; 2004:p. 106.

Physician Workforce Studies with a distributional focus

www.ruralmedicaleducation.org