Research By the Ages

Robert C. Bowman, M.D.

Introduction:

The graduation of physician researchers is a major goal of most if not all medical schools. The factors regarding graduation or researchers have received a great deal of policy and medical education attention, but less evidence-based work.

Methods:

The American Medical Association Masterfile was used to compare the graduation of physician researchers by medical school, by medical school type, and by birth origin student characteristics (age and location) to assess the impact of admissions of various types of students and the impact of state education investments.

Results:

Older medical students are more likely to choose research careers. Generally the schools graduating the most physician researchers also graduated the most younger and older medical students into eventual research careers. There were 22 medical schools that graduated half of the physician researchers. In these 22 research schools when younger graduates chose research at high rates, older graduates had even higher rates of research choice. In the 22 top research schools out of state, younger, urban born, and foreign born admissions resulted in more research graduates. In the 103 other allopathic schools, none of these various types of students were associated with improvements in research graduation rates. Even in the 22 schools, the research graduates were not retained and those from east coast schools were the most likely to move over 1000 miles away to other locations.

Discussion:

Medical schools must do individual work to understand their admissions. There are three medical careers that apparently require more life experience, dedication, and maturity. These include office-based primary care (especially rural locations), psychiatry, and research. In each of these careers the percentages of career choice increase with age. Those who are most oriented toward science and technology are most likely to choose subspecialties, especially when they are younger. A medical education system that does not understand how to graduate more researchers, a clear top priority for most deans, really does not understand admissions.

Introduction:

The lure of National Institutes of Health funding is too much to ignore as schools face declining state funding and federal funding. Schools have attempted to attract more potential physician researchers to their schools. There are different methods that schools use to recruit potential researchers, including preprofessional efforts involving high school and college students, competing for the national pool with early admissions and scholarships, or recruiting older students who may have already demonstrated research prowess.

Methods:

The AMA Masterfile has data involving self-designation of research as a primary practice activity. This data was matched to medical school of origin and age of the physician at graduation from medical school. The data involved the allopathic classes graduating 1987 – 1994. Inclusion of 1995 - 2000 graduates progressively decreased the percentages indicating research careers. This indicates a delay factor involving the length of training, delay in choosing research, and a delay in recording this career change in the Masterfile data. The data used should not be considered complete. It does not include those who do research part time or those who have failed to fill out physician career survey forms. Research and academic physicians do have a history of poor membership in the AMA and poor return of such forms. However the data is collected consistently for the graduation years indicated and should be comparable across medical schools and ages of the graduates.

Data on the zip code location of the researchers was also mapped by latitude and longitude and compared to the medical school zip code to see if researchers were retained within 60 miles of their medical school training location. The graduation of researchers is also compared to high school and college graduation rates for states.

Results:

There were 1480 physician graduates of 1987 to 1994 who designated a research career as their primary physician activity as recorded in the AMA Masterfile of 2004. Half of the physician researchers of the nation graduate from 22 out of 125 allopathic medical schools or 22 out of 150 total medical schools (including osteopathic and Caribbean).

Private schools were more likely to graduate researchers, but the effect of research graduation is limited to the top research schools and there is little or no effect in 70 – 90% of medical schools. Those who choose research careers tend to be older at graduation from medical school. Those less than 30 at graduation from medical school choosing research involved 0.90% of the graduates compared to 0.98% of older graduates. The effect of admissions of out of state students to increase research graduations is also limited to the few top research schools

In the top quartile private allopathic research schools (n=11) the younger graduates chose research at 2.8%, the older (over 29 at graduation) graduates at 4.4%. The effect of age also applied at a lower level to the other private schools. For the top quartile 11 private schools, out of state born students were more likely to choose research careers at 3.3% for out of state compared to 1.7% for instate born. In the other 3 quartiles there was less effect 1.48% for out of state to 1.17% instate, 0.93% to 0.71% instate, and a reversal from 0.40% out of state to 0.57% instate or slightly higher for instate born. The students born in counties of over 1 million and those born in foreign nations were also more likely to become researchers. In the other 3 quartiles, the most urban born students and the foreign born students had the same probability of a research career or a lower probability.

In the top quartile public allopathic 22 research schools, the younger students chose research at 0.8% compared to older students at 1.4%. The public school top quartile out of state born students chose research at 1.13% compared to 0.65% for those born instate. For the lower 3 quartiles involving the differences were minimal at 0.73% to 0.62%, 0.48% to 0.44%, and 0.54% to 0.38%. The students born in counties of over 1 million and those born in foreign nations were also more likely to become researchers. In the other 3 quartiles, the most urban born students and the foreign born students had the same probability of a research career or a lower probability.

In over 75% of US medical schools, admissions of younger students, foreign born students, the most urban students, and out of state students to improve research graduation does not work.

Schools graduating more physician researchers tend to also graduate both younger and older researchers.

In the top research medical schools, 14 of 17 had twice as much success with older graduates. When the rank order was by schools who have had reasonable success with younger researchers, the older graduate advantage remained with 9 of 14 demonstrating twice the success with older researchers.

The schools graduating the most researchers have the highest MCAT scores (+0.76), the most National Institutes of Health Funding (+0.68), and the greatest graduate medical education emphasis (+0.62 correlation between number of residents and fellows divided by medical student numbers). The top research schools graduate 2 – 8 researchers a year. The other 104 schools graduate a physician researcher about every other year. The schools in states with the lowest college graduation rates graduated the lowest levels of researchers (most schools in the South and Midwest). There were fewer researchers from  newer public allopathic schools, osteopathic schools, Caribbean schools, and traditional black schools.

Retaining physician researchers at or near a medical school is a difficult challenge. Researchers graduating from the highest percentage research schools in the eastern metropolitan corridor (Maryland and Pennsylvania to New York and Massachusetts) were the most likely to be over 1000 miles away from their medical school as of 2004 locations (+0.44 correlation). Selecting medical students for research purposes may not work out well even for the top schools, but attending such schools may work out for graduates of these top schools who may be heavily recruited as they demonstrate potential.

Researchers were not born in the states where medical school applicants had the highest MCAT scores (only +0.072 correlation) and were less likely to come from states with a greater percentage of rural population (-0.34). They tended to have birth origins in states with the highest income levels (+0.48) and states with the highest percentage of bachelor’s degrees (+0.51).

States with a wider gap between high school and college graduation rates (ratio HS graduation 1986 divided by college graduation rate) were least likely to graduate researchers (-0.44), but were the most likely to graduate family physicians (+0.68). The low research schools graduate more family physicians, particularly when the high school graduation rate is relatively higher than the college graduation rate. Breadth of education graduates family physicians and depth of education graduates researchers and subspecialists.

Discussion:

Medical schools not only need to spend more time and effort studying admissions policies regarding physician distribution, they also need to understand policies that result in more physician researchers.

The major finding of the study is the great difference between the top research schools and the great majority of medical schools. The efforts that work for the top schools do not tend to work for other schools. Even the top schools do not appear to get the desired outcomes regarding retention of alumni as researchers and direct improvement of National Institutes of Health research funding increases.

This study has significant limitations due to secondary data collection. The numbers of actual researchers are likely limited to a select few with full time research assignments, but the study should assist with individual admissions committee efforts to obtain more resources and participation to help shape the desired outcomes.

There are some key questions to address:

Can medical schools admit more students of a certain type to improve the possibility of graduating more researchers?

Success in recruitment and graduation of physician researchers is limited to the top 22 research schools. Attempts to recruit younger, out of state, or higher MCAT scoring students have largely failed for the other 100 allopathic schools.

Older physician graduates are more likely to choose research careers.

Do schools retain the researchers that they graduate?

The retention rate within 60 miles of the medical school is about 23% and the eastern schools with the highest research graduation rates had the most researchers locating over 1000 miles away.

Who benefits from admission of research-likely students?

It appears that the major beneficiaries of the current admissions process are the students themselves. They often receive full scholarships and may attract the attention of other schools hoping to build their research efforts, given their prestigious medical school graduation status.

Are there “side-effects” of research-oriented admissions?

The schools with the highest graduation of researchers, those with the highest MCAT scores, and those with the most NIH and GME funding all graduate the fewest physicians into family medicine, rural locations, and underserved primary care. The schools that admit older, instate, rural born, and lower income students (who share the common characteristic of lower MCAT score) in the attempt to graduate more family physicians also graduate more primary care physicians, more psychiatrists, more rural physicians of all types, and more office-based primary care physicians for poverty locations.

Each 1 point higher MCAT for a school results in 1 additional researcher every other year. The same 1 point higher MCAT school graduates 4 percentage points or 6 – 9 fewer family physicians each and every year and such schools also graduate fewer for rural locations and office-based poverty locations.

Admissions of younger students or early admissions may not represent a good expenditure of resources. The major reason to do so may be competition rather than graduation of researchers. Internal recruitment of older students interested in research appears to be a successful method at many schools with some schools better than others in this area.

Medical schools and physicians must understand useful competition and the limits of admissions and competition, particularly regarding younger students who are likely to be less mature and less likely to know their eventual career choice.

Primary care and research careers The pipeline to a physician researcher is even longer than that of underserved physicians. Current policies appear to benefit the students who receive special admissions, but not the schools that grant such admissions. The policies that result in more researchers graduate fewer physicians who distribute to needed rural and primary care careers. The broader admissions policies that result in more family physicians also graduate more rural physicians, more rural physicians, and more physicians for office-based primary care poverty locations.

 www.ruralmedicaleducation.org