Increasing the production of generalists: a computer simulation

The Association of American Medical Colleges set a goal for US medical schools to produce 50% of their graduates as generalists. Methods for achieving this goal have not been established.

We developed a computer simulation to gauge the effect of interventions that could occur at five different points during medical school education. Each intervention would improve by 10% the proportion of trainees choosing a generalist career.

Our model predicts that 24.1% of 1992 medical school graduates will ultimately become generalists. Any of the interventions we evaluated increased generalist production to the 25%-26.5% range. If class size is held constant and if a 10% annual improvement occurred in all of the other intervention points for 3 years, our model still predicts only 44.8% generalists.

Computerized "what-if" analysis can effectively project specialty outcomes after various interventions. The best intervention increased generalist production 2.6 times more than did less effective interventions. The loss of primary care residents into subspecialty fellowships is substantial. Even if 100% of current medical graduates could be directed into family practice, internal medicine, and pediatric residencies in their current proportions, only a 43.7% generalist outcome would be achieved because of attrition of internists and pediatricians into subspecialties. Family practice is the most efficient producer of generalists but currently lacks either the capacity or the marketability to be the nation's sole supplier. More effective interventions are needed to achieve a 50% generalist outcome.