The subspecialization rate of third year internal medicine residents from 1992 through 1998

The appropriateness of U.S. physician workforce size and the proportion of generalists versus specialists have long been debated. Difficulty collecting reliable data and varying methodologies complicate clear analysis of workforce questions.

This work examines the rate at which internists subspecialized during the 1990s. It also compares two approaches for estimating subspecialization rates: (a) following resident classes longitudinally ("cohort" approach), and (b) comparing 1st year fellowship (F-1) class size to the previous year's 3rd-year resident (R-3) class size (F-1/R-3).

Data were collected through the American Board of Internal Medicine's tracking program. Physicians completing their R-3 year in 1992 through 1998 were the participants. The proportion of each R-3 group that eventually entered subspecialty training was examined. Demographic data for those entering subspecialty training and those who did not were compared. Subspecialization rate estimates for the cohort and F-1/R-3 approaches were also compared.

The number of internists increased, whereas the number entering subspecialty training declined. Men were more likely to enter a subspecialty than women. International medical school graduates were more likely to enter a subspecialty than U.S. medical school graduates. University-based residency program trainees were more likely to enter a subspecialty than community hospital program trainees. Those entering subspecialty training tended to be younger and score higher on the internal medicine certification examination than those who did not. Almost identical estimates where produced by the cohort and F-1/R-3 approaches.

There was a downward trend in the rate at which internists entered subspecialty training during the 1990s. The two methodologies examined produced similar results.