The National Study of Internal Medicine Manpower: XX [1] gives detailed insight into the changing demographics in internal medicine. I believe, however, that a slight error was made in the interpretation of the demographics; it becomes even more important to discuss this issue because recommendations are often based on these interpretations.

According to Table 2 in the article, 99% of U.S. medical school graduates are in group D, 43% of whom go directly into subspecialty training (the highest percentage among all the groups). The authors have attributed this high percentage in both groups C and D to the possibility of an exceptionally high rate of foreign national international medical graduates (FNIMGs) in groups C and D entering subspecialty training. This assumption appears completely incorrect for group D; I find it hard to believe that 1% of FNIMGs in group D would account for a major part of the 43% of group D graduates entering subspecialties.

The authors further attempt to identify programs with high numbers of FNIMGs, which are likely to be particularly vulnerable to reduction under a reformed system of graduate medical education. This identification, according to the authors, was done because the FNIMGs have a higher rate of subspecialization. If the data are reviewed carefully, the programs in group D should be seen as particularly vulnerable to reductions under a reformed system of graduate medical education.