US graduate medical education, 2000-2001.

For the last three quarters of a century, the American Medical Association's national collection of graduate medical education (GME) data has evolved in its scope and methods. This year's GME survey involved new technology. The National GME Census for 2000-2001, jointly administered by the American Medical Association and the Association of American Medical Colleges, was part of an Internet-based product called GME Track. Because of technical problems, data collection was less complete than in previous years. Similar to the 1999-2000 survey, we observed an increase in the number of subspecialty programs, with 79 more than last year (2.1% increase), and a decrease in the number of specialty programs, with 40 (0.9%) fewer. Parallel to this continuing trend was a decrease in the number of graduates of US medical schools who were matched into primary care residencies, particularly family practice programs (20% decrease compared with 1996-1997). The number of graduates of osteopathic medical schools training in allopathic programs continued to rise, increasing 7.9% from last year. Numbers of Hispanic and Asian graduates from US allopathic medical schools (USMDs) in graduate year 1 (GY1) positions increased numerically to 887 and 2356, respectively, and proportionally by 7.2% and 17.3%, respectively. Although the number of white USMDs in GY1 positions increased, their proportion decreased slightly among those with known race or ethnicity from 72.2% to 71.7%, and the number of black USMD GY1 residents, numbering 859, declined from the previous year. Although we observed an overall decline in the average number of on-duty hours expected of residents in their first year in a program (from 55 in 1996-1997 to 54 in 2000-2001; P<.001), the average number of hours reported by the majority of programs that typically report the most on-duty hours did not decrease. The issues of resident work hours and the diversity and specialty distribution of the physician workforce continue to foster debate.