The biomedical sciences employ race as a descriptive and analytic category. They use race to describe differences in rates of morbidity and mortality and to explain variations in drug sensitivity and metabolism. But there are problems with the use of race in medicine. This article identifies a number of the problems and assesses some solutions. The first three sections consider how race is defined and whether the racial data used in biomedical research are reliable and valid. The next three sections (...) explain why racial variation in disease, including genetic disease, is not evidence that race is biological. The final section explains how a proper understanding of the role of race in medicine bears on public policy. (shrink)