To the Editor.— We read with dismay the statement regarding human immunodeficiency virus (HIV)-infected physicians issued by the American Medical Association (AMA) on January 17,1991, and reported in the New York Times on January 18, 1991 (p A17). This statement recommends that HIV-infected physicians abstain from performing invasive procedures that pose "an identifiable risk of transmission" or inform their patients of their infection. The AMA also advises physicians who are "at risk of acquiring HIV infection" and who perform "invasive procedures" to be tested for HIV infection. We agree that the safety of patients is paramount when there is a known risk. We believe, however, that the AMA's recommendations are unwarranted, imprecise, and not in the public interest. First, the risk of HIV transmission from an infected physician to a patient is not identifiable. A serosurvey of patients of an HIV-infected surgeon failed to identify a single HIV-infected patient without