An active surveillance system was used to identify all residents with hemophilia in six U.S. states (Colorado, Georgia, Louisiana, Massachusetts, New York, and Oklahoma). A hemophilia case was defined as a person with physician-diagnosed hemophilia A or B and/or a measured baseline factor VIII or IX activity (FA) of 30% or less. Case-finding methods included patient reports from physicians, clinical laboratories, hospitals, and hemophilia treatment centers. Once identified, trained data abstractors collected clinical and outcome data retrospectively from medical records. Among cases identified in 1993-1995, 2,743 were residents of the six states in 1994, of whom 2,156 (79%) had hemophilia A. Of those with available FA measurements, 1,140 (43%) had severe (FA < 1%), 684 (26%) had moderate (FA 1%-5%), and 848 (31%) had mild (FA 6%-30%) disease. The mean and median age was 25.4 and 23 years, respectively. The age-adjusted prevalence of hemophilia in all six states in 1994 was 13.4 cases/100,000 males (10.5 for hemophilia A and 2.9 for B). The prevalence by race/ethnicity was 13.2 cases/100,000 among white, 11.0 among African American, and 11.5 among Hispanic males. Application of age-specific prevalence rates from the six surveillance states to the U.S. population resulted in an estimated national population of 13,320 cases of hemophilia A and 3,640 cases of hemophilia B. For the 10-year period 1982-1991, the average incidence of hemophilia A and B in the hemophilia surveillance system (HSS) states was estimated to be 1 in 5,032 live male births.