Methods of surveillance for HIV infection at U.S. sentinel hospitals

Abstract

The U.S. sentinel hospital surveillance system for human immunodeficiency virus (HIV) infection includes approximately 40 short-stay hospitals located in 31 metropolitan areas in the United States and Puerto Rico. Several hospitals began testing in late 1986, and additional sentinel hospitals have since been recruited. At each sentinel hospital, anonymous, unlinked testing for antibody to HIV is conducted monthly on 300 blood specimens, selected systematically and stratified by age of the patient. Specimens are excluded from patients whose reason for hospital visit on that occasion was for a medical condition associated with HIV infection or with risk factors for HIV infection, in order to limit the expected overrepresentation of HIV-infected persons among hospital patients compared with the general catchment population of the hospital. The incidence of acquired immunodeficiency syndrome (AIDS) in metropolitan areas with sentinel hospitals has been approximately twice the incidence of AIDS in the entire United States. However, while absolute levels of HIV seroprevalence should therefore be interpreted with caution, trends in the age-, sex-, and race-specific HIV seroprevalence at sentinel hospitals likely reflect trends in the communities served by the hospitals. Although concentrated in areas disproportionately affected by AIDS, sentinel hospitals will contribute seroprevalence data over time that reflect the impact of HIV infection across all age and behavioral risk groups. Sentinel hospitals will also constitute a key surveillance system to help integrate the age group-specific and risk group-specific findings from other activities in the CDC family of seroprevalence surveys.