A BRIEF ORIGINAL CONTRIBUTION Repeat Negative Human Immunodeficiency Virus (HIV) Testing in San Francisco: Magnitude and Characteristics

The authors assessed the characteristics of repeat human immunodeficiency virus (HIV) testers at publicly funded sites in San Francisco. During 1992-1993, 3 1 % of all HIV tests were performed on persons testing negative for the third time or more. Persons with greater numbers of prior negative tests were less likely to test HIV-positive. Repeat negative testers were more likely than first-time negative testers to be homosexual or bisexual males, homosexual or bisexual injection drug users (IDUs), or heterosexual IDUs. Repeat testers who seroconverted were more likely to be in these same transmission categories than repeat testers who remained negative. Because of the similarities in risk profile between those most likely to retest and those most likely to seroconvert, attempts to limit repeat testing must proceed cautiously. Am J Epidemiol 1995;142:719-23. acquired immunodeficiency virus; HIV Human immunodeficiency virus (HIV) antibody testing serves a vital function by counseling individu-als to reduce their risk of HIV infection and by refer-ring infected persons for treatment. The Centers for Disease Control and Prevention spends over $100 million annually on HIV testing, representing one-fourth of the national HIV prevention budget (1). Competition for limited prevention funds has prompted debate on the role of publicly funded HIV testing. The importance of identifying and counseling HIV-infected persons is clear: to encourage them to seek treatment and to reduce behaviors that would transmit HIV. However, the impact of testing HIV-negative persons is less clear. Most studies have shown no reduction in risk behavior due to testing HIV-negative (2-6). Some investigators have sug-gested that receipt of a negative test result after en-gaging in high risk practices may reinforce a message that the behavior was not risky (3, 4, 7). Others have questioned the value of repeat testing both for persons Received for publication July 20,1994, and in final form Decem