OVERSTATING SOCIAL NETWORKS' ABILITY TO DIAGNOSE NEW CASES OF HIV

Kimbrough et al.1 describe a peer-driven, network-oriented strategy as highly efficient for HIV case finding. The study found a 5.6% prevalence rate among tested participants, but did not discuss the possibility that some of these positives may be repeat testers. Our experience in New York City shows that partners and network associates with known HIV diagnoses may not always disclose their HIV status to testers. In 2008, New York City's Department of Health and Mental Hygiene (DOHMH) funded 5 community-based organizations to conduct social network testing using a similar incentive structure to Kimbrough et al. Of 147 associates tested for HIV between January and June 2008, 8 were identified as newly positive based on self-report. We linked the identified positive cases to DOHMH's name-based registry of persons with HIV/AIDS—the HIV/AIDS Registry System (HARS)—and found that 3 cases were previously diagnosed. This caused the HIV prevalence rate for newly reported cases to drop from 5.4% (8 of 147) to 3.4% (5 of 147). Similarly, of 299 partners field-tested through DOHMH's HIV partner notification activities from February 19, 2008 to June 30, 2009, 23 tested HIV positive but only 16 were newly diagnosed as per HARS check, reducing the HIV prevalence rate for newly reported cases from 7.7% to 5.3%. “Waterfront Scene,” by Pino Janni, depicts the East River docks in New York City and was all about finding work for longshoremen as well as for artists. From the recent Smithsonian Art Museum exhibition, “1934: A New Deal for Artists.” Printed with permission. Pino Janni, “Waterfront Scene,” 1934, oil on canvas, Smithsonian American Art Museum, Transfer from the U.S. Department of Labor. Hanna et al.'s larger analysis further documented that HIV-positive tests cannot be assumed to represent new diagnoses.2 By matching Western blot tests reported in New York City to HARS, Hanna et al. found that only 36.8% of Western blots were new diagnoses and 63.2% were repeat tests. The amount of repeat testing varied by testing setting, with a higher proportion of repeat testers in community-based organizations (68.8%) than in private medical offices (41.7%) and health department clinics (35.6%). Our experience is that patients retest for a number of reasons, including denial of their HIV status, entry into medical care, or desire for the incentive given at testing. Some of these reasons preclude disclosure of a patient's known status to testers. Though social networking may indeed be an effective way to identify undiagnosed HIV infections, evaluation of these and other testing programs needs to account for repeat testing's impact on case-finding yield. Health department surveillance data can provide information on the true case-finding efficacy of HIV testing programs. However, access to HARS information is limited to state and local health departments. We recommend that the evaluation of testing strategies include collaboration with health departments to account for repeat testing.