HIV incidence among repeat HIV testers at a county hospital, San Francisco, California, U.S.A.

Timothy A. Kellogg, Willi McFarland, Jennifer L. Perlman, Hillard Weinstock, Stephanie Bock, Mitchell H. Katz, Julie L. Gerberding, David R. Bangsberg

Research output: Contribution to journalArticlepeer-review

26 Scopus citations


Objectives: To estimate HIV incidence, characterize correlates of HIV seroconversion, and monitor temporal trends in HIV transmission among patients repeatedly tested for HIV by a county hospital in San Francisco. Design: Retrospective longitudinal study. Methods: HIV incidence was retrospectively calculated among persons voluntarily tested for HIV antibody more than once at San Francisco's county hospital or one of its affiliated satellite community clinics between 1993 and 1999. Linkage of HIV test results in computerized databases identified "seroconverters" as individuals who had a negative antibody test followed by a positive test. The interval between tests was used as the person-time at risk. Cox proportional hazards analysis identified correlates of HIV seroconversion. Results: A total of 84 HIV seroconversions were identified among 2893 eligible patients repeatedly tested for HIV antibody over a cumulative 5860 person-years (PYs) (incidence of 1.4 per 100 PYs, 95% confidence interval [CI]: 1.2-1.7). The majority of seroconversions (71 [84.5%]) were among injection drug users (IDUs) (incidence of 2.0 per 100 PYs, CI: 1.6-2.4). HIV incidence was highest among men who have sex with men (MSM) who were also IDUs (incidence of 3.8 per 100 PYs, CI: 2.7-5.1) and lowest among non-IDUs, heterosexual men, and non-IDU women (incidence of 0.3 per 100 PYs, CI: 0.1-0.6). In multivariate analysis, correlates of HIV seroconversion were age 25 to 29 years (hazard ratio [HR] = 3.9, CI: 2.4-6.3), MSM (HR = 2.9, CI: 1.9-4.4), and IDU (HR = 3.2, CI: 1.8-5.8). Overall, no temporal trend in annual HIV incidence was noted during the study period; however, HIV incidence among MSM IDUs increased from 2.9 per 100 PYs in 1996 to 4.7 per 100 PYs in 1998. Conclusions: The rate of seroconversion in this hospital and affiliated clinic population is unexpectedly high. Moreover, HIV transmission among IDU patients has not decreased over the last several years. The San Francisco county hospital provides a high-risk sentinel population to monitor emerging trends in HIV transmission, especially among IDUs, and presents multiple opportunities for prevention interventions, because these patients are being seen repeatedly by clinicians.

Original languageEnglish (US)
Pages (from-to)59-64
Number of pages6
JournalJournal of Acquired Immune Deficiency Syndromes
Issue number1
StatePublished - Sep 1 2001
Externally publishedYes


  • HIV incidence
  • Hospitals
  • Injection drug use
  • San Francisco
  • Sentinel surveillance

ASJC Scopus subject areas

  • Infectious Diseases
  • Pharmacology (medical)


Dive into the research topics of 'HIV incidence among repeat HIV testers at a county hospital, San Francisco, California, U.S.A.'. Together they form a unique fingerprint.

Cite this