Accuracy of Self-Reported HIV Testing Among Late Adolescents: Findings From a Sample of Young African-American Women Seeking Services at Sexual Health Clinics

Background: Self-reported behavior is the cornerstone of measurement in HIV/STD and prevention research, as well as guides clinical practice during sexual health visits, yet little is known about its accuracy. This study examined the accuracy of self-reported HIV testing among a sample of young African-American women seeking services at sexual health clinics in the Southeastern U.S. Methods: Data were collected via ACASI at 4 time points (every six months for 18-months) from 374 African-American adolescent females (14-20 years) seeking services at sexual health clinics. Self-reported HIV testing history, sociodemographic, family, psychosocial, and laboratory-confirmed STD results were collected during each assessment. Contingency tables determine consistency/accuracy of self-reported HIV testing over the 18-month period. Multivariate logistic regression identified predictors of inconsistent reporting. Results: Of the 374 participants who completed an ACASI at all 4 time points, 73 (19.5%) inconsistently reported having had an HIV test at Time 1 and subsequently reported they had never taken an HIV test at one or more of the other time points. Testing positive for a laboratory-confirmed STD (AOR = 2.59, P = .001), having older sex partners (AOR = 1.54, P = .017), and fear of negotiating condom use (AOR = 1.11, P = .008) were significantly predictive of inconsistent reporting. Discussion: Marked inconsistency was observed between self-reported HIV testing history across an 18-month time period. STD history, partner characteristics, and condom negotiation skill were predictors of inaccurate/inconsistent reporting. While self-report remains necessary, it may not be sufficient to accurately assess young women's HIV testing history.