RELIABILITY AND VALIDITY OF HIV SELF-REPORTED RESULT IN DETERMINING HIV PREVALENCE IN A REGION WITH HIGH HIV PREVALENCE: WESTERN KENYA

Background: Several studies have been done around HIV self-reported result, there is still a paucity of data on the reliability and validity of data from self-reported HIV result. We sought to assess the reliability and validity of self-reported HIV result in western Kenya, a region with high HIV prevalence. Methodology: We used a prospective cohort study design. Participants aged 15 years or older were recruited to the study using simple random technique. Eligible participants were taken through a behavioral interview and later followed up for HIV testing. Results: A total 5955 respondents were interviewed and later tested for HIV within one year, and 99.9% reported having an HIV test. Agreement between self-reported and actual HIV result was 96.9% with a Cohen's kappa of 0.869(95%CI= 0.8501 - 0.8874). Males (kappa = 0.874, 95%CI 0.841 - 0.907) presented slightly higher agreement than women (kappa = 0.866, 95%CI 0.843 - 0.889). The sensitivity of HIV self-reported result was 0.81, 95%CI (0.784 -0.836) and a specificity of 0.997, 95%CI (0.995-0.998) with an accuracy of 0.97, 95%CI (0.965-0.974). HIV self-reported result had a higher positive predictive value of 0.978, 95%CI (0.964-0.987). Knowing partner's HIV status, testing previously as couples and being 45 – 49 years of age had significantly increased odds of correctly reporting HIV status. Conclusion: Self-reported positive status provides an accurate measure of HIV status, hence self-reported sero-positivity should be treated as HIV positive for purposes of surveillance and equally for inclusion into interventions which require HIV positive individuals.