Abstract
INTRODUCTION: Lack of universal, annual testing for human immunodeficiency virus (HIV) in health facilities suggests that expansion of HIV testing and counselling (HTC) to non-clinical settings is critical to the achievement of national goals for prevention, care and treatment. Consideration should be given to the ability of lay counsellors to perform home-based HTC in community settings. METHODS: We implemented a community cluster randomized controlled trial of home-based HTC in Sisonke District, South Africa. Trained lay counsellors conducted door-to-door HIV testing using the same rapid tests used by the local health department at the time of the study (SD Bioline and Sensa). To monitor testing quality and counsellor skill, additional dry blood spots were taken and sent for laboratory-based enzyme-linked immunosorbent assay (ELISA) testing. Sensitivity and specificity were calculated using the laboratory result as the gold standard. RESULTS AND DISCUSSION: From 3986 samples, the counsellor and laboratory results matched in all but 23 cases. In 18 cases, the counsellor judged the result as indeterminate, whereas the laboratory judged 10 positive, eight negative and three indeterminate, indicating that the counsellor may have erred on the side of caution. Sensitivity was 98.0% (95% CI: 96.3-98.9%), and specificity 99.6% (95% CI: 99.4-99.7%), for the lay counsellor field-based rapid tests. Both measures are high, and the lower confidence bound for specificity meets the international standard for assessing HIV rapid tests. CONCLUSIONS: These findings indicate that adequately trained lay counsellors are capable of safely conducting high-quality rapid HIV tests and interpreting the results as per the kit guidelines. These findings are important given the likely expansion of community and home-based testing models and the shortage of clinically trained professional staff.