P089 Accuracy of Interpretation and Home Test Kit Result Reporting for Screening of Human Immunodeficiency Virus Infection

Background Men who have sex with men (MSM) and individuals identifying as Black or Hispanic/Latino experience the largest burden of Human Immunodeficiency Virus (HIV) infection. The OraQuick In-Home HIV Test (OraSure Technologies©, Pennsylvania, USA) is the only HIV self-test approved by the US Food and Drug Administration. Self-testing can supplement HIV prevention to increase identification of infections among at-risk groups. During a larger study on social media and HIV prevention, we assessed the accuracy of participants’ interpretations of their results. Methods We recruited Black and Latino MSM between 18–30 years through advertisements on internet-based social media, informational sites, and dating sites. Participants ordered a free OraQuick self-test. They tested and interpreted their results following kit instructions. Participants submitted a test kit photograph to a secure online platform to report their results. Two trained researchers reviewed the photographs and interpreted results independently. We calculated the proportion of agreement and kappa coefficient between reviewers and between reviewers and participants. Results We enrolled 271 participants, 191 (70%) ordered a kit, 159 (83%) used it. Of those, 113 (71%) submitted readable test result images. Among those submitting images, 71.3% were Black and 30.4% were Hispanic/Latino, mean age 25 years (SD 3.6). The proportion of agreement in result interpretation between reviewers was 100% (113/113), kappa coefficient 1.0. The proportion of agreement in result interpretation between participants and reviewers was 97.3% (110/113). Of the concordant results, 101 (89.4%) were negative, 7 (6.2%) were positive, 2 (1.8%) were invalid. Of the 3 discordant results, 2 participants interpreted their positive results as invalid and 1 participant interpreted their positive result as negative. The kappa coefficient was 0.85 (95% CI 0.67–1.0). Conclusion Most participants submitted results online. The level of agreement of interpretation between participants and researchers was good. Prevention programs could use self-testing during disruptions of care.