Preferences and Usability of self-test kits for human immunodeficiency virus (HIV) among gay and bisexual men (PUSH study)

This was a non-blinded randomised crossover trial with equal numbers of participants randomised to two arms using a computer-generated sequence created by an independent biostatistician. Men were recruited from participating sexual health clinics and asked to test themselves using two HIV self-testing (HIVST) kits (one oral, and one blood-based) in a private room at the clinic following manufacturer-supplied written, graphic, and/or video instructions. One arm received the oral HIVST first, followed by the blood-based HIVST, and the other the reverse. A study nurse observed the participants during self-testing to note if they performed all the necessary steps correctly or made any errors as per a pre-designed checklist. The study nurse did not provide any instructions or advice on how to do the tests to mimic real-world conditions. The oral test required participants to swab their upper and lower gums, and the blood-based test required participants to prick their finger using a built in lancet and collect a drop of blood. No time limits were placed on viewing instructions or conducting the tests, however the study nurses noted if the test results were read in the recommended time period. During the time when the tests were incubating (15-20 minutes) participants were asked to complete part 1 of a questionnaire that captured socio-demographic characteristics, HIV (including conventional and self-testing) and sexually transmitted infection (STI) testing history, and sexual behaviour. After self-testing, participants completed part 2 of the questionnaire which captured ease of use for each self-testing kit overall and for the different steps involved, and ranking of self-tests in the order of preference. The entire process was completed in approx 45-60 minutes and there was no washout period between tests.