RESUMEN
OBJECTIVES: To compare the usability and acceptability of oral fluid- and blood-based HIV self-test kits among men who have sex with men in Australia. DESIGN: Randomised crossover trial. SETTING, PARTICIPANTS: Gay, bisexual, and other men aged 18 years or older who have sex with men, who attended two metropolitan sexual health clinics in Sydney and Melbourne, 7 January - 10 December 2019. MAIN OUTCOME MEASURES: Ease of use of HIV self-test kits; preferred HIV self-test type; difficulties encountered during HIV self-testing. RESULTS: 170 men were recruited (median age, 34 years; interquartile range, 29-43 years); 144 identified as gay (85%), 96 were born outside Australia (57%). Participants were more likely to report the oral fluid HIV self-test was easy to use than the blood-based self-test (oral fluid, 99%; blood, 86%; odds ratio [OR], 3.0; 95% confidence interval [CI], 1.4-6.6). The oral fluid test was preferred by 98 participants (58%; 95% CI, 50-65%), the blood-based test by 69 (41%; 95% CI, 33-48%). Difficulties with the oral fluid test kit identified by observing nurses included problems placing the buffer solution into the stand (40 of 170 participants, 24%) and not swabbing both gums (23 of 169, 14%); difficulties with the blood-based test kit included problems filling the device test channel (69 of 170, 41%) and squeezing the finger firmly enough to generate a blood drop (42 of 170, 25%). No participant received an invalid result with the oral fluid self-test; two of 162 participants (1%) received invalid results with the blood self-test. After adjusting for age, education level, and ethnic background, characteristics associated with higher odds of using HIV self-testing in the future were overseas birth (adjusted OR, 3.07; 95% CI, 1.42-6.64), and self-evaluated ease of use and confidence in using the kits. CONCLUSION: It is important to provide options for obtaining both oral fluid- and blood-based HIV self-tests. The usability and acceptability of both kits were high, but the ease of use and perceived accuracy influenced test kit preference.