HIV self-test performance evaluation among priority populations in rural Mozambique: Results from a community-based observational study

Background In 2021, Mozambique initiated community-based oral HIV self-testing (HIVST) to increase testing access and uptake among priority groups, including adult males, adolescents, and young adults. Within an HIVST pilot project, we conducted a performance evaluation assessing participants' ability to successfully conduct HIVST procedures and interpret results. Methods A cross-sectional study was performed between February-March 2021 among employees, students (18­24 years of age), and community members, using convenience sampling, in two rural districts of Zambézia Province, Mozambique. We quantified how well untrained users performed procedures for the oral HIVST (Oraquick®) through direct observation using a structured checklist, from which we calculated an HIVST usability index (scores ranging 0­100%). Additionally, participants interpreted three previously processed anonymous HIVST results. False reactive and false non-reactive interpretation results were presented as proportions. Bivariate analysis was conducted using Chi-square and Fisher exact tests. Results A total of 312 persons participated (131[42%] community members, 71[23%] students, 110[35%] employees); 239 (77%) were male; the mean age was 28 years (standard deviation 10). Average usability index scores were 80% among employees, 86% among students, and 77% among community members. Main procedural errors observed included "incorrect tube positioning" (49%), "incorrect specimen collection" (43%), and "improper waiting time for result interpretation" (42%). From the presented anonymous HIVST results, 75% (n = 234) correctly interpreted all three results, while 9 (3%) of study participants failed to correctly interpret any results. Overall, 36 (12%) gave a false non-reactive result interpretation, 21 (7%) a false reactive result interpretation, and 14 (4%) gave both false non-reactive and false reactive result interpretations. Community members generally had lower performance. Conclusions Despite some observed testing procedural errors, most users could successfully perform an HIVST. Educational sessions at strategic places (e.g., schools, workplaces), and support via social media and hotlines, may improve HIVST performance quality, reducing the risk of incorrect interpretation.