Oral fluid HIV self-test performance in rural South Africa
Deville et al.
Type of approach
Type of assistance
General population: General population in rural areas
UNAIDS HIV prevalence (2017)
The adequacy of packaging and labelling, direct sample collection, test performance, and reading and interpretation of oral fluid test results by untrained, unsupervised users was evaluated. Study participant oral fluid results were compared with the same test conducted by a blinded trained professional. Participants were recruited from an HIV Counselling and Testing outreach programme in the Moutse area, Limpopo.
Summary of findings
Of the total 1391 study participants, 113 were included in the Sensitivity Analysis Population and 1278 were included in the Specificity Analysis Population. Three percent (48/1391) of participants were excluded due to the interpretation by the participant of a self-test result as invalid (10%, 5/48), not sure/do not know (88%, 42/48), or refused/ambiguous (2%, 1/48). This resulted in a Test System Failure Rate of 3.45% (95% CI 2.56%-4.55%). Overall, the sensitivity in the untrained users was 99.02% (101/102) and specificity 100.0% (1241/1241). The 113 positive test results were confirmed by an HIV rapid test. 1193 participants (84.6%) had at least one observation of difficulty or error with one or more of the test steps. Only four tests did not provide a valid result.
Willingness to pay
Willingness to pay details
Overall, the sensitivity in the untrained users was 99.02% (101/102) and specificity 100% (1241/1241). 1193 participants (84.6%) had at least one observation of difficulty or error with one or more of the test steps. Only four tests did not provide a valid result.
Linkage to prevention, care and treatment