20th International AIDS Conference - Melbourne, Australia


WEPE034 - Poster Exhibition

Can laypersons in high-prevalence South Africa perform an HIV self-test accurately?

M. Dong1,2, R. Regina2, S. Hlongwane2, M. Ghebremichael1, D. Wilson3, K. Dong1,2

1Massachusetts General Hospital, Ragon Institute of MGH, Harvard & MIT, Cambridge, United States, 2ITEACH SA, Pietermaritzburg, South Africa, 3Edendale Hospital, Medicine, Pietermaritzburg, South Africa

Background: HIV self-testing (HST) is receiving increased attention globally, with the recognition that new, cost-effective strategies are needed to achieve universal HIV testing. A number of studies of various designs and sample populations have investigated HIV self-testing, but this is the first to examine accuracy of unsupervised blood-based self-testing, including interpretation of participants'' own results, by laypersons of limited education from a resource-limited, high HIV prevalence population.
Methods: A cross-sectional study is currently being conducted with a target of 230 participants. As of January 2014, 195 enrollees from three representative communities in KwaZulu-Natal, South Africa have completed the study. Participants tested using a self-test kit that included commercially available rapid test components (finger-prick whole blood rapid test, lancet, pipette, alcohol swab) and illustrated instructions designed specifically for self-testing, supported by a telephone helpline. To assess accuracy of self-test performance, two confirmatory tests were performed for comparison: HCT performed by trained staff, and laboratory ELISA. To assess accuracy of result interpretation, participants and study staff read HST results independently.
Results: 193 (98.97%) out of 195 participants had a valid self-test result. Sensitivity and specificity of the self-test were 96.97% (CI: 84.18% to 99.49%) and 100% (95% CI: 97.67 to 100%), respectively. 192 (98.46%) out of 195 participants interpreted their result correctly (1 of 33 positive results was interpreted as negative, 1 of 2 invalid results was interpreted as negative, and 1 negative result was interpreted as “unsure of result”. 32.3% called the helpline for assistance during self-testing.
Conclusions: HST with a blood-based rapid test is highly accurate when performed by laypersons in representative high-burden, limited-education settings in KZN, South Africa, when packaged with illustrated instructions, accompanied by a telephone helpline. Blood-based rapid tests represent a cost-effective alternative to oral fluid tests and thus may be ideal for scale-up of HST in resource-limited settings, but safe disposal of infectious waste must be addressed. Additionally, working with manufacturers to customize components for a self-testing kit may improve both safety and ease of use.

Download the e-Poster

Back to the Programme-at-a-Glance