Background South Africa has about 11% of the total population living with HIV, the largest to date for any country. Facility-based HIV testing has reached only 50% South Africans because of fear of visibility leading to stigma, embarrassment and discrimination. Alternative strategies like self-testing for HIV may improve engagement, but evidence is limited. For self-testing to be successful, knowledge regarding the process, clear instructions about how to conduct, interpret and seek linkages to counselling and staging is essential.

Methods We created an internet-based HIV self-testing programme with a popular oral HIV test. The programme had built-in content for counselling, personal risk staging, instructions to self-test, and to seek counselling and referral. We also created an equivalent paper version and evaluated both programmes in 251 health care professionals working at University of Cape Town, South Africa. The tested internet programme was converted into an interactive, engaging smartphone HIV self-test application. The application was piloted for design, content and comprehension in 12 young adults (aged 18–25 years).

Results Internet and paper-based self-testing programmes were well received (91.3%) by participants with overall preference for self-testing reported at 100%. User feedback on the smartphone application was incorporated after pilot evaluation and the following were improved: (a) a user centred design and layout, (b) colourful interface with clear instructions, (c) clarity of content for comprehension, (d) built-in features for expanded access, and (e) overall presentation. After six iterations, a prototype Android application was developed.

Conclusion High preference to self-test facilitated the use of the internet and paper-based programmes. This indicates that if validated self-tests are presented with clear instructions to self-test and built-in confidential linkages to counselling and treatment are provided, many more individuals will opt for HIV self-testing. These programmes and the smartphone application will be useful for the scale-up of unsupervised self-testing initiatives in literate populations worldwide.