A mixed methods investigation of implementation barriers and facilitators to a daily mobile phone sexual risk assessment for young women in Soweto, South Africa

Background The HIV epidemiology in South Africa reveals stark age and gender disparities, with young women being the most vulnerable to HIV acquisition in 2017. Evaluation of HIV exposure is a challenge in HIV prevention research. Intermittent in-clinic interviewer-administered risk behaviour assessments are utilised but may be limited by social desirability and recall biases. We piloted a mobile phone application for daily self-report of sexual risk behaviour in fifty 18-25 year old women at risk of HIV infection enrolled in HIV Vaccine Trials Network 915 (HVTN 915) in Soweto, South Africa. Through a mixed-methods investigation, we explored barriers and facilitators to completing daily mobile phone surveys among HVTN 915 study participants and staff. Methods We analysed quantitative data on barriers and facilitators to mobile phone study completion collected during the larger HVTN 915 study as well as two post-study focus group discussions (FGDs) with fifteen former participants with a median age of 24 years (IQR 23-25) and six individual in-depth interviews (IDIs) with HVTN 915 staff. FGDs and IDIs utilised semi-structured interview guides, were audio-recorded, transcribed verbatim and translated to English. After coding, thematic analysis was performed. Results The main facilitator for daily mobile phone survey completion assessed across 336 followup visits for 49 participants was the daily short message system (SMS) reminders (93%, 312/336). Across 336 visits, 31/49 (63%) retained participants reported barriers to completion of daily mobile phone surveys: forgetting (20%, 12/49), being too busy (19%, 11/49) and the survey being an inconvenience (15%, 9/49). Five main themes were identified during the coding of IDIs and FGDs: (1) facilitators of mobile phone survey completion, such as daily SMS reminders and follow up calls for non-completers; (2) barriers to mobile phone survey completion, including partner, time-related and technical barriers; (3) power of incentives; (4) response bias in providing sensitive information, and (5) recommendations for future mobile phone based interventions. Conclusion Despite our enthusiasm to use innovation to optimise sexual risk assessments, technical and practical solutions are required to improve implementation. We recommend further engagement with participants to optimise this approach and to further understand social desirability bias and study incentives in sexual risk reporting.