Evaluating Measures of HIV Risk Perception with a Mixed Method Approach Among Adolescent Girls and Young Women Using PrEP in Kenya

Introduction: Pre-exposure prophylaxis (PrEP) for HIV prevention is a promising HIV prevention option for young African women. The evidence of effective PrEP use in this population group is mixed, and has been tied to poor adherence. Discordance between assessed HIV risk and the actual experience of risk may underlie the dissonance observed between epidemiologic risk, risk perception, uptake of and adherence to prevention measures. We examined the contribution of research methods used in characterising HIV risk to this dissonance by contrasting HIV risk awareness as assessed by standard structured questionnaires and qualitative in-depth interviews. Methods: We conducted secondary analysis of data from a trial of African women aged 18 to 24 years interested in taking PrEP. Enrollment took place at two sites in Thika and Kisumu, Kenya, and participants were followed up for a period of 24 months between 2017 and 2020. We conducted a multi-step mixed methods assessment of HIV risk awareness and related decision-making behaviour. First, a quantitative analysis of survey data with descriptive statistics to characterize the cohort, proportion testing for trend and generalised estimating equations to assess the association between predictors of risk and risk perception. Secondly, an inductive content analysis of interview data for narrative themes; and lastly, joint display methodology to summarise findings from the two analytical methods. Results: We enrolled 350 young women with a median age of 21 years (IQR 19, 22). At baseline, 182 (52%) participants reported feeling at risk for HIV. In multivariable analysis, reporting a lot of HIV worry was significantly associated with higher odds of risk perception (OR 1.95 [CI 1.31, 2.91]), while significantly lower odds were associated with visits at month 12 (OR 0.30 [CI 0.18, 0.48]) and 24 (OR 0.21 [CI 0.09, 0.20]), living with other family or employer (OR 0.56 [CI 0.33, 0.95]) and being single with a steady partner (OR 0.48 [CI 0.33, 0.70]). Qualitative data from 75 serial interviews provided insight into the changes observed over time in report of both risk awareness and sexual behavioural choices. We identified three major themes from the combined qualitative and quantitative analysis: risk dynamism, behavioural risk patterns and the influence of the social environment. HIV risk awareness and decision making was both intentional and contextual; driven primarily by HIV literacy, PrEP use, assessment of a partner’s HIV risk profile and the self-agency to exercise held knowledge. Effective risk mitigation was contingent on intimate partner relationship dynamics and was hindered by the fear of intimate partner violence. Supportive social environments enabled disclosure and promoted the self-agency to effectively mitigate risk. Knowledge of partner status and PrEP use emerged as the primary determinants of risk awareness and related decision making, informing both risk perception and mitigation. Conclusion: We observed complementarity and contrast between the two quantitative and qualitative data in how HIV risk was defined and rationalised, and identified factors at the individual, partnership and societal levels that informed risk awareness and related decision-making. Interview data expounded on the quantitative findings and revealed a more dynamic and rationalised experience of risk than was observed in survey data. The experience of HIV risk by young African women was not haphazard, but rather was rationalised based on HIV and PrEP literacy, current knowledge of partner status and attitudes, and prevailing sexual and economic needs. HIV related decision making was dependent not just on having access to preventive measures, but also on the agency and wherewithal to actualize known mitigation measures. Further work on how knowledge of partner HIV status and PrEP use can be leveraged to support HIV prevention and other sexual reproductive health programs targeting young women is warranted.