Influence of attendance at a clinic service for high-risk women on HIV prevention: A longitudinal study in Kampala, Uganda

Background: HIV/AIDS continues to be a major public health threat. Globally, there has been notable progress towards reaching people for HIV testing and treatment, as well as in the development of effective prevention tools. But preventing new HIV infections remains a key challenge, particularly among populations at the most risk including female sex workers and their sexual partners. Reducing HIV transmission in these populations requires combined HIV prevention options not only to be available, but also sufficiently used over time. However, regular attendance at clinic facilities, and use of HIV services by female sex workers is affected by persistent and common barriers including high mobility, stigmatisation, and discrimination in health-care settings. The aim of this thesis is to examine the effect of irregular exposure to a package of HIV prevention interventions among women at high risk of HIV in Uganda. Methods: A cohort of HIV-negative women at high risk of HIV was scheduled to attend dedicated clinic services once every three months within the Good Health for Women Project in Kampala, Uganda. An initial cohort of women was enrolled between April 2008 and May 2009 and followed-up (cohort 1). Alongside cohort 1, a second cohort was enrolled from January 2013, and followed-up (cohort 2). At each scheduled clinic visit, participants were offered a combination HIV prevention package that included risk reduction counselling, STI management, and HIV testing. Based on these data, I analysed (i) trends in HIV incidence following enrolment and the influence of missed scheduled visits on HIV incidence trends. In addition, multiple imputation of time-to-event was used to assess the influence of attrition over time on incidence; (ii) the association of the number of missed scheduled visits with subsequent HIV risk; (iii) the association of the number of missed scheduled visits with subsequent proximate determinants of HIV risk (sexually transmitted infections (STIs), alcohol use, inconsistent condom use with paid sex). The analyses were censored at 29th August 2017. Results: Of the 3084 HIV-negative women enrolled, 2206 (71.5%) had at least one followup visit. HIV incidence declined rapidly following enrolment from 6.1/100 person-years in ≤6 months to 2.0/100 at year 3 in cohort 1 (p-value<0.001), and from 3.8/100 to 1.8/100 in cohort 2 (p-value=0.04). HIV incidence was associated with the prior number of missed scheduled visits (adjusted hazard ratio=1.40; 95%CI 0.93-2.12 for 1-2 missed visits versus none, and 2.00 (95%CI 1.35-2.95) for ≥3 missed visits versus none; p-trend=0.001). Missed visits were associated with increased detection of STIs, and increased reporting of daily 3 alcohol use at the subsequent visit attended, but not with reported consistency of condom use with paid sex. Conclusion: The increased HIV risk when high-risk women interrupted their regular exposure to combination HIV prevention was most likely attributed to untreated STI. But the absence of a concurrent change in pattern of inconsistent condom use with paid sex could additionally suggest influence of contextual factors on HIV risk. HIV programmes need to routinely monitor the individual-level exposure to and use of complementary HIV interventions among high-risk women, as well as actively facilitate their attendance of effective intervention packages at regular intervals. For those unable to regularly attend services at health facilities, HIV programmes must adopt innovative ways to deliver combination intervention packages, incorporating STI management and risk reduction counselling and using mobile clinics, clinic vouchers, or even leveraging technology.