Building Key Populations HIV Treatment Cascades in Data-Scarce Environments: Towards a participatory stakeholder methodology for cascades construction, adoption, and utilization
IntroductionRecent HIV key populations (KP) surveillance studies in South Africa, including female sex workers (FSW) and men who have sex with men (MSM), demonstrate the disproportionate burden of HIV they bear compared to the general population. The national response for KP has lagged due to relatively scarce KP data focused narrowly on urban areas. We adopted a participatory data triangulation approach with stakeholders to overcome the challenges of KP program planning in KP data-scarce districts. Here we describe our methodology for achieving consensus on population size estimates (PSE) and treatment cascade indicator estimates derived from FSW and MSM surveillance data and applied across the country.\n\nMethodsThe South African National AIDS Council (SANAC) convened the group; technical advisors from the University of California San Francisco (UCSF) facilitated; and stakeholders from government, non-government, academic, and KP-led advocacy organizations involved in program implementation and research provided input through three in-person meetings covering four phases of work: surveillance data analysis; cascades data extraction; presentation for feedback; and data extrapolation.\n\nResultsTechnical advisors presented eight cascades (three FSW, five MSM) to stakeholders, recommending data-informed extrapolation factors for each population. Stakeholders adopted recommendations by consensus with few adjustments. FSW cascades displayed high awareness of HIV status and steep breakpoints towards ART uptake; MSM cascades displayed less HIV status awareness, but relatively good ART uptake, with metropolitan areas displaying better uptake than rural districts.\n\nConclusionThe participatory process enabled KP stakeholders to vet disparate data sources against programmatic experience and recommend consistency in cascades data; participatory triangulation of additional surveillance and program data will follow. The considerable time and resource investments in this process had downstream benefits, including consistency in sub-national HIV implementation plans. We recommend this consensus-based approach as a transparent, consistent, and sound methodology for cascades construction in KP data-scarce environments.