Karibu na sirenji (Near the Syringe): Applying mixed methods to characterize the HIV risk environment and gender differences among persons who inject drugs in Nairobi, Kenya

In Kenya, people who inject drugs (PWID) are considered a key population that are disproportionately affected by the HIV and hepatitis C (HCV) epidemics, with prevalence estimates reaching upwards of 19-25% and 11-36%, respectively. Kenya’s national HIV program scale-up for key populations has resulted in significant reductions in HIV incidence among PWID, but service gaps remain with only 43% of PWID aware of their HIV status, 68% of PWID living with HIV (PWID-LH) taking antiretroviral treatment (ART), and 64% of those on ART are virally suppressed. HIV care outcomes improved with the introduction of opioid agonist therapy (OAT) clinics that provide integrated methadone and HIV treatment services; however, OAT service access among PWID remains low (26%). Ecological frameworks, like the Modified Social Ecological Model and HIV risk environment frameworks, provide a holistic approach to understanding HIV, HCV and drug-related outcomes by looking beyond individual-level factors to examine the surrounding structural factors (e.g., political, economic, community, social) that are largely outside of an individual’s control, but often influence individual-level risk behaviors (e.g., syringe sharing, sexual risks, engagement in care). While ecological frameworks have been applied in other PWID-LH settings, their application has been limited in Kenya, yet, may be beneficial to current HIV strategies. To better understand environmental influences that affect HIV risk and service uptake, we applied a mixed methods design to (1) qualitatively characterize HIV and HCV barriers and facilitators to care among PWID through the lens of peer educators, including resource recommendations to improve service uptake; (2) qualitatively describe the Nairobi-specific HIV risk environment surrounding PWID-LH; and (3a) quantitatively identify distinct subgroups of PWID-LH based on demographic and risk environment factors using cluster analysis techniques in order to (3b) assess whether empirically derived clusters are associated with suboptimal care (i.e., discontinued care or treatment, virally unsuppressed) among PWIDLH