Background
In Mali, the HIV cascade is 59-53-unknown. People in artisanal small-scale gold mining zones (ASGMZ) are highly exposed to HIV and represent a potential “bridge” for infectious diseases including HIV. We aimed to assess the effects of proximity community-based HIV services on access, retention in care and health.
Methods
The ANRS-12392/Sanu Gundo non-randomized interventional trial was conducted in 2021-2022 in the Malian ASGMZ of Diassa (control) and Kofoulatiè (intervention). Community-based HIV testing was offered in both sites. People newly diagnosed in the intervention site received proximity community-based services including ART; whereas those in the control site were referred to conventional public hospitals. Individual and clinical characteristics including viral load (VL) were collected at different time points (J0, M1, M3, M6, M9 and M12). Outcomes were: time to linkage-to-care (i.e. time to ART initiation); retention in care at M12 (i.e. time to last visit); availability of VL assessment (yes/no); VL evolution at M12; and perceived health status (good/excellent vs. bad). Cox’s model was used for linkage-to-care and retention in care. VL assessment and perceived health status were analyzed using random-effects model to control for repeated measures. Sampling bias between arms was corrected using inverse propensity score weighting. All analyses compared arms (control/intervention) and were adjusted for individual characteristics.
Results
9786 HIV tests were carried out (control site: 56.7% and intervention site: 43.2%). The control group included 81 HIV+ participants, and 89 were in the intervention group. HIV prevalence of 1.5% 95%CI[1.2-1.9] and 2.2% 95%CI[1.8-2.6] were estimated for the control an intervention sites, respectively (0.8%[0.7-1.1] at national level). Linkage-to-care was faster in the intervention group (aHR:1.8; 95%CI:1.03-3.2). Retention in care at M12 was shorter in the intervention group (aHR:2.1; 95%CI:1.3-3.5). No difference was found between arms concerning VL assessment (p=0.4), nor the evolution of VL (p=0.5). Participants in the intervention group were less likely to perceive good/excellent health status (aOR:0.41; 95%CI:0.23-0.73).
Conclusions
Community-based services improved HIV testing and linkage-to-care in ASGMZ. Retention and availability of viral load assessment did not differ between groups. Coordination of community-based and conventional public services would ensure the continuum care and contribute to the improvement of the HIV cascade.