BACKGROUND: In Mali, artisanal gold-mining zones (AGMZ) are characterized by mass circular migration, convergence of key-populations and the lack of adapted healthcare services. This contributes to the potential risk of “bridging” infectious diseases including HIV. We aimed to assess the extent of HIV in AGMZ and the effectiveness of proximity community-based services for early antiretroviral treatment (ART) initiation.
METHODS: ANRS-12392/Sanu Gundo is an ongoing non-randomized interventional study launched in December-2020 in two Malian AGMZ (Diassa and Kofoulatiè). Community-based activities including HIV testing were offered during 5 months. Demographic and activity-related characteristics were recorded during HIV testing. People newly diagnosed were either offered with proximity community-based HIV services including psychosocial and peer-support activities (intervention group); or referred to public clinics for classic care (control group). Individual, behavioral and clinical characteristics were collected at different time points during 12 months. HIV prevalence was estimated. Effectiveness of the intervention on ART initiation one month after diagnosis (M1) was assessed using a Probit logistic model.
RESULTS: Overall, 9,785 people were tested for HIV. Prevalence in both Diassa (1.51% [1.22%-1.87%]) and Kofoulatiè (2.15% [1.76%-2.64%]) was higher than the national prevalence (0.8% [0.6%-1.0%], UNAIDS). Among the 170 people diagnosed with HIV, 81 (47.6%) were included in the control group and 89 (52.3%) in the intervention group. Data was available for 149 PLHIV at M1 (control: 70 and intervention: 79). Median age[IQR] was higher in the control group (40[31-50] years vs. 27[22-35], p<0.001). Median viral load[IQR] at diagnosis indicated poorer health in the control group (2,503[839-103,330]copies/mL vs. 1,238[839-27,906]copies/mL, p=0.039). Probit estimation without confounders showed that PLHIV in the control group were 70.6% (p<0.001) less likely to be on ART one month after diagnosis. Including covariates indicated that early ART initiation was more likely for older PLHIV (13%, p=0.008) and non-gold-miners (19%, p=0.006). The intervention effectiveness on early ART initiation persisted after adjusting for covariates (- 80.2% for control group, p<0.001).
CONCLUSIONS: The Sanu Gundo intervention was effective for early ART initiation to improve linkage-to-care often disrupted by cyclical migration and the lack of adapted healthcare services. Community-based approaches can effectively promote access and retention in care to limit HIV dissemination.

Download the e-Poster (PDF)