Fig. 4. Counterfactual modelling scenarios predicting the effect of interventions to increase HIV suppression in men on incidence reductions in women.

a,b, Estimated additional number of men with HIV in the census-eligible population in round 18 that already had suppressed virus (light grey), those who would have achieved viral suppression in the counterfactual intervention scenarios (colour), and those who would have remained with unsuppressed virus in the counterfactuals (dark grey) (posterior median: bars; 95% credible interval: error bars). c, Reduction in incidence in women of the census-eligible population in round 18 under the counterfactual targeted scenarios (posterior median: bars; 95% credible interval: error bars). d, Estimated incidence rates among women in the census-eligible population in round 18 (black solid line) and the counterfactual scenarios (colour), with incidence rates among men in round 18 shown as reference (black dashed line) (posterior median: lines; 95% credible intervals: ribbons). Throughout all subfigures, estimates are based on data from n = 15,053 participants in survey round 18, including n = 110 individuals in the incidence cohort in round 18, n = 432 individuals with HIV and with measured viral load in round 18, and n = 61 heterosexual source–recipient pairs in rounds 16–18, and information inferred through hierarchical models from all individuals in earlier rounds.