Fig. 6: Transported effect estimates of MDA to non-trial areas during the intervention year, including covariates used in the transportability analysis.

a, Covariates used in the transportability analysis for each commune. Coloured points indicate covariates included in the transportability models after screening for collinearity, data sparsity, association with malaria case counts and feature selection using elastic net regression. Blue- and yellow-shaded points indicate the standardized mean difference for a given covariate in a given commune, calculated as the mean in a non-trial commune minus the mean in the trial site divided by the pooled s.d. White points indicate covariates that did not pass the screening. b, Transported effect for each commune (coloured points), expressed as the percentage reduction in malaria incidence (1 − the ratio of cases in the MDA versus control arm) during the transmission season of the intervention year, using doubly robust transportability models. Analyses were restricted to communes with a population size of ≤152 people per 100 m², those with a predicted probability of trial participation of >0.75, and communes where SMC was offered during the trial period. Separate analyses were performed for each commune using monthly data from July to December (n = 366 per commune). Analyses used doubly robust transportability models. Vertical coloured lines indicate the 95% CI for each commune, obtained from a non-parametric bootstrap that resampled commune-months 1,000 times with replacement. Communes are sorted according to the transported effect size. The grey solid line indicates the original trial estimate (55%; 95% CI: 28%, 71%). Asterisks indicate CIs that were truncated for data visualization (lower bound < −200% and/or upper bound > 100%).