Fig. 4: Validation analysis using IPTi trial data.

All panels show the predicted proportion infected over time since the SP dose among those at risk of new infection for the Macete et al. trial²⁰ (a, b) and the Gosling et al. trial¹² (c, d). In panels a and c, dots represent the observed proportion of infections, with error bars denoting 95% binomial confidence intervals (Clopper-Pearson method) around the observed proportion. The sample size for plots a and c was 1497 for Mozambique (SP = 747 and placebo = 750) and 639 for Tanzania (SP = 319 and placebo = 320). Solid lines and shaded areas denote the model fit to the data (posterior median) and associated 95% Credible Intervals. Dashed lines show the predicted proportion infected given the estimated protection parameters from the main analysis and estimated frequency of each genotype (56% dhps GEA and 44% dhps AKA in the Mozambique trial²¹, and 53.6% dhps GEA and 41.1% dhps GEG in Northern Tanzania¹³). Panels b and d show the predicted proportions of infection with each genotype for each treatment. c dhps: dihydropteroate synthase.