Artificial Intelligence Causal Effect Estimation of Artemisinin Combination Therapy Adherence Under Variable Mobile Health Reminder Delivery Logging: A Target Trial Emulation Evaluation of Healthcare Resistance Selection

Variable logging of mobile health (mHealth) reminder delivery complicates causal inference about artemisinin combination therapy (ACT) adherence and its downstream selection of drug-resistant Plasmodium falciparum. The design assumed a between-cluster standard deviation of 6.0, a residual standard deviation of 4.5, and an intraclass correlation of 0.640. Power was estimated across 2000 replications per design cell for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The analysis identified 52 clusters as the number required to detect a 5 percentage point effect at 80% power. The framework specifies how inverse-probability-of-logging weights, clone-censor-weight estimation, and g-computation can be embedded within a BACI contrast to recover adherence effects that are robust to reminder-logging incompleteness, and it links those effects to a resistance-selection criterion defined over the distribution of sub-therapeutic drug exposure. The contribution is a formalised design and a power-calibration result that together specify the conditions under which mHealth reminder programmes can be causally evaluated for resistance-relevant adherence outcomes.