Artificial Intelligence Causal Inference of Community Clinic Malaria Rapid-Test Stock Routing Under Unmeasured District Resupply Gaming: A Healthcare Instrumental Variable Stockout Evaluation

Community-clinic malaria rapid diagnostic test (RDT) stockouts are routinely attributed to consumption and lead-time variation, yet district resupply officers who reallocate stock across clinics in response to informal signals create a form of unmeasured confounding that standard stockout regressions cannot separate from true need. The design is evaluated analytically and by replication across 2000 design cells under a declared variance structure with between-cluster SD = 6.0, residual SD = 4.5 and intraclass correlation = 0.64. Power was estimated for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points; 54 clusters are required for 80% power at a 5-percentage-point effect. The contribution is a design-based identification argument and a power calculus that make explicit the cluster counts at which route-based instruments separate gaming from need, together with the boundary conditions under which the exclusion restriction fails.