Artificial Intelligence Doubly Robust Inference of Community Clinic Hepatitis B Birth-Dose Administration Under Unmeasured Cold-Chain Failure Confounding: A Healthcare Instrumental Variable Safety Evaluation

Hepatitis B birth-dose administration within 24 hours of delivery is the single most effective intervention against perinatal transmission, yet community clinics in low-resource settings face unmeasured cold-chain failures that confound naive comparisons of administration and seroconversion. The estimator combines an outcome regression with an inverse-probability-weighted instrumental variable moment condition, achieving consistency if either the treatment or the outcome model is correctly specified. A before-after-control-impact design is characterised under a declared variance structure with between-cluster standard deviation 6.0, residual standard deviation 4.5, and intraclass correlation 0.64; 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 design requires 52 clusters to detect a 5 percentage point effect at 80 percent power. The framework provides a defensible inferential route for safety evaluation where randomisation is infeasible and cold-chain failure cannot be directly measured.