Artificial Intelligence Counterfactual Triage of Obstetric Referral Transport Under Variable Road Network Imputation Completeness: A Sensitivity-Bounded Evaluation of Healthcare Access Equity

Obstetric referral transport in low-resource settings depends on road networks that are incompletely mapped, and triage algorithms trained on such networks may encode access inequities rather than resolve them. The framework separates two sources of bias: missingness in the road graph and confounding in the referral decision. A sensitivity bound is derived that expresses the maximum equity loss attributable to imputation incompleteness as a function of the intraclass correlation of access outcomes across clusters. The design evaluation used a declared variance structure with between-cluster standard deviation 6.0, residual standard deviation 4.5, and intraclass correlation 0.64, 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. Power reached conventional levels at 54 clusters for a 5 percentage point effect. The framework proposes that equity-aware triage must be evaluated against imputation-conditional bounds rather than point estimates of access gain, and the design analysis indicates that cluster counts below 54 cannot reliably distinguish equitable from inequitable triage at policy-relevant effect sizes.