Artificial Intelligence Triage for Obstetric Hemorrhage Under Variable Blood Bank Cross-Match Latency: A Distributionally Robust Evaluation of Healthcare Transfusion Escalation Readiness

Obstetric hemorrhage remains a leading cause of maternal mortality, and artificial intelligence triage tools that recommend transfusion escalation must operate under blood bank cross-match latencies that vary unpredictably across facilities and over time. The framework is theoretical and methodological: no observational data are used or claimed. The analysis characterises the operating characteristics of a before-after-control-impact design under a declared variance structure with between-cluster SD = 6.0, residual SD = 4.5, and intraclass correlation = 0.640. Power was estimated across 2000 design cells for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The design required 52 clusters to detect a 5-percentage-point effect at 80% power. The framework proposes a tractable robust policy, identifies the conditions under which distributional ambiguity reverses the ranking of escalation strategies, and specifies the design parameters under which a cluster-level evaluation can credibly detect readiness improvements. The contribution is a formal bridge between distributionally robust optimisation and the practical evaluation of AI triage in transfusion-dependent obstetric care.