Artificial Intelligence Drug Interaction Alerting Under Variable Pharmacy Dispensing Log Granularity: A Causal Mediation Evaluation of Healthcare Override Patterns

Clinical decision support systems that flag potential drug-drug interactions depend on pharmacy dispensing logs whose temporal granularity varies across institutions, yet the consequences of that variation for alert burden and clinician override behaviour remain poorly characterised. The framework is operationalised as a before-after-control-impact design with a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64), and its operating characteristics are characterised 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 analysis indicates that 52 clusters are required to detect a 5 percentage point effect at 80% power. The contribution is a design-theoretic account of how log granularity enters the causal chain from dispensing record to override decision, together with the identification conditions under which mediation can be separated from confounding by institution-level prescribing culture.