Quantifying Global Heterogeneity in the Causal Effects of Digital Medication Systems on Error Propagation: A Bayesian Counterfactual Study

Abstract: - Digital medication systems are widely adopted to improve medication safety, yet their effectiveness varies substantially across healthcare settings. This study develops a hierarchical Bayesian counterfactual framework to quantify global and context-specific causal effects of digital medication systems on medication error propagation across sequential care pathways. Medication processes are modeled as structured causal networks linking prescribing, verification, dispensing, administration, and monitoring. Results demonstrate a significant global reduction in cumulative error propagation under digital exposure; however, pronounced heterogeneity is observed across institutions. Three distinct causal regimes emerge: digitally aligned environments achieving comprehensive error attenuation, partially aligned settings exhibiting mediation-dependent improvements, and misaligned systems showing negligible response. Upstream medication processes account for the majority of safety gains in aligned institutions, while downstream effects depend on workflow coherence and organizational readiness. Nonlinear threshold behavior reveals that meaningful benefits arise only after critical alignment levels are reached. These findings reposition digital medication systems as context-sensitive structural modifiers rather than universal safety solutions and highlight the necessity of precision implementation strategies tailored to local sociotechnical conditions.