Text

Introduction: ????Medication errors are a leading cause of preventable harm in hospitals, with pharmacists and physicians central to risk mitigation via medication review and prescription approval. However, variability in these practices and documentation hinders quality improvement and research. Routine clinical data offer research potential but are subject to inconsistencies and confounding [1], [2]. To prepare for a multicenter interventional study, we modeled the standard-of-care medication management workflows across 14 German university hospitals, focusing on pharmacist-physician interactions to identify confounders and define key dependent variables.

Methods: Fourteen university clinics engaged in structured workshops to map their medication management processes using Business Process Model and Notation (BPMN) and Decision Model and Notation [3]. Workflows from prescribing through pharmacist review, physician approval, and dispensing were diagrammed, with all relevant data inputs and outputs (e.g., prescriptions, labs, medication histories, clinical notes) linked to each step. Iterative alignment of individual models yielded a consolidated BPMN standard-of-care model. Variations in workflow timing and data availability were cataloged as potential confounders, while variables tied to process outcomes were earmarked as dependent variables for the planned study.

Results: The finalized BPMN model captured end-to-end medication management across all sites, detailing sub-processes such as admission medication reconciliation and electronic prescription verification. Data flow mapping revealed that critical information – like updated medication lists and allergy records – became available primarily post-pharmacist review, with slight inter-site timing differences. Key confounders included variability in routine pharmacist-led reviews and differences in electronic prescribing systems. We identified dependent variables, including the count and resolution rate of drug-related problems, time to order approval, and documentation completeness at discharge. Documented data-quality issues, such as incomplete medication lists at care transitions, further informed model refinement.

Conclusion: Collaborative BPMN modeling effectively standardized medication management processes across multiple university clinics, enhancing understanding of interdisciplinary workflows and data interoperability. Clearly identifying critical data points, potential confounders, and dependent variables significantly improves readiness for conducting robust multicenter interventional studies utilizing routine healthcare data. This structured process modeling approach underscores its utility in improving data quality, comparability, and ultimately patient safety through standardized medication management practices.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.

References