Answers are grounded in this paper and cite their sources.
Benchmarking is commonly used in health care to compare performance among providers. However, medication safety has largely been exempt from this type of benchmarking: using voluntarily reported medication safety events for cross-institutional performance comparisons is discouraged because these data are not directly comparable. The member hospitals of the Alliance of Dedicated Cancer Centers (ADCC) launched a Patient Safety Organization (PSO)-supported Medication Quality and Safety Dashboard project to examine whether this cohort was similar enough that benchmarking could illuminate medication safety vulnerabilities and best practices. In this retrospective analysis, the authors evaluated institutional demographics, pharmacy volume metrics, and deidentified medication safety event reports from existing systems; developed a severity crosswalk; standardized key event definitions; normalized event rates per 10,000 doses dispensed or administered; and calculated a near-miss event (NME) ratio as a marker of reporting culture. Despite these efforts, the study revealed that heterogeneity in reporting platforms, system configurations, data capture practices, and institutional size produced wide variability and significant outliers in event volumes and rates, precluding meaningful cross-institutional performance comparisons. The authors' findings reaffirm long-standing cautions against benchmarking voluntarily reported event data across institutions, even among seemingly similar cancer centers. However, they found practical value from internal use of the NME ratio to assess reporting culture and error interception, and from collaborative work to standardize taxonomy for medication safety events that goes beyond the medication-use phase and medication type. They identified best practices for using internal medication-related safety data. These lessons can be applied to the incorporation of AI as more tools and literature support become available.
No takes yet. Share an insight, caveat, or question.
Michaud et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: