SHM Converge Scientific Abstract ePoster Gallery-Site-Level Analysis of Diagnostic Process Failures: A Novel Quality Improvement Tool

Pdf Summary

This abstract describes a systems engineering method for helping hospitals identify their own diagnostic safety problems. Diagnostic errors are common in hospitalized patients, especially among those who experience rapid response events or ICU transfers, but most hospitals lack a practical way to analyze where diagnostic process failures occur locally.<br /><br />Using data from the AHRQ-funded UPSIDE study, the authors examined confirmed diagnostic error cases from 29 hospital sites. They mapped failures across the Diagnosis Error Evaluation and Research (DEER) taxonomy, then reweighted the events to account for differences in case sampling across hospitals. The data were normalized by both the number of failure points in each domain and the number of cases per institution. This allowed comparison of local patterns with system-wide trends.<br /><br />The team visualized the results using heat maps and Pareto diagrams. At the system level, the most frequent failure domains were Assessment, Test Ordering and Interpretation, History-taking, and Patient Monitoring. Assessment was also a major contributor at the site level, but other domains varied significantly across hospitals, including Testing, Consultation, and Physical Exam. This variation suggests that individual hospitals may have different diagnostic weakness patterns even within the same overall network.<br /><br />The authors conclude that their approach offers a practical tool for hospitals to benchmark local diagnostic failures against broader data, prioritize improvement efforts, and design targeted interventions. Future work will test whether interventions guided by this analysis can reduce diagnostic errors.

Asset Subtitle

Peter Barish

Meta Tag

Author List Andrew D. Auerbach, Anuj K. Dalal, Jeffrey L. Schnipper, Katie E. Raffel, Marie Leeson, Molly A. Kantor, Peter Barish, Sumant Ranji, Tiffany M. Lee

Category Innovations

Concept Diagnostic Error

Concept Diagnostic Process Failure

Concept UPSIDE Study

Concept DEER Taxonomy

Concept Hospitalized Patient

Distinguished Non-Finalist

Presenter Organization University of California, San Francisco

Presenting Author Peter Barish

Track Quality Improvement

Keywords

diagnostic errors

hospital safety

systems engineering

DEER taxonomy

diagnostic process failures

ICU transfers

rapid response events

heat maps

Pareto diagrams

diagnostic improvement

Diagnostic Error

Diagnostic Process Failure

UPSIDE Study

DEER Taxonomy

Hospitalized Patient