A human error taxonomy for analysing healthcare incident reports: assessing reporting culture and its effects on safety performance

The present paper reports on a human error taxonomy system developed for healthcare risk management and on its application to evaluating safety performance and reporting culture. The taxonomy comprises dimensions for classifying errors, for performance-shaping factors, and for the maturity of reporting culture contained in incident reports. Applying several dimensions in the taxonomy, we propose on the one hand two safety performance measures, i.e., the rate of near-miss reporting and the rate of near-miss detection by safety procedure, and on the other, measures for diagnosing reporting culture including average descriptive depth in reports. We applied the taxonomy to a total of 3749 incident cases collected from two Japanese hospitals, which were at different stages of patient safety activities: Hospital A initiated organisation-wide initiatives several years before the survey period, while such safety-related activities had just commenced in Hospital B. The hospitals also differed in their reporting rates of incidents per nurse: 3.05 (A) vs. 0.65 (B). Results show that the taxonomy can identify differences between these hospitals both in terms of safety performance and reporting culture. In addition, a correlation trend was observed between these two measures.