<![CDATA[Incident reporting is a fundamental component of organizational learning in patient safety. However, Kaizen-based continuous improvement programs designed to strengthen incident reporting are often evaluated using report counts, which cannot distinguish improvements in reporting culture from increases in actual patient harm. This study aimed to develop and validate an analytical framework that differentiates these effects and establish a systematic review protocol for its application. The observed reporting rate was decomposed into three components: exposure, the true incident rate, and reporting propensity. A severity-stratified Reporting Propensity Index (RPI) was developed using partial identification bounds, alongside a behavioral reporting model, an assessment of bias in before-and-after study designs, and a sustainability half-life measure. Four Monte Carlo simulation studies encompassing 52 scenarios were conducted to evaluate estimator performance, and a systematic review protocol was developed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. Under the specified assumptions, the RPI estimator was unbiased and achieved nominal confidence interval coverage, whereas the conventional total-count estimator exhibited bias of up to ?0.35 log units. When assumptions were violated, the partial identification bounds contained the true parameter in at least 97.4% of simulation replications. Uncontrolled before-and-after designs overestimated intervention effects, consistent with theoretical predictions, while reusing baseline-period data in moderator analyses increased the false-positive rate to 20.7%. These findings suggest that severity-stratified incident reporting, analyzed using the RPI framework, may provide hospitals with a more reliable measure of reporting propensity and support evidence-based decisions regarding the implementation and evaluation of Kaizen-based patient safety improvement programs.]]>
Copyrights © 2026
