Medical Incident Reports Analysis

Ensuring patient safety is the highest priority in clinical settings; however, no matter how strict the preventive measures are, medical accidents and incidents cannot be completely avoided. When an incident occurs, it is reported as an incident report. A large number of incident reports are submitted daily, and among them are valuable data that can serve as clues for preventing future incidents. However, reviewing all of this massive amount of data requires tremendous time and effort, and traditionally, attention has been focused only on cases with a high incident level. As a result, important hints contained in minor or seemingly insignificant reports may be overlooked. In particular, for text data, staff in the medical safety management department must individually read and interpret each report, making the analysis subjective and preventing objective evaluation. In this study, we extracted keywords from incident reports at Shimane University Hospital and explored methods for efficiently analyzing incident reports using text mining techniques.