An Investigation on Patient Incident Reports: Association Rule Mining Approach

Abstract

The occurrence of error events in incident reporting systems leads to deterioration of performance in patient safety. Investigating the critical pattern to identify human or system factors in the problem process is crucial for healthcare organizations. However, little is known about incident reports in general ward settings. To analyze error events in an incident reporting system widely used in Taiwan, and to explore various levels of severity assessment and critical attributes leading to. The data consisted of error events (n = 738), including 13 types of reported errors, reported by one of the best regional hospitals in Taiwan in 2016–2018. The association rules are used to extract professionals' and patients' related sources of risk in the general ward. The important attributes including types of reported errors, gender, and working shifts result in serious and major severeness. Female inpatients who suffered from unexpected cardiac arrest events have less severeness comparing to male inpatients. Unexpected cardiac arrest events represent important events in types of reported errors, suggesting the professionals in healthcare organizations should consider an intervention for preventing the event.