利用Apriori演算法探討異常通報事件關聯規則之初探

A Preliminary Study on the Association Rules of Medical Incident Report by Using Apriori Algorithm

本研究依照病人安全通報常發生事件，選擇跌倒事件、管路事件、和檢查／檢驗／病理切片，找出病人安全事件通報之人、事、時、地、與嚴重性之關聯性。資料樣本採取2016、2017和2018年度病人安全事件通報，去除沒有傷害的事件通報。使用Apriori篩選過程設定為提升度大於1、信賴度設定90%、支持度設定10%。分析結果依據支持度以30%以上、20%到未滿30%之間、10%到未滿20%做分類。研究結果以最高支持度之項目顯示異常事件通報最容易發生在檢查檢驗事件且多數無陪伴者在場。隸屬單位為醫事檢驗科的檢查檢驗事件發生通報之危險因子為無陪伴。隸屬於護理科之管路事件在發生通報之危險因子為男性、無陪伴、通報者單位疏失、嚴重度1、65歲以上。隸屬於護理科之跌倒事件發生通報之危險因子為大夜、小夜。嚴重度2之檢查檢驗事件發生通報之危險因子為其他單位疏失、其他單位疏失及無陪伴。

"In this study, inpatient falls, accidental tube removal, and examination /inspection / pathology events were selected to ascertain the relationshipsamong the person, event, time, place, and severity from the data. The datasamples were taken from 2016, 2017, and 2018 of the Taiwan Patient-SafetyReporting system by excluding the events without harm. The study usedthe Apriori algorithm to screen the process by setting up the lift of greaterthan 1, the confidence of 90%, and the support of 10%. The analysis resultswere classified according to the values of support with more than 30%, 20%to less than 30%, and 10% to less than 20%.The results showed that the incident with the highest support value camefrom the patient in inspection and the majority of patients were withoutaccompanying persons. The risk factor of the examination events of themedical laboratory was from the patient without accompanying persons.The risk factor of the accidental tube removal events of the nursingdepartment was found to be male patients whose ages were 65 years old ormore without accompanying persons along with the negligence of thereporting unit with the severity of one. The risk factor of a fall incident ofthe nursing department was associated with night shifts. Finally, the riskfactor of the severity of two of the examination events was associated withthe negligence of other units or the negligence of other units and the patientwithout accompanying persons."