本研究目的在訂定精神科急性病房跌倒、暴力、自傷（殺）三項護理品質指標之閾值，並分析影響三項品質指標之意外事件發生率的相關因素，以作為擬定改善措施之參考。本研究採回溯世代研究(retrospective cohort study)設計，以台灣北部某醫學中心於2007年1月1日至2012年12月31日期間住精神科急性病房共99床之病人為研究樣本；資料來源為取樣期間資訊室住院病人資料檔、醫療品質管理中心異常事件通報電子檔及護理業務指標資料檔，採檔案分析的方式串聯相關資料庫的檔案後以Excel軟體進行資料轉檔、除誤、串檔與歸戶等處理，再以SPSS統計軟體進行描述性與推論性統計。研究結果顯示，跌倒、暴力、自傷（殺）發生率分別為0.03%-0.28%、0.02%-0.07%、0.003%-0.06%，以六年平均發生率設定醫學中心精神科急性病房跌倒閾值為0.11%、暴力閾值為0.05%、自殺閾值為0.02%；「病人年齡」、「佔床率」、「直接護理時數」等因素與意外事件發生率達統計上顯著差異。建議同屬性病房可建立護理人力資料庫，並依每日病人動態適時、機動性調整支援人力。

This study analyzed the threshold of falls, self-harm (suicide), and violence among acute psychiatric patients and analyzed the related factors for accident rates as the basis for future interventions. A retrospective case-control study design was used in this study. Data underwent archival analysis and were retrieved from the electronic archival files of Taipei Veterans General Hospital from over 2007-2012. The information covers 99 beds, which was the total bed number of the acute psychiatry ward. Accident-related events included falls, violence, and self-harm as recorded in the patient information system and unusual event notification database. Nursing care information was adapted from the nursing index database. Each event threshold was established by using the mean incidence over the 6 years. The factors related to accident-related nursing indicators higher than the threshold were analyzed through multiple linear regression. The average rate of occupied beds was 91.95% ± 0.05%. The incidences of falls, violence, and self-harm were 0.03%-0.28%, 0.02%-0.07%, and 0.003%-0.06%, respectively. The threshold of fall incidences was set at 0.11%. The threshold of violence incidences was set at 0.05%, and the threshold of self-harm incidences was set at 0.02%. Significant differences were observed for age, occupied rate, and nursing hours with the incidence of accidents. The results suggest that a flow of nursing staff resources, according to the bed occupancy rate, should be established to adjust manpower and reduce the incidence of accident-related events in acute psychiatric wards.