跌倒是某醫學中心異常事件的首位，佔所有病人安全通報事件的四分之一，跌倒後有將近一半的病人發生輕度到中重度不等的傷害。醫院原版16 項跌倒風險篩檢表過於繁瑣，且敏感性、特異性效益不佳。本研究旨在簡化南台灣某醫學中心病人跌倒風險之篩檢表。本研究採回溯性病例對照方式，以醫學中心原有16項跌倒危險評估內容收集資料，收集2015 年全年住院當中跌倒通報之病人169人，平均年齡63.9 歲，對照組採科別分層隨機抽樣，控制年齡及科別變項，人數1：1 方式，收集住院當中沒有跌倒的病人169 人，總共完成338 人資料收集。以SPSS 20.0 統計軟體，進行兩組描述性統計、卡方檢定及診斷性試驗進行資料分析。結果呈現338 位個案中男性211 人，女性127 人，年齡大多分佈在52.0~80.3歲之間，卡方檢定分析出7 項跌倒危險因素，分別為步態不穩、過去一年曾跌倒、頭暈/ 眩、不認為自己會跌倒、下床需人協助、使用鎮靜安眠藥及輕瀉劑，做為簡版跌倒風險篩檢表之項目。ROC 曲線求得最佳AUC.725，以一項一分，≧ 2 分收案的敏感度83.4%、特異性43.8%、陽性預測值59.7%、陰性預測值72.5% 及Youden index .40 最好。本研究之簡版跌倒風險篩檢表經修訂後應用在臨床護理，做為護理師入院評估及每日身體評估中篩選跌倒高風險病人之依據。

Falling is the most frequent accident that occurs in our hospital, accounting for more that 25% of all case reports on patient safety incidents. Almost half of fall patients exhibit minor, moderate, or severe injuries. An original 16-item fall-risk screening tool was found to be complicated and the outcomes of sensitivity and specificity were ineffective. The purpose of this study was to simplify the risk screening tool used for predicting patient falls in a medical center in southern Taiwan. We used the 16-item screening tool and conducted a retrospective case-control study based on data collected regarding 169 cases of inpatient falls reported at the medical center in 2015. A control group was selected from a stratified random sample of patients treated at various departments. Controlling for age, we frequency matched a sample of 169 inpatients who did not fall during hospitalization at a ratio of 1:1, obtaining a total sample of 338 cases. SPSS Version 20.0 was employed to perform a descriptive statistical analysis of the two groups and to conduct chi-square tests and a diagnostic test. The 338 inpatients comprised 211 men and 127 women, with ages mostly ranging from 52.0 to 80.3 years. The results of the chi-square analysis revealed the following seven risk factors for falling: unsteady gait, past history of falls, dizziness/vertigo, patients overestimating their ability of forgetting their limitations, requiring assistance in ambulation, being prescribed sedatives or hypnotics, and taking laxatives. The simplified 7-item version of the fall-risk screening tool was found to be appropriate through an analysis of the receiver operating characteristic curve (ROC), the result of the analysis showed an area under the curve of .725. Based on the fall risk screening score with 2 points as the cutoff value, the sensitivity, specificity, positive predictive value, negative predictive value, and Youden index for detecting patients at a high risk of falling were 83.4%, 43.8%, 59.7%, 72.5%, and .40, respectively. The fall-risk screening tool could be used by clinical nurses as a procedure in patient admissions and daily physical assessments to identify patients at a high risk of falling.