目標：參考Andersen第一與第二階段醫療服務利用行為模式，擬欲探討2000至2009年精神疾病社區復健使用者之個人特質，並且分析使用精神疾病社區復健病患歷經指標出院一年期間，其往後一年内再入院風險及相關危險因子。方法：採回溯性世代設計，資料出處為國衛院所發行之全民健康保險資料庫1999至2009年「居家照護特殊需求檔」，研究對象為2000至2007年精神疾病社區復健使用者為新收個案，共25,218人。以Cox比例風險迴歸模式預測精神疾病社區復健使用者在指標出院後經歷一年期間，其往後一年内再入院之風險。結果：研究發現以Cox迴歸模式在出院後一年内精神疾病社區復健使用次數為8-14次與S15次的再入院風險較使用1-7次者低，女性、中區分局病患、精神分裂症患者與住院小於55天等變項，再入院之風險較低。結論：精神疾病社區復健使用次數越少之個案，其爾後再入院之風險比使用次數多者還高，值得健康照護相關人員進ー步思考影響其使用的因素，另外，其使用的模式及其與再入院之關係為何？亦值得未來進ー步探究。（台灣衛誌2013 ； 32(6) ： 586-598)

Objectives: With reference to phase I and phase II of Andersen’s behavioral model of health service utilization, the aims of this study were first to investigate the characteristics of community rehabilitation care usersl and then to analyze the frequency of readmission and its risk factors within one year after discharge from a hospital for the treatment of mental illness. Methods: A retrospective cohort study design was used. Data analyzed in this study were retrieved from the ‘home care specific files’ of the National Health Insurance Research Database issued by the National Institutes of Health between 1999 and 2009. A total of 25,218 people were new cases of community rehabilitation care between 2000 and 2007. In order to forecast these community rehabilitation care users’ risk of readmission within one year after discharge, the Cox proportional hazard model was utilized. Results: We found that the adjusted hazard ratio was significantly lower for groups of 8-14 times users and 15 or more-times users of community rehabilitation care than for 1-7 times users (0.73 & 0.72 respectively) in the Cox regression model. Furthermore, the risk of readmission was lower for females, clients who were insured by sub-bureau in the middle of Taiwan, a diagnosis of schizophrenia, and a length of hospital stay less than 55 days. Conclusions: The lower the frequency of psychiatric patients’ using community rehabilitation care, the higher the risk of readmission. (Taiwan J Public Health. 2013;32(6):586-598)