目標：本研究目的為了解養護機構對住民實施身體約束之比率及其相關因素。方法：本研究為橫斷面研究設計，以抽取機率與機構人數多寡成比例(Probability proportional to size, PPS)的方法從某市178所養護機構中選取20家共848名住民為本研究樣本。住民約束情形與疾病史資料分別以觀察及調閱病歷方式蒐集之，而身體功能、平衡能力、與肌力之資料亦是透過觀察或測量而獲得。本研究利用χ^2與simple t test比較受約束與未受約束住民特性之差異，也利用邏輯斯迴歸模式分析與住民接受約束或多重約束之顯著相關因素。結果：機構住民被約束及多重被約束的比率分別為74.1% (95% confidence interval, CI=71.2-77.0%)和47.8% (95% CI=39.9%-55.6%)。相較於無約束住民，約束住民有顯著較高中風、失智、與壓瘡的比率，入住時間也較久。此外，被約束住民的日常生活功能(Activity of Daily Living, ADL)、認知功能檢查(Mini Mental State Examination, MMSE)、左右腿及腳的肌力得分都顯著地較無約束住民差。住民被約束的原因以預防跌倒佔最多數(79.8%)。多變量分析發現，ADL得分增加與較低的被約束勝算(Adjusted Odds Ratio, AOR=0.98, 95% CI=0.97-0.99)及被多重約束(AOR=0.95, 95% CI=0.93-0.97)有顯著相關；此外，家屬簽署同意書與「甲等」機構住民被約束的勝算比也顯著較高。結論：相較於歐美國家，本研究顯示台灣機構住民身體被約束比率明顯偏高，本研究也發現約束比率高低與家屬簽同意書、機構評鑑等級及住民日常生活身體功能有顯著相關。

Objectives: To investigate the rate of and factors associated with the use of physical restraints among residents of long-term care institutions. Methods: This is a cross-sectional study involving 848 residents sampled from 20 residential institutions using probability proportional to size (PPS) across a total of 178 residential institutions in one city. Information on the use of physical restraints and disease history was collected using observation and chart review. Physical functioning, balance and leg muscle power were also measured and recorded. Χ^2 and simple t test were used to compare differences in characteristics between residents who were and were not physically restrained. A logistic regression model was used to identify factors significantly associated with odds of any type of physical restraint use or multiple physical restraint use. Results: The rate of any physical restraint use and multiple physical restraint use was estimated at 74.1% (95% confidence interval, CI=71.2-77.0%) and 47.8% (95% CI=39.9%-55.6%). Compared to those not physically restrained, residents who were physically restrained had a higher rate of cerebrovascular disease, dementia and pressure sores, and had been resident in the institution for longer. Restrained residents had a significant deficit in their scores for activity of daily living (ADL), Mini Mental State Examination (MMSE) and muscle power of both legs. The leading reason given for using physical restraints was "preventing a fall" (79.8%). Multivariate analysis showed that an increased ADL score was significantly associated with reduced odds of the use of any physical restraint (Adjusted Odds Ratio, AOR=0.98, 95% CI=0.97-0.99) or the use of multiple physical restraints (AOR=0.95, 95% CI=0.93-0.97). Residents from institutions with a second grade by accreditation, whose families had signed an agreement for physical restraints, also showed a significantly increased AOR for the use of physical restraints. Conclusions: The use rate of physical restraint was higher in this Taiwanese residential sample compared to rates found in Europe and United States. Use rate varied within the institution and was associated with signed agreement and poorer physical functioning as indicated by the residents' ADL scores.