2015-08-012024-05-13https://scholars.lib.ntu.edu.tw/handle/123456789/650945摘要：老化使人反應遲緩、肌力降低、平衡功能下降，再加上骨質疏鬆，容易使老人因低能量創傷（例如 跌倒）而造成髖部骨折。台灣人口老化速度非常快，若高齡人口持續成長，可以預見的是髖部骨折發 生率會持續攀升。老人在髖部骨折後之死亡率較同年齡者之平均死亡率為高，此外髖部骨折後也可能 會產生各種併發症，例如心血管疾病、感染、壓瘡、失能、心理疾病等，因此，髖部骨折預期將會對 人口老化的台灣造成重大的醫療及社會負擔。而探求如何改善老人髖部骨折的預後就顯得更加的重 要。先前曾有研究評估復健治療對髖部骨折預後的影響，雖然部分研究結果顯示復健治療對髖部骨折 預後可能有改善的趨勢，然而至今仍尚未有定論。由於先前各研究之研究設計與研究對象有相當大的 差異，因此目前難以對復健是否能改善髖部骨折後的預後有所定論，此外，由於先前的研究大多是以 單一醫院為基礎，也較難外推其結果，而到目前為止應用大型健康資料庫對於髖部骨折後復健之療效 評估的相關研究仍非常有限，故本研究計晝擬使用台灣全民健保資料庫為研究資料來源，進行復健對 髖部骨折預後之療效評估。本研究計晝擬採用世代縱向追蹤研究設計探討復健治療是否可降低髖部骨 折後死亡以及相關併發症的發生。因本研究計晝屬於觀察性研究，沒有對於是否接受復健治療進行隨 機分派，為了減少對於接受復健治療與否的潛在性的選擇性偏差，本計晝將採用傾向分數配對 (propensity score matching)的分析方式。傾向分數（propensity score)是指一位髖部骨折病人基於其 臨床上所有可以觀察到的因素之下，此病人接受復健治療之機率。本研究將使用8 digit to 1 digit的 greedy運算方式進行配對，藉由傾向分數配對，可以把干擾因子的影響降到最低。統計分析將使用 stratified Cox proportional hazard regression評估復健對於老人髖部骨折預後事件風險的影響。本研究計 晝將有助於釐清復健治療對於老人髖部骨折預後之效益，對於協助臨床醫療決策與資源分配將具有重 要參考價值。<br> Abstract: Aging may lead to a higher risk of osteoporosis and a decline in agility, muscle strength, and balance function, which can predispose the elderly to hip fracture by low energy trauma (e.g. fall). As other developed countries, Taiwan has been confronted with the crisis of aging. It can be expected that the burden of hip fracture will increase with the growing of aged population. Hip fracture is associated with a higher mortality. Moreover, patients with hip fracture are subject to complications such as cardiovascular diseases, infections, pressure ulcer, disability, and psychological disorders, leading to enormous medical and socioeconomic burden. It is therefore of paramount importance to investigate how to improve the prognosis of hip fracture. Previous studies have evaluated the effects of rehabilitation on the prognosis of hip fracture. While some of these studies suggested a tendency of favorable effects of rehabilitation on hip fracture, the evidence remains inconclusive. These conflicting findings from previous studies may rise from the heterogeneity in terms of study design and study population. In addition, since most of the previous studies were conducted on the basis of a single center, the generalizability may be limited. The objective of the present study is to conduct a population-based, propensity-score matched, longitudinal follow-up study to evaluate the effects of rehabilitation on the prognosis of hip fracture in the elderly. Our study will use the National Health Insurance database and the mortality registry from the Ministry of Health and Welfare, Taiwan. The propensity score is the probability of receiving rehabilitation given the observed clinical characteristics, including demographic variables, medical co-morbidities, geographic and socioeconomic variables. Matching on propensity scores can minimize potential selection bias and therefore establish comparable rehabilitation group and reference group. The 8-to-1 digit greedy matching algorithm will be used to identify the propensity score-matched reference subjects. Stratified Cox proportional hazard regression matched on propensity score is used to estimate the effects of rehabilitation on the mortality and medical complications following hip fracture. This study is expected to provide valuable information for medical decision making and resource allocation in elderly patients with hip fracture.髖部骨折復健風險因子預後研究傾向分數Hip fractureRehabilitationRisk factorsOutcome studyPropensity score