邏輯斯與長期資料之聯合模型-腦中風病患之姿態性低血壓實例研究

Joint modeling of logistic and longitudinal data - Applications to stroke patients with orthostatic hypotension

蔡昊澐(hao-Yun Tsai)

指導教授 ： 黃怡婷

國立臺北大學/商學院/統計學系/碩士(2008年)

摘要

姿態性低血壓(OH)是一種心血管疾病， 假使罹患腦中風的病患在開始接受神經復健時， 若引發姿態性低血壓， 常會導致病患不適的症狀，或甚至趺倒而造成骨折， 病患也可能因而延遲復健與治療的時間， 不僅增加病患的痛苦， 也增加醫療成本的負擔。 在臨床上， 醫生一直沒有辦法有效的評估出病患是否有姿態性低血壓的毛病， 本論文希望利用某醫院提供的臨床資料來找出與姿態性低血壓有關的危險因子。 由於資料包含病患重複觀察的收縮壓， 直接使用邏輯斯迴歸模型較不適當的，Tsiatis(1995)提出的二階段模型和Tsiatis及Wulfsohn(1997)提出的聯合模型皆可建立存活時間和長期觀測的危險因子之關連， 本論文提出合併長期觀測的收縮壓、 危險因子及姿態性低血壓狀態的二階段模型和聯合模型， 且提供修改後之二階段模型和聯合模型之係數估計及其大樣本之分配， 利用蒙地卡羅模擬分析評估參數估計值的正確性。

關鍵字

姿態性低血壓 ； 聯合模型 ； 二階段模型 ； 邏輯斯迴歸模型 ； 隨機效應模型 ； 中風

並列摘要

Orthostatic hypertension (OH) is one of the cardiovascular diseases. If a patient having stroke also has OH, then it is possible to have a higher chance to fall or syncope during the recovery. This may cause this patient to have possible fracture and the burden of medical cost therefore increases. How to diagnosis OH is clinically important. However, there is no obvious clinical method. This thesis uses a clinical data to identify potential clinical factors that are associated with OH. Since this data include repeatedly observed systolic pressures, patient's basic characteristics, clinical symptoms and the OH status, the logistic regression is not appropriate. The two-stage model proposed by Tsiatis (1995) and the joint model proposed by Tsiatis and Wulfson (1997) can be used to model a sequence of predictors and survival jointly. A modified two-stage model and a modified joint model are proposed to incorporate the sequence of systolic pressure, risk factors and the status of OH. The large sample properties of estimators of model parameters are derived. Monte Carlo simulations are also performed to evaluate the accuracy of these estimators.

並列關鍵字

orthostatic hypotension ； joint model ； logistic regression ； random effect model ； two stage model ； stroke

參考文獻

counts in patients with AIDS. Journal of the American Statistical Association 90: 27-37.

Hsieh, F., Tseng, Y. K. and Wang, J. L. (2006). Joint

Approach Revisit. Biometrics 62: 1037-1043.

(1998). Convergence properties of the Nelder-Mead simplex

Models for Longitudinal Data. Biometrics 38: 963-974.