雙變量存活分析用於配對世代研究

Bivariate Survival Analysis Applied to Matched-Pair Study

張維之(Wei-Chih Chang)

指導教授 ： 王維菁

國立交通大學/理學院/統計學研究所/碩士(2018年)

摘要

本篇論文啟發自一篇有關腎臟病之醫學研究，探討曾經中風者是否日後罹患慢性腎臟病的風險亦會提高。透過配對研究的方法將中風者歸類為暴露組會遇到的挑戰是“中風”本身亦為動態的事件，情況相較於傳統的配對方法更複雜，如何將中風發生的時間納入分析是值得探究的議題。論文先回顧了雙變量存活分析衡量相關性的參數、疾病死亡模型與半競爭風險資料。再介紹文獻使用的配對方法。我們利用Savignoni et al. (2014) 提出的配對法，在半競爭風險的架構下以Clayton模型為假設，提出一個新的方法來估計暴露組對非暴露組的風險比值。並透過不同模擬情境確認此方法之正確性。

關鍵字

雙變量存活分析 ； Clayton 模型 ； 風險比值 ； 疾病死亡模型 ； 配對 ； 半競爭風險模型

並列摘要

The thesis is motivated by the medical research which investigated whether the occurrence of stroke increases the hazard of developing chronic kidney diseases. There are several approaches to tackle the problem. One direction is to study the association between the two event times based on semi-competing risks data under the illness-death model. Another approach is to conduct matched pair designs which compare those with and without stroke. The latter approach however is more complicated than traditional matched pair studies since the exposure status is determined by a random event. Under the framework of semi-competing risks, we propose a new method to estimate the hazard ratio of exposed and non-exposed groups, which takes the exposure time into account, under the Clayton model assumption. Simulations are implemented to verify the validity of the proposed methodology.

並列關鍵字

bivariate survival analysis ； Clayton model ； hazard ratio ； illness-death model ； matched-pair ； semi-competing risks model

參考文獻

1.Genest, C., & Rivest, L. P. (1993). Statistical inference procedures for bivariate Archimedean copulas. Journal of the American statistical Association, 88 (423), 1034-1043.

2.Holt, J. D., & Prentice, R. L. (1974). Survival analyses in twin studies and matched pair experiments. Biometrika, 61(1), 17-30.

3.Savignoni, A., Giard, C., Tubert-Bitter, P., & De Rycke, Y. (2014). Matching methods to create paired survival data based on an exposure occurring over time: a simulation study with application to breast cancer. BMC Medical Research Methodology, 14(1), 83.

4.Shinozaki, T., Mansournia, M. A., & Matsuyama, Y. (2017). On hazard ratio estimators by proportional hazards models in matched-pair cohort studies. Emerging Themes in Epidemiology, 14(1), 6.

5.Wang, W. (2003). Estimating the association parameter for copula models under dependent censoring. Journal of the Royal Statistical Society: Series B, 65(1), 257-273.