Survival Analysis Methods for Transplantation Studies

描述(由申请人提供)：该项目开发用于分析多个事件时间的模型。这份提案考虑了两个项目。第一个项目提出了一类称为调制更新过程的随机过程模型回归分析的新方法。它们可用于模拟代表患者疾病进展中的特定事件的多个事件时间。在这个项目中，我们提出了加速失效时间和转换模型的扩展。我们还考虑了这些模型中的参数估计。第二个项目考虑配对竞争风险回归模型。这类模型出现在生物医学和精算随访研究中。在这样的研究中，匹配(或阻断)作为一种方法来控制预后变量对由时间和失败类型表示的结果的可变性的影响。提出了一类分析配对竞争风险的随机过程模型。这两个项目还开发了用于分析来自骨髓和肾脏移植研究的后续数据的应用程序。我们考虑根据移植前的特征和随访史来预测患者移植后的经历。

DESCRIPTION (provided by applicant): The project develops models for analysis of multiple event times. Two projects are considered in this proposal. The first project proposes new methods for regression analysis of a class of stochastic process models called modulated renewal processes. They can be used to model multiple event times representing specific episodes in the disease progression of a patient. In this project we propose extensions of accelerated failure time and transformation models. We also consider parameter estimation in these models. The second project considers matched pair competing risk regression models. Models of this kind arise in biomedical and actuarial follow-up studies. In such studies, matching (or blocking) serves as a method to control for effects of prognostic variables on the variability of the outcomes represented by time and failure type. We propose a class of stochastic process models for analysis of matched pair competing risks. Both projects develop also applications to analysis of follow-up data from bone marrow and kidney transplant studies. We consider prediction of patient post-transplant experience based on pre-transplant characteristics and follow-up history.