Abstract：
Many non-fatal events can be considered recurrent in that they can occur repeatedly over time, and some researchers may be interested in the trajectory and relative risk of non-fatal events. With the competing risk of death, the treatment effect on the mean number of recurrent events is non-identifiable since the observed mean is a function of both the recurrent event and terminal event processes. In this paper, we assume independence between the non-fatal and the terminal event process, conditional on the shared frailty, to fit a parametric model that recovers the trajectory of, and identifies the effect of treatment on, the non-fatal event process in the presence of the competing risk of death. Simulation studies are conducted to verify the reliability of our estimators. We illustrate the method and perform model diagnostics using the Carvedilol Prospective Randomized Cumulative Survival trial which involves heart-failure events.
摘要：
许多非致命事件可以被认为是复发性的，因为它们可以随着时间的推移反复发生，一些研究人员可能对非致命事件的轨迹和相对风险感兴趣。在竞争的死亡风险下，由于观察到的平均值是复发事件和终末事件过程的函数,因此治疗对平均复发事件的影响是无法确定的.在本文中,我们假设非致命事件和终端事件过程之间的独立性，以共同的脆弱为条件，为了拟合恢复轨迹的参数模型，并确定治疗对，在存在竞争性死亡风险的情况下的非致命事件过程。进行了仿真研究，以验证我们的估计器的可靠性。我们使用涉及心力衰竭事件的卡维地洛前瞻性随机累积生存试验来说明方法并进行模型诊断。
