Statistical Analysis for Mixed Recurrent-Event and Panel-Count Data

描述（由申请人提供）：摘要本申请涉及混合复发事件和面板计数数据的新分析策略的开发和应用。虽然重复事件数据和面板计数数据都是由重复事件过程产生的，但它们具有不同的观测系统。在前者中，对受试者进行连续观察，而在后者中，仅在离散时间点对受试者进行观察。因此，复发事件数据记录复发事件的所有发生时间，而小组计数数据仅记录观察时间点之间的事件。可能在单个队列中，一些受试者具有复发事件数据，而其他受试者具有小组计数数据，或者每个受试者在某些时期具有复发事件数据，而在其他时期具有小组计数数据。在长期随访研究中，这些混合数据并不罕见。目前的普遍做法是近似或简化这些复杂的数据，从而可能导致误导性的结论。有一个迫切需要开发直观的，高效的，计算上可行的方法来分析复杂的数据在事件历史研究。本计画拟利用儿童癌症存活者纵向研究的资料：（1）发展均值函数的非参数估计及混合资料的非参数两样本比较方法：（2）发展比例均值模型的半参数估计方程及加性率模型的半参数估计方程，以进行回归分析; 3）将目标2中的方法扩展到多变量混合复发事件和面板计数数据。这些方法对于复杂事件历史数据的研究可能具有很强的统计和临床相关性。

DESCRIPTION (provided by applicant): Abstract This application addresses the development and application of new analytic strategies for mixed recurrent-event and panel-count data. While recurrent-event data and panel-count data are both generated from recurrent-event processes, they have different observation systems. In the former, subjects are observed continuously and in the latter, subjects are observed only at discrete time points. Consequently, recurrent-event data record all occurrence times of recurrent events, while panel-count data record only the events between observation time points. It is possible that in a single cohort, some subjects have recurrent-event data while others have panel-count data, or every subject has recurrent-event data during some periods and has panel-count data during other periods. It is not unusual to have these mixed data in long-term follow-up studies. The current common practice is to approximate or simplify these complex data, resulting in potentially misleading conclusions. There is an urgent need to develop intuitive, efficient, and computationally feasible methods for analyzing complex data in event history studies. This project proposes to use data from the renown longitudinal Childhood Cancer Survivor Study (CCSS) to: 1) develop both a nonparametric estimation of the mean function and a procedure of nonparametric two-sample comparison for these mixed data; 2) Develop a semiparametric estimating equation-based method for a proportional mean model and a semiparametric estimating equation-based method for an additive rate model for regression analysis; and, 3) Extend the methods developed in Aim 2 for multivariate mixed recurrent-event and panel-count data. These approaches potentially have strong statistical and clinical relevance for the study of complex event history data.