Regression Analysis of Incomplete Data from Event History Studies with the Proportional Rates Model

• Activity Number: 313
• Type: Contributed
• Date/Time: Tuesday, August 2, 2016 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Guanglei Yu* and Liang Zhu and Jianguo Sun and Leslie L. Robison
• Companies: University of Missouri - Columbia and St. Jude Children's Research Hospital and University of Missouri and St. Jude Children's Research Hospital
• Keywords: Incomplete data ; Marginal model ; Multiple imputation ; Proportional rates model

Abstract

This paper discusses regression analysis of a type of incomplete data, mixed recurrent and panel count data, arising from event history studies with the proportional rates model. By mixed data, we mean that each study subject may be observed continuously during the whole study period, continuously over some study periods and at some time points, or only at some discrete time points. If all subjects are observed continuously, one will have complete data, which are usually referred to as recurrent event data. On the other hand, if all subjects are observed only at discrete time points, the observed data are incomplete and commonly referred to panel count data. For the problem, we present a multiple imputation-based estimation procedure and one advantage of the proposed marginal model approach is that it can be easily implemented. To assess the performance of the procedure, an extensive simulation study is conducted and indicates that it performs well for practical situations and can be more efficient than the existing method. The methodology is applied to a set of mixed data from a longitudinal cohort study.