Estimating Functionals of Multistate Processes Under Cross-Sectional Sampling

• Activity Number: 286
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Micha Mandel*+
• Companies: Hebrew University of Jerusalem
• Keywords: biased sampling ; truncation ; competing risks

Abstract

Multi-state processes are commonly used to model the dynamics of subjects in a population. A motivating example is of patients in a hospital who may acquire infections before they are discharged from the hospital. In this talk, I consider a cross-sectional sampling of independent multi-state processes in which data are subject to length-bias and censoring, and develop estimators for functionals of the underlying distribution such as the probability of each of the paths of the process, and cumulative incidence functions of competing events. Consistency and asymptotic normality of the estimators are proved under various assumptions on the model and the sampling design. The method is applied to infection data obtained in a cross-sectional study of patients hospitalized in intensive care units.