Robust Inference in Semiparametric Discrete Hazard Models for Randomized Clinical Trials

• Activity Number: 113
• Type: Topic Contributed
• Date/Time: Monday, August 1, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Health Policy Statistics
• Author(s): Vinh Nguyen*+
• Companies: University of California at Irvine
• Keywords: Survival analysis ; Censoring ; Discrete time ; Model misspecification ; Robust inference

Abstract

Time-to-event data in which failures are only assessed at discrete time points are common in many clinical trials. Examples include oncology studies where events are observed through periodic screenings such as CT scans. When the survival endpoint is acknowledged to be discrete, common semi-parametric methods for the analysis of observed failure times include the discrete hazard models (e.g., the discrete-time proportional hazards and the continuation ratio model) and the proportional odds model. In this manuscript, we consider estimation of an average treatment effect in discrete hazard models when the semi-parametric assumption is violated. Building on previous work for the continuous-time proportional hazards model we demonstrate that the estimator resulting from these discrete hazard models is consistent to a parameter that depends on the underlying censoring distribution. An estimator that removes the depedence on the censoring mechanism is proposed and its asymptotic distribution is derived. Basing inference using the proposed methodology allows for statistical inference that is reproducible and scientifically meaningful. Simulation is used to assess the proposed methodology.