Bayesian Threshold Regression for Informatively Censored Current Status Data

• Activity Number: 318
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Michael Pennell*+ and Tao Xiao
• Companies: Ohio State University and University of Maryland/Ohio State University
• Keywords: Carcinogenicity study ; Data augmentation ; Interval censoring ; First hitting time model ; Time to tumor ; Wiener process

Abstract

In carcinogenicity studies in animals, tumors are not discovered until the time of examination and hence time to tumor is left censored; this is known as current status data. Sometimes, the examination time is not independent of the event time. For example,an exam may have occurred because the animal died due to a cause related to tumor development. In this case, survival analysis methods which assume independent censoring would result in biased inferences. To address this issue, we propose a Bayesian approach which jointly models time to event and time to censoring using latent Wiener processes which fail once they hit a boundary value. Using data augmentation, we sample the unobserved event time and values of the latent processes for those subjects who do not experience an event. Informative censoring is accounted for by modeling time to censoring using two latent health processes: one independent of the event of interest and the other dependent. In addition to being a conceptually appealing model, our model does not require the assumption of proportional hazards. We demonstrate our method using time to lung tumor data from a National Toxicology Program study.