Multilevel Modeling of Follow-Up Studies with Missing Data

• Activity Number: 536
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Monica Michelle Bennett*+ and John Seaman and James Stamey
• Companies: Baylor University and Baylor University and Baylor University
• Keywords: Missing data ; Multilevel regression ; Bayesian inference

Abstract

Follow-up studies are often used to assess the rates of incidence of a particular event, such as death or injury, among different groups. The observed counts are usually assumed to follow a Poisson distribution. The analysis is complicated when some of the records indicating the cause of the incident have been destroyed or lost. If we ignore the missing data then both the overall event rates and the uncertainty in the model parameters will be underestimated. To help overcome this problem, we develop a Bayesian multilevel Poisson model to account for the missing data. The model is developed so that it estimates the rates of incidence, as well as the probability of missing information.