Bias Correction for Rare Events in Self-Controlled Case Series Design of Vaccine Safety

• Activity Number: 533
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section for Statistical Programmers and Analysts
• Author(s): Chan Zeng*+ and Jason M. Glanz and Sophia R. Newcomer and Stanley Xu
• Companies: Kaiser Permanente Colorado and Kaiser Permanente Colorado and Kaiser Permanente Colorado and Kaiser Permanente Colorado
• Keywords: Bias correction ; Penalized maximum likelihood ; Rare events ; Vaccine safety study ; Self-controlled case series

Abstract

The self-controlled case series(SCCS)method is used to examine the association of vaccination with adverse events using only data from patients who experienced such events. Conditional Poisson regression models are used to estimate relative incidence and they perform well in large or medium case samples. However,in most vaccine safety studies, the adverse events are rare and the MLE method is often biased. Several bias correction methods have been studied in logistic regression, but none of these methods have been examined in the SCCS design. In this study, two bias correction approaches- penalized maximum likelihood and bias reduction after maximum likelihood estimation-are evaluated in small samples under SCCS design. Simulation studies show that performance of bias correction depend on number of cases, strength of association, and the ratio of the risk period to the observation period