An Application of Bayesian Approach for Testing Non-inferiority Case Studies in Vaccine Trials

Abstract

Non-inferiority designs are often used in vaccine clinical trials to show a test vaccine or a vaccine regimen is not inferior to a control vaccine or a control regimen. Traditionally, the non-inferiority hypothesis is tested using frequentist methods, e.g., comparing the lower bound of 95 % confidence interval with a pre-specified non-inferiority margin. The analyses are often based on maximum likelihood methods. Recently, Bayesian approaches have been developed and considered in clinical trials due to advances in Bayesian computation such as Markov chain Monte Carlo (MCMC) methods. Some of the advantages of using Bayesian methods include accounting for various sources of uncertainty and incorporating prior information which is often available for the control group in non-inferiority trials. In this chapter, we will illustrate the use of Bayesian methods to test for non-inferiority with real examples from vaccine clinical trials. Consideration will be given to issues including the choice of priors or incorporating results from historical trial, and their impact on testing non-inferiority. The pros and cons on using Bayesian approaches will be discussed, and the results from Bayesian analyses will be compared with that from the traditional frequentist methods.