- This event has passed.
Title: Non-inferiority testing under generalized Poisson distribution
Venue: ISRT seminar room
Speaker: Md Abu Manju, PhD
Department of Mathematics and Computer Science
Eindhoven University of Technology, Eindhoven, The Netherlands
Abstract: In recent years, non-inferiority studies for count data have been increasingly used in the evaluation of new test methods or new treatments (e.g., microbiological test methods, particle counters or drugs, vaccines) and the Poisson distribution is commonly assumed as it provides a standard framework for the analysis of count data. A generalization of the Poisson distribution, referred to as the generalized Poisson distribution (GPD), which models not only overdispersion, but also underdispersion, may be more appropriate in certain applications, including comparison of microbiological test methods. We therefore propose tests for the assessment of non-inferiority and sample size calculation procedures under GPD. Asymptotic likelihood ratio test (LRT), Wald and Exact conditional tests are derived, and the type I error rate and statistical power are computed for these three tests based on simulations. In terms of type I error rate and statistical power, LRT and Wald test perform similarly, but the Exact test is conservative and has less type I error rate and power than the LRT and Wald test. Finally, expressions are derived for calculating the sample sizes that yield sufficient power to test the non-inferiority of the new methods or treatments based on Wald test and LRT.