Yulia Kozhevnikova

Title: BLOST: Bayesian Longitudinally Ordinal Sequential Trial Design for Evaluating Respiratory Disease Treatments
Date: Friday, March 7th
Time: 11:00am
Location: Zoom
Supervised by: Dr. Haolun Shi

Abstract: Given the urgent need to develop effective treatments for possible respiratory diseases, the traditional framework of clinical trial design may be inadequate for the rapid progression required in such cases. We have developed a Bayesian Longitudinally Ordinal Sequential Trial (BLOST) framework to optimize drug development processes and enhance resource efficiency in response to acute respiratory outbreaks. The design presents a Bayesian model based on longitudinally observed ordinal outcomes, which accounts for patient heteerogeneity and facilitates information borrowing across the time points. Our sequential framework is designed to compare the experimental treatment with a standard one through an extensive simulation study. We consider three analytical approaches: a standard Bayesian method based on Hamiltonian Monte Carlo, an enhanced version applying Bayesian model selection, and a conventional frequentist approach. Frequentist Type I and Type II error rates are maintained through parameter calibration. The conducted simulation study demonstrates improved operating characteristics of our design over conventional methods and illustrates its practical application.