Optimizing vaccine trial design: A novel promising zone approach with adaptive Interim analysis for managing delayed effects

Title : Optimizing vaccine trial design: A novel promising zone approach with adaptive Interim analysis for managing delayed effects

Abstract:

Vaccines can exhibit varying delayed effects among individuals, posing challenges in trial design and analysis when using the log-rank test. This can lead to significant power loss and complexities in making interim decisions in adaptive designs. In this presentation, we introduce a novel promising zone design with an interim analysis approach. We utilize a piecewise proportional hazard model with random lag time to model individual survival and propose an adaptive design tailored for vaccines with diverse delayed effects. Our methodology includes solutions for calculating conditional power and adjusting the critical value for the log-rank test using interim data.

By categorizing the sample space into unfavorable, promising, and favorable zones based on survival parameter re-estimations and interim analysis results, we adjust sample sizes accordingly. Through simulations, we illustrate that our approach yields higher overall power compared to fixed sample designs and is comparable to matched group sequential trials. Furthermore, we validate that adjusting the critical value effectively manages the type I error rate. Lastly, we provide practical recommendations for implementing our method in cancer immunotherapy trials, emphasizing its potential to enhance trial efficiency and decision-making processes.

Biography:

Dr. Depeng Jiang is an associate professor in the Department of Community Health Sciences at the University of Manitoba and currently serves as the Director of the Biostatistical Consulting Unit at the George & Fay Yee Centre for Healthcare Innovation. With extensive experience in providing statistical consulting services to diverse clients and training students and researchers in statistical analysis methods, Dr. Jiang focuses on conducting innovative research with a significant impact on quantitative methods in medical and behavioral sciences. His research interests primarily revolve around longitudinal analysis and multilevel models, person-centered statistical approaches such as latent class analysis and growth mixture models, structural equation models, clinical trial design, and mental health program evaluation.