Advanced Statistical Modeling for Complex Biomedical and Clinical Data

This session presents recent methodological advances for analyzing complex and high-dimensional data structures, with emphasis on longitudinal processes, survival and recurrent event modeling, and modern evidence synthesis. The talks cover new joint modeling strategies for longitudinal change points and recurrent events, as well as Bayesian retrospective time-scale joint models for linked longitudinal–survival outcomes. Advances in survival methodology will include bounded hazard ratio models for evaluating time-to-treatment effects, methods for reconstructing individual patient data in competing-risks settings, and microsimulation frameworks for assessing adaptive strategies. Additional contributions feature genetic-algorithm-derived single-index modeling for longitudinal data, meta-analytic methods for medians using asymmetric pseudo-likelihoods, and exact Poisson two-stage designs for landmark time-to-event evaluations. Overall, these presentations highlight timely and practical solutions for modeling nonlinear trajectories, dynamic treatment processes, heterogeneous data sources, and complex event structures. The session will appeal to researchers working in longitudinal analysis, survival modeling, Bayesian methodology, microsimulation, and statistical design and analysis of complex studies