Chasing the Frontier: Balancing Innovation and Validation in LLM-Driven Reviews
24 March, 1 PM (GMT), Online
Large language models continue to improve on a time-scale measured in months. Even if model development stopped today, the integration of these tools into systematic literature reviews (SLRs) would continue to transform workflows for years to come. In this presentation Dr. Disher will provide a summary of the current GenAI in SLR landscape, discuss how the nature of LLM development and applications has disrupted the academic evaluation/application landscape, and provide suggestions for how we can adapt in order to allow us to stay on the frontier while maintaining rigor and safety.
About Tim Disher:
About Tim: Tim Disher, PhD, RN, is the founder of Sandpiper Analytics, where he helps life-science and academic partners navigate the evolving intersection of systematic literature reviews, advanced statistical evidence synthesis, and emerging AI technologies. With over 12 years of experience spanning network meta-analysis, indirect treatment comparisons including IPD-based methods, and causal inference from real-world data. Tim specializes in developing rigorous, transparent analytic solutions that keep pace with rapid innovation statistical analysis and AI capabilities.
Tim has been an active contributor to the global HEOR community, including extensive methodological work presented at ISPOR covering Bayesian NMA, surrogate endpoint validation, longitudinal treatment-effect modeling, effect-modifier assessment, proportional-odds model development, and machine-learning–enabled approaches to indirect comparisons. His applied and methodological research is published widely in outlets such as JAMA Pediatrics, Pediatrics, Cochrane, and Value in Health.