Epistemic AI Analysis Reveals Strategic Business Decisions as the Leading Driver of Phase 3 Clinical Trial Terminations

NEW YORK — (January 8, 2026) — In a comprehensive longitudinal study published in Nature Reviews Drug Discovery, Epistemic AI has identified a widening gap between clinical trial volume and successful completion. The research shows that while global trial initiations have climbed steadily since 2013, the rate of premature terminations in critical late-stage phases has grown even more rapidly.The analysis, conducted with EpistemicGPT and the Epistemic Knowledge Graph, underscores a critical pivot in pharmaceutical decision-making. Mapping over ten years of trial outcomes, Epistemic AI found that business-led catalysts and portfolio rebalancing are now the most frequent triggers for Phase 2 and 3 terminations.Key Research FindingsIn an analysis of 3,180 terminated trials between 2013 and 2023, the study identifies a stark trend: the rate of late-stage terminations has doubled over the last decade, jumping from 11% to 22%. While operational hurdles and safety concerns remained remarkably stable, the surge was driven primarily by “Strategic and Business” factors. Those factors such as M&A activity and portfolio reprioritization, now outpace clinical efficacy as the leading cause of trial failures.Heather Bowling, PhD, lead author, and VP of Scientific Development at Epistemic AI, indicated that the goal of the study was “to draw attention to the termination trends which may inform some decision-making and ultimately accelerate treatments for patients."Reducing Sunk Costs in Drug DevelopmentPhase 3 trials are notoriously resource-intensive, with costs ranging between $50-250MM per asset. Consequently, it would be far more advantageous to terminate unsuccessful programs during the less resource-intensive Phase 2 stage. In the article the researchers examine the specific reasons for terminations in hopes of shedding light on how these factors are translating into everyday drug development practices. Understanding these drivers can be critical for optimizing any pipeline and resource allocation.“Our multi-agent AI platform allows researchers to look beyond the individual trial to see the broader connections across the trial landscape," said Stefano Pacifico, CEO and founder of Epistemic AI. “By identifying these high-frequency themes, and we can do this for many other connections not only in trial termination reasons, Epistemic AI helps sponsors bridge the gap between better science and better planning, and results.”AuthorshipThis analysis has been published in Nature Reviews Drug Discovery (2025) titled "Analysis of phase II and phase III clinical trial terminations from 2013 to 2023." Authors: Heather Bowling, Arianna Cocucci, Da Chen Emily Koo, and Richard K. Harrison.https://www.nature.com/articles/d41573-025-00208-6