The ASA Biopharmaceutical Section Regulatory-Industry Statistics Workshop is held annually in Rockville, MD to provide biostatisticians with the opportunity to collaborate and expand the frontiers of statistical knowledge. The 2025 ASA conference centered around the future of statistics in the era of artificial intelligence (AI) and machine learning (ML). Two statisticians attended on behalf of PROMETRIKA to explore current AI/ML research efforts and use cases from statisticians across industry, academia, and regulatory agencies. Across a three-day series of short courses, speaker sessions, panels, roundtables, and poster sessions, a clear message emerged that AI/ML has a strong potential to improve statistical frameworks throughout the drug development life cycle.
Aside from the established benefit of AI in drug discovery and operational efficiencies, presenters at the ASA conference emphasized novel applications in pre-clinical, clinical, and post-market studies. ML has been utilized to replace or minimize the use of animal-based pharmacokinetic studies, and leveraged for dose range finding, site selection, subject recruitment and retention, data management, processing of big data from real world data or digital health technologies, optimization of drug manufacturing conditions, pharmacovigilance studies, safety monitoring for adverse drug experiences, and more. The FDA has increasingly encouraged the use of real-world data when appropriately applied to clinical trials, but external data often proves messy, complex, or even incomparable to internal data. When implemented properly, AI can extract key features of these large datasets in hybrid or post-market trials. PROMETRIKA has experience working with trial designs that include external control groups, which are valuable when the standard randomized controlled trial poses challenges in rare diseases, slow study recruitment, or resource limitations.
Conference attendees also convened to discuss how the role of a statistician may diverge from or overlap with the role of a data scientist. While statisticians and data scientists may prioritize different aspects of the clinical trial, ML techniques fundamentally build upon statistical principles. The implementation of ML models thus requires the oversight and expertise of statisticians to ensure that they satisfy statistical assumptions and are both dynamic and fit-for-use. The FDA encourages early regulatory engagement for any novel methodologies to maintain transparency for representative, accurate, and traceable models. Looking ahead, statisticians may begin to step into the role of a data scientist or work alongside them for continued progress and development.
PROMETRIKA is dedicated to remaining innovative in the wake of emerging methodologies that promise improved outcomes for patients and our pharmaceutical partners. Our team is continually discussing how to integrate new technologies into our everyday workflow while maintaining a high standard of data integrity and quality.
