The Pfizer/Northeastern/ASA 2nd Symposium on Risks and Opportunities of AI in Pharmaceutical Medicine (AIPM) was successfully conducted on June 5, 2023, at the Portland Ocean Gateway in Portland, Maine.
The symposium is a timely and invaluable platform for distinguished statisticians, data scientists, regulators, and other professionals to address the challenges and opportunities of AI in pharmaceutical medicine; to foster collaboration among industry, academia, regulatory agencies, and professional associations; and to propose recommendations with policy implications for proper implementation of AI in promoting public health.
The theme of this year's conference was AI and Causality, and featured Nobel laureate Guido Imbens of Stanford University as the keynote speaker.
The symposium, which was kicked off with remarks by David Madigan (Northeastern University), Ron Wasserstein (American Statistical Association), and Kannan Natarajan (Pfizer Inc.), was attended both in person and virtually.
The carefully planned sessions addressed pertinent AI topics ranging from personalized medicine to use of real-world evidence (RWE) in medical research.
In one of the morning sessions, Tala Fakhouri of the US Food and Drug Administration (FDA) presented regulatory perspectives on AI in drug development, whileTianxi Cai of Harvard University addressed issues of reproducibility and interoperability of RWE. Sheraz Khan (Pfizer Inc.) shared results of a recent study involving prediction of respiratory illness from voice.
The talk by Michael Katehakis (Rutgers University) concerned the use of reinforcement learning for individualized treatments in clinical practice and trials, while Elizabeth Stuart (Johns Hopkins University) highlighted issues associated with combining EHR and clinical trial data to understand treatment effect heterogeneity. Bruce Church (Aitia) illustrated a causal AI tool using a cardiovascular disease case study. Anant Madabhushi (Emory University) shared ongoing work on AI for personalized medicine, and Michael Lu (Massachusetts General Hospital and Harvard Medical Schoo) presented results on AI use to predict risk from chest X-Rays and CTs. Lastly, Emre Kıcıman, Microsoft Research, described validation of critical assumptions associated with a causal analysis suite for Health Analyses.
A highlight of the symposium was a panel discussion, moderated by David Madigan, in which Aloka Chakravarty (FDA), Haoda Fu (Eli Lilly), George M.Hripcsak (Columbia University), Nigam Shah (Stanford University) and Mark van der Laan (UC Berkeley), exchanged views on a range of issues relating to causality and AI in biomedical research and development. The symposium was then concluded with a few remarks by Louisa Smith of Northeastern University.
Following the symposium, attendees were invited to a networking evening reception that also featured Rai Winslow (Northeastern University), who gave a stimulating presentation on "engineering" the future of healthcare.
Planning is now already underway for AIPM 2024, which is anticipated to address, among other topics, the impacts of emerging technologies, such as auto-GPT and other generative AI models, on healthcare research and delivery.
Further details of the 2023 symposium can be found at:
------------------------------
Demissie Alemayehu & David Madigan, Co-Chairs
------------------------------