PANEL 2 - AI FOR CLINICAL DECISION MAKING:

• What validation criteria must AI systems meet for regulatory bodies (such as the FDA, PDMA, or EMA), standardization organizations (such as NIST), or funding organizations (such as the NIH) to use the diagnostics and recommendations as reliable, accurate, and safe?
• What intermediate steps can the healthcare ecosystem stakeholders take to facilitate the development and operationalization of such standards?

Panelists:

• Dean Sittig, PhD, Biomedical Informatics, University of Texas Health Science Center 
• Gretchen Purcell Jackson, MD, PhD, FACS, FAMCI, FAMIA, President and Board Chair, AMIA, Vice President and Scientific Medical Officer and Associate Professor, Intuitive Surgical, Vanderbilt University Medical Center
• Sayon Dutta, MD, MPH, Assistant Professor of Emergency Medicine, Massachusetts General Hospital, Physician Lead, Emergency Medicine and Clinical Decision Support, Mass General Brigham Digital
• Session Chair: Steven E. Labkoff, MD, FACP, FACMI, FAMIA, Global Head, Clinical and Healthcare Informatics, Collaborating Scientist, Division of Clinical Informatics, Beth Israel Deaconess Medical Center.

 Please add on this page your comments that you would like to see discussed at this session.

View "AI in Healthcare: Risk Management, Trust, and Liability. Exploring Healthcare Risk and Risk Management in the AI World"

https://www.dcinetwork.org/library/recording-ai-healthcare-risk-managem…

AI's integration into healthcare continues to expand, raising pertinent questions, particularly regarding how to prevent errors and the course of action when errors occur. In the prevailing paradigm, the clinician assumes the ultimate role of decision-maker, consequently bearing the final accountability for their recommendations and subsequent outcomes. In contrast, the mechanisms by which AI systems formulate conclusions often remain opaque, making it notably challenging, if not implausible, to deconstruct the decision-making processes within these systems.  How do we, as a society, figure out what role AI systems can play responsibly inside the healthcare ecosystem, and if things go awry, then what?

This dilemma prompts an essential question: where should the responsibility rest? For example, what ramifications ensue when an AI system examining pathology slides overlooks a cancer diagnosis or a clinical decision support system steers a clinical course of action in error, and its guidance is followed?

We have assembled a distinguished panel of experts specializing in healthcare risk management to unwind these issues. Our panel encompasses the leader of the Harvard Risk Management Foundation, a legal health care risk expert well-versed in Managed Care, and a lawyer and consultant specializing in complex relationships, including those in the health information technology space.

This upcoming webinar pledges to explore a plethora of intricate inquiries, potentially illuminating a path to unraveling and comprehending the multifaceted challenges raised by the amalgamation of risks and benefits accompanying the advent of AI systems within the healthcare ecosystem.

Dear Colleagues. Echoing Yuri that it truly was a tremendous event together - I look forward to the action steps ahead together. Should you be interested - here's a short post that revisits an earlier 2019 MIT Sloan Management Review (yes, I realize MIT and not Harvard... however we can blend the two institutional insights I'm sure) examining Three People-Centered Design Principles for AI. Thoughts and feedback welcomed - including what was on the mark in 2019 and what should we consider going forward? https://www.linkedin.com/pulse/from-2019-updated-2023-three-people-cent… Onwards and upwards together, and with highest regards, -d. -- Principal, LeadDoAdapt Ventures, Inc. & Distinguished Fellow Henry S. Stimson Center, Business Executives for National Security