National Academy of Medicine Workshop Explores Generative AI in Healthcare

The National Academy of Medicine recently held a workshop on October 25th, focusing on Generative AI and Large Language Models in Health and Medicine. The event served as a platform for health system executives and industry stakeholders to discuss various issues surrounding the governance, regulation, and deployment of AI in healthcare.

The Evolution of AI from Tool to Assistant

A notable point of discussion at the workshop was the transition of AI from a mere tool to a valuable assistant within the healthcare sector. Dr. Vincent Liu, a senior research scientist at Kaiser Permanente’s Northern California Division of Research, highlighted this transformation and its implications. While AI initially functioned as a controlled tool, it has now evolved into an intelligent assistant that can fulfill multiple roles, such as a reference librarian, medical resident, translator, patient liaison, and scribe.

In this new phase, healthcare providers are learning to interact with AI as a colleague. Dr. Liu emphasized the importance of understanding how to direct AI effectively, ensuring that the interaction is maximally efficient for future healthcare applications.

Unlocking AI’s Potential

Dr. Nigam Shah, a professor of Medicine at Stanford University and the chief data scientist for Stanford Health Care, urged participants to consider the reasons behind the previous shortcomings of AI in healthcare. He emphasized the interplay between machine learning models, policies, and the capacity to take action, as well as the net benefits of those actions.

Dr. Shah explained that many predictive models had been developed in the past for various healthcare applications, such as population health, readmission predictions, and sepsis predictions. However, the lack of appropriate policies and work capacity designs often hindered the realization of the potential benefits. He expressed concern that the healthcare industry must learn from these previous experiences and prepare for the challenges posed by generative AI, which can significantly accelerate processes and create additional complexities.

Creating an Assurance Lab

Dr. Gil Alterovitz, the Department of Veterans Affairs’ chief AI officer and director of the VA National Artificial Intelligence Institute, shared the VA’s approach to handling AI. Several years ago, the VA initiated its AI strategy and established a task force and an AI working group to bring together various offices that leverage AI.

The VA also created an agency-wide trustworthy AI framework and a list of AI use cases. It established a collaborative, shared AI governance structure that spans different medical centers and feeds into the national level.

Dr. Alterovitz stressed the importance of creating an “assurance lab” where the interplay between models, work capacity, and policies for generative AI could be analyzed. He highlighted the need for industry trials and transparency, aiming to improve the AI deployment process within healthcare.

Potential Applications of Generative AI

Jackie Gerhart, M.D., a family medicine physician and clinical informaticist at Epic, presented specific use cases for generative AI in healthcare. These cases included generating concise chart summaries tailored to various user types and specific instances. Another application, “Messaging Made Easy,” focused on generating draft responses for patient inquiries, aiming to expedite response times for clinicians.

Dr. Steven Waldren, chief medical informatics officer at the American Academy of Family Physicians, discussed the impact of AI on documentation. He reported over a 70 percent reduction in documentation time for doctors using AI solutions, with the addition of generative AI expected to enhance efficiency further. One solution focused on creating problem-oriented summaries for patients, reducing physician time by 60 percent.

The workshop provided a comprehensive overview of generative AI’s potential in healthcare. It shed light on the transformative shift from AI as a tool to AI as a collaborative assistant and emphasized the importance of creating a safe space for analyzing AI models’ interplay with work capacity and policies to ensure successful deployment in the healthcare sector. The participants discussed practical applications and highlighted the potential benefits of generative AI in various healthcare scenarios.

Dave Graff

Leave a Reply

Your email address will not be published. Required fields are marked *

Next Post

Dell Collaborates with Meta to for Deployment of Llama 2 Models

Thu Nov 2 , 2023
In a strategic partnership, Dell Technologies is teaming up with Meta to facilitate the deployment of Meta’s Llama 2 models on Dell’s generative AI (GenAI) IT infrastructure. This collaboration is aimed at simplifying the implementation of generative AI solutions for Dell customers, enabling them to harness the power of large […]
Chris Kelly

You May Like