Reimagining OSCE Grading and Medical Education at UT Southwestern

The Next Generation of Evidence-Based Clinical AI with Dr. Travis Zack

Jesse chats with Dr. Travis Zack, Chief Medical Officer at OpenEvidence and Assistant Adjunct Professor and AI Researcher at the University of California, San Francisco. Together, they discuss OpenEvidence's mission and its rapid adoption among physicians; how OpenEvidence differs from general-purpose AI platforms; Travis' background and role at OpenEvidence; the rationale behind the organization's free access model and its long-term vision; partnerships between OpenEvidence and health systems; the organization's impact on clinical decision-making, physician practice, and medical education; the future of OpenEvidence and clinical AI; and much more. Jesse also reflects on this key headline: Rhode Island passes AI scribe opt-out law. In our Resource Link segment, we list one valuable resource you might want to check out. To view this link, subscribe, or find out more information about our podcast, visit www.HealthcareAIPioneers.com [https://www.HealthcareAIPioneers.com]. Want to be a sponsor, marketing partner, or guest, or provide feedback on the podcast? Email us at Info@HealthcareAIPioneers.com [Info@HealthcareAIPioneers.com]. Music by Turning Pages [https://pixabay.com/users/cfl_turningpages-50915814/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843] from Pixabay [https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843].

Implementing AI in Radiology and Clinical Decision-Making with MedStar Health

Jesse chats with Dr. Raj Ratwani, Vice President of Scientific Affairs at the MedStar Health Research Institute, Director of the MedStar Health National Center for Human Factors in Healthcare, and a Professor of Emergency Medicine at Georgetown University School of Medicine; and Dr. Ross Filice, Chief of Imaging Informatics, Director of Research in the Department of Radiology, and Professor at MedStar Georgetown University Hospital. Together, they discuss MedStar Health's integration of an AI clinical decision support tool, "AI Doc," into radiology care; human factors and conditions under which clinical AI tools are likely to work; how radiology residents at Georgetown are allowed to use AI Doc; concerns surrounding deskilling in AI-assisted clinical environments; post-deployment surveillance and monitoring of AI Doc; how to build AI governance structures in healthcare settings; whether AI could eventually replace radiologists; and much more. Jesse also reflects on this key headline: AMA Adopts Policy Pushing Back on AI Creep in Medicine. In our Resource Link segment, we list one valuable resource you might want to check out. To view this link, subscribe, or find out more information about our podcast, visit www.HealthcareAIPioneers.com [https://www.HealthcareAIPioneers.com]. Want to be a sponsor, marketing partner, or guest, or provide feedback on the podcast? Email us at Info@HealthcareAIPioneers.com [Info@HealthcareAIPioneers.com]. Music by Turning Pages [https://pixabay.com/users/cfl_turningpages-50915814/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843] from Pixabay [https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843].

Reimagining OSCE Grading and Medical Education at UT Southwestern

Jesse chats with Dr. Thomas Dalton, Associate Professor of Internal Medicine and Geriatrics at UT Southwestern Medical Center, and Dr. Andrew Jamieson Assistant Professor in the Lyda Hill Department of Bioinformatics at UT Southwestern Medical Center and Principal Investigator of Jamieson Lab. Together, they discuss objective structured clinical examinations, how UT Southwestern created a model to help solve problems with OSCEs, inter-rater reliability in the human baseline in their research data compared with benchmarks, how the resulting system was operationally deployed for real students being graded on real exams, what happens to human graders who are no longer needed for OSCE scoring, applying multimodal AI to video recordings of student physical exams, how AI assessment tools translate from medical school into graduate medical education, where AI-powered assessment could be headed, and much more. Jesse also reflects on this key headline: OpenAI launches ChatGPT for clinicians. In our Resource Link segment, we list one valuable resource you might want to check out. To view this link, subscribe, or find out more information about our podcast, visit www.HealthcareAIPioneers.com [https://www.HealthcareAIPioneers.com]. Want to be a sponsor, marketing partner, or guest, or provide feedback on the podcast? Email us at Info@HealthcareAIPioneers.com [Info@HealthcareAIPioneers.com]. Music by Turning Pages [https://pixabay.com/users/cfl_turningpages-50915814/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843] from Pixabay [https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843].

Improving Lean Methodology and Clinical Analysis at UCSF and ZSFG

Jesse chats with Dr. Christopher (Toff) Peabody, Associate Clinical Professor of Emergency Medicine at the University of California San Francisco, Founder and Director of the UCSF Acute Care Innovation Center, and Associate Chief Medical Officer for Performance Excellence at Zuckerberg San Francisco General Hospital; and Dr. Lucas Zier, Interventional Cardiologist and Pulmonary Hypertension Specialist at ZSFG and UCSF Health. Together, they discuss the implementation of an LLM embedded in ZSFG's lean performance improvement process, how combining lean methodology with an LLM works in practice, overcoming data quality challenges with the LLM, a predictive model for readmissions for heart failure, a generative AI tool that summarizes patients' social and behavioral determinants of health, closing the gap between what a clinician intuitively knows they want from an AI tool and what they can actually articulate precisely enough for a developer to build, and much more. Jesse also reflects on this key headline: NBC News investigation reveals OpenEvidence is now used by approximately 65% of U.S. physicians. In our Resource Link segment, we list one valuable resource you might want to check out. To view this link, subscribe, or find out more information about our podcast, visit www.HealthcareAIPioneers.com [https://www.HealthcareAIPioneers.com]. Want to be a sponsor, marketing partner, or guest, or provide feedback on the podcast? Email us at Info@HealthcareAIPioneers.com [Info@HealthcareAIPioneers.com]. Music by Turning Pages [https://pixabay.com/users/cfl_turningpages-50915814/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843] from Pixabay [https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843].

A Close Look at AI Drafting and Documentation at Mayo Clinic

Jesse chats with Dr. Heather Heaton, Associate Professor of Emergency Medicine, Consultant, and Vice Chair of Clinical Systems Oversight at Mayo Clinic; Dr. Shant Ayanian, Academic Hospitalist and Data Scientist in the Division of Hospital Internal Medicine at Mayo Clinic; and Angie Griffin, Director of Strategy and Innovation for Clinical Systems at Mayo Clinic. Together, they discuss a study evaluating Epic's LLM-based hospital course drafting tool against human-written summaries across 100 real hospitalizations at Mayo; whether AI performance compared to humans is a win for AI, a damning statement about human baseline quality, or both; design or workflow interventions that can prevent the "rubber stamp" problem of clinicians signing AI-generated documentation without review; which tools in Mayo's growing AI portfolio get rigorous independent evaluation; Mayo's governance structure for clinical AI tools; recommendations for health systems that considering Epic's AI hospital course drafting tool; where AI-assisted clinical documentation could be headed; and much more. Jesse also reflects on this key headline: Researchers at Harvard and Beth Israel Deaconess found that OpenAI's o1 model outperformed two attending internal medicine physicians on real-world emergency department clinical reasoning. In our Resource Link segment, we list one valuable resource you might want to check out. To view this link, subscribe, or find out more information about our podcast, visit www.HealthcareAIPioneers.com [https://www.HealthcareAIPioneers.com]. Want to be a sponsor, marketing partner, or guest, or provide feedback on the podcast? Email us at Info@HealthcareAIPioneers.com [Info@HealthcareAIPioneers.com]. Music by Turning Pages [https://pixabay.com/users/cfl_turningpages-50915814/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843] from Pixabay [https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=479843].