♡ Code to Bedside: May 2023 Newsletter

May 2023

Director's Brief

Steven Lin, MD

Executive Director of HEA₃RT

Welcome! In this issue, we are excited to spotlight our partnership with four major academic institutions as part of the newly formed National Primary Care AI/ML Learning Collaborative, as well as a collaboration with the Division of Oncology to improve advance care planning with AI/ML. We also present preliminary results from our ongoing AI/ML-powered skin health screening project in the San Jose community with Google and Santa Clara Family Health Plan. See what we’re working on and join us!

Collaborator Spotlight

National Primary Care AI/ML Learning Collaborative, led by HEA₃RT, aims to advance a shared vision of AI/ML for primary care

Despite being the largest U.S. health care delivery platform, primary care is at risk of being left behind in the research, development, and implementation of AI/ML applications. Although primary care accounts for 52% of all care delivered – more than all other specialties combined – only 3% of FDA-approved AI/ML tools are intended for primary care, and less than 1% of AI/ML research funding goes to primary care.

The American Board of Family Medicine awarded grants to four family medicine departments to catalyze AI/ML research that has foundational and translational importance to primary care: University of Houston, University of Pittsburgh, University of California San Diego, and University of Texas Health Science Center at San Antonio. Over the next five years, a national learning collaborative formed by these departments, in collaboration with HEA₃RT as the coordinating center, will generate key learnings that could be applied broadly to support primary care AI/ML research and implementation across the country. The objective of the National Primary Care AI/ML Learning Collaborative is to strengthen primary care with AI/ML technologies designed by primary care, for primary care.

Project Spotlight

Proactively identifying oncology patients for advance care planning using artificial intelligence

In a recent survey of Californians about attitudes towards death and dying, almost 80% of respondents indicated that they would like to speak with their doctor about end-of-life care, but only 7% had ever had a doctor bring up the subject. Though conversations about end-of-life care, also known as advance care planning, have the potential to measurably improve health outcomes, these conversations happen too infrequently and too late, resulting in suboptimal care delivery.

To address this challenge, providers within the Stanford inpatient oncology service recently implemented an AI-enabled workflow to facilitate the timely identification of patients for advance care planning conversations prior to discharge from inpatient oncology. However, despite the theoretical potential of this tool, little is known of the human factors from patients and providers which influence the feasibility and acceptability of the AI-enabled workflow.

HEA₃RT is partnering with Dr. Ali Raza Khaki and the Stanford Division of Oncology to gather patient and provider perspectives surrounding the implementation of AI-enabled workflows in inpatient oncology–as well as the role of AI-technology in patient care delivery more generally– in a series of qualitative interviews. We look forward to using the insights gleaned from this study to inform a more nuanced approach to the implementation of novel health technologies across all patient care contexts.

Affiliate Faculty Spotlight

Ali Raza Khaki, MD

Faculty

Dr. Ali Raza Khaki is a medical oncologist and clinical assistant professor at Stanford University School of Medicine. As a practicing oncologist, Dr. Khaki treats patients with all forms of genitourinary cancer, and spends time attending on the inpatient oncology service at Stanford Hospital.

In addition to his clinical practice, Dr. Khaki is closely involved in clinical oncology research. His research interests include using real-world evidence to investigate the utilization of novel therapies in patients with genitourinary cancers, and exploring health care utilization and cost for patients with advanced cancer. Dr Khaki is currently collaborating with HEA₃RT on a study aiming to better understand the feasibility and acceptability of an AI-enabled workflow for identifying and flagging patients for advance care planning in inpatient oncology settings.

Outside of work, Dr. Khaki enjoys spending time with his family (including coaching his 5 and 7-year old kids little league baseball) and is an avid Seattle sports fan (go Seahawks!).

Team Member Spotlight

Julie Lee, MD, MPH

Clinical Informatics Fellow

Dr. Julie Lee is a Clinical Informatics Fellow and internal medicine physician. Her motto as an informatician is "functionality meets aesthetics," as human-centered design should not go remiss when building clinical digital tools. She grew up in Los Angeles and received her BA in Psychology from Columbia University, MPH from Yale University, and MD from the University of Buffalo.

At Stanford, Julie has led projects in clinical decision support and workflow improvement including integration of the PDMP Report into the electronic health record and refining the design, intake, and life cycle of Best Practice Advisory alerts. She is collaborating with HEA₃RT to learn more about clinical application of AI, specifically on comprehensively assessing the usability and feasibility of a population health management tool with AI risk prediction. She is also committed to making sure technology can reduce physician burnout and has piloted innovative tools to standardize InBasket response, triage, and action.

In her free time, Julie enjoys creating digital illustrations, discovering new food places, and teaching her cat new tricks.

Recent Events

Stanford 21st Annual Community Health Symposium

April 5, 2023 (Stanford, California)

HEA₃RT, Google, and Santa Clara Family Health Plan presented research titled: "Bridging the Gap Between Community, Academia, and Industry to Understand the Usefulness of an AI-Powered Dermatology Tool Among Everyday Users."

National Primary Care AI/ML Learning Collaborative Convening

May 10, 2023 (Virtual)

HEA₃RT led the second virtual convening of the National Primary Care AI/ML Learning Collaborative with researchers from University of Houston, University of Pittsburgh, University of California San Diego, and University of Texas Health Science Center at San Antonio.

Santa Clara Family Health Plan Older Adult Resource Fair

May 20, 2023 (San Jose, California)

HEA₃RT, Google, and SCFHP conducted their AI-powered skin health screening study at the community health fair – the fourth in a series of community-based participatory research events.

Recent Publications & Presentations

Voice assistants’ responses to questions about the COVID-19 vaccine: national cross-sectional study – JMIR Formative Research. 2023;7:e43007. Published February 8, 2023.

Advancing artificial intelligence and machine learning for primary care and population health – UCSF Department of Family and Community Medicine Research Meeting, San Francisco, CA. Presented March 17, 2023.

Bridging the gap between community, academia, and industry to understand the usefulness of an AI-powered dermatology tool among everyday users – Stanford 21st Annual Community Health Symposium, Stanford, CA. Presented April 5, 2023.

Advancing artificial intelligence and machine learning for primary care and population health – Virginia Tech Carilion School of Medicine Expert Conversation, Roanoke, VA. Presented April 19, 2023.

Can advance care planning AI models support better end-of-life care? – American College of Physicians Annual Meeting, San Diego, CA. Presented April 29, 2023.

Feasibility and acceptability of a U.S. national telemedicine curriculum for medical students and residents: multi-institutional cross-sectional study – JMIR Medical Education. 2023;9:e43190. Published May 8, 2023.

Integrating risk prediction tools with actionable insights into value-based health care – Stanford STARR Informatics Summit, Stanford, CA. Presented May 11, 2023.

Generative AI and the future of primary care: a case study – Stanford Generative AI and Medicine Seminar, Stanford, CA. Presented May 18, 2023.

The future of primary care and population health with artificial intelligence and machine learning – University Health Network Artificial Intelligence Seminar, Toronto, ON, Canada. Presented May 24, 2023.