PB0419 - Artificial Intelligence and Venous Thromboembolism: Talking to the Experts

Sunday, June 23, 2024

13:45 – 14:45 ICT

Room: Exhibition Hall

Co-Author(s)

AG

Anuranita Gupta

Internal Medicine Resident
Beth Israel Deaconess Medical Center
Boston, Massachusetts, United States
Barbara D. Lam, MD

Fellow
Beth Israel Deaconess Medical Center
Boston, Massachusetts, United States
SZ

Sabrina Zerbey

Clinical Research Assistant
Division of Hemostasis and Thrombosis, Beth Israel Deaconess Medical Center
Boston, Massachusetts, United States
WR

William Robertson

Associate Professor of Emergency Healthcare
Weber State University
Ogden, Utah, United States
RR

Rachel P. Rosovsky

Director of Thrombosis Research in Division of Hematology
Massachusetts General Hospital
Boston, Massachusetts, United States
LL

Leslie Lake

President
National Blood Clot Alliance
New York City, New York, United States
LD

Laura Dodge

Assistant Professor of Obstetrics, Gynecology and Reproductive Biology
Beth Israel Deaconess Medical Center
Boston, Massachusetts, United States
AA

Alys Adamski

Epidemiologist
Centers for Disease Control and Prevention
Atlanta, Georgia, United States
NR

Nimia Reyes

Medical Officer
Centers for Disease Control and Prevention
Atlanta, Georgia, United States
KA

Karon Abe

Branch Chief
Centers for Disease Control and Prevention
Atlanta, Georgia, United States
IV

Ioannis Vlachos

Co-Director of the Bioinformatics Program, Cancer Research Institute
Beth Israel Deaconess Medical Center
Boston, Massachusetts, United States
Jeffrey I. Zwicker, MD

Chief of Hematology Service
Hematology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, USA
New York, New York, United States
MS

Mara Schonberg

Assistant Professor
Beth Israel Deaconess Medical Center
Boston, Massachusetts, United States

Presenting Author(s)

Rushad Patell, MBBS

Assistant Professor
Beth Israel Deaconess Medical Center
Boston, Massachusetts, United States

Background: Venous thromboembolism (VTE) is a significant cause of preventable death among hospitalized patients. Artificial intelligence (AI) may be useful in improving VTE management. We previously conducted a national survey to explore informatician attitudes towards using AI for VTE management and interviewed a subset for the current study.

Aims: To identify major themes in the perceived strengths and weaknesses of AI in improving VTE management. This involves exploring strategies and challenges to guide future research and implementation.

Methods: We conducted 30–60-minute semi-structured interviews via videoconference which were recorded and transcribed. A codebook was developed, and two coders separately reviewed the interviews using inductive analysis. Overarching themes were agreed upon by discussion of the common codes between coders. Representative quotes are presented to illustrate themes.

Results: Of the informaticians surveyed, 24 agreed to be interviewed (6 clinician informaticians, 9 data scientists, 3 biomedical/computational biologists and 6 other designations). Participants reported that AI models can increase accuracy and efficiency of clinical practice and can serve as decision support and management tools. Participants also reported that VTE management specifically can be managed with AI given that it is a clearly defined problem. Challenges included subpar quality of training data, potential for model inaccuracy, and ethical concerns of bias and privacy. There were multiple suggestions for implementation of AI. Participants noted that AI should be integrated into the electronic health record, that it should have human oversight when being used, and should be able to explain how it came to its conclusion. Almost all participants agreed that patients would prefer interacting with a human over an AI tool.

Conclusion(s): Informaticians see AI as a promising tool for VTE management. Challenges to implementing AI will have to be addressed to create ethical and accurate models that reduce the burden on healthcare providers.