SHM Converge Scientific Abstract ePoster Gallery-Everything, Everywhere and All at Once: Exploring Hospitalists' Experiences Using Artifical Intelligence Tools for Work

Pdf Summary

This study explored how hospitalists are currently using artificial intelligence (AI) tools, what barriers they face, and where AI may be useful in the future. Researchers surveyed and held seven semi-structured focus groups on May 10, 2024 with members of the Hospital Medicine Reengineering (HOMERuN) Research Network. In total, 40 participants from 28 institutions joined the focus groups, and 32 completed the survey. Most survey respondents were physicians.<br /><br />The findings showed wide variation in AI use, with many participants describing a cautious but growing interest in adoption. Common barriers included lack of trust, concerns about reliability, absence of clear guidelines or standardization, high cost, and uncertainty about whether AI actually saves time or improves productivity. Participants also raised concerns that AI could affect clinical judgment, reduce critical thinking, and introduce errors or bias into decision-making, especially if used for documentation or reasoning.<br /><br />Despite these concerns, participants were enthusiastic about AI’s potential to reduce administrative burden, improve efficiency, support clinical decision-making, and help address workforce strain and burnout. Suggested use cases included pre-populating notes, scribing histories, generating differential diagnoses, supporting clinical decisions, creating discharge summaries, helping with coding and billing, assisting with follow-up tasks, and supporting research and population health work.<br /><br />Overall, the study found that hospitalists see real promise in AI, but adoption is limited by trust, cost, and lack of standards. The authors conclude that AI may improve care quality and provider satisfaction, but broader and safer implementation will require clearer guidance and attention to workflow, bias, and privacy concerns.

Asset Subtitle

Khooshbu Dayton

Meta Tag

Author List Amy Yu, Angela Alday, Angela Keniston, Catherine Callister, Gopi J. Astik, Kendall Rogers, Khooshbu Dayton, Kirsten N. Kangelaris, Marisha Burden, Matthew Sakumoto, Michelle Knees

Category Research

Concept Artificial Intelligence

Concept Hospitalist Role

Concept Trust

Concept Reliability

Concept Efficiency

Distinguished Non-Finalist

Presenter Organization University of Colorado School of Medicine

Presenting Author Khooshbu Dayton

Track Technology in Hospital Medicine

Keywords

artificial intelligence

hospitalists

AI adoption

clinical decision support

administrative burden

barriers to implementation

trust and reliability

workflow efficiency

burnout reduction

healthcare technology

Artificial Intelligence

Hospitalist Role

Trust

Reliability

Efficiency