SHM Converge Scientific Abstract ePoster Gallery-Pilot Study of Artificial Intelligence (Ai) Scribes on Hospitalists' Time Allocation and Patient-Provider Interactions

Pdf Summary

This pilot study examined whether AI scribes could reduce documentation burden and improve patient-provider interaction for hospital medicine providers. Using a stepped-wedge time-motion design across two academic medical centers and two community hospitals in one health system, researchers observed direct care hospitalists enrolled in an AI scribe pilot. Encounters were measured at serial time points using TimeCaT software to assess documentation time, direct patient care time, and provider eye contact time with patients. Analyses compared baseline to later time points and also explored differences by provider gender.<br /><br />Overall, AI scribe use was associated with a decrease in documentation time by 2.3 minutes and an increase in patient care time by 1.8 minutes, though these changes were not statistically significant. Eye contact with patients increased significantly by 3.3 minutes overall. When stratified by gender, female providers showed larger improvements: documentation time decreased by 5.5 minutes, patient care time increased by 1.9 minutes, and eye contact time increased by 4.7 minutes, with the latter statistically significant. Male providers showed smaller, non-significant changes.<br /><br />The findings suggest that AI scribes may help reduce documentation burden and allow more direct patient engagement, especially for some subgroups of providers. However, the study was small and underpowered, limiting the ability to detect significant effects. It was also limited to hospital medicine providers in one health system using a single AI scribe application.<br /><br />The authors conclude that these results are hypothesis-generating and support further research, including qualitative studies on provider experience and future evaluation of patient perspectives during AI-assisted clinical encounters.

Asset Subtitle

Sharmila Tilak

Meta Tag

Author List Jeffrey L. Schnipper, Sharmila Tilak

Category Research

Concept AI Scribe

Concept Documentation Time

Concept Direct Patient Care Time

Concept Eye Contact

Concept Documentation Burden

Distinguished Non-Finalist

Presenter Organization Brigham and Women's Hospital

Presenting Author Sharmila Tilak

Track Technology in Hospital Medicine

Keywords

AI scribes

documentation burden

hospital medicine

patient-provider interaction

time-motion study

TimeCaT software

eye contact

patient care time

provider gender differences

stepped-wedge design

AI Scribe

Documentation Time

Direct Patient Care Time

Eye Contact

Documentation Burden