Related Experiment Video
Updated: Jul 5, 2025
07:51
Hydra, a Computer-Based Platform for Aiding Clinicians in Cardiovascular Analysis and Diagnosis
Published on: September 26, 2018
7.6K
Trust in Machine Learning Driven Clinical Decision Support Tools Among Otolaryngologists
Hannah Chen1, Xiaoyue Ma2, Hal Rives1
1Sean Parker Institute for the Voice, Department of Otolaryngology-Head and Neck Surgery, Weill Cornell Medicine, New York, New York, USA.
The Laryngoscope
|January 17, 2024
Summary
Otolaryngologists are influenced by machine learning clinical decision support tools (ML-CDST), particularly when lacking diagnostic confidence. Specific explanations from ML-CDST increase physician trust and diagnostic judgment revision.
Area of Science:
- Medical Informatics
- Artificial Intelligence in Medicine
- Otolaryngology-Head & Neck Surgery
Background:
- Machine learning-driven clinical decision support tools (ML-CDST) are nearing integration into clinical practice.
- This study examines the impact of ML-CDST on diagnostic judgment within Otolaryngology-Head & Neck Surgery.
Purpose of the Study:
- To investigate whether ML-CDST influence otolaryngologists' diagnostic decisions.
- To assess the effect of ML-CDST explanation types on diagnostic judgment.
Main Methods:
- Otolaryngologists participated in a virtual experiment involving human-AI interaction.
- Participants diagnosed 12 video-stroboscopic exams and then revised diagnoses after ML-CDST output.
- ML-CDST explanations varied: none, general, or specific logic.
Main Results:
- Physicians were significantly more likely to change their diagnosis when reporting lower confidence (p=0.001).
- A specific explanation of the ML-CDST logic increased diagnostic judgment revision (p=0.048).
- Forty-five otolaryngologists were recruited for this study.
Conclusions:
- Otolaryngologists demonstrate susceptibility to ML-CDST diagnostic recommendations, especially when uncertain.
- Specific explanations enhance otolaryngologists' trust in ML-CDST outputs.
- The explainability of AI tools is crucial for clinical adoption.
