Related Experiment Video
Updated: Aug 5, 2026
02:37
Robot-Assisted Transcanal Endoscopic Ear Surgery for Congenital Cholesteatoma
Published on: December 15, 2023
Comprehensive Evaluation of AI Consent Forms in Otolaryngologic Surgery
Sholem Hack1, Rebecca Attal1, Armin Farzad2
1City St. Georges University London School of Medicine, Program Delivered by University of Nicosia at the Chaim Sheba Medical Center Ramat Gan Israel.
Summary
Large language models (LLMs) can generate clearer surgical consent forms. AI-generated forms were perceived as clear and accurate, though further clinical validation is needed.
Area of Science:
- Medical Informatics
- Artificial Intelligence in Healthcare
- Patient Communication
Background:
- Surgical consent documents often exceed average health literacy levels.
- Large language models (LLMs) present a potential solution for creating clearer, procedure-specific consent forms.
Purpose of the Study:
- To evaluate the clarity, clinical accuracy, and acceptability of consent forms generated by GPT-4 and Claude for otolaryngologic procedures.
Main Methods:
- Generated 20 AI consent forms (10 GPT-4, 10 Claude-2.1) using standardized prompts.
- Physicians rated forms for accuracy and usability; 300 adults rated clarity, trust, and comfort.
- Readability assessed using Flesch-Kincaid Grade Level (FKGL); AI forms compared to national templates.
Main Results:
- AI forms received high clarity ratings from lay participants.
- Claude showed numerically higher scores for clarity and signing comfort, but differences were not statistically significant.
- Experts rated GPT-4 forms as more accurate (p=0.034).
- AI forms had lower FKGL than templates, though slightly above target reading levels (8.8-9.4).
Conclusions:
- AI-generated consent forms are perceived as clear and clinically complete in a non-clinical setting.
- Trade-offs exist between perceived clarity and clinical detail depending on the AI model.
- Prospective clinical and legal validation in diverse patient populations is necessary.