Applications of Generative Artificial Intelligence for Strabismus Surgery Video-Based Education.

OBJECTIVE

To develop educational artificial intelligence (AI)-generated videos for patients undergoing strabismus surgery and assess patient perceptions of these videos.

DESIGN

Prospective interventional study.

SUBJECTS: Adult patients with strabismus seen at University of California San Diego who were evaluated and subsequently underwent strabismus surgery.

METHODS

A pipeline of AI software including Claude Sonnet, Sora, Visla, and Tavus were used to generate a script educating patients on strabismus surgery, create avatars of real ophthalmologists, and AI-generated video clips. Surveys on Google Forms were created to (1) preoperatively assess whether patients could discern whether the videos were AI-generated and educational and (2) postoperatively whether the video remained educational. These videos were disseminated to patients once consent for strabismus surgery was obtained, and surveys were given to patients preoperatively and postoperatively.

MAIN OUTCOME MEASURES: Patient perceptions and educational value of the video were assessed using Likert scales and yes versus no questions, which were averaged or enumerated, respectively. Subjective comments were also thematically assessed.

RESULTS

Thirty patients who underwent strabismus surgery were included in this study. The mean age was 54.0 ± 19.8 years with the majority undergoing surgery for the first time (n = 23 [76.7%]). Patients responded that these videos improved their understanding of strabismus (4.3 ± 0.9), strabismus surgery (4.2 ± 1.1), and eased their concerns (4.5 ± 0.8). Additionally, patients generally did not recognize that the video was AI-generated (4.3 ± 1.1), responded that they would watch more AI-generated videos in the future (4.7 ± 0.7), and would recommend AI-generated content to others (4.7 ± 0.65). When surveyed postoperatively, this sentiment persisted. Subjectively, patients generally emphasized how this video was both educational and reassuring.

CONCLUSIONS

In this proof-of-concept study, patients generally felt AI-generated videos for strabismus surgery added educational value. Additionally, most patients perceived the AI-generated content positively. Future studies with larger sample sizes and outcome-based and validated educational endpoints are needed to prospectively evaluate the clinical impact of AI-generated educational content on other conditions and interventions. This approach may open new opportunities to improve patient understanding and ability to contribute to their care.

FINANCIAL DISCLOSURES: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.