An Analysis Of Chatgpt As Educational Resources For Patients About Refractive Surgery

To assess the suitability and comprehensibility of the medical information offered by ChatGPT-4, an Artificial intelligence-driven conversational search engine, concerning commonly performed refractive surgeries such as LASIK, PRK, and SMILE.  

Artificial intelligence search engine-based study.

We compiled comprehensive lists of frequently asked questions regarding the definition, prevalence, visual manifestation, diagnostic techniques, surgical and non-surgical treatment alternatives, postoperative details, complications associated with surgery, and visual prognosis of LASIK, PRK and SMILE. Each question was posed three times on the ChatGPT-4 online platform. The data for this study were collected on January 01, 2024, and two cataract and refractive specialists independently evaluated the adequacy of the responses. The content's readability was evaluated using Readable, an internet program created for this purpose.

The average Flesch Kincaid Grade Level and Flesch Reading Ease Score were 11.25±0.4, and 34.3±3.75 for PRK, 11.48±0.96 and 37.74±5.23 for LASIK, 11.69±0.87 and 36.6±6.89 for SMILE. The ratings suggest that the answers are challenging for the typical individual to read as the scores were below 70, and the questions required a college-level education to comprehend the content.

Almost most of the responses produced by ChatGPT-4 were consistently correct. On the other hand, ChatGPT in its present state do not provide a source of accurate information. The credibility and readability of responses of AI-based vehicles should be improved as they will be widely used in medicine. It is important to inform patients, physicians, and laypersons about the constraints associated with using these instruments for health and eye-related counseling.