Assessing the accuracy and readability of ChatGPT 4.0's original and simplified responses to common patient questions regarding periacetabular osteotomy

Abstract

Abstract Purpose The study aimed to evaluate the accuracy, comprehensiveness, and readability of responses generated by ChatGPT 4.0 to 30 common patient questions about the Bernese periacetabular osteotomy (PAO). Methods Two fellowship‐trained orthopaedic surgeons specializing in hip preservation selected thirty questions from a prior study identifying common PAO questions on social media. Each question was entered into ChatGPT 4.0, and the surgeons independently graded responses. Responses were evaluated using an established grading system. Accuracy and comprehensiveness were assessed based on the concordance of response content with current literature. Readability was analysed by calculating the Flesch‐Kincaid Grade Level and Flesch‐Kincaid Reading Ease. Results Regarding accuracy and comprehensiveness, 98.3% of responses were graded as “excellent” or “satisfactory, requiring minimal clarification.” Readability analysis revealed an average Flesch‐Kincaid Grade Level corresponding to an 11th‐grade reading level (11.09 ± 1.47) and a mean Reading Ease score requiring college level reading comprehension (39.12 ± 8.25) for original responses, 8th‐grade reading level (8.16 ± 1.46) requiring high school to college level reading comprehension (51.53 ± 9.62) for simplified responses, and 7th‐grade reading level (7.09 ± 1.23) requiring high school level reading comprehension (62.46 ± 7.48) for 6th grade responses. Conclusion ChatGPT 4.0 offered excellent or satisfactory answers to the most common questions surrounding PAO. Asking ChatGPT 4.0 to simplify or respond at a specific reading level may increase the readability of responses. The 4.0 model has shown the potential to be a valuable adjunct for patient education, though the readability may need to be improved via simplified responses. Level of Evidence Level N/A.

Ähnliche Arbeiten