Readability of patient education materials in hand and wrist surgery: Can ChatGPT make a difference?

Abstract

The internet is a common resource among orthopaedic patients. Many orthopaedic Patient Education Materials (PEMs) are written above the recommended 6th-grade reading level. This study aims to evaluate the readability of online hand and wrist PEMs and determine whether AI can enhance their readability. 135 hand and wrist surgery PEMs were identified from the British Society for Surgery of the Hand, the American Society for Surgery of the Hand, and the American Academy of Orthopaedic Surgeons websites. Readability was assessed using eight formulae. The mean Reading Grade Level (RGL) was compared to the 6th and 8th-grade reading levels. 27 articles were input into ChatGPT with the following command: “Rewrite the following patient education materials using a 6th-grade reading level. The material should be ready to publish in the patient education section of a healthcare website.” The readability was then reassessed. The mean RGL of the 135 PEMs was 9.89 (range = 7.2-15), significantly higher than the 6th-grade level by 3.89 grade levels (95% CI, 3.7 -8; P<0.001) and the 8th-grade level by 1.89 grade levels (95% CI, 1.7 – 2.08; P<0.001). The 27 pre-simplification articles had a mean RGL of 10.4. The ChatGPT-simplified PEMs’ mean RGL was 8.54, exceeding the requested 6th-grade level by 2.54 grade levels (95% CI, 2.13 – 2.96; P<0.001). The simplified articles’ mean RGL was significantly lower than the original articles (95% CI, 1.2 – 2.49; p<0.001). Hand and wrist PEMs from leading healthcare organisations are written above the recommended levels. ChatGPT shows potential as an accessible tool to improve readability. Integrating AI into the creation and review process of PEMs could be a valuable strategy for improving readability.

Ähnliche Arbeiten