Improving the readability of trauma patient education materials: a ChatGPT solution demonstrated using materials by the Orthopaedic Trauma Association

ABSTRACT Introduction: ChatGPT is an artificial intelligence language model capable of understanding, contextualizing, and generating human-like text. The purpose of this study was to assess the ability of ChatGPT to rewrite orthopaedic trauma patient education materials at the recommended sixth grade level. Methods: The academic grade level of each of the 41 Orthopaedic Trauma Association (OTA/AO) online patient education articles was evaluated using the Flesh-Kincaid Grade Level (FKGL) and Flesch Reading Ease (FRE). Each article was then provided to ChatGPT along with instructions to simplify the readability of the text to a sixth grade level. The FKGL and FRE of the ChatGPT revised articles were calculated and compared with the original articles. Two orthopaedic trauma surgeons assessed the content of the revised articles and categorized them as “accurate,” “refinable,” or “insufficient” based on the preservation of information from the original articles. Results: ChatGPT significantly reduced the FKGL (8.2 ± 1.1‒5.7 ± 0.5, P < 0.001) and increased the FRE (65.5 ± 6.6‒76.4 ± 5.7, P < 0.001) of the OTA/AO patient education articles. Twenty-nine (70.7%) revised articles were accurate without modifications. Three (7.3%) articles required minor modifications, and 9 (22%) articles required substantial edits. Conclusion: ChatGPT can be used to simplify and enhance the readability of patient education materials. The average readability of the OTA/AO educational articles was changed from an eighth grade to a fifth grade level. However, nearly a third of the ChatGPT revised articles required revisions due to content omissions thus highlighting the importance of expert review. Level of Evidence: NA.