Enhancing Health Literacy in Endourology: Using Artificial Intelligence to Improve Readability of Patient Education Materials

Abstract

Purpose Patient education materials (PEMs) enhance healthcare access and inclusivity, particularly for individuals without clinical backgrounds. However, many PEMs exceed recommended readability levels. This study evaluated whether artificial intelligence (AI)-assisted editing using ChatGPT-4o (OpenAI, San Francisco, CA, USA) could improve the readability of endourology PEMs related to prostate cancer, nephrolithiasis, bladder cancer, and kidney cysts. Methods Twenty-one publicly available PEMs from the American Urological Association (AUA) were analyzed. Each document was uploaded into ChatGPT-4o with instructions to rewrite the text to an eighth-grade reading level or lower while preserving content and word count. Readability of original and AI-modified PEMs was assessed using ReadabilityFormulas.com across six validated indices: Flesch-Kincaid Grade Level (FKGL), Simple Measure of Gobbledygook (SMOG), Gunning-Fog Index (GFI), Coleman-Liau Index (CLI), Automated Readability Index (ARI), and Flesch Reading Ease (FRE). Pre- and post-AI readability scores were compared using paired two-tailed t-tests. Results Across all indices and disease categories, AI modification significantly improved readability. Unmodified PEMs were written at approximately the ninth to 11th grade level, whereas AI-modified versions were reduced to the fourth to sixth grade range (p < 0.001 for all comparisons). AI-modified PEMs also demonstrated reduced variability across readability scores, indicating improved standardization. Conclusions ChatGPT-4o significantly improved the readability of AUA endourology PEMs, aligning them with established health literacy recommendations. AI-assisted editing represents a scalable and standardized approach to improving patient comprehension and accessibility of urologic education materials.

Ähnliche Arbeiten