The Understandability and Readability of Online Patient Education Materials and ChatGPT Responses Regarding...

IntroductionXerostomia is a common condition that can impair quality of life. Google and ChatGPT are increasingly used for patient education, though the readability and understandability of this information remain unclear. This study evaluated commonly asked questions about xerostomia and assessed the quality of online and AI-generated patient educational material (PEM).Methods"SEO minion" was used to identify Google People Also Ask (PAA) questions from four primary search terms. Questions were categorized using Rothwell's classification and entered into ChatGPT. Online sources and ChatGPT responses were evaluated using the Flesch Kincaid Grade Level (FKGL), where higher scores indicate higher reading grade level; Flesch Reading Ease (FRE), where lower scores indicate more difficult readability (0 (complex) to 100 (easy)); JAMA benchmark criteria, where higher scores indicate greater source quality (0-4); and Patient Education Materials Assessment Tool (PEMAT), where higher scores indicate greater understandability (0-100).ResultsA total of 117 unique PAA questions were identified, most of which were value-based (55.6%) or fact-based (42.7%). Online sources had a mean JAMA score of 1.09, mean FKGL of 9.87, mean FRE of 51.37, and mean PEMAT score of 79.92. ChatGPT responses performed worse, with a mean FKGL of 15.24, mean FRE of 30.14, and mean PEMAT score of 72.33.ConclusionsBoth online PEM and ChatGPT responses on xerostomia exceeded recommended reading levels, with ChatGPT generating substantially less readable content. Although many materials met understandability thresholds, their elevated reading level may still limit accessibility. More readable, reliable, and patient-centered educational materials are needed.