Evaluating chatGPT's effectiveness in answering Turkish questions from chronic kidney disease patients

Objective: This study aims to evaluate the responses of artificial intelligence (AI)-based conversational platforms to questions posed by individuals with Chronic Kidney Disease (CKD) regarding their condition, treatment processes, and daily life. In this context, the alignment of the platforms' responses with current clinical guidelines was examined, and the accuracy, clarity, comprehensiveness, and limitations of the content were analyzed. Materials and Methods: A total of 100 CKD patients followed by the Department of Nephrology at Cerrahpaşa Medical Faculty were asked about issues they were curious about concerning their disease and what they would like to ask their physicians. A total of 291 questions were collected, the questions were separated to 4 groups and questions on the miscellaneous category were eliminate remaining 121 questions were evaluated Each question was submitted to two AI-based conversational platforms (ChatGPT-4o and Gemini 1.5) in 2025. The responses were independently evaluated by three nephrology specialists using the DISCERN and QAMAI scales. Descriptive statistics as well as appropriate comparative tests (Chi-square and Fisher's exact test) were used for statistical analysis. Results: ChatGPT was found to be significantly superior to Gemini in terms of clarity (p=0.006), goal achievement (p=0.011), and information currency (p<0.001). It also outperformed Gemini in the QAMAI dimensions of clarity (p<0.001) and usefulness (p=0.009). However, both models were inadequate in providing source citations. Conclusion: ChatGPT demonstrated stronger and more comprehensible content in patient education. AI-based tools may serve as supportive resources in informing patients and can also help reduce physicians' repetitive informational workload, thereby contributing to the mitigation of physician burnout. Keywords: Chronic Kidney Disease, artificial intelligence, ChatGPT, Gemini, outpatient care