Kacer Emine Ozdemir, Ipekten Funda
Department of Pediatrics, Aksaray University, Faculty of Medicine, Aksaray, Turkey.
Department of Biostatics, Erciyes University, Faculty of Medicine, Kayseri, Turkey.
J Paediatr Child Health. 2025 Jan;61(1):60-65. doi: 10.1111/jpc.16710. Epub 2024 Oct 29.
Artificial intelligence (AI) systems hold great promise in improving medical care and health problems.
We aimed to evaluate the answers by asking the most frequently asked questions to ChatGPT for the prediction and treatment of fever, which is a major problem in children.
The 50 questions most frequently asked about fever in children were determined, and we asked them to ChatGPT. We evaluated the responses using the quality and readability scales.
While ChatGPT demonstrated good quality in its responses, the readability scale and the Patient Education Material Evaluation Tool (PEMAT) scale used with materials appearing on the site were also found to be successful. Among the scales in which we evaluated ChatGPT responses, a weak positive relationship was found between Gunning Fog (GFOG) and Simple Measure of Gobbledygook (SMOG) scores (r = 0.379) and a significant and positive relationship was found between FGL and SMOG scores (r = 0.899).
This study sheds light on the quality and readability of information regarding the presentation of AI tools, such as ChatGPT, regarding fever, a common complaint in children. We determined that the answers to the most frequently asked questions about fire were high-quality, reliable, easy to read and understandable.
人工智能（AI）系统在改善医疗保健和健康问题方面具有巨大潜力。
我们旨在通过向ChatGPT询问有关发热预测和治疗的最常见问题来评估其答案，发热是儿童的一个主要问题。
确定了关于儿童发热最常被问到的50个问题，并向ChatGPT提问。我们使用质量和可读性量表评估其回答。
虽然ChatGPT的回答质量良好，但用于评估网站上出现的材料的可读性量表和患者教育材料评估工具（PEMAT）量表也很成功。在我们评估ChatGPT回答的量表中，冈宁雾度（GFOG）和简化晦涩度测量（SMOG）分数之间存在弱正相关（r = 0.379），Flesch–Kincaid年级水平（FGL）和SMOG分数之间存在显著正相关（r = 0.899）。
本研究揭示了关于人工智能工具（如ChatGPT）提供的有关儿童常见病症发热的信息的质量和可读性。我们确定，关于发热最常被问到的问题的答案质量高、可靠、易于阅读和理解。
