From Artificial Intelligence to Real-Life Practice: Can ChatGPT be a Guide About Pediatric Dysphagia?
International Journal of Child Health and Nutrition, cilt.15, sa.1, ss.37-43, 2026 (ESCI, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 15 Sayı: 1
- Basım Tarihi: 2026
- Doi Numarası: 10.6000/1929-4247.2026.15.01.4
- Dergi Adı: International Journal of Child Health and Nutrition
- Derginin Tarandığı İndeksler: Emerging Sources Citation Index (ESCI), Scopus
- Sayfa Sayıları: ss.37-43
- Anahtar Kelimeler: ChatGPT, pediatric dysphagia, reliability, safety, usefulness
- Sağlık Bilimleri Üniversitesi Adresli: Evet
Özet
Introduction: The widespread use of Artificial Intelligence (AI)-based tools has significantly simplified access to medical information. Pediatric dysphagia, or difficulty swallowing in children, is among the commonly queried topics because of its close relationship with feeding safety, nutritional intake, and growth outcomes. This study aims to evaluate the reliability, usefulness, and safety of responses generated by Chat Generative Pre-Trained Transformer (GPT) regarding pediatric dysphagia. Methods: A set of thirty carefully selected questions covering various aspects of pediatric dysphagia, including general information, risk factors, diagnosis, treatment, and follow-up, was prepared based on clinical data, digital trends, and frequently asked questions from health websites. These questions were submitted to ChatGPT (version 4.0), and the responses were independently evaluated by two experts using a 4-point Likert-type scale (1: lowest, 4: highest) to assess reliability, usefulness, and safety. Additionally, the readability of each response was measured using the Flesch-Kincaid Grade Level test, which estimates the educational level required to comprehend the text. The implications of this readability score for caregiver comprehension were also interpreted. Results: ChatGPT’s responses received high scores overall, with average ratings of 3.73 for reliability, 3.87 for safety, and 3.87 for usefulness. The average Flesch-Kincaid Grade Level was 13.03, indicating suitability for university-level readers. This suggests that while the responses are accurate and informative, their linguistic complexity may limit accessibility for some caregivers. Conclusion: ChatGPT shows promise as a supportive tool in providing basic information about pediatric dysphagia. However, to ensure accurate and personalized medical evaluation, these AI-generated responses must be verified through professional clinical review. Given that pediatric dysphagia directly affects nutritional intake, growth, and feeding safety, validated AI-based guidance tools could help caregivers recognize feeding problems earlier and seek appropriate medical care promptly.