Assessment of Large Language Models for Tracheostomy Care in Pediatrics: ChatGPT vs Claude
Journal of Pediatric Health Care, cilt.40, sa.4, ss.488-492, 2026 (SCI-Expanded, SSCI, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 40 Sayı: 4
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.pedhc.2025.11.017
- Dergi Adı: Journal of Pediatric Health Care
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI), Scopus, CINAHL, EMBASE, MEDLINE
- Sayfa Sayıları: ss.488-492
- Anahtar Kelimeler: artificial intelligence, large language models, pediatric nursing, Pediatric tracheostomy
- Ankara Yıldırım Beyazıt Üniversitesi Adresli: Evet
Özet
Introduction This study aimed to evaluate the accuracy, completeness, clarity, and relevance of responses provided by two freely available large language models (LLMs)—ChatGPT-4.o and Claude 4.0 Sonnet—to caregiver-focused questions about pediatric tracheostomy. Methods Seventeen clinically relevant questions were submitted to each LLM, and responses were evaluated by nine pediatric otolaryngologists across four domains using a 5-point Likert scale. Results ChatGPT-4.o showed significantly higher accuracy (Mean = 4.49, p = .029), while Claude 4.0 Sonnet achieved better completeness scores (Mean = 4.25, p = .022). No significant difference was found in clarity or clinical relevance. ChatGPT-4.o responses were more readable (FRES = 63.78, FKGL = 6.41) compared to Claude (FRES = 36.93, FKGL = 10.44). Conclusion Preliminary findings suggest that both LLMs may offer supplementary educational value for tracheostomy caregivers, but their role in urgent situations was not evaluated. Use should remain under clinical oversight, and further research is needed.