Journals / Anatolian Journal of Emergency Medicine / 2026 / Cilt: 9 Sayı: 1

The Role of Large Language Models in Pediatric Emergency Medicine: Accuracy and Decision-Support Potential of ChatGPT

Pediatrik Acil Tıpta Büyük Dil Modellerinin Rolü: ChatGPT'nin Doğruluğu ve Karar Destek Potansiyeli

Çağlar Kuas(ESKISEHIR OSMANGAZI UNIVERSITY, SCHOOL OF MEDICINE, DEPARTMENT OF SURGICAL MEDICAL SCIENCES, DEPARTMENT OF ACADEMIC EMERGENCY MEDICINE)0000-0003-1283-9420
Ertuğ Günsoy(University of Health Sciences,Van Training And Research Hospital, Emergency Medicine Department)0000-0002-9653-3236
Mustafa Emin Çanakçı(ESKISEHIR OSMANGAZI UNIVERSITY, SCHOOL OF MEDICINE, DEPARTMENT OF SURGICAL MEDICAL SCIENCES, DEPARTMENT OF ACADEMIC EMERGENCY MEDICINE)0000-0001-9015-1782
Murat Çetin(UNIVERSITY OF HEALTH SCIENCES, İZMİR DR. BEHÇET UZ HEALTH RESEARCH CENTER FOR PEDIATRIC DISEASES AND SURGERY)0000-0002-0667-7966
Volkan Ercan(ESKISEHIR OSMANGAZI UNIVERSITY, SCHOOL OF MEDICINE, DEPARTMENT OF SURGICAL MEDICAL SCIENCES, DEPARTMENT OF ACADEMIC EMERGENCY MEDICINE)0000-0002-9734-5843
Engin Özakın(ESKISEHIR OSMANGAZI UNIVERSITY, SCHOOL OF MEDICINE, DEPARTMENT OF SURGICAL MEDICAL SCIENCES, DEPARTMENT OF ACADEMIC EMERGENCY MEDICINE)0000-0003-4301-5440
Nurdan Acar(ESKISEHIR OSMANGAZI UNIVERSITY, SCHOOL OF MEDICINE, DEPARTMENT OF SURGICAL MEDICAL SCIENCES, DEPARTMENT OF ACADEMIC EMERGENCY MEDICINE)0000-0002-3532-1803
Pages
39–43
Publication Date
2026-03-29
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

These findings indicate that LLMs could complement, but not replace, human expertise.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:Aim: To assess the accuracy and decision-support potential of ChatGPT in pediatric emergency practice by comparing its performance with human responses to structured multiple-choice questions.
📊
Key Findings & Data:Material and Methods: This cross-sectional study used 100 randomly selected questions from Pediatric Emergency Medicine: Just the Facts, Second Edition.
💡
Conclusion & Impact:These findings indicate that LLMs could complement, but not replace, human expertise.
🔬Design:Modeling & Simulation
📍Scope:Comprehensive Sample
⚙️Variables:ChatGPT · büyük dil modelleri · pediatrik acil tıp

Abstract

Aim: To assess the accuracy and decision-support potential of ChatGPT in pediatric emergency practice by comparing its performance with human responses to structured multiple-choice questions. Material and Methods: This cross-sectional study used 100 randomly selected questions from Pediatric Emergency Medicine: Just the Facts, Second Edition. The GPT-4o model was tested without prior prompts, and its answers were compared with the reference solutions and human accuracy rates reported in the source. Accuracy rates were calculated and compared using z-tests. Correlation between ChatGPT and human performance was analyzed with Spearman’s test. Results: ChatGPT answered 85 of 100 questions correctly, achieving an accuracy of 85% (95% CI: 78.0–92.0), which was significantly higher than the mean human accuracy of 54% (95% CI: 50.8–57.4) (p < 0.001). Topic-based analysis showed that ChatGPT’s accuracy ranged from 75% to 100%, while human accuracy ranged from 30% to 65%, with higher variance. Among the 15 questions answered incorrectly by ChatGPT, 60% were case-based; the average correct human response rate for these was 35 ± 17%. A moderate positive correlation was observed between human and ChatGPT performance (ρ = 0.40, p < 0.001). Conclusion: ChatGPT demonstrated high accuracy on structured pediatric emergency questions, suggesting potential as a supportive tool in decision-making and education. While its strengths lie in knowledge-based tasks, limitations remain in complex case-based reasoning. These findings indicate that LLMs could complement, but not replace, human expertise. Prospective studies are warranted to evaluate real-world integration in pediatric emergency care.

Özet

Amaç: Bu çalışmada, yapılandırılmış çoktan seçmeli sorularda ChatGPT’nin doğruluğu insan yanıtlarıyla karşılaştırılarak, çocuk acil pratiğinde karar destek potansiyeli değerlendirildi. Gereç ve Yöntemler: Kesitsel çalışmada Pediatric Emergency Medicine: Just the Facts, Second Edition kitabından rastgele seçilen 100 soru kullanıldı. GPT-4o modeli ek komut verilmeden test edildi. Yanıtlar kaynak cevaplarla ve literatürde bildirilen insan doğruluk oranlarıyla karşılaştırıldı. Doğruluk oranları hesaplandı; z-testi ile karşılaştırma, Spearman testi ile korelasyon analizi yapıldı. Bulgular: ChatGPT 100 sorunun 85’ine doğru yanıt verdi (%85; GA %78,0–92,0). İnsan doğruluk oranı ortalama %54 (GA %50,8–57,4) olup fark istatistiksel olarak anlamlıydı (p < 0,001). Konu bazlı analizde ChatGPT’nin doğruluğu %75–100, insan yanıtlarının doğruluğu %30–65 arasında değişti. ChatGPT’nin yanlış yanıtladığı 15 sorunun %60’ı vaka bazlıydı; bu sorularda insan doğruluk oranı %35±17 idi. İnsan doğruluk oranları ile ChatGPT performansı arasında orta düzeyde pozitif korelasyon bulundu (ρ = 0,40; p < 0,001). Sonuç: ChatGPT, bilgi yoğunluğu yüksek pediatrik acil sorularında yüksek doğruluk göstermiştir. Bulgular, modelin eğitim ve karar desteğinde tamamlayıcı bir araç olabileceğini göstermektedir. Bununla birlikte, vaka bazlı muhakemede sınırlılıkları devam etmektedir. ChatGPT ve benzeri modellerin klinik kullanıma entegrasyonu için ileriye dönük çalışmalara ihtiyaç vardır.

🔖

Cite This Article

Kuas, Ç., Günsoy, E., Çanakçı, M. E., Çetin, M., Ercan, V., Özakın, E., Acar, N. (2026). The Role of Large Language Models in Pediatric Emergency Medicine: Accuracy and Decision-Support Potential of ChatGPT. Anatolian Journal of Emergency Medicine, Cilt: 9 Sayı: 1, 39–43. https://doi.org/10.54996/anatolianjem.1776853

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (11)

  1. Berg H ten, van Bakel B, van de Wouw L, et al. ChatGPT and Generating a Differential Diagnosis Early in an Emergency Department Presentation. Ann Emerg Med. 2023 Sep;S019606442300642X.
  2. Preiksaitis C, Ashenburg N, Bunney G, et al. The Role of Large Language Models in Transforming Emergency Medicine: Scoping Review. JMIR Med Inform. 2024 May 10;12:e53787.
  3. Hoppe JM, Auer MK, Strüven A, Massberg S, Stremmel C. ChatGPT With GPT-4 Outperforms Emergency Department Physicians in Diagnostic Accuracy: Retrospective Analysis. J Med Internet Res. 2024 Jul 8;26:e56110.
  4. Kaboudi N, Firouzbakht S, Shahir Eftekhar M, et al. Diagnostic Accuracy of ChatGPT for Patients’ Triage; a Systematic Review and Meta-Analysis. Arch Acad Emerg Med. 2024 Jul 30;12(1):e60.
  5. Del Monte F, Barolo R, Circhetta M, et al. Diagnostic efficacy of large language models in the pediatric emergency department: a pilot study. Front Digit Health. 2025;7:1624786.
  6. Ramgopal S, Varma S, Gorski JK, Kester KM, Shieh A, Suresh S. Evaluation of a Large Language Model on the American Academy of Pediatrics’ PREP Emergency Medicine Question Bank. Pediatr Emerg Care. 2024 Dec;40(12):871–5.
  7. AccessEmergency Medicine. About AccessEmergency Medicine. McGraw Hill. Web site. Available at: https://accessemergencymedicine.mhmedical.com/ss/about.aspx. Accessed 20 July 2025.
  8. Cascella M, Montomoli J, Bellini V, Bignami E. Evaluating the Feasibility of ChatGPT in Healthcare: An Analysis of Multiple Clinical and Research Scenarios. J Med Syst. 2023 Mar 4;47(1):33.
  9. Liu J, Wang C, Liu S. Utility of ChatGPT in Clinical Practice. J Med Internet Res. 2023 Jun 28;25:e48568.
  10. Fatima A, Shafique MA, Alam K, Fadlalla Ahmed TK, Mustafa MS. ChatGPT in medicine: A cross-disciplinary systematic review of ChatGPT’s (artificial intelligence) role in research, clinical practice, education, and patient interaction. Medicine (Baltimore). 2024 Aug 9;103(32):e39250.
  11. Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44-56.