|
|
Journal of The Korean Society of Emergency Medicine 2026;37(3): 143-150. |
| 응급의료 상황에서 간호사와 ChatGPT의 중증도 분류 일치도 연구 |
김미연1
, 황용2
|
1원광대학교병원 응급의료센터
2원광대학교 의과대학 응급의학교실 |
| Inter-rater agreement of Korean Triage and Acuity Scale between emergency nurses and ChatGPT in the emergency room |
| Mi Yeon Kim1
, Yong Hwang2
|
1Emergency Medical Center, Wonkwang University Hospital, Iksan, Korea
2Department of Emergency Medicine, Wonkwang University School of Medicine, Iksan, Korea |
| Correspondence |
Yong Hwang ,Tel: 063-859-1171, Fax: 063-859-2129, Email: b612sky2@naver.com,
|
|
|
Received: April 28, 2025; Revised: August 28, 2025 Accepted: February 25, 2026. Published online: June 30, 2026. |
|
|
|
| ABSTRACT |
Objective:
This study assessed inter-rater agreement between the Korean Triage and Acuity Scale (KTAS) classifications performed by emergency room (ER) nurses and those generated by ChatGPT (OpenAI, San Francisco, CA, USA) for patients presenting to emergency medical centers.
Methods:
This retrospective observational study analyzed patients who visited an emergency department in 2023. Among 34,098 cases, 910 patients were randomly selected. The agreement between ER nurses and ChatGPT was evaluated using the severity classification variables obtained from the National Emergency Department Information System. The inter-rater agreement was assessed using Cohen’s kappa and the intraclass correlation coefficient (ICC).
Results:
The overall agreement between ER nurses and ChatGPT was moderate, with a Cohen’s kappa of 0.566. Subgroup analyses revealed kappa values of 0.589 and 0.375 for adults and pediatric patients, respectively. For the disease- related classifications, the Cohen’s kappa was 0.513, while the other group showed a Cohen’s kappa of 0.582. The ICC for overall agreement was 0.553.
Conclusion:
ChatGPT revealed moderate agreement with ER nurses in the KTAS classification. These findings suggest that ChatGPT has potential as a supportive tool in emergency triage within clinical practice. |
| Key words:
Generative artificial intelligence; KTAS; Triage; Consensus; Hospital emergency service |
|
|
|
PDF Links
Full text via DOI 
Download Citation 
