What question did this study set out to answer?

The aim is to evaluate ChatGPT's performance in sepsis diagnosis, treatment, and mortality prediction.

April 18, 2026Asian Journal of Surgery0 citationsOpen Access

Performance of ChatGPT in artificial intelligence-supported clinical decision-making for sepsis: A multidisciplinary expert evaluation

JZJinmei ZhangSun Yat-sen University ZTZhiwei TangSun Yat-sen University LTLili TangZunyi Medical University

Key Points

The aim is to evaluate ChatGPT's performance in sepsis diagnosis, treatment, and mortality prediction.
Evaluation of ChatGPT responses to 33 sepsis-related questions.
Assessment of response quality based on expert ratings.
Analysis of strengths and limitations in complex cases.
78.8% of ChatGPT's responses received a high-quality rating.
Identified limitations in handling complex and personalized cases.
Future improvements in accuracy, personalization, and data security are necessary.

Abstract

• This study examines ChatGPT's potential and challenges in sepsis diagnosis, treatment, and mortality prediction. • ChatGPT responded to 33 sepsis-related questions; 78.8% were rated highly, with limitations in complex and personalized cases. • Despite limitations, ChatGPT shows potential in sepsis care, with future work needed to improve on accuracy, personalization, and data security.

AIに質問

Bookmark

View Full Paper

Cite This Study

Zhang et al. (Wed,) studied this question.

synapsesocial.com/papers/69e3215140886becb654087d https://doi.org/https://doi.org/10.1016/j.asjsur.2026.03.151

AIに質問

Bookmark

View Full Paper