Pre-trained ChatGPT for report generation in automated microbial identification and antibiotic susceptibility testing systems

Liping Hu ¹, Xiaohong Xu ², Yingting Zhuang ³

⁰Department of Laboratory Medicine, Xiamen Chang Gung Hospital Hua Qiao University, Xiamen, People's Republic of China.

Scientific Reports +

|

October 16, 2025

View abstract on PubMed

Summary

Large language models like ChatGPT can improve automated microbial identification and antibiotic susceptibility testing (ID/AST) reports. Prompt engineering significantly enhanced ChatGPT

Area Of Science

Clinical Microbiology
Artificial Intelligence in Healthcare
Medical Informatics

Background

Automated microbial identification and antibiotic susceptibility testing (ID/AST) systems are crucial in clinical microbiology.
Standardizing report generation from these systems remains a significant challenge.
Large language models (LLMs) offer potential for enhancing report consistency and objectivity.

Purpose Of The Study

To evaluate the effectiveness of ChatGPT in generating standardized microbiology reports for automated ID/AST systems.
To compare AI-generated reports with those produced by clinical microbiologists (CM).
To assess the impact of structured prompt engineering on AI report quality.

Main Methods

A prompt engineering framework was developed for ChatGPT, guided by Clinical & Laboratory Standards Institute (CLSI) guidelines.
Eight clinical cases were analyzed, comparing reports from CM, ChatGPT before prompt training (ChatGPT_BT), and ChatGPT after prompt training (ChatGPT_AT).
Performance was evaluated across five dimensions: accuracy, relevance, objectivity, completeness, and clarity.

Main Results

ChatGPT_BT showed higher relevance and completeness than CM (p < 0.0001).
ChatGPT_AT demonstrated significant improvements across all five dimensions compared to CM (p < 0.001).
ChatGPT_AT also showed notable improvements over ChatGPT_BT in relevance, objectivity, completeness, and clarity (p < 0.05).

Conclusions

Structured prompt engineering significantly enhances ChatGPT's ability to generate high-quality microbiology reports.
ChatGPT shows substantial potential as an assistive tool for clinical microbiologists.
AI-assisted reporting can improve the objectivity, clarity, completeness, relevance, and accuracy of automated ID/AST system reports.

Keywords:

Automated ID/AST systems ChatGPT Clinical microbiologists LLM