Clinical Report: Assessment of Large Language Models in Tuberculosis Q&A Systems

Overview

This study evaluates the performance of five prominent large language models in answering questions related to pulmonary tuberculosis, focusing on readability, quality, and trustworthiness.

Background

Pulmonary tuberculosis remains a significant global health issue, with high incidence rates and substantial mortality. Effective health education is crucial for improving patient adherence to treatment and clinical outcomes.

Data Highlights

No numerical data or trial data were provided in the source material.

Key Findings

• The study systematically evaluated five large language models: Doubao, DeepSeek, Wenxin Yiyan, Tongyi Qianwen, and ChatGPT.
• Evaluation criteria included readability, quality, and reliability of the models' responses to tuberculosis-related questions.
• Patients' knowledge of tuberculosis significantly impacts their treatment adherence and outcomes.
• Traditional health education methods are limited in scope and reach.

Clinical Implications

The evaluation of these models can guide the selection of reliable information sources for patients.

Conclusion

The study provides a foundational assessment of large language models in the context of tuberculosis health education.

Related Resources & Content

1. WHO, Global Tuberculosis Report 2025 -- Global gains in tuberculosis response endangered by funding challenges
2. WHO, Consolidated Guidelines on Tuberculosis: Module 4 -- Treatment and care
3. WHO, Recommendations for TB Diagnosis -- WHO recommends near point-of-care tests, tongue swabs, and sputum pooling for TB diagnosis
4. Frontiers in Medicine — Assessing multiple-choice question quality in internal medicine: a comparative analysis of three large language models against expert consensus
5. Frontiers in Psychiatry — Assessing Large Language Model Responses to Pediatric Depression FAQs: A Cross-sectional Study on Readability, Accuracy, and Sentiment
6. npj Digital Medicine — The evaluation illusion of large language models in medicine
7. Journal of Medical Internet Research (JMIR) — Benchmark Integrity and Reasoning-Trace Errors in Medical Question Answering With Large Language Models: Mixed Methods Study With Sparse Autoencoders
8. Global gains in tuberculosis response endangered by funding challenges
9. WHO consolidated guidelines on tuberculosis: module 4: treatment and care
10. WHO recommends near point-of-care tests, tongue swabs, and sputum pooling for TB diagnosis