What does this research mean for the field?

Large language models (LLMs) can generate accurate and comprehensible health education materials for discharged patients with diabetes, but they require further refinement to ensure actionability and safety. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

The study aims to assess the effectiveness of health education materials generated by large language models for patients with diabetes post-discharge.

March 13, 2026

Evaluating Large Language Models-Generated Health Education Materials for Discharged Patients with Diabetes: A Comparative Analysis.

Key Points

The study aims to assess the effectiveness of health education materials generated by large language models for patients with diabetes post-discharge.
Comparison of educational materials generated by ChatGPT-4o, Doubao 1.5, and DeepSeek R1.
Evaluation of content for accuracy, comprehensibility, actionability, and safety.
Analysis of potential applications for diabetes education.
All models generated accurate and understandable materials.
Concerns were noted regarding the actionability and safety of the generated content.
Recommendations suggest the need for further refinement of materials to enhance personalization and engagement.

Abstract

While ChatGPT-4o, Doubao 1.5, and DeepSeek R1 generate accurate and comprehensible materials, concerns remain regarding their actionability and safety. These findings suggest that LLMs should be used as auxiliary tools in diabetes education, requiring further refinement for personalized and actionable content.

Evaluating Large Language Models-Generated Health Education Materials for Discharged Patients with Diabetes: A Comparative Analysis.

Key Points

Abstract

Cite This Study