PulseTrendingJournal ClubResearchersJournalsExplore
Instagram
HomeTrendingJournal ClubExplore
Synapse
⌘+K
Synapse
May 20, 2026Journal of Diabetes Science and Technology

Evaluating the Accuracy, Quality, and Reproducibility of AI Chatbot-Generated Diabetes Self-Management Education and Support Plans Aligned With ADCES7

View Full Paper
Ask AI
Bookmark
Share

Authors

TYTeng‐Hung YuHHHui‐Chun HsuYLYau‐Jiunn Lee

Discussion

Loading...

Member takes

Overview

Randomized trial evaluates AI-generated diabetes support plans, indicating potential for scalable education tools.

Key Points

  • This study assesses the quality, accuracy, and reproducibility of diabetes self-management plans generated by an AI chatbot.
  • Evaluated eleven virtual patient profiles across five key diabetes self-management time points.
  • Used structured prompts with retrieval-augmented generation to generate plans assessed with DSMES Process Evaluation Checklist.
  • Engaged ten certified diabetes educators to score the plans and examined internal consistency, inter-rater reliability, and reproducibility.
  • Mean domain scores ranged from 29.0 to 42.7, with high scores in clarity (85.1%) and goal-directedness (81.1%).
  • Moderate performance in accuracy (74.9%) and feasibility (75.3%) observed, but limitations in patient engagement noted.
  • Reliability was high with a Cronbach's α of 0.889 and reproducibility moderate with intra-assay coefficients of 0.04 to 0.15.

Cite This Study

Yu et al. (2026) studied this question.

synapsesocial.com/papers/6a0d4f92f03e14405aa9af25https://doi.org/10.1177/19322968261447304
View Full Paper
Ask AI
Bookmark
Share