Scholarlink

Background <p>Artificial intelligence (AI)–based interventions are increasingly integrated into diabetes management, but the certainty of their clinical benefits remains uncertain.</p> Methods <p>We conducted a systematic review and meta-analysis of 13 studies assessing AI-based interventions for personalized diabetes care. Primary outcomes included glycated hemoglobin (HbA1c), fasting blood glucose (FBG), body mass index (BMI), lipid profiles, and self-management capacity. The certainty of evidence was assessed using the GRADE framework and compared with a recent scoping review of 45 studies published in the <i>Journal of Medical Internet Research</i>.</p> Results <p>Meta-analysis showed a significant reduction in HbA1c (MD − 0.35%, 95% CI − 0.60 to − 0.10; GRADE: moderate). For FBG, BMI, and lipid outcomes, effect estimates were imprecise and graded low to very low. Narrative findings suggested improvements in self-management and quality of life. Comparison with the scoping review confirmed consistent benefits in HbA1c reduction and self-management across digital tools; however, both reviews highlighted limitations, including small sample sizes, short follow-up periods, and heterogeneity.</p> Conclusion <p>AI-based interventions hold promise for enhancing glycemic control and patient self-management in diabetes care; however, current evidence is limited. Future large, multicenter randomized controlled trials with more extended follow-up periods and cost-effectiveness analyses are needed to establish the long-term effectiveness and sustainability of this approach.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The application of AI-based interventions in diabetes personalized management: a systematic review and meta-analysis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月16日

查看全文
  • Qingqing Hu,
  • Jing Xu,
  • Yiran Peng,
  • Jue Wang,
  • Guanghui Huang

摘要

Background

Artificial intelligence (AI)–based interventions are increasingly integrated into diabetes management, but the certainty of their clinical benefits remains uncertain.

Methods

We conducted a systematic review and meta-analysis of 13 studies assessing AI-based interventions for personalized diabetes care. Primary outcomes included glycated hemoglobin (HbA1c), fasting blood glucose (FBG), body mass index (BMI), lipid profiles, and self-management capacity. The certainty of evidence was assessed using the GRADE framework and compared with a recent scoping review of 45 studies published in the Journal of Medical Internet Research.

Results

Meta-analysis showed a significant reduction in HbA1c (MD − 0.35%, 95% CI − 0.60 to − 0.10; GRADE: moderate). For FBG, BMI, and lipid outcomes, effect estimates were imprecise and graded low to very low. Narrative findings suggested improvements in self-management and quality of life. Comparison with the scoping review confirmed consistent benefits in HbA1c reduction and self-management across digital tools; however, both reviews highlighted limitations, including small sample sizes, short follow-up periods, and heterogeneity.

Conclusion

AI-based interventions hold promise for enhancing glycemic control and patient self-management in diabetes care; however, current evidence is limited. Future large, multicenter randomized controlled trials with more extended follow-up periods and cost-effectiveness analyses are needed to establish the long-term effectiveness and sustainability of this approach.