Related Experiment Video
Updated: Jul 10, 2026
05:53
Measuring Frailty in HIV-infected Individuals. Identification of Frail Patients is the First Step to Amelioration and Reversal of Frailty
Published on: July 24, 2013
Evaluation of the Risk Prediction Model for Frailty in Diabetic Patients: A Systematic Review and Meta-Analysis
Qing Chen1, Meiling Yang1, Mengmeng Chen1
1School of Nursing, Guangzhou Medical University, Guangzhou, Guangdong, China, gzhmc.edu.cn.
Journal of Diabetes Research
|July 9, 2026
Summary
This review evaluated frailty prediction models in diabetes patients, finding good performance but methodological flaws. Future models need more rigor and transparency for clinical use.
Area of Science:
- Gerontology
- Endocrinology
- Biostatistics
Background:
- Frailty is common in diabetes patients, increasing risks of disability, hospitalization, and mortality.
- Existing frailty prediction models for this population lack clear methodological quality, performance, and applicability assessments.
Purpose of the Study:
- To systematically review and meta-analyze existing frailty prediction models for patients with diabetes.
- To evaluate their methodological quality, predictive performance, and clinical applicability.
Main Methods:
- Searched multiple databases (PubMed, Embase, etc.) for eligible prediction model studies up to April 2026.
- Two reviewers screened studies, extracted data, and assessed quality using CHARMS and PROBAST tools.
- Performed meta-analyses of pooled model discrimination (AUC) and common predictors.
Main Results:
- Included 19 studies with 36 frailty prediction models; AUC values ranged from 0.703 to 0.975.
- Random forest models showed the highest discrimination (AUC=0.975). Key predictors included age, depression, ADL, nutrition, diabetes duration, physical activity, polypharmacy, HbA1c, cognition, and marital status.
- All studies had a high risk of bias due to reporting limitations, but overall applicability was high.
Conclusions:
- Frailty prediction models for diabetes patients show promise for clinical utility and predictive performance.
- Significant methodological limitations and high risk of bias were noted across studies.
- Future models require enhanced rigor, external validation, and transparent reporting for improved reliability and implementation.
