Predictive Modelling of 3‐Month Glycemic Target Achievement in Type 2 Diabetes Mellitus Using Artificial Intelligence and Continuous Glucose Monitoring: A Retrospective Cohort Study
Authors' conclusion
The XGBoost model estimates 3-month glycemic target achievement within this treatment pathway and was implemented as a web-based clinical decision-support tool.
Study context
- Population
- Patients with T2D after short-term intensive insulin pump therapy followed by physician-selected maintenance treatment
- Participants
- 1,079 patients
- Exposure
- SVM_RBF vs XGBoost models for predicting 3-month glycemic target achievement
- Outcome
- 3-month glycemic target achievement
- Follow-up
- 3 months
Methodological detailsmachine learning prediction model development and validation, training, validation and test sets (6:2:2), XGBoost selected for implementation as less sensitive to compression of correlated continuous glucose monitoring (CGM) predictors and supports TreeSHAP; implemented as web-based tool
- Machine learning prediction model development and validation
- Training, validation and test sets (6:2:2)
- XGBoost selected for implementation as less sensitive to compression of correlated CGM predictors and supports TreeSHAP; implemented as web-based tool
Main finding
As reported in the abstract
Test-set AUC
| Group | Value | AUC |
|---|---|---|
| SVM_RBF | 0.926 | 0.926 |
| XGBoost | 0.917 | 0.917 |
Value
SVM_RBF0.926
XGBoost0.917
AUC
SVM_RBF0.926
XGBoost0.917
- SVM_RBF vs XGBoost: paired DeLong p = 0.521
Observational data: associations, not effects.
Further findings
Leading predictors (SHAP)
- Leading predictors: diabetes duration, age, fasting C-peptide, mean amplitude of glucose excursions (MAGE), body mass index (BMI) and time in range (TIR)
Incremental value of adding CGM (nested logistic regression)
- Continuous net reclassification improvement
- 0.308, p = 0.013
- Integrated discrimination improvement
- 0.0062, p = 0.029
- No improvement in AUC or categorical net reclassification improvement
Original publication
Predictive Modelling of 3‐Month Glycemic Target Achievement in Type 2 Diabetes Mellitus Using Artificial Intelligence and Continuous Glucose Monitoring: A Retrospective Cohort Study. Diabetes Obes Metab. Published 11 August 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →