Development and Validation of a Machine Learning Model to Predict Weekly Risk of Hypoglycemia in Patients with Type 1 Diabetes Based on Continuous Glucose Monitoring
Authors' conclusion
Prediction models based on real-world CGM data can be used to predict the risk of hypoglycemia in the forthcoming week. The models showed good performance in both the internal and external validation cohorts.
Study context
- Population
- T1D patients with long-term continuous glucose monitoring (CGM) monitoring+253 validation patients
- Participants
- 205 patients
- Outcome
- Excessive hypoglycemia risk the following week (TBR >4% and upper TBR 90th percentile limit)
- Follow-up
- Following week
Methodological detailsmachine learning (XGBoost), internal test, external validation in two independent cohorts
- Machine learning (XGBoost)
- Internal test
- External validation in two independent cohorts
Performance
As reported in the abstract
ROC-AUC
| Test or model | ROC-AUC (test dataset) | ROC-AUC (external validation) |
|---|---|---|
| XGBoost | 0.83-0.87 (95% CI 0.83-0.88) | 0.81-0.90 |
XGBoost
ROC-AUC (test dataset)0.83-0.87 (95% CI 0.83-0.88)
ROC-AUC (external validation)0.81-0.90
Further findings
- Most discriminative features included low blood glucose index, glycemic risk assessment diabetes equation (GRADE), hypoglycemia, TBR, waveform length, CV and mean glucose during the previous week
Original publication
Development and Validation of a Machine Learning Model to Predict Weekly Risk of Hypoglycemia in Patients with Type 1 Diabetes Based on Continuous Glucose Monitoring. Diabetes Technol Ther. Published 29 May 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →