Development of Machine Learning Models for the Identification of Elevated Ketone Bodies During Hyperglycemia in Patients with Type 1 Diabetes
Authors' conclusion
The novel approach for identifying elevated ketone levels in patients with type 1 diabetes utilized in this study indicates that CGM could be a valuable resource for the early prediction of patients at risk of developing DKA.
Study context
- Population
- Patients with T1D with ketone body measurements and continuous glucose monitoring (CGM) data (over 30,000 days of wear time)913 ketone samples
- Participants
- 138 patients
- Comparison
- Random forest vs logistic regression classifier
- Outcome
- Identification of elevated ketone bodies (≥0.6 mmol/L) during hyperglycemia
Methodological detailssupervised binary classification machine learning, random 70%/30% training/test split at patient level
- Supervised binary classification machine learning
- Random 70%/30% training/test split at patient level
Main finding
As reported in the abstract
AUC for elevated ketone levels
| Group | Value | AUC |
|---|---|---|
| Random forest (RF) | 0.836 (90% CI 0.783–0.886) | 0.836 |
| Logistic regression (LR) | 0.710 (90% CI 0.646–0.77) | 0.710 |
Value
Random forest (RF)0.836 (90% CI 0.783–0.886)
Logistic regression (LR)0.710 (90% CI 0.646–0.77)
AUC
Random forest (RF)0.836
Logistic regression (LR)0.710
Original publication
Development of Machine Learning Models for the Identification of Elevated Ketone Bodies During Hyperglycemia in Patients with Type 1 Diabetes. Diabetes Technol Ther. Published 8 March 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →