Blood Glucose Prediction Algorithms Require Clinically Relevant Performance Criteria Beyond Accuracy
Authors' conclusion
Blood glucose prediction algorithms require evaluation and optimization criteria beyond accuracy to better support optimal treatment decisions.
Study context
- Population
- OhioT1DM dataset (T1D)
- Comparison
- Five blood glucose prediction techniques evaluated with the composite glucose prediction metric (CGPM)
- Outcome
- CGPM (RMSE, temporal gain, geometric mean of glycemic event prediction)
Methodological detailsPareto frontier analysis, custom loss function
- Pareto frontier analysis
- Custom loss function
Main finding
As reported in the abstract
- The data-driven model with the lowest root mean squared error (RMSE) performed poorly on glycemic event prediction (RMSE bias toward target-range predictions)
- Ridge regressor trained with the custom loss function improved event prediction
Original publication
Blood Glucose Prediction Algorithms Require Clinically Relevant Performance Criteria Beyond Accuracy. Diabetes Technol Ther. Published 29 April 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →