To evaluate the effectiveness of machine learning models in predicting glycemic response in type 2 diabetes patients over 6 to 18 months.
Key Findings:
Logistic regression showed modest discrimination with an AUC of 0.53.
Tree-based models outperformed logistic regression with AUCs of 0.69 for random forest and 0.67 for XGBoost.
Baseline HbA1c and 3-month glycemic measures were the strongest predictors of response.
Interpretation:
Machine learning models may enhance prediction of glycemic response in type 2 diabetes, supporting clinical decision-making for treatment intensification.
Limitations:
Models are not yet suitable as stand-alone clinical tools.
Substantial unexplained interindividual variability in glycemic trajectories.
Conclusion:
While preliminary, the findings suggest that machine learning could aid in identifying patients at high risk for inadequate glycemic control.
AACE Endocrine AI
is published by Conexiant under a license arrangement with the American Association of Clinical Endocrinology, Inc. (AACE®). The ideas and opinions expressed in
AACE Endocrine AI do not necessarily reflect those of Conexiant or AACE. For more information, see Policies.