Predictors of Metformin Failure: Repurposing Electronic Health Record Data to Identify High-Risk Patients
Authors' conclusion
Baseline HbA1c was the strongest predictor of metformin failure and additional factors substantially improved performance suggesting that routinely available clinical data could be used to identify patients at high risk of metformin failure who might benefit from closer monitoring and earlier treatment intensification.
Study context
- Population
- Patients with ≥1 abnormal diabetes screening test who initiated metformin, 3 sites (Arizona, Mississippi, Minnesota); 48% female, mean age 57 ± 14 years
- Participants
- 22,047 metformin initiators
- Outcome
- Metformin failure: lack of target HbA1c (<7%) within 18 months of index or start of dual therapy
- Follow-up
- 18 months
Methodological detailsmulticentre, prediction model (XGBoost)
- Multicentre
- Prediction model (XGBoost)
Main finding
As reported in the abstract
Metformin failure prediction
Rate of metformin failure
43%
C-index, XGBoost model with baseline HbA1c, age, sex and race/ethnicity
0.731
95% CI 0.722, 0.740
C-index with addition of other clinical factors
0.745
95% CI 0.737, 0.754P < .0001
Observational data: associations, not effects.
Original publication
Predictors of Metformin Failure: Repurposing Electronic Health Record Data to Identify High-Risk Patients. J Clin Endocrinol Metab. Published 7 January 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →