Machine learning for prediction of chronic kidney disease progression: Validation of the Klinrisk model in the CANVAS Program and CREDENCE trial
Authors' conclusion
The Klinrisk machine learning model, using routinely collected laboratory data, was highly accurate in its prediction of CKD progression in the CANVAS/CREDENCE trials. Integration of the model in electronic medical records or laboratory information systems can facilitate risk-based care.
Study context
- Population
- Patients with T2D in the pooled CANVAS/CREDENCE trials
- Comparison
- Klinrisk machine learning model vs Kidney Disease Improving Global Outcomes (KDIGO) heatmap
- Outcome
- Prediction of chronic kidney disease (CKD) progression (≥40% eGFR decline or kidney failure), assessed by AUC, Brier scores and calibration
- Follow-up
- Up to 3 years
Methodological detailspooled CANVAS/CREDENCE trials
- Pooled CANVAS/CREDENCE trials
Performance
As reported in the abstract
Klinrisk performance for CKD progression · 1 and 3 years
| Test or model | AUC 1 year | AUC 3 years | Brier score 1 year | Brier score 3 years |
|---|---|---|---|---|
| Klinrisk | 0.81 (95% CI 0.78-0.83) | 0.88 (95% CI 0.86-0.89) | 0.020 (0.018-0.022) | 0.056 (0.052-0.059) |
Klinrisk
AUC 1 year0.81 (95% CI 0.78-0.83)
AUC 3 years0.88 (95% CI 0.86-0.89)
Brier score 1 year0.020 (0.018-0.022)
Brier score 3 years0.056 (0.052-0.059)
- Klinrisk vs KDIGO heatmap: P < 0.01 — improved performance with Klinrisk at every interval
Original publication
Machine learning for prediction of chronic kidney disease progression: Validation of the Klinrisk model in the CANVAS Program and CREDENCE trial. Diabetes Obes Metab. Published 28 May 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →