Development and validation of 10‐year risk prediction models of cardiovascular disease in Chinese type 2 diabetes mellitus patients in primary care using interpretable machine learning‐based methods
Authors' conclusion
Using routinely available predictors and ML-based algorithms, this study established 10-year CVD risk prediction models for Chinese T2DM patients in primary care. The findings highlight the importance of renal function indicators, and variability in both blood pressure and HbA1c as CVD predictors, which deserve more clinical attention. The derived risk prediction tools have the potential to support clinical decision making and encourage patients towards self-care, subject to further research confirming the models' feasibility, acceptability and applicability at the point of care.
Study context
- Population
- Chinese T2D patients aged ≥18 years without history of cardiovascular disease (CVD) or end-stage renal disease, managed in public primary care clinics in 2008, followed up until December 2017
- Participants
- 141,516 patients
- Exposure
- Machine learning (ML) models vs Cox regression models
- Outcome
- Incident CVD (10-year risk prediction; discrimination and calibration)
- Follow-up
- 10 years
Methodological detailsrisk prediction model development and validation, 10-year follow-up, sex-specific models, random split into development and validation samples, machine learning-based methods, Cox regression models for comparison
- Risk prediction model development and validation
- 10-year follow-up
- Sex-specific models
- Random split into development and validation samples
- Machine learning-based methods
- Cox regression models for comparison
Main finding
As reported in the abstract
| Group | Harrell's C statistic |
|---|---|
| ML models | >0.80 |
ML models
Observational data: associations, not effects.
Further findings
Incident CVD · median follow-up 9.75 years
- Patients developing CVD
- 32 445 (22.9%)
Most important predictors
- Age, T2D duration, urine albumin-to-creatinine ratio (ACR), estimated glomerular filtration rate (eGFR), systolic BP variability and HbA1c variability were the most important predictors
- Nonlinear effects of several predictors, particularly U-shaped effects of eGFR and BMI
Original publication
Composed by AI from the published abstract and checked against it. How Briefs are made →