Heart failure & cardiomyopathies
Predictive Models Aid Prognostication: Secondary Analysis Integrating Model and Physician Prognostic Estimates in Heart Failure
Authors' conclusion
Integrating SHFM predictions with physician judgment improved accuracy.
Study context
- Population
- Heart failure (HF) outpatients with LVEF ≤40%, Canadian multicentre cohort
- Participants
- 1,643 outpatients
- Comparison
- Seattle HF Model (SHFM) plus physician estimates (integrated random forest survival model) vs SHFM alone and physician estimates alone
- Primary endpoint
- Predictive value (discrimination by C-statistic, calibration, risk reclassification, net benefit) for 1-year mortality
- Follow-up
- 1 year
Methodological detailsprognostic accuracy, stratified by specialty
- Prognostic accuracy
- Stratified by specialty
Primary result
As reported in the abstract
Discrimination (C-statistic) for 1-year mortality
| Group | C-statistic |
|---|---|
| SHFM | 0.76 (95% CI: 0.72-0.80) |
| Cardiologists | 0.75 (95% CI: 0.71-0.79) |
| Family doctors | 0.72 (95% CI: 0.66-0.78) |
C-statistic
SHFM0.76 (95% CI: 0.72-0.80)
Cardiologists0.75 (95% CI: 0.71-0.79)
Family doctors0.72 (95% CI: 0.66-0.78)
- Integrating SHFM and physician predictions (cardiologists): C-statistic, 95% CI 0.78-0.86; significantly improved, excellent calibration
- 0.82
- Integrating SHFM and physician predictions (family doctors): C-statistic, 95% CI 0.83-0.91; significantly improved, excellent calibration
- 0.87
Further findings
1-year mortality (whole cohort)
- 1-year mortality
- 9% (95% CI: 8%-11%)
Risk reclassification by integrated model
- Without events, better risk-classified vs cardiologists
- 71% (95% CI: 70%-72%)
- Without events, better risk-classified vs family doctors
- 60% (95% CI: 58%-61%)
- With events, misclassified vs cardiologists
- 45% (95% CI: 58%-63%)
- With events, misclassified vs family doctors
- 11% (95% CI: 25% to 3%)
Clinical net benefit
- Integrated model led to higher clinical benefit
Original publication
Predictive Models Aid Prognostication: Secondary Analysis Integrating Model and Physician Prognostic Estimates in Heart Failure. JACC Adv. Published 24 October 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →