Heart failure & cardiomyopathiesMeta-analysis
Risk prediction models for incident heart failure: a systematic review and meta-analysis
- systematic review and meta-analysis
- random-effects meta-analyses
Authors' conclusion
The PREVENT, ARIC and MESA risk scores demonstrate promising performance and should be prioritised for further validation and progression to impact studies.
Study context
- Population
- Derivation or validation studies for incident heart failure (HF) prediction modelsstudies representing 238 models, from 24,531 publications identified
- Participants
- 76 participants
Pooled result
As reported in the abstract
Discrimination among models validated in at least one cohort
models validated in at least one cohort
64
models with moderate discrimination
four
models with high discrimination
eight
Further pooled results
Calibration
- In patients with low predicted risk, calibration may have been adequate
Most promising models
- PREVENT, Atherosclerosis Risk in Communities (ARIC) and Multi-Ethnic Study of Atherosclerosis (MESA) models most promising for further validation and impact studies
Risk of bias
- Risk of bias high in assessments
- 82.9%
Characteristics of externally validated models
- derived using machine learning
- 14
- incorporated novel biomarkers such as proteomics and polygenic risk scores
- six
- included measures of social determinants of health as predictors
- nine
Original publication
Risk prediction models for incident heart failure: a systematic review and meta-analysis. Heart. Published 20 April 2026. doi:10.1136/heartjnl-2025-327118
Summarised from the published abstract; every number is checked against it