Heart failure & cardiomyopathies
Prediction Models for Heart Failure in the Community: A Systematic Review and Meta-Analysis
Authors' conclusion
Prediction models for estimating risk of incident HF in the community demonstrate excellent discrimination performance.
Study context
- Population
- Community-based cohorts; multivariable models derived, validated and/or augmented for incident heart failure (HF) prediction36 studies · 59 models
- Comparison
- Atherosclerosis Risk in Communities (ARIC) vs GRAM vs PCP-HF white men vs PCP-HF white women vs RETAIN
- Outcome
- Discrimination (summary c-statistic) of HF prediction models
Methodological detailsBayesian meta-analysis, MEDLINE and EMBASE searched from inception to 3 November 2022
- Bayesian meta-analysis
- MEDLINE and EMBASE searched from inception to 3 November 2022
Pooled result
As reported in the abstract
Summary c-statistic
ARIC
0.802
summary c-statistic · 95% CI 0.707–0.883
GRAM
0.791
summary c-statistic · 95% CI 0.677–0.885
PCP-HF white men
0.820
summary c-statistic · 95% CI 0.792–0.843
PCP-HF white women
0.852
summary c-statistic · 95% CI 0.804–0.895
RETAIN
0.839
summary c-statistic · 95% CI 0.748–0.916
Further pooled results
Discrimination in cohorts with a uniform prediction window
- ARIC risk score and PCP-HF models had significant summary discrimination
Risk of bias (PROBAST), certainty of evidence, clinical impact
- Model results at high risk of bias
- 77%
- Certainty of evidence low
- No model had a clinical impact study
Original publication
Prediction Models for Heart Failure in the Community: A Systematic Review and Meta-Analysis. Eur J Heart Fail. Published 31 July 2023. doi:10.
Summarised from the published abstract; every number is checked against it