Risk prediction models for heart failure admissions in adults with congenital heart disease: Methodological issues

We read the paper authored by Marelli et al. [ [1] Cohen S. Liu A. Wang F. Guo L. Brophy J.M. Abrahamowicz M. et al. Risk prediction models for heart failure admissions in adults with congenital heart disease. Int. J. Cardiol. 2020; 322: 149-157 Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar ]. published in International Journal of Cardiology. In a retrospective cohort study, all patients with congenital heart disease (CHD) were included to assess the risk prediction model to identify Adult Congenital Heart Disease (ACHD) patients at high risk for Heart Failure Hospitalization (HFH). Four models were established by using binary logistic regression. Performance of models was compared by discrimination and calibration. Also, the authors used bootstrapping methods to assess the internal validity. As result, the overall cumulative incidence of HFH from age 18 to 65 was 12.6% and C-statistic for final model was 0.93.