Methods for Updating a Risk Prediction Model for Cardiac Surgery: A Statistical Primer [1]
Siregar and colleagues review the methods for updating existing risk prediction models, an approach that allows models that have lost their predictive power to be adjusted to new clinical situations. When appropriate, this approach can be more efficient than creating an entirely new model. They consider five methods: intercept recalibration, logistic recalibration, model revision, closed test procedure, and Bayesian modeling. The authors then illustrate the application of these methods, using data from the Netherlands Heart Registry to update the EuroSCORE II model.