A simple method to adjust clinical prediction models to local circumstances

Abstract

When a clinical prediction model in new patients does not show adequate performance, an alternative to developing a new model is to update the prediction model with new data. The updated model will be based on more patient data, and may yield better risk estimates.