The aim of this Letter to the Editor was to respond to a comment highlighting potential statistical biases in an analysis of our recently published article. We therefore specified the method for selecting the model variables in order to limit overfitting, then we used the Firth method to control the sparse data bias, and finally for checking internal validity we used bootstrapping methods. In total, the conclusions of our model were not changed by these new analyses.
Keywords: Bootstrapping; Firth method; Methodology; Multiple logistic regression; Predictive performance.