Introduction to regression models

Regression models are widely used. They are viewed by clinicians as a passport for a publication in major journals. The expressions: multivariate, logistic regression and Cox model are synonymous with complexity and appear both magical and seductive. Altogether, most clinicians do not read the results of regression models with a critical eye. Our aim is to describe the most common regression models in a unifying framework, and to explain how to interpret each model. We hope that investigators will then be able to discuss the models with the statisticians.