Statistical modelling for thoracic surgery using a nomogram based on logistic regression

Abstract

A well-developed clinical nomogram is a popular decision-tool, which can be used to predict the outcome of an individual, bringing benefits to both clinicians and patients. With just a few steps on a userfriendly interface, the approximate clinical outcome of patients can easily be estimated based on their clinical and laboratory characteristics. Therefore, nomograms have recently been developed to predict the different outcomes or even the survival rate at a specific time point for patients with different diseases. However, on the establishment and application of nomograms, there is still a lot of confusion that may mislead researchers. The objective of this paper is to provide a brief introduction on the history, definition, and application of nomograms and then to illustrate simple procedures to develop a nomogram with an example based on a multivariate logistic regression model in thoracic surgery. In addition, validation strategies and common pitfalls have been highlighted.