chapter  8
Data driven approaches I: conventional statistical inference methods, including linear and logistic regression
ByTiziana Rancati, Claudio Fiorino
Pages 43

86Multivariable methods of statistical analysis are commonly used in oncology. This type of analysis considers multiple variables to predict a single outcome, with multivariable methods exploring the relation between two or more predictors (independent variables) and one outcome (dependent variable). These models serve two purposes: (1) they can predict the value of the dependent variable for new values of the independent variables and (2) they can help describe the relative contribution of each independent variable to the dependent variable, controlling for the influences of the other independent variables. Linear regression and logistic regression are among the most popular multivariable models. They have some mathematical similarities but differ in the expression and format of the outcome variable. In linear regression, the outcome variable is a continuous quantity, such as blood pressure or plasma level of a biomarker. In logistic regression, the outcome variable is usually a binary event, such as alive versus dead, or case versus control. In this chapter, linear and logistic regression will be presented from basic concepts to interpretation.