ABSTRACT
Prognostic or risk models evaluate the association between an outcome (depen-
dent variable)—such as disease, response, and survival-and one or more pre-
dictive or prognostic factors (independent variables). Such models are used for a
number of purposes, including to (1) improve our understanding of a disease
process, (2) improve the design and analysis of clinical trials, (3) generate risk
stratification, (4) assist in comparison of outcome between treatment groups in
nonrandomized studies by allowing adjustment for case mix, (5) define risk
groups based on prognosis, (6) predict disease outcome more accurately or par-
simoniously, and (7) act as a guide for clinical decision making including
treatment selection and patient counseling (1). In general, there are two major
types of models depending upon the purpose of the study: 1. explanatory models
and 2. predictive models.