The modeling modules in EPICURE are generalized regression programs in which a response of interest is modeled as a function of one or more covariates. As the first step in using these EPICURE risk regression modules, you must indicate the input variables that describe the response to be modeled. These variables are called key variables. The nature of the key variables depends on the specific module being used. Certain variable names, for example, cases, are recognized as key variables when input variable names are initially defined with the NAMES command or read from a BSF or CSV file header. However, in each regression program there are commands that can be used for explicit specification of key variables. Key variables can be changed at any time during a session.