PEANUTS is used for fitting proportional hazards models to ungrouped survival data using partial likelihood methods. The standard PEANUTS key variables are a binary-case indicator and an exit-time variable. PEANUTS can be used with left-truncated data, that is, data with variable entry times. In this case it is also necessary to specify the entry-time variable. PEANUTS models can include stratification variables, which are specified with the STRATA command and removed with the NOSTRATA command. (Stratification is discussed later in this chapter.)
PEANUTS can also be used analyze case-cohort data. Fitting case-cohort data requires the use of one or more stratification variables to define the sampling strata and two additional key variables: one to indicate cases occurring among subjects who were not included in the sampled cohort and a second that specifies the sampling fraction within each of the strata. There are no default names for the outside-the-cohort indictor or the sampling fraction variables.
The case-indicator variable is coded as 1 to indicate the occurrence of the event of interest (often called “failures”) and 0 to indicate that the event did not occur (“censoring”) PEANUTS recognizes either cases or event as default names for the case-indicator variable. The CASES or EVENT command is used for explicit specification of the case-indicator variable. The CASECOHORT command is used to indicate that a case-cohort analysis is to be performed and to specify the name of the outside-the-cohort indicator and stratum sampling fraction variables. The command CASECOHORT CLEAR @ is used to clear the case cohort settings. This command does not change other aspects of the model specification.
The default name for the exit-time variable (failure time for cases and censoring time for non-cases) is time. The TIME or EXIT command is used for explicit specification of the time variable. When working with left-truncated data, you must specify the entry time explicitly using the ENTRY command. There are no default names for the outside-the-cohort indicator and sampling fraction variables in case-cohort analyses.