Proportional Hazard Model with Time-Dependent Covariates - Quick Tab

Select the Quick tab of the Proportional Hazard Model with Time-Dependent Covariates dialog box to access options to specify the variables/covariates used in the proportional hazard model with time-dependent covariates.

Variables (survival times, censoring, [optional] grouping). Click the Variables button to display the standard variable selection dialog box for specifying the variable(s) with the survival times, the censoring indicator variable, and an optional grouping variable (for a stratified, by-group analysis).Use the options in the Independent variables (covariates) group box (see below) to specify the independent variables for the analysis.

Specifying survival times. There are three ways to specify survival times. Select one variable with survival times (e.g., number of weeks surviving), select two variables containing the start and end dates respectively, or select six variables containing dates. Specifically, these variables should contain the month (1 to 12), day (1 to 31), and year when the particular observation began (e.g., when the patient was admitted to the hospital), and the month, day, and year when the observation was terminated (due to death/failure or censoring, e.g., when a patient was dismissed from the hospital). While processing the data, Survival Analysis will compute the number of days that elapsed between dates and perform the analysis on this measure. Note that if the value of the year is less than 100, Statistica will automatically assume that the year refers to the 20th century; for example, the year 88 will be converted to 1988. The censoring indicator variable should contain the integer codes or text labels that uniquely identify complete and censored observations.

The (optional) variable with group codes should contain the integer codes or text labels that uniquely identify to which group (sample) each observation belongs. If no grouping variable is specified, a regular analysis will be performed on all data. If a variable is specified here, then a stratified analysis will be performed. In a stratified analysis, separate regression models are first fit to each group and the log-likelihoods for those models are summed up. This log-likelihood is then compared to that of the overall model (collapsed across groups).

Code for complete responses. Enter the codes or text labels that were used in the censoring indicator variable to uniquely identify complete (uncensored) observations in the Code for complete responses field. To review all codes in the respective variable, double-click on this field (or press the F2 key on your keyboard) to display the Variable dialog box.  

Code for censored responses. Enter the codes or text labels that were used in the censoring indicator variable to uniquely identify incomplete (censored) observations in the Code for censored responses field. To review all codes in the respective variable, double-click on this field (or press the F2 key on your keyboard) to display the Variable dialog box.  

Independent variables (covariates). Use the Independent variables group box to specify the covariates in the model. In the Covariate column, enter a label for the covariate; that label will be used in the results spreadsheets. In the Expression column, define the respective covariate by entering an arithmetic expression.

Syntax. The syntax rules for the expressions are identical to those for spreadsheet formulas. In addition to the available arithmetic operators and functions, the expression interpreter will also recognize the symbol t_ or T_ (i.e., letter t and the underscore) to denote the survival times. By using this symbol, you can define time-dependent covariates. Shown below are a few examples:

Covariate

=

Expression

timdep

 

v2*(log(t_)-5.4); time-dependent

Covar2

 

Age*(Group=1) - Age*(Group=2)

FactorA

 

FactorA

FactorB

 

FactorB

AXB

 

FactorA*FactorB

As illustrated in the first example (timdep = ... ), you can include comments in the expressions by placing them after the expressions, separated by a semicolon. Variables may be referenced by name, or via the standard Vxxx convention where xxx is the respective variable number in the data file. The second expression illustrates the use of logical operators: If a case belongs to Group 1 (where Group is a variable name), then Covar2 is defined as +Age; if a case belongs to Group 2, then Covar2 is defined as -Age. The third and fourth expressions define the fixed covariates FactorA and FactorB; the fifth expression defines their interaction. Note that in the examples above, only the first expression actually defines a time-dependent covariate; while the others make no reference to the survival times (i.e., they do not include the symbol t_ or T_).

Editing expressions. To edit expressions, you can highlight them, and then either double-click on them or press the F2 key on your keyboard to place the cursor in the respective field. The standard Windows conventions for copying (Ctrl+C) and pasting (Ctrl+V) are also supported. You can use the arrow keys to move up and down through the list of expressions. To add an expression, press the Insert key on your keyboard; to delete an expression, highlight it and then click the Delete highl. expressn button, or press the Delete key on the keyboard. Note that you can also save the expressions, and later use them in future analyses.

Delete highl. expressn. Click the Delete highl. expressn button to delete the highlighted text in the Expression column.

Select variables. Click the Select variables button to display the standard variable selection dialog box, which allows you to choose the variables for the expression. The variables (i.e., their names) that you select in this dialog will be moved to both the Covariate column and the Expression column. Thus, by default the selected variables will be defined as fixed covariates.

Open expressions. Click the Open expressions button to display the standard Open File dialog in which you can select a previously saved file with the expressions. Those files are by default saved with the file name extension *.ini.  Note that Statistica will automatically append the expressions in the file to the ones that were previously entered.

Save expressions. Click the Save expressions button to display the standard Save As File dialog box in which you can save the current expressions in a file; by default Statistica will suggest the file name extension *.ini.  

Codes (for groups). Click the Codes button to display the Select Codes for Grouping Variable dialog box, which allows you to specify the codes or text labels that were used in the grouping variable to uniquely identify to which one of the two groups each observation belongs. These codes can only be specified if a stratified (by-group) analysis was requested, that is, if a grouping variable was selected via the Variables button (see above). To review all codes in the respective variable, double-click on this field. The default selection is all codes.