In a proportional hazards model, if failure events can be attributed to independent causes, it may be appropriate to define the cause of failure for each observation. Defining multiple causes of failure will produce separate coefficients for each cause. It is equivalent to creating a separate model for each cause, and treating failures from other causes as censored observations.
To define the cause of failure:
Each cause will then appear in the outcome table below the outcome variable name. Selecting a specific cause will display that cause’s coefficients in the explanatory table.