Customizing the Kaplan-Meier Survival Plot


Additional Dynamic Variables

PROC LIFETEST passes to the survival plots a number of dynamic variables that contain summary statistics. Some of these dynamic variables are used when you call the %SurvivalSummaryTable macro. Table 23.1 and Table 23.2 list these additional dynamic variables for the Kaplan-Meier curves and the life-table curves, respectively. These dynamic variables are not declared in the templates for the survival curves, but you can declare them and use them to enhance the default plots.[21] The names of the dynamic variables depend on the STRATA= suboption of the PLOTS=SURVIVAL option: STRATA=INDIVIDUAL produces a separate plot for each stratum, and STRATA=OVERALL produces one plot that has overlaid curves.

Table 23.1: Additional Dynamic Variables for Stat.Graphics.ProductLimitSurvival

STRATA=

Dynamic

Description

OVERLAY

StrValj

Label for the jth stratum

 

NObsj

Number of observations in the jth stratum

 

NEventj

Number of events in the jth stratum

 

Medianj

Median survival time of the jth stratum

 

LowerMedianj

Lower median survival time of the jth stratum

 

UpperMedianj

Upper median survival time of the jth stratum

 

PctMedianConfid

Confidence of the median intervals, in percentage

INDIVIDUAL

NObs

Number of observations

 

NEvent

Number of events

 

Median

Median survival time

 

LowerMedian

Lower median survival time

 

UpperMedian

Upper median survival time

 

PctMedianConfid

Confidence of the median intervals, in percentage


Table 23.2: Additional Dynamic Variables for Stat.Graphics.LifetableSurvival

STRATA=

Dynamic

Description

OVERLAY

StrValj

Label for the jth stratum

 

NObsj

Number of observations in the jth stratum

 

NEventj

Number of events in the jth stratum

INDIVIDUAL

NObs

Number of observations

 

NEvent

Number of events




[21] Because the number of dynamic variables is a function of the number of strata, the template definition cannot automatically contain the correct number of dynamic variables.