01-25-2018 10:24 AM
I'm trying to make a table of survival and a corresponding graph, both of which need to include the # of patients at risk at various time points. For my table, I am using the LIFETEST ODS output dataset PRODUCTLIMITESTIMATES for the # of patients at risk for my table, but then I am using, per the recommendation here https://blogs.sas.com/content/graphicallyspeaking/2011/11/28/survival-plot-with-external-at-risk-tab..., the SURVPLOT dataset to add the # of patients at risk to my survival graph. However, for patients whose survival time is exactly the value of the timelist I am using, these datasets produce different #s.
For example, this code:
ods output productlimitestimates = PLE
survivalplot = SURVPLOT;
proc lifetest data = sashelp.bmt plots=survival(atrisk=0 to 200 by 10) timelist=0 to 200 by 10;
time T * Status(0);
For the ALL group for time=110, the PLE dataset has a value of 30 for the variable LEFT, but the SURVPLOT dataset has a value of 31 for the variable AtRisk.
Does anyone know why this is like this? Am I wrong in expecting both of these datasets to have the same values for Left and AtRisk for all timepoints?