Does anyone know of a method to compare Harrel's
c-statistics for censored data? I guess a
modification of the wilcoxon rank-sum test might work,
however, does anyone have more detail? Thanks.
For what it's worth, I know of a method for defining confidence intervals
for the difference between 2 Harrell C statistics, based on the jackknife,
but I haven't had time to implement it in Stata. However, a quick and dirty
solution might possibly be to use the jackknife using -jknife- in official
Stata, together with Bill Gould's -stcstat- command (type -findit harrell
c- to find out more).
