note on generalisations of the concordance index for survival data

Abstract

Concordance index (c-index) was adapted to survival data by Harrell (1982). In its basic form, the index depends on censoring, however the issue can be effectively dealt with. More importantly, Harrell's c-index cannot be used with time-varying effects and/or time-dependent covariates, and several generalisations were proposed. We look at some of them, explore their differences, point to a basic difference between these generalisations, and strongly favour one type of generalisation.