Question

I have matched case control data – 1 case matched to two controls. My N is ~ 580 and I have 168 matched groups. About 14 of these groups are missing 1-2 members, with cases and/or controls missing. I’ve been trying to figure out how R’s clogit function – from the Survival package – handles these missing groups, if at all, and haven’t found anything.

Answer 1

clogit actualy reframes the question to fit the inferential engine in coxph and the algorithm can handle varying numbers of cases and controls per stratum. The code includes this line that constructs a new coxph formula:

newformula[[2]] <- substitute(Surv(rep(1, nn), case), list(case = formula[[2]], 
        nn = nrows))