Hi there, I tried to fit a penalized spline for a continuous risk factor in recurrent events data. I found that I can include both pspline and frailty terms in coxph. So I use code like fit1 <- coxph(Surv(start, end, event)~pspline(age, df=0) + male + white +frailty(id), data.all) It yields results for both age spline and frailty variance. Next I tried to draw the plot for age effect. I used termplot function termplot(fit1, terms=c(1), se=TRUE) However, it gives me some error message: Error in dimnames(pred) <- list(dimnames(x)[[1]], termname) : length of 'dimnames' [2] not equal to array extent I don't know why. However, if I delete the frailty option, and use only fit2 <- coxph(Surv(start, end, event)~pspline(age, df=0) + male + white, data.all) I can generate the plot by termplot(fit2, terms=c(1), se=TRUE) So it seems to me that adding the frailty() option makes the "termplot" function fail to work? Do you know how to draw the plot for my function fit1? Thanks a lot! Lei At 09:53 AM 7/20/2009, you wrote:> I would be willing to chair the session. > Terry Therneau >