I'm using a Cox model with 100k subjects and 1624 events to model the effects of a treatment with respect to 49 covariates, 127df.
Due to proportional hazard violation of some continuous variables, I've added restricted cubic splines using rcs(). Based on the p-values provided by cox.zph (which I've decided at 0.01 due to my sample size), the variables that previously violated the PH assumption no longer do.
However, it seems worth investigating this further instead of simply relying on the p-values in cox.zph. However, my understanding tells me that visualization of Schoenfeld plots after rcs() transformation has little value in doing this.
My questions thus are:
How, then, would one investigate further if the PH assumption is truly satisfied after applying restricted cubic splines?
Can this be visualized in a convenient way similar to how scaled Schoenfeld plots are used?
rcs(x)''is significant butrcs(x)'andrcs(x)are not), how would you suggest tackling this? I don't recall having seen this described anywhere – JED HK Aug 23 '22 at 13:21rcs()(especially because most continuous variables were significantly nonlinear, according topspline()and Poisson plots. Any thoughts here Professor Harrell? This should be the final piece of my analysis, help appreciated – JED HK Aug 23 '22 at 13:59rcs()that very much suggest nonPH (and in some cases more so than before the addition of splines)? – JED HK Aug 23 '22 at 19:18plotmethod forcox.zphhas a y-axis scale parameter and set all plots to use the same scale. If things are standardized (as with usingterms) the scale might be -2 to +2. – Frank Harrell Aug 24 '22 at 11:21ylim, it does have this and I've been using it to resize the plot to capture its shape properly (if the limits are too far out it appears flat, if the limits are too small the entire line cannot be seen). The magnitude between the vars varies greatly. – JED HK Aug 24 '22 at 13:55