topic Re: Percentile of continuous hazard ratio in SAS Procedures

Percentile of continuous hazard ratio

openatom — Sun, 22 Mar 2020 16:06:18 GMT

I'm using SAS 9.4. I ran a proc phreg with a continuous predictor (bp) and I am trying to get a percentile for this HR. It looks like phreg has a "percent=" option, but I can't figure out how to use it correctly. Could anyone advise? Here is the basic phreg code I'm working with:

proc phreg data=new;

model date*censor(0)=bp;

run;

Re: Percentile of continuous hazard ratio

PaigeMiller — Sun, 22 Mar 2020 16:13:28 GMT

I am trying to get a percentile for this HR

What does HR mean? How does this relate to PROC PHREG? Are you trying to run PROC PHREG on percentiles? Or do you want to determine a percentile of something in the output of PROC PHREG?

Re: Percentile of continuous hazard ratio

openatom — Sun, 22 Mar 2020 19:09:18 GMT

HR= hazard ratio

I would like the percentile of a hazard ratio for a continuous variable (blood pressure).

My understanding is that when I run a proc phreg if HR=1.06, with a continuous predictor, the HR tells me for every one unit increase in that variable (blood pressure), the HR increases by 1.06. I need a way to find percentiles of this HR since it is a continuous variable.

Re: Percentile of continuous hazard ratio

PaigeMiller — Sun, 22 Mar 2020 19:37:13 GMT

I'm not an expert on PROC PHREG, but I know how to find percentiles. Nevertheless, it's not clear to me what set of numbers you want percentiles for. If you have only the number 1.06, there's no such thing as a percentile of a single number. So I am asking for even more details on what output you get from PROC PHREG that could produce percentiles; it can't be a single number; it has to be a collection of numbers, explain further.

When you say "percentiles", do you really mean "confidence interval"?