topic Re: Proc lifetest with multiple imputation in Statistical Procedures

Proc lifetest with multiple imputation

Haemoglobin17 — Mon, 31 Jul 2023 15:08:40 GMT

Dear all,

I have been trying to calculate survival estimates probabilities and standard error from a dataset where I performed multiple imputation.

I wrote the following code, because, for each dataset imputed, I would like to obtain a survival curve for treated and another for untreated. Of course, Have is sorted by group.

The problem is that, every time I run the code, SAS is not responding and I have to restart the software... Have you ever run something similar? I think that the problem is with the statement strata because, without it works fine.

Thanks!!

proc lifetest data= Have outsurv=Want;

time followup*event(0);

by group;

strata _Imputation;

weight sw;

run;

Re: Proc lifetest with multiple imputation

OsoGris — Mon, 31 Jul 2023 15:50:14 GMT

Have you tried "strata _imputation_" ?

Re: Proc lifetest with multiple imputation

Haemoglobin17 — Mon, 31 Jul 2023 16:12:38 GMT

Hi @OsoGris, yes , this is just a typo!

Re: Proc lifetest with multiple imputation

SAS_Rob — Mon, 31 Jul 2023 16:18:19 GMT

You should not put the _IMPUTATION_ variable on the STRATA statement. It should appear on the BY statement so that each data set gets an independent analysis.

Re: Proc lifetest with multiple imputation

Haemoglobin17 — Tue, 01 Aug 2023 09:40:58 GMT

Hi @SAS_Rob ,

If I put _Imputation_ in the by statement I will obtain five (in my case) survival curves, one for each imputed dataset, while I would like to obtain two curves for each imputed dataset, one for treated and the other for untreated, for each imputed dataset.

I can divide the imputed dataset in 5 sub-dataset based on the value on _Imputation_ and apply the previous script without the strata statement.

However, I was wondering if there is a faster way to obtain the same result without running 5 different proc lifetest..

Re: Proc lifetest with multiple imputation

OsoGris — Wed, 02 Aug 2023 13:05:42 GMT

If you want curves for treated (events?) and untreated (censored?) separately then it seems you would have to run two PROC LIFETEST steps for each imputed data set. That would mean 10 PROC LIFETEST steps. I don't see any other way around this.

Re: Proc lifetest with multiple imputation

SAS_Rob — Wed, 02 Aug 2023 19:18:52 GMT

If you want two separate curves per imputed data set then you should place both variables on the BY statement.

BY _IMPUTATION_ GROUP;