Help with PROC SEQDESIGN Warning Message

jwaller — Wed, 17 Feb 2021 13:42:32 GMT

First let me preface that I am new to using PROC SEQDESIGN. I am designing a CT and determining a sample size with 1 interim look. The interim look is to examine whether to stop to reject the null or fail to reject the null. The sample size determined with the OverallStudy design statement is similar to the sample size determined using PROC POWER so I feel comfortable that I am on the right track. However, I am receiving a warning message in PROC SEQDESIGN that states:

" WARNING: The expected number of events 102.02, derived from the MODEL=TWOSAMPLESURV option, is not the same as the value 105.03 derived from the maximum information specified or derived in the procedure. The SAMPLESIZE statement is not used."

Can someone point me in a direction of why and what I might try to get the sample size to run for both the OverallStudy design and the StopToRejectAccept design in the code below:

proc seqdesign altref=0.641854 boundaryscale=pvalue plots=all;

OverallStudy: design nstages=1 alpha=0.05 beta=0.10;

StopToRejectAccept: design nstages=2 method=errfuncobf info=cum(55 110)

alpha=0.05 beta=0.10 stop=both (betaboundary=nonbinding);

samplesize model=TWOSAMPLESURV

(nullhazard=0.114 hazard=0.06 ref=hazard accrual=uniform acctime=180 foltime=180

loss=exp(hazard=0.025));

run;

Thanks so much for any help you can give!

Jenn

Re: Help with PROC SEQDESIGN Warning Message

StatDave — Wed, 17 Feb 2021 15:55:28 GMT

Try running PROC SEQDESIGN separately with only one DESIGN statement in each run.

Re: Help with PROC SEQDESIGN Warning Message

jwaller — Wed, 17 Feb 2021 16:02:34 GMT

Gracias! Worked perfectly. I don't know why I didn't think to try that combo. I appreciate your quick response.

Jenn

topic Help with PROC SEQDESIGN Warning Message in Statistical Procedures

Help with PROC SEQDESIGN Warning Message

Re: Help with PROC SEQDESIGN Warning Message

Re: Help with PROC SEQDESIGN Warning Message