1
$\begingroup$

I wonder if there is any advice for performing a sample size calculation for survival data using the one-sample log rank test with interims (ie group sequential) - ie with O'Brien-Fleming or some other alpha-spending technique (or if there are alternative methods for interims in one-arm survival studies then that is of interest too - I know about Simon's two stage method of course)

I was planning to base my sample size calculation on simulation using the excellent one-sample log rank calculator here (except I was thinking of making the p-value one-sided by dividing by 2) http://www.ms.uky.edu/~mai/splus/LogRank1.r

$\endgroup$
1
  • $\begingroup$ I think you can get an estimate using simulation studies. Check out R package 'powerSurvEpi' and 'LogrankPower' $\endgroup$
    – Ronna
    Commented Oct 12, 2019 at 18:23

1 Answer 1

2
$\begingroup$

My guess is that this hasn't been really worked out. Most researchers size the study to achieve 0.9 power at the final analysis, relegating interim analyses to just 'hope for the best' in detecting an early, large difference. The early look will have low power because of fewer events and because of the conservative group-sequential cutoff. A full Bayesian method would help a bit, because there is no penalty for multiple looks. Because of no penalty, you can look as often as desired without planning for it.

$\endgroup$
1
  • $\begingroup$ many thanks, Professor Harrell, its good to know I'm not missing anything obvious, many if need be I'll turn to simulation with something a bit ad hoc in the "hope for the best" vein that you mention. $\endgroup$ Commented Sep 29, 2017 at 13:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.