Interim analyses using seqdesign for Non-inferiority

[AndreyMyslivets]

Hello everybody!

Could you please help me with calculations and script?

The main endpoint analysed using proportion difference.

I have the following information:

1. alpha = 0.05 one-sided;

2. power = 0.80;

3. a non-inferiority margin equal -0.20 (greater is better);

4. an actual difference between treatment groups (test drug - reference drug) planned as -0.05

5. groups weight is 1:1;

6. an overall response rate of reference drug is 38% from articles.

7.  stopping rules O’Brien-Fleming ;

8. number of stages is 2:

      1 interim at 75% of patients

      1 final

I have a draft skript:

proc seqdesign altref=<...>

               pss

               stopprob

               errspend

               ;

   OneSidedOBrienFleming: design

            nstages = 2

            method = obf

            alpha = <...>

            beta = 0.20

            info = cum(0.75 1);                        

   samplesize model=twosamplefreq (TEST=PROP <...> );

run;

Could You please help to impliment the correct data instead of <...>

Thanks!

Andrey

[Damir]

Dear andrey,

i have the same problem with a margin equal to -0.04 and 0 for actual difference between treatement.

Did you find a solution to your problem ?

Thinks

Damir

[kekeke]

As I know, SEQDESIGN does not currently have options to directly support group sequential non-inferiority, superiority, or equivalence trials.