# Thread: sample size calculation non-inferiority trial

1. ## sample size calculation non-inferiority trial

Dear Sir
I need your help with a sample size calculation. We are planning to do a phase 3 non-inferiority randomised study comparing chemotherapy regimen A and B. We expect the 5 year overall survival with A to be 50% and B to be 40%. With a non-inferiority margin of 10%, power of 80% and p value 0.5, how many patients do we need to recruit in each arm. Sealed envelope gives infinite
Regards
Venkat

2. ## Re: sample size calculation non-inferiority trial

Do an internet search for:

Non-Inferiority Tests for Two Survival Curves Using Cox’s Proportional Hazards Model

3. ## The Following User Says Thank You to hlsmith For This Useful Post:

venkat (10-17-2015)

4. ## Re: sample size calculation non-inferiority trial

We are planning to do a phase 3 non-inferiority randomised study comparing chemotherapy regimen A and B.
You are planning a trial with real patients? Where treatment is chemotherapy? And you rely on a webforum like this here, in order to determine how many human beings you will have to treat? No biostatistician involved in a chemotherapy trial with humans? That is surely surprising.

And what sealed envelope are you referring to?

With kind regards

K.

5. ## The Following User Says Thank You to Karabiner For This Useful Post:

venkat (10-17-2015)

6. ## Re: sample size calculation non-inferiority trial

I was just hoping this was a hypothetical, say homework - but agreed on the dubious nature.

The envelope contains infinite...love!

7. ## The Following User Says Thank You to hlsmith For This Useful Post:

venkat (10-17-2015)

8. ## Re: sample size calculation non-inferiority trial

Dear all
Thank you for your valuable comments. I am clinician and not a statistician. Yes this is a chemotherapy trial involving real patients. The trial is currently an idea and needs ethics review and approval. I need to know whether it is practical to do the trial at my hospital. If I have to randomise 1000 patients it is not feasibleas I will not be able to recruit these many patients in 5 years, but if it involves 100-150 patients then it can be done. I have got differing opinions from 2 bio-statisticians, one says we need a sample size of about 150 and the other says 900 and both of them are confident of their calculations.
The link for sealed envelope (online sample size calculator is given below)
https://www.sealedenvelope.com/power...y-noninferior/

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts