Bioequivalence and Bioavailability Forum • power…

power… [Design Issues]

posted by Helmut – Vienna, Austria, 2012-04-13 18:17 (5186 d 18:18 ago) – Posting: # 8421
Views: 5,155

Hi John!

❝ It's Friday! :-D :smoke:

Still smoking in the US?!

❝ […] in case something does happen during the study and more than half the subjects went wazoo (food poisoning or others..). We need at least n=12 to submit... I ran a study in my previous company with 36 subjects and 18 subjects were removed from the study due to food poisoning in the clinic (post dose, not sure if it was really food poisoning since the other subjects were fine).

Wow, that’s a weird story.
According to ICH-E9 we have to justify the sample size (not to speak about ICH-E6 and ICH-E8). FDA wants us to plan a BE study with a power of 80–90%. It’s the job of the IEC to protect the safety of subjects. If they are compos mentis they should reject a study planned with a power of 99.92% (CV 15%, T/R 95%, n=36) with a justification like “there might be ⅔ dropouts resulting in 83% power with n=12”.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Complete thread:

Metadate CD sampling jag009 2012-04-12 18:11
- Metadate CD sampling: 24 hours Helmut 2012-04-12 18:45
  - Metadate CD sampling: 24 hours jag009 2012-04-12 23:44
    - Metadate CD sampling: 24 hours Helmut 2012-04-13 00:43
      - Metadate CD sampling: 24 hours jag009 2012-04-13 15:59
        
        power…Helmut 2012-04-13 16:17
        Sounds fine ElMaestro 2012-04-13 16:35
      - Metadate CD sampling: 24 hours jag009 2012-04-18 15:57
        
        Metadate CD sampling: 24 hours ElMaestro 2012-04-18 16:35
        
        Metadate CD sampling: 24 hours jag009 2012-04-18 17:09