Bioequivalence and Bioavailability Forum

Fancy smoothing? [RSABE / ABEL]

posted by Helmut  – Vienna, Austria, 2014-07-22 01:13 (3949 d 12:35 ago) – Posting: # 13286
Views: 13,146

Hi ElMaestro,

❝ Hmmmm that's a relevant question. Perhaps they just wanted to demonstrate internal validity of the sim algos? I also use GraphPad Prism for graphing, though an earlier version. It does not stack or offset curves, so they really do seem to have gotten a little hump of sorts.

Below a comparison: Calculated, 10⁴, and 10⁶ sim’s. The red dotted line on top is the signi­fi­cance limit (0.05373 for 10⁴ and 0.05036 for 10⁶).



I tried also to repeat Japan’s bizarre method (Fig. 2 of the paper):



A lot of ‘noise’, but essentially I could reproduce the reported inflation (con­verging at ~7.5%). Inflation with this method is a textbook example (unadjusted multiple testing). I’m happy that finally someone dem­on­strated it. The inflation is what to expect from pooling two groups, where the size of the second one is 50% of the first: 0.05+0.05/2=0.075… Voilà.
For my experiences in 日本 see the end of this post.

❝ One conclusion to draw is that the whole scaling business might not be that smart when it comes to type I errors.

If EMA’s method is concerned, yes.

❝ At the more general level another proposal is that before a group of guideline authors agree on a new requirement they should investigate the requirement by specific simulations rather than just follow their instincts or adopt concepts that were made for a different null hypothesis […]

Exactly. It’s not only statistics which sucks. The PK group single-handed invented a bunch of PK metrics whose relevance and sensitivity to formulation differences are not supported by a single publication. What the heck is AUC_τ/2? We had three conferences on the MR GL, hun­dreds of pages of comments… Guess what? Felt like talking to a brick wall.

❝ […] ("0.0294" .... do I need to say more?).

No.

—
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:

• EMA: α-inflation – the sus­pi­cion begins to mount Helmut 2014-07-21 02:04 
  • EMA: α-inflation – the sus­pi­cion begins to mount ElMaestro 2014-07-21 09:24
    • Why simulate a simple AB|BA at all? Helmut 2014-07-21 16:14
      • Why simulate a simple AB|BA at all? ElMaestro 2014-07-21 20:50
        • Fancy smoothing?Helmut 2014-07-21 23:13
          • Arbitrary smoothing? d_labes 2014-07-22 08:43
            • Noise… Helmut 2014-07-22 13:31
              • TSD Japonica… d_labes 2014-07-22 16:06
                • TSD Japonica… Helmut 2014-07-22 17:50
              • Noise debugged d_labes 2014-07-24 13:47
                • Noise debugged Helmut 2014-07-24 15:07
                  • Shit happens d_labes 2014-07-24 16:09
        • Commentary d_labes 2014-07-22 08:18
          • Commentary nobody 2014-09-15 08:24
            • Commentary Helmut 2014-09-15 10:12
        • IBE/PBE = Two-Stage Helmut 2014-09-29 03:33
          • IBE/PBE = Two-Stage ElMaestro 2014-09-29 15:16
            • Sequential Design = Inflation likely… Helmut 2014-09-30 15:03