Bioequivalence and Bioavailability Forum

Hi John,

❝ The only reason I use it is because I am too lazy to write SAS codes for determining half-life.

See Ref #1 in the post above. Of course, no option to select data points during visual inspection of fits. The 100% agreement with WinNonlin the authors proudly reported clearly means only the “automatic” method based on the maximum R²_adj.

❝ Heck they don't even have an option to run partial+full replicate BE analysis yet (I mean within the program, not using your worksheet).

Yep. If I wouldn’t told them I guess they wouldn’t know until today that RSABE exists and what it means. Took me ~10 years to convince them that the reported power is for the dead and buried 80/20 rule. Since v6.4 we have – additionally – power for the TOST. Wow!
BTW, if using SAS one should know what to do as well. Today I got a response from the large Indian CRO (back-story: useless post-hoc power 29% but reported by SAS with 100%). The answer of the CRO’s “statistician” was:

Further, we have checked the analysis part pertaining to power calculation once again and also veri­fied with the results from WinNonlin software Version 5.3. And, it proves that the results pro­vided earlier are ok.

2×2=5… How I love the word “verify”. Rubbish in, rubbish out.

Convincing the “Far Side” to implement a workaround (!) for unequal variances in parallel designs was even tougher. Took me merely 13 years.