Bioequivalence and Bioavailability Forum

Dear Nee!

❝ After estimating the PK parameters, there are certain subjects which don't have their Rsq values greater than 80%. So these are subjects are not used for the statiscal calculation of AUC0-inf.

Revise your SOP. An arbitrary cut-off is meaningless ( search the forum). I guess you mean an R² of 0.8 – not 80%? Or are you mixing something up here and are talking about AUC_t/AUC_∞ ≥80% of AUC_∞?

❝ […] one entire sequence (RRT) is not available. While we are analyzing in SAS, we are getting an error "Average is not estimable".

Since the minimum sample size for the FDA in a replicate design is 24, one sequence consists of 8 subjects. Very low chances that you see the same PK behavior in exactly one sequence… How many subjects are actually left for the analysis?

❝ We want to know whether this error is occuring because of the missing sequence.

Maybe (I don’t have SAS). I didn’t get an error with an example data set in Phoenix… Maybe you can try to change the coefficients of the remaining sequences to 0.5 or try Proc Mixed instead of Proc GLM.