Bioequivalence and Bioavailability Forum

Dear Friends,

I have just completed M. Pharm. and have joined CRO.

I try to understand statistics for reviewing PK results.

I have come across the outlier topic. Suppose, we run the studentized residual test using PROC GLM in SAS for outlier detection in ln-transformed pharmacokinetic parameters, say lnCmax, for two-way crossover design and if we found more than one extreme values which are higher than critical value.

In this case, should we consider all extreme values as outliers?

OR

Should we remove the most extreme one from the dataset and re-run the test again to judge other outlier values?

Does studentized residual test detect single outlier at single run?

or

can studentized residual test detect more than one outliers at single run?

Thanks in advance for your suggestion.

Best Regards,
Sunil