Bioequivalence and Bioavailability Forum

Hello khaoula,

❝ I try to check the results of a bioequivalence study has already been validated by SAS, with Kinetica4.4.1 because we had problem with this version of kinetica: CV very low and sequence effect wrong (I TRY TO EXPLAIN THAT TO MY SUPERVISOR by testing others BE study validated)

If Kinetica does not give the same result as SAS, does that then imply that Kinetica is wrong? How did you validate with SAS?

If Kinetica does not test the MS for sequence against the between-MS for subject, then that might not in itself be a problem. The 90% CI will be the same anyway. It is only a postulate that the sequence MS cannot be tested against the model residual residual. Some people would say it can and will give you lecture about within- and between-factors and the nature of the residual.

Between us - don't tell anyone: ElMaestro is very indifferent, since it does not affect the confidence interval.

Hi Khaoula,

as ElMaestro wrote, using SAS does not guarantee that results are “correct” – one has to use the right code as well. The forum is full of examples of wrong code…

❝ […] Kinetica4.4.1 […] I try to introduce only the value of Cmax but I have not been able to do it, could someone explain me how to do it ?, the procedure of the software indicate how tu entrer all the informations (concentrations and time of all subject)

I don’t have v4.4.1. In v5.0.10 there is an “Assistant” which allows import of structured data (that’s how I recalculated your study). Please read Kinetica’s manual to find out whether this assistant was already implemented in v4.x. If yes, you can import Excel-files or CSV-files (data separator comma, semi­colon, tabulator, or blank) written in any text-editor. Example:

Subject Period Sequence Treatment Cmax
   1       1      RT        R      …  
   1       2      RT        T      …  
   …       …      …         …      …  
   n       1      TR        T      …  
   n       2      TR        R      …  

If this does not work, maybe a workaround is possible. I guess (!) Kinetica would accept datasets with only two timepoints per subject. In odd rows enter time and concentration zero and in even rows the timepoint and concentration of C_max. Performing standard NCA should give you the structure which you can use in BE. Untested example:

Subject Period Sequence Treatment Time Conc
   1       1      RT        R       0    0 
   1       1      RT        R       …    … 
   1       2      RT        T       0    0 
   1       2      RT        T       …    … 
   …       …      …         …       …    … 
   …       …      …         …       …    … 
   n       1      TR        T       0    0 
   n       1      TR        T       …    … 
   n       2      TR        R       0    0 
   n       2      TR        R       …    …