Bioequivalence and Bioavailability Forum

HI Dixit,

excluding single measurements (1) from PK analysis is a rather different topic than excluding a whole patient (2).

@(1) If a PK-measurement increases within one patient at a certain time point, this measurement might be exclude from calculation of half-life.

@(2) If a patient does violate prespecified criteria, he may be excluded from the analysis (however, often a ITT and PP analysis is performed).

For both cases: if there is a documented violation obviously influencing the measurement(s), data can be excluded.

In all cases a rationale must be given. Exclusion of data because of departure of (say) more than 3 SD from the global mean is never accepted.
Those analyses can be used in best case as secondary analyses to investigate in amount of influence of the outliers.

Regards,
JPL