ABSTRACT
In the developmental process for many drugs, the drug product is administered under controlled conditions to healthy, normal individuals or to the targeted patient population. It is reasonable to question whether the distribution of the estimates of a drug concentration in a blood specimen might be approximated by the normal distribution. In manufacturing a drug product, it is common to collect and analyze specimens from the mix of active and inactive ingredients, the blend. The resulting drug concentration values can be categorized by three different class variables: analyst, day, and specimen. If we had several different lots of a drug product and analyzed enough units from each, we could detect as statistically significant even a 1% difference in the potency between the lots. The determination of a drug concentrations in plasma specimens requires the construction of a calibration response curve.
