|
Abstract:
|
In the pharmaceutical industry, setting and maintaining specifications for drug products and drug substances are crucial. Very often Cpk (Process Capability index) is used. Cpk is most meaningful when the process is under statistical control and a large amount of data exists, which is hard to achieve during development. For small sample size data during development, Pfizer published a method using robustness contours instead of using Cpk, which are easy to plot and to visualize; however, the Pfizer method doesn't take into account the impact of sample size. We propose a Tolerance Cpk plot that uses the tolerance factor to account for uncertainty from limited sample size. It is applicable widely, for both 2-sided (Assay) and 1-sided (Impurities, Moisture, Particle Size, etc.) specifications. An easy-to-use Excel macro tool was developed. Case studies using simulated data were examined. The tool will be very useful for setting specification limits and evaluating process capability with limited data.
|
