Bikash Chatterjee, president of Pharmatech Associates (Foster City, CA), a consulting firm, uses the following recipe to implement PAT-enabled continuous improvement initiatives at pharmaceutical and biotech companies. According to Chatterjee, anyone contemplating a PAT program to ensure process and product predictability should consider this blueprint.
I. Characterization Activities:
1. Process Development
A. Baseline the Equipment:
a. engineering studies to characterize sources of variation
b. Qualify the equipment- IQ and OQ
B. Control Range Exploration
a. Evaluate where the edge of failure is for each unit operation so meaningful process characterization studies can be performed without compromising statistical rigor
2. Range Finding
A. Design statistically unbiased screening studies to understand where the gross process range is for acheiving product performance. This exercise also helps identify critical process variables as they relate to key output variables for each unit operation
II. Variance Reduction and Robustness
A. Based upon the reduced variable range from step 2, optimize the manufacturing space for each critical output varaible so as to maximize the potential for an acceptable output, i.e. find the sweet spot in the process.
B. Confirmation Studies and Training Runs
Once the final manufacturing range has been identifed , confirm the upper, lower and center-point limits of the process ranges to confirm that all product is within the expected variability.
These basic steps are essential precursors to any firm contemplating PAT as a next step in their operating and quality systems since it is essential to understand the contribution of each variable to the process and product variablity measured at the end of the process. Real time control and in-process testing and analytics are implicit in the output from this process.