 |
PHARMACEUTICAL INDUSTRY :
Manufacturing Process Improvement |
|
| |
Client: |
|
A major US generic Pharmaceutical Company
|
| |
|
|
|
| |
Problem: |
|
The client had developed a process for formulating a significant new product and had successfully completed three validation batches. Of the first five manufacturing batches, three were rejected. Each batch contained $250,000 of raw materials. The process consisted of blending, granulation, milling, and blending and compression. |
| |
|
|
|
| |
Approach: |
|
A Tunnell team consisting of a formulation subject matter expert, a statistician, a data analyst, and a project manager formed a team with the client to develop a solution. A review of the data and interviews with the key personnel strongly suggested that the problem was in the granulation stage. The team developed a Design of Experiment (DoE) to be run in equipment 25% the size of the commercial equipment. The DoE would determine the critical process parameters and describe the interactions between those parameters. A second set of experiments was run to optimize the process. Finally, using the process understanding gained from the small equipment runs and input from the experts, the product was successfully manufactured in commercial scale granulation equipment. |
| |
|
|
|
| |
Results: |
|
The team determined that the amount of water and power level to determine the end point of mixing were critical to success in the process. The DoE data showed that the original water amount would not always produce acceptable results and that increasing the water amount by less than 4% (500 g) would be optimum for the process. In the course of experimentation, effects of granulation and compression parameters that could have resulted in batch failures were identified and quantified. Parameter ranges were established that avoided these problems. The client was able to easily validate the process. Based on the production forecast, $7M in cost due to discarded batches would be avoided and supply was assured.
|
| |
|
|
|
| |
 |
|
|
| |
|
|
|