 |
PHARMACEUTICAL INDUSTRY :
Action and Alert Limit Development |
|
| |
Client: |
|
A major global
pharmaceutical company with facilities in the U.S. and Puerto Rico
developing and manufacturing ethical drugs in a variety of dosage
forms including parenteral, solid and liquid.
|
| |
Problem: |
|
Results from
an FDA audit revealed that product and process specifications were
established and in use based on arbitrary, non-statistically based
rationales. The citation required that the company develop a statistically
sound set of action and alert limits based upon actual process capabilities
for both attribute (e.g. packaging line defects) and variable (e.g.
fill volume) types of process performance parameters.
|
| |
Approach: |
|
Working in partnership
with a cross-functional client team made up of representatives from
QA, QC, Manufacturing, Technical Services and others, Tunnell designed
and implemented a process for baselining process performance, evaluating
process variation for stability and predictability, determining process
capability (e.g. s, Cpk), evaluating appropriateness of current manufacturing/packaging
parameter targets (e.g. fill volume), and making recommendations for
changes to current action and alert limits.This approach also included
development of the protocols for establishing these limits as well
as the forms and formats required. Because the data was not readily
available, the Tunnell team "mined" the necessary data from the batch
records.
|
| |
Results:
|
|
Over $1 million
saved in first year from improved filling yields
Identified opportunities for process variation reduction thereby reducing
number of OOS results and manufacturing deviation reports generated
Developed an FDA defensible approach for establishing action and alert
limits for new products on a go-forward basis, thereby avoiding future
FDA citations in this area
Formulated recommendations for ongoing real time trending of critical
process data thereby enabling the client to monitor process performance
and react in a timely fashion to process shifts and trends
|
| |
 |
|
|
