cGMP perspective
That said, the cGMPs still call for companies to have firm control of their processes, statistical or otherwise. Being out-of-statistical-control,
but within specifications is not acceptable from a cGMP perspective because there is a higher probability of manufacturing
defective product. Companies are required to show continuing due diligence, provide data and analysis to prove that the process
is being brought into a state of statistical control, or provide a scientific and statistical study to explain why it cannot
be controlled.
To summarize, for control charts, practical significance and OOT go together, and statistical significance and statistical
control go together. There are four possible combinations, the control charts for which can be viewed together at http://PharmTech.com/ controlcharts:
Figure 1: Not out-of-trend and not out-of-statistical-control.
|
1. Not OOT and not OOSC: the ideal situation (see Figure 1)
Figure 2: Out-of-statistical-control but not out-of-trend. Fails Western Electric rule number 2 (7 points on the same side
of the center line).
|
2. Not OOT but OOSC: cannot yet see a practical change, but failed one or more WE8 rules (see Figure 2)
Figure 3: Not out-of-statistical-control but is out-of-trend. Passes all 8 Western Electric rules.
|
3. OOT but not OOSC: can see a practical change that is not yet statistically significant (see Figure 3)
Figure 4: Out-of-trend and out-of-statistical-control.
|
4. OOT and OOSC: can see a practical change that is statistically significant (see Figure 4).
Finally, any of these four combinations can be OOS or not OOS as the specification concept is independent of trend (practical
significance) and statistical significance.
Lynn D. Torbeck is a statistician at Torbeck and Assoc., 2000 Dempster Plaza, Evanston, IL 60202, tel. 847.424.1314, Lynn@Torbeck.org
,
http://www.torbeck.org/.
References
1. FDA, Guidance for Industry: Process Validation: General Principles and Practices (Rockville, MD, Nov. 2008)
2. FDA, Guidance for Industry: Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production (Rockville, MD, Oct. 2006)
3. L.D. Torbeck, Pharm. Technol.
34 (3), 21–23 (1999)
4. Matrixx Initiatives, Inc. v. Siracusano, 131 S. Ct. 1309, (2011)
5. L.D. Torbeck, Pharm. Technol.
34 (3), 76 (2010).
|
