For some companies, manufacturing is a crapshoot. There is
little basis for confidence that that the right equipment, materials,
components, people and records will be at the right place and the right time to
begin manufacturing. And once manufacturing has begun, a crystal ball is as
good as anything to divine whether the product batch will be released on
Manufacturing is a continuing exercise of muscling and
cajoling the product from one stage to another while the jealous QA department
chalks up one deviation after another that sends the minions scattering about
to do the minimal essential work required to release the product to market.
Deviations and the associated CAPAs begin to amass and
backlogs soon become the ominous specter that one would just as soon explain
away just as Scrooge did his night visitors.
This scenario is nothing less than a testament to an
out-of-control situation, and deviation investigations are nothing more than
slight-of-hand or putting the best face on recurring, nagging and relentless
There is always time to crank out the legalistic deviation
investigations, but never enough time for leaders to emerge to mindfully gain
control of the situation through conscious understanding of the underlying
causes for the mayhem and carve a path forward out of the wilderness. Rather
management decries the onerous GMPs that burden the business while patients are
desperately waiting for your magnanimous supply of lifesaving drug.
Warning Letters are fraught with observations that point to
issues with the deviation investigation system and the ability to solve
problems once and for all. Here are my top ten reasons that the
deviation/investigation/CAPA system fails:
1. R&D throws half-baked processes over the fence to
operations, saying that they did their part to file a submission to the FDA
ahead of the competition. Now it’s time for Operations to take over in the
spirit of continuous improvement.
2. The event occurs and is opened in the system, but there
are delays in reporting time-sensitive information. It is not until the time
gets close to release the product that open deviations surface and stand in the
way. Unfortunately, no one has accurate information to assemble the facts for a
3. Products impacted by the deviation are not evaluated
based on scientific data, and the perimeter is not drawn wide enough to include
anything beyond the current batch—and certainly not to anything released into
the market place.
4. The true root cause or most probable cause is not
determined by a disciplined approach, and is not documented. The root cause
reads more like a problem statement, or the symptoms of the problem. The bottom
of the problem is not reached, and will surely stop short if the true cause
touches on a politically sensitive subject.
5. Significant deviations cannot be distinguished from all
the minor issues flooding into the system. When every single documentation
error is entered into the system with its corresponding investigation and CAPA,
it’s no wonder that there is a backlog and meaningless activities working at
6. Deviations are seen as singular events, rather than yet
another example of a problem that is already identified as a deviation in the
system. There is no “look back” to determine if the new event is already
covered by a current effort, or the evidence of an ineffective CAPA.
7. Inappropriate ownership for deviations. The areas
responsible for the deviation have no incentive to permanently resolve the
problem, because they know that QA knows the “right way” to write it up to get
the lot released. QA “rationalizes” rather than independently assesses the
“justification.” (See The QA Pharm 10/30/10.)
8. Performance metrics are not presented to management for
review in an action, decision-oriented forum where assignments are made,
followed-up at subsequent reviews, and individuals by-name held accountable for
9.Because TrackWise (an excellent system) provides
visibility to problems and serves as a tracking tool, junk gets into the system
for those two reasons only (visibility/tracking) when in actuality they are
neither a deviation, nor a CAPA. Events going into data management system must
meet the definition of a deviation, and there cannot be a CAPA without first
having a root cause.
10. There is a tolerance for repeated problems.
How do you score against these top ten reasons?
The QA Pharm