Can only fools use foolproof systems?
"We were a contract manufacturer that was proud of our warehouse control and computer inventory system from a well-known vendor,"
explains our Agent-In-Place. "Company management assured our clients and FDA that personnel had been adequately trained and
that the control scheme was 'foolproof.'
"Things went smoothly until a data-entry clerk logging quarantined material into the system reversed two numbers: 57,282 items
were received, but the number was entered as 52,782!
"When the warehouse manager reviewed the inventory, he noted that only a portion of the received materials had been quarantined.
Because the other portion of the batch wasn't in quarantine, he assumed that the entire lot shouldn't have been in quarantine.
He manually changed the inventory status of the entire lot from 'quarantined' to 'approved.' Before the lot had been formally
evaluated and released, it was issued and used in manufacturing.
"Of course, when quality assurance inspected the lot, they found problems with it. Not knowing that some had already been
used, they rejected the lot."
Foaming for release
"Our group had helped out in validating the new production line on an orally disintegrating tablet project, and dissolution
results were typically greater than 95% active ingredient released in 10 minutes," notes our Agent-In-Place. "We had already
settled in on the next project when a lot was rejected. The site noted that the dissolution testing of this lot showed low-release
results that were also highly variable, which was unusual for the product.
"I was the dissolution specialist on the original project, so I went to the site to troubleshoot the method. As a starting
point, I observed a repeat dissolution test being conducted on the lot in question. When the bath covers were removed, I found
foam crawling up the paddle shafts.
"The quality control laboratory analyst was shocked when I sampled the foam and told me, 'That isn't in the method!' I
replied, 'We have now moved from QC into R&D and investigation.'
Analysis of a sample of the foam showed high API content. It was known that the tablets were prone to foam formation, and
a simethicone suspension was used to suppress foam in the manufacturing process. Further investigation into the simethicone
suspension showed that the simethicone settles out if not agitated immediately before aliquoting to the batch, which apparently
had not been done in this case. This allowed the tablets to melt into foam on the surface and prevent release of the API."
Critical melding point
"We have a large-volume parenteral that is filled into glass bottles," describes our Agent-In-Place. "The original product
had metal slings attached directly to the bottles so that they could be hung up-side down to administer the product directly
from the bottle. To improve the speed of packaging and lower the price of the package, we changed to a new plastic sling.
"But then, we started receiving complaints. Nurses reported to us sling breakage at the hanging location. Our investigation
showed that the breakage was associated with a bad meld where the injected plastic met at the furthest point from the injected
site. We ended up having to make another sling change, this time to a type that did not have the meld right at the most critical