The Batch That Failed Sterility: Release Pressure and a Silent Workforce
At a contract manufacturer of sterile surgical kits, the plant director's annual bonus is weighted 40% to on-time-in-full delivery. A batch of 4,000 units failed the in-process sterility check. A retest on a second sample passed, and the director has asked the quality manager, who holds release authority, to sign off so the order ships Friday. Internal data complicate the picture: 38% of operators agreed in the last anonymous survey that 'raising a problem here damages your career,' and two of the last four audit findings were closed without attached evidence. The quality manager has put the objection in writing and meets the plant director and the HR business partner tomorrow.
1. The plant director wants the batch released on the retest. What should the quality manager do, and what must the record show?
Read the worked answer
Do not release on the retest alone.
Reasoning: a failed sterility check is a signal about the process, not only about the sample. A passing retest cannot distinguish among three explanations - sampling error, test-method variability, or genuine intermittent contamination in the process - and only the third is a patient-safety hazard. Signing off without discriminating among them converts an unknown into an assurance, which is the one thing release authority exists to prevent.
Actions: place the batch on hold; open a documented nonconformance and investigation covering the sampling plan, method suitability, bioburden trends, and environmental monitoring; quarantine the lot. Release only through a formal deviation or concession approved under the quality system by someone senior to the quality manager, with written technical justification and, where customer or regulatory agreements require it, notification and acceptance. If shipping without that path is demanded, escalate in writing rather than argue verbally.
The record must show the objection, the data reviewed, who decided, on what rationale, and what was communicated externally. An email trail helps, but the deviation file is what an auditor will read.
Keep the bonus question separate: the director's incentive design is a system problem for HR and the executive team, and mixing it into the release decision weakens a technical judgment by making it look personal.
2. The survey result and the closed audit findings suggest the written policy is not the culture. What would you change, and what would you refuse to count as progress?
Read the worked answer
Treat culture as observable leader behavior and system incentive, not as a communications problem.
Diagnose before acting. The 38% is an aggregate; break it down by shift, unit, and supervisor, and read exit interviews. Establish who closed the two findings without evidence and under what time pressure. A finding closed without evidence is not a documentation lapse - it is a statement about what the organization rewards.
Changes that alter behavior:
- Incentives. Rebalance the plant director's scorecard so delivery cannot pay out while a batch sits on hold for a safety question.
- Response time. Measure the interval from a raised concern to a visible, documented answer. Silence teaches futility faster than any policy.
- Audit integrity. Reopen the two findings, re-verify a sample of others, and report the result upward.
- Protection. Confirm the reporter is not exposed, and say publicly what happened.
Refuse to count as progress: training completion rates, poster campaigns, a revised policy document, or a survey score in the quarter after a campaign. All are inputs. Progress looks like more near-miss reports, faster closure, fewer repeat findings, and supervisors who escalate without being asked.