Deviations Are Not the Problem. The Response Is.
In any GMP manufacturing environment, deviations will happen. A process parameter drifts outside its validated range. A raw material fails an incoming test. A batch record is completed incorrectly. These events are not evidence of a dysfunctional organisation. They are the normal noise of complex manufacturing.
What matters is what happens next. A deviation that is investigated thoroughly, attributed to a genuine root cause, and closed with a corrective action that actually prevents recurrence is a system working as it should. A deviation that is closed with a vague root cause and a CAPA that amounts to retrain the operator is a system generating paperwork, not quality.
Regulators know the difference. So do sponsors reviewing their CDMO’s quality metrics.
The Deviation Life Cycle
| Stage | What Happens | Common Failure Mode |
| Detection | Operator or automated system identifies an event outside expected parameters | Detection delayed because monitoring is inadequate or parameters are not well defined |
| Immediate containment | Affected material quarantined; process halted if necessary; batch placed on hold | Production continues before the scope of the deviation is understood |
| Initial assessment | Quality unit classifies severity; determines whether batch can continue | Severity underestimated; critical deviation treated as minor to avoid disruption |
| Root cause investigation | 5-Why analysis, fishbone diagram, or fault tree analysis identifies the underlying cause | Root cause attributed to human error without investigating why the error occurred |
| Impact assessment | All potentially affected material and batches are reviewed | Impact assessment limited to the single batch; historical batches not reviewed |
| CAPA development | Corrective action addresses root cause; preventive action reduces recurrence risk | CAPA is procedural (update the SOP) rather than systemic (fix the system that allowed the deviation) |
| Effectiveness check | CAPA is verified to have worked after implementation | Effectiveness check not scheduled; CAPA closed before evidence of effectiveness gathered |
Root Cause Analysis: Beyond Human Error
The most common root cause in pharmaceutical deviation investigations is human error. It is also the least useful root cause attribution there is.
Attributing a deviation to human error tells you nothing about why that error occurred or how to prevent it. It closes the investigation without answering the real question. Was the procedure unclear? Was the operator undertrained? Was the workstation poorly designed? Was there time pressure that encouraged shortcuts? Was the check system inadequate?
A genuinely useful root cause investigation works backwards from the event through the contributing factors to the system-level condition that made the error possible. That is the condition that needs to be fixed.
Out-of-Specification (OOS) Results: A Special Category
An out-of-specification analytical result requires a specific investigation process governed by the FDA’s guidance on OOS results and EU GMP Chapter 6 on quality control. The process has two phases: a laboratory investigation to determine whether the OOS result is a genuine product failure or a laboratory error, followed by a full-scale investigation if the laboratory phase cannot identify an assignable cause.
The distinction matters because the two scenarios have different consequences. A laboratory error that is identified and confirmed allows the original sample to be retested (with restrictions). A genuine OOS result triggers a full investigation including review of the manufacturing record, assessment of related batches, and a decision about batch disposition.
Invalidating an OOS result without a confirmed, documented, and justified assignable laboratory cause is a major regulatory finding. Inspectors specifically look for OOS investigations where the outcome was invalidation without adequate justification.
CAPA: The System That Should Drive Improvement
Corrective and preventive action (CAPA) is the mechanism by which deviations feed back into the quality system and generate genuine improvement. A well-managed CAPA system tracks the status of all open actions, links CAPAs to their originating deviations, and monitors whether closed CAPAs have actually achieved their intended effect.
The most common CAPA failure mode is the procedural fix. The batch record instruction was unclear, so the CAPA is to update the batch record instruction. This may be necessary, but it is rarely sufficient. If an operator did not follow a clear instruction, the question is why. If the instruction was genuinely unclear, the deeper question is why it was unclear and how other instructions in the same system should be reviewed.
How Ardena Manages Deviations Across Its Network
Ardena’s quality management system applies consistent deviation investigation and CAPA standards across all sites, with centralised oversight of quality metrics and trend analysis. Sponsors receive timely notification of deviations affecting their programmes and have visibility into the investigation timeline and closure status through the programme’s quality governance structure.
