Root Cause Analysis and CAPA Effectiveness in GMP Investigations

Corrective and Preventive Action (CAPA) is a defined regulatory expectation for medical device manufacturers, and a routine one in pharmaceutical GMP more broadly. Source

A recurring deviation is usually a sign that the original root cause analysis identified a symptom rather than the actual cause — the CAPA "fixed" something, but not the thing that was really driving the problem.

Why root cause analysis often stops too early

• The investigation stops at the first plausible explanation instead of testing whether that explanation actually accounts for the evidence.
• Human error is cited as the root cause without asking what system or process condition made that error possible or likely.
• The analysis focuses on the specific incident without checking whether the same conditions exist elsewhere and could recur.

A more defensible investigation sequence

1. 1Describe the deviation precisely, including what was expected versus what actually happened.
2. 2Gather objective evidence before forming a conclusion, not after.
3. 3Test the leading hypothesis against the evidence — does it actually explain everything observed, or only part of it?
4. 4Identify the systemic or process-level condition that allowed the immediate cause to occur.
5. 5Define a corrective action that addresses that systemic condition, not just the immediate incident.
6. 6Verify effectiveness after implementation — closing the CAPA is not the same as confirming it worked.

Signs a CAPA program is investigating symptoms, not causes

• The same deviation type keeps recurring despite CAPAs being closed against it.
• "Retraining" or "reminded the operator" appears repeatedly as the corrective action.
• CAPA effectiveness checks are scheduled but rarely actually completed or reviewed.
• Investigations are completed quickly but with little documented evidence-gathering.

Questions to ask about a CAPA program

• How many CAPAs this year were reopened because the deviation recurred?
• Do our investigations distinguish immediate cause from systemic cause?
• Are effectiveness checks actually being completed, or just scheduled?
• Is there a backlog of open investigations, and if so, why?