How to Conduct a Robust Root Cause Investigation for CAPA
An effective CAPA investigation does more than name a cause. It shows how you arrived at the root cause. Your conclusion may well be accurate, but if the reasoning behind it is unclear, stakeholders may have questions.
A strong root cause investigation make your logic clear to others. They show what the problem was and why it occurred. They include evidence that supports your explanation. They also explain why you ruled out other possibilities.
Why “Root Cause” Alone is not Enough
Root cause investigators often face pressure to close investigations quickly and meet time-to-close expectations. That urgency can lead to conclusions that name a cause without fully explaining the reasoning behind it. Even when the conclusion is correct, it still needs to answer the question stakeholders will ask next: how do we know?
When the link between your evidence and conclusion is unclear, the investigation may seem incomplete. It may also look like a subjective opinion.
Such an investigation is difficult to defend. CAPAs may be less effective if people cannot clearly tie the result to the evidence. Various stakeholders – auditors included – may interpret your findings differently.
Showing how you determined the root cause builds confidence in both the conclusion and any actions that follow.
Good Problem Solving Starts with Facts
A robust root cause investigation begins with a clear problem statement that describes exactly what you observed. The statement should identify :
- the object. This means the equipment, component, product, campaign, batch, or sample in question.
- the defect: how the object did not conform to expectations. For example: Batch 12345 was out of specification for low theoretical yield.
The investigation should also explain why the event matters:
- What was the impact on product quality, OEE, maintenance cost and labor, or customers?
- In addition to the current impact, the investigation should consider historical performance and potential future trends.
- What potential risks could we face if we don’t find the root cause or if the defect happens again?
This information helps determine how deep the investigation should go and how robust the CAPA should be. They may for example be systemic causes underlying the defect.
| CONCERN | CURRENT IMPACT | FUTURE IMPACT | TIME FRAME |
|---|---|---|---|
| Batch 12345 OOS for Low Theoretical and Reconciliation Yield | Batch on hold $52,000.00 | Customer service Product expires | 5 months |
Use Facts to Draw Comparisons
Usually, root cause investigators document where they observed the problem, and the system or process affected.
Although dates and times matter, it is more important to confirm what happened before and during the event. Use time-stamped evidence whenever possible.
This might include process conditions, equipment states, operator actions, alarms, maintenance activity, environmental conditions, and procedural steps.
Investigations should also document the boundaries, or extent, of the problem. These boundaries show:
- what related things could have shown the same change in performance but did not
- how else the object could have failed but did not
- where else the event could have occurred but did not
- at what other point in the object’s life or process someone could have observed the change but did not.
These boundaries make comparisons possible. In complex investigations, comparisons guide the search for possible causes.
In all investigations, they help explain why you ruled out some possible causes. They also keep the assessment focused on evidence rather than speculation.
| IS | IS NOT | |
|---|---|---|
| WHAT | Product ABC Batch 1A2345 | Product DEF Batches 1A2343 and 1A2344 |
| Low Theoretical and Reconciliation Yield | Assay, Dissolution, Coating Defects | |
| WHERE | BDR-3 | BDR-4,-5,-6,-7,-8 |
| WHEN | 07/2/2026 | 06/29/2026 |
| During Coating Reconciliation During the 3rd and Last Batch in Campaign | During Compression Yield During Blend Yield During First and Second Batches |
Evaluate Possible Causes against the Facts
When documenting possible causes, it is important to remember that the cause of one problem is often another problem.
Following the 5 Whys approach, keep investigating.
Continue until you find the decision, condition, or control failure that started the chain of causes.
Write possible causes as specific causal problems. Each cause should include a clear mechanism. This mechanism should explain how it led to the performance change.
General statements such as “operator error” or “equipment failure” do not explain the mechanism. They are therefore not specific enough to support an effective CAPA.
An example: “The spray arm was too far away from the tablet bed. As a result, the tablets received too little coating. This reduced weight gain.”
This does state both the causal problem (insufficient coating, reduced weight gain) and the mechanism (Spray arm distance too far)
However, if that had happened, you would also expect coating defects. If on inspection, there were no coating defects, you could rule out this possible cause.
By contrast: “Running at a low adjust-weight range can make the tablet’s total weight less than the core tablet’s theoretical weight. This can cause the actual measure to be higher than the theoretical measure.”
This also states the causal problem and the mechanism. However, in this case, a batch record review showed a clear difference. Batch 3 did run at a lower weight than batches 1 and 2 – so this was the true cause.
From Findings to Confidence
Showing clearly how you arrived at the true cause has many benefits:
- Your conclusion becomes a defensible explanation
- Corrective actions are more targeted, as they follow on from the evidence
- Communication improves because stakeholders can follow the logic
- Audit readiness improves because the investigation not only shows the conclusion, but also explains why it is justified
The strongest investigations do not simply state the root cause; they make the evidence path visible.
Continue Exploring
Related Articles
Why Jumping to Solutions without Finding the Root Cause Costs Organizations more in the Long Run
8D is not a problem solving method?
Beyond 5 Whys: a 5-minute problem solving lesson
Go Deeper
Find out more about Kepner-Tregoe’s Root Cause Investigation training and learn our tried-and-tested problem solving process in depth
How Gerresheimer proved that the “obvious cause” was not to blame, leading to a surprising true cause and timely resolution
