Glossary · Corrective and Preventive Action

CAPA

CAPA manages, as one procedure, the corrective action that fixes a problem that occurred and the preventive action that keeps it from recurring.

How it connects

  1. Root cause

    Find underlying causes

  2. Corrective action

    Address the actual issue

  3. Preventive action

    Prevent potential issues

  4. Verify effect

    Review results and update standards

Use the review to improve the next cycle

Definition

Definition

CAPA manages, as one procedure, the corrective action that fixes a problem that occurred and the preventive action that keeps it from recurring.

In more detail

When a defect appears, two things must happen: decide what to do with the affected product, and find the cause so it does not come back. The first is correction, the second prevention.

CAPA becomes an empty formality most often because root-cause analysis stops at "operator error". If the structure that allowed the mistake remains, the problem returns. The same happens when the action taken is only a revision to a procedure document.

Effective CAPA checks what happened afterwards. Confirming with data that the same class of nonconformance actually declined after the action is what closes the loop.

How it is used on the floor

  • Track one case from nonconformance through cause analysis, action and effectiveness check.
  • Count recurrence by type to decide when the root cause needs revisiting.
  • Handle customer complaints and internal nonconformances in the same framework.

Correction is not corrective action

Correction disposes of the affected product — rework, scrap or concession. Corrective action removes the cause so it does not recur. Closing a case after correction alone guarantees the problem comes back.

Related terms

Terms that make more sense read together.

Related products

What IYULAB provides in this area.

Quick Answers

CAPA — frequently asked

Questions that come up when evaluating CAPA.

Should every nonconformance raise a CAPA?

No. Select on impact and recurrence. Raising one for everything turns the procedure into paperwork and buries the cases that matter.

What method should we use for cause analysis?

5 Whys and cause-and-effect diagrams are common. The method matters less than not stopping at human error and following through to structural cause.

How is effectiveness verified?

Track the occurrence count of the same type over a defined period after the action. This only works if nonconformance data is classified consistently.

How would this apply to your plant?

A look at your equipment and current workflow is usually enough to scope it.

Talk to an engineer
KR KO