Maturity model

CAPA maturity

Almost every site can close a CAPA. Far fewer can show that the action removed the cause and that the problem stopped. This model tracks that distance.

In one line

CAPA maturity is measured by whether problems stop recurring, not by closure rate.

The five levels

What each level looks like here

Level 1

Reactive

The process runs when something forces it to.

Actions are recorded after events, often as retraining, with no confirmed root cause and no check that anything changed.

Typical weakness. Closure is the objective; effectiveness is not assessed at all.

Risk. Recurrence is guaranteed and the record documents the pattern.

Level 2

Controlled

Procedures exist and are followed, mostly.

A CAPA procedure exists with due dates and approvals. Root cause is attempted but often stops at human error.

Typical weakness. Effectiveness is confirmed as 'implemented' rather than measured.

Risk. Repeat findings, which an inspectorate treats far more seriously than a first.

Level 3

Systematic

The process is designed, resourced and measured.

Investigation depth is proportionate to risk, causes are confirmed with evidence, and effectiveness criteria are defined before implementation.

Typical weakness. Backlog control is reactive — the queue is worked by age rather than by exposure.

Risk. High-exposure items sit behind easy ones because progress is measured by count.

Level 4

Integrated

It connects to the rest of the quality system.

Backlog is triaged by regulatory exposure, trends feed preventive action, and CAPA connects to change control and validation impact.

Typical weakness. Preventive action still lags corrective action in practice.

Risk. Known weaknesses recorded but not acted on until they produce an event.

Level 5

Optimized

It improves itself, and the improvement holds.

Effectiveness is verified against pre-defined data after closure, intake is repaired when it produces backlog, and recurrence is trended as a system metric.

Typical weakness. Sustaining the discipline when volume rises.

Risk. Reversion under pressure, which shows first in effectiveness-check quality.

Expected controlsWhat a site at level 3 or above should be able to show.
Pre-defined effectiveness criteria
Measurable, with the data source and review date named before implementation.
Confirmed root cause
Established by evidence rather than by plausible narrative.
Exposure-based triage
The queue ordered by regulatory exposure, not by age or ease.
Intake repair
The process that generated the backlog is fixed, not just the queue drained.
Recurrence trending
Repeat events tracked as a system metric visible to management.
Moving up

What actually shifts the level

  • Audit your own closed CAPAs for whether an effectiveness criterion existed before implementation. The proportion is usually the fastest diagnostic available.
  • Count how often retraining appears as the action against similar events — that count is the argument for redesign rather than more training.
  • Re-triage the open queue by exposure once, and keep it that way.
  • Add the question 'should the system have caught this earlier?' to every investigation, because it usually has a second cause.
Regulatory basis

Where the underlying requirements sit

Limits

This is a framework, not a classification

These five levels are a GMPConsultant.nl framework for structuring a conversation about capability. They are not an official regulatory classification, no authority recognises or issues them, and a maturity level is not a statement of GMP compliance.

Next step

Score this domain rather than estimate it

The GMP Health Index scores seven domains and maps each to a maturity level, in about ten minutes. It calculates in your browser and stores history on your device only.