CAPA in a GMP environment: are we fixing the root cause or only the symptom?

Documenting a deviation is mandatory, understanding it is the real work. Without a sufficient investigation of the underlying root cause, there is a risk of taking actions that only treat the symptoms. This increases the risk that similar deviations will occur again.

In many quality systems, every deviation is properly recorded, assessed and closed on time. Even so, it reappears in a similar form a few months later, on a different line or in a different shift. This may indicate that the underlying root causes were not investigated sufficiently or that the CAPAs derived from them were not effective.

Chapter 1 of the EU GMP Guide requires an appropriate level of root cause analysis for deviations. Where human error is suspected or identified as the cause, this has to be justified. In doing so, it must be ensured that possible process, procedural and system-based errors have not been overlooked. In addition, the effectiveness of the actions taken has to be monitored and assessed. Root cause analysis, the assessment of human error and the effectiveness check in particular are where weaknesses can occur in practice.

Where root cause analysis starts

It begins with a precise description of the event: what happened, when, where, to what extent, and what should have happened. Only once the expected and the actual state are clearly separated is it worth looking for causes. Settling on an assumption too early in a root cause analysis risks overlooking other possible causes.

There are proven tools for the search itself. The 5 Whys method helps to identify possible underlying causes through repeated questioning; these should then be verified against the available information and data. The Ishikawa diagram groups possible influences by man, machine, method, material, measurement and environment. In complex cases, a fault tree analysis shows how several causes interact.

How deep the analysis goes depends on the risk. According to ICH Q9(R1), the level of effort, formality and documentation of the quality risk management process should be commensurate with the level of risk. The depth of a deviation investigation should likewise be set based on risk. While a less extensive root cause analysis may be sufficient for a low-risk deviation, an OOS result in a release test, for example, requires a thorough investigation. This should also include checking whether other batches, products or equipment could be affected by the same cause.

Why the analysis often ends too early

A typical weakness is a result that is settled too quickly. "Human error" as the cause and "employee retrained" as the action are often used to justify closing a deviation without the underlying causes having been investigated sufficiently. Why was the error possible? Was the SOP ambiguous, the working environment unfavorable, the time too short, or the step in the batch record easy to miss?

An example from practice: in production, a form was filled in using an outdated version. An employee had printed it from the document management system, although a newer version was already in effect. This was noticed during the QA review, and a deviation was documented.

The investigation showed that, in addition to the wrong version being selected, the system settings also played a role: older document versions were still accessible in the system. The display of older versions and the role-based access permissions were therefore adjusted to reduce the risk of outdated documents being used again.

Then there is time pressure. Internally defined deadlines often apply to processing and closing deviations, while a thorough root cause analysis takes time. Under this pressure, there is a risk that the first plausible explanation is accepted prematurely as the final root cause. Another point is looking at cases in isolation: if deviations are only ever considered individually, connections between seemingly unrelated cases can be missed. Regular trend analyses of deviations and CAPAs help to identify recurring patterns and possible common causes.

Data is often missing, too. Without sufficient and relevant information, such as measured values, logbooks, audit trails or interviews with the people involved, the root cause analysis may remain limited to assumptions. The identified causes should be supported by traceable evidence wherever possible. If the actual cause cannot be determined unambiguously despite an appropriate investigation, the most likely causes should be assessed on the basis of the available findings, the remaining uncertainty documented, and suitable actions derived from this.

What makes a CAPA effective

In practice, three terms are often confused in connection with CAPA: correction, corrective action and preventive action. A correction eliminates a detected deviation, for example through permissible rework. Immediate actions to contain possible effects, such as provisionally quarantining an affected batch, are something different again. A corrective action aims to eliminate the cause of a deviation that has occurred and to prevent it from recurring. A preventive action, by contrast, is intended to eliminate the causes of potential deviations before they occur.

An effective action addresses the proven cause. If the cause lies in an ambiguous SOP, training on the same SOP changes little. Actions that specifically fix system-related weaknesses can be particularly effective: a clearer work instruction, a technical lock, an additional verification step or an adjusted equipment setting. The less an action depends on the memory of individual people, the more robust it is.

Every action needs an owner, a due date and an assessment of possible new risks. If a CAPA requires a change to a GMP-relevant process, it has to be checked whether this must be managed through the formal change control procedure. A large number of actions for a single deviation should also be questioned critically. What matters is that each action is traceably directed at an identified cause or a relevant contributing factor.

How to tell that the action has worked

An implemented action is not yet an effective one. The effectiveness check answers a separate question: has the action achieved its defined objective and sufficiently reduced the risk of the deviation recurring? This requires criteria defined in advance and as measurable as possible, an appropriate observation period and suitable data against which effectiveness can be assessed.

Suitable criteria should be measurable and aligned with the objective of the respective CAPA. Examples include the absence of comparable deviations over a number of batches set based on risk, a stable trend in a critical process parameter, or demonstrably improved execution of a revised work step. If the problem persists, the reasons for the lack of effectiveness have to be investigated and, where necessary, the original root cause analysis and the defined actions reassessed.

An effective CAPA system is not defined only by deviations being closed on time. What matters is whether the underlying causes are investigated appropriately, suitable actions are implemented and their effectiveness is demonstrated. This is how recurring deviations can be reduced in a targeted way, and how it can be shown traceably during inspections that the pharmaceutical quality system contributes to continual improvement.

Image notice: The illustrative and header images were generated using AI and do not depict real people, places, or events. The portrait photographs show real people.