Regulations and compliance · Practical guide

GxP revalidation: when do you need new evidence?

An update is one possible trigger for revalidation. New interfaces, changed master data rules, additional sites or a new decision supported by the system can also exceed the scope of existing evidence. Start with how the system is intended to be used.

Table of contents

At a glance

When does a GxP system need revalidation?

New or supplementary evidence is needed when a change, new intended use or finding calls prior suitability into question. Annex 11 calls for controlled changes, including configuration changes, in section 10, and periodic assessment of the validated state in section 11. Affected functions, data and risks determine scope. Not every change requires a full retest, and an unchanged version number does not establish that actual use remains suitable.

Scope and Applicability

The term "revalidation" is used differently across companies. What is essential are traceable impact assessments, required testing, and release decisions, not the name of the document package.

Sources: [1]

The justification and scope should be decided separately.

An impact assessment describes which assumptions from previous validation remain valid and which require further evaluation. It may justify the continued use of existing evidence. A general statement of "no impact" without reference to functions, data, and dependencies, however, is difficult to verify.

The Periodic Review can identify gaps, but it does not replace timely control of a specific change. Conversely, a regular review does not automatically trigger a complete revalidation project. Its outcome may also confirm that existing controls and documentation remain appropriate.

Sources: [1]

Revalidation from effect instead of from a keyword

A change may require additional evaluations without triggering a full reimplementation. Conversely, a minor technical adjustment can have significant operational implications. Therefore, start with the intended use, affected data, and potential impacts. The term "patch," "configuration," or "major release" is merely indicative.

Annex 11 addresses controlled changes and regular assessment of the validated state. This does not lead to a blanket, comprehensive annual revalidation cycle. The required evidence shall be based on the type of change, available knowledge, and remaining risks. The term "revalidation" should be used clearly and consistently within the organization, rather than applying the same label to all evaluations.

New insights can trigger testing requirements: recurring errors, insufficient test evidence, changed usage, or findings during a Periodic Review. It is not merely a new software version that alters the assessment. An unchanged system that suddenly controls a different quality process requires a new evaluation.

Sources: [1]

Create an event trigger matrix with potential test conclusions

Capture the reason, affected functions, existing evidence, and open questions. Then determine which evidence remains appropriate and what additional information is required. The decision should explain why the selected scope adequately addresses the implications. A blanket repetition of all previous tests may still fail to identify relevant new risks.

Own work aid: Develop an trigger matrix with potential inspection conclusions
TriggerOwn key questionPossible additional evidence
New Use Which new decision is supported? Enhanced requirements and process validation
Functionality change Which critical logic is changed? Targeted Functional and Regression Testing
Data migration What values and relationships are transferred? Alignment and Reconstruction
Role change What actual actions are affected? Positive and negative access cases
Recurrent error What is the common root cause? Root cause investigation and corrective action effectiveness
Review finding Which previous statement is inadequate? Rationale closure of the specific gap

Sources: [1]

Example: minor parameter change with broad impact

A common conversion factor is updated. Technically, this affects only one configuration line. Functionally, several reports and interfaces use the same factor. Testing only the first known report does not fully capture the impact.

Determine dependent functions and critical usages. Use known values against independently derived expected results. Evaluate appropriate normal and boundary cases, as well as recipient effects. Existing validation evidence for independent functions may be reused. The scope is justified by dependency, not by the number of modified lines.

Evaluate historical data when the reason for change is a correction of an error. A new correct calculation does not necessarily confirm that previous decisions were unaffected. The scope and required actions are assessed professionally. A planned technical change may be classified differently than a correction to a previously incorrect rule.

Own work aid – Example: minor parameter change with broad impact
Test stepActionExpected finding
Dependency Joint use of factor recording Affected scope known
Reference calculation Calculate normal and limit values Independent specified values available
Functional test Affected expenses generate New regulation works correctly
Regression Review relevant common processes No unexplained adverse effect
History Evaluate previous use Error correction and planned replacement distinguished

Sources: [1]

Existing evidence supports continued use

Review the status, scope, and prerequisites of previous tests. A validation conclusion may remain valid if the function and relevant environment have remained unchanged. A merely similar test designation is insufficient. A clear link to the current requirement and the actual system must be maintained.

Supplier documentation can support the evaluation. However, verify that it covers your specific version, configuration, and usage. A general release report does not confirm every customer-specific interface. Missing information may necessitate additional internal assessments or a limited usage decision.

Deliberately avoid repeating sections with a justified rationale. This does not require an extensive, individual analysis of every screen. A clear mapping of changes, dependencies, and testing scope delivers greater value than a generic statement "no impact" without substantiation.

Sources: [1]

Regression, Requalification and Process Validation differ

An infrastructure change may require appropriate qualification and operational documentation. A changed functional responsibility may necessitate different tests. A role change pertains to access rights. These tasks may occur within a single project, but should address their respective questions. A successful server startup does not constitute validation of correct result calculation.

Select critical end-to-end scenarios when multiple components interact. Include failure paths where changes could affect their behavior. A regression test using only normal data may be insufficient when the intervention involves restart or exception handling.

The batch release decision consolidates all required supporting evidence. Identify open findings and usage limitations. A global status "revalidated" without defined scope is largely inconclusive. The decision should indicate which usage has been confirmed for each specific version.

Sources: [1]

Identify and control validation requirements in ongoing operations

Link Change Control, Incident Management, and Periodic Review. Recurrent disruptions may indicate that a previous control is insufficient. A review can identify missing evidence or altered usage. The information should be consolidated into a single evaluation pathway, rather than remaining isolated in separate lists.

Identify professional and technical responsibility. A system owner should be able to perform an impact assessment and know when QA or other expertise is required. A template may assist in this process; it does not replace the assessment. The handover should be validated through a specific change scenario.

Sources: [1]

Working template: derive additional test scope from a change

First describe the specific reason and the relevant technical impact. In the example, a commonly used factor is modified. The number of modified lines is not an indicator of overall scope. Identify dependent calculations, reports, and recipients, as well as potential historical consequences.

Link each required statement to existing or new evidence. An appropriate prior validation can remain in place if the conditions and functions have not changed. A new dependency requires appropriate supplementation. The decision should justify that the testing scope is both sufficient and directly aligned with actual risks.

Document the intentionally non-repeated areas in a clear and understandable format. A general statement of "no impact" without supporting evidence is weak. A complete copy of all previous tests may inadvertently introduce new risks. Quality lies in the justified attribution of effects and evidence.

Sample filling example – internal work aid
Evaluation fieldExample entryDecision on testing
Trigger Common conversion factor is amended. Substantive impact over technical labeling.
Dependency Two reports and one interface leverage the factor. Identify what is affected.
Existing evidence Independent functions remain unchanged. Appropriate evidence justified continued use.
New inspection Reference values and critical edge cases. Confirm altered efficacy.
Regression Relevant common processes. Identify unknown adverse effects.
History Change is intended assignment or a correction for error. Evaluate previous use suitability.

Sources: [1]

Next steps

  • Describe the change or new use specifically.
  • Review existing requirements and evidence of continued validity.
  • Define additional tests based on impact assessment.
  • Document decisions and outcomes regarding further use.
Further reading:

Project experience and collaboration

Identify the additional evidence a change requires

With the system owner and process owners, we identify which conclusions about suitability a change affects. A configuration correction with few dependencies may need focused testing. Changed use or uncertain interactions call for a broader assessment. We examine whether existing evidence still applies and justify any additional checks. A new version number or an annual date alone determines neither the need for full revalidation nor that further work can be omitted.

Frequently asked questions about Revalidation triggers

Is there a general annual mandatory revalidation requirement?

No. Annex 11 requires controlled changes (No. 10) and regular assessment of the validated state (No. 11), not a general annual full revalidation of all tests. Additional evidence arises from usage, changes, findings, and risks. Maintain a justified scope, rather than equating a calendar-based schedule with complete revalidation.

Sources: [1]

Can a patch be released without a functionality test?

This depends on a justified assessment of impact and available evidence. A blanket waiver solely due to the designation "Security Patch" is not substantiated.

Sources: [1]

Is a successful Periodic Review considered a new validation?

Not automatic. It evaluates the current state and can trigger actions, but is not equivalent to the repetition of the entire project.

Sources: [1]

Can a configuration change trigger more validation requirements than a software patch?

Yes. The impact on the process is decisive. A single threshold or common factor can influence multiple quality decisions. In contrast, a patch may affect only a limited, independent scope. The impact assessment should consider specific functions, data, and dependencies, not merely the technical designation of the change.

Can a Periodic Review trigger additional tests?

Yes, if altered usage, insufficient evidence, or relevant findings are identified. The scope is derived from these observations. A review is not automatically a complete revalidation. It may necessitate targeted actions, additional substantiation, or a more in-depth assessment.

What should an impact assessment explicitly justify?

The selected testing scope should be linked to the affected functions, data, dependencies, and potential consequences. Identify appropriate existing evidence and remaining gaps. The technical change title alone is insufficient. A subsequent reviewer should be able to understand the decision without additional verbal explanation.

Sources and review date

Annex 11, Sections 10 and 11 link controlled changes to the assessment of the continued valid state. The impact matrix is an independent work aid for selecting additional evidence, not a fixed test frequency or a mandatory annual comprehensive review.

  1. European Commission: EU GMP Annex 11, Computerised Systems (2011)§§ 10 and 11: Modifications and regular evaluation; no general, comprehensive annual test repetition. Current version as of EudraLex dated 14.10.2026; no universal project or deadline mandate.

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Determine the need for additional substantiation requirements selectively

We support you in assessing the current situation, clarifying responsibilities, and implementing appropriate evidence within operations. The type of support that is most effective depends on your system, your team, and the specific objective.

  • Computer system validation →

    Bring your GxP systems into validated operation, from laboratory software, MES and eQMS to ERP. We define the scope, plan risk-based tests using GAMP 5 and hand the evidence to QA for release.

Send a project enquiry Customer testimonials →