Table of contents
At a glance
What should you do if a GMP-relevant audit trail was disabled?
Treat the finding as a possible data integrity gap. Contain further risk and preserve available records and configuration evidence under control. Identify the scope and potential affected period, assess data and product decisions, and define interim controls. Controlled reactivation is only part of the response. Investigate the cause, CAPA, effectiveness and remaining uncertainty.
The technical assessment does not yet indicate what has occurred.
A disabled audit trail is initially a specific control finding. It does not automatically prove that data has been manipulated. Nor does it conclusively demonstrate that historical data remains unaltered. Assess which events the function should capture, which records may be affected, and whether other reliable traces are available. The evaluation must clearly identify both substantiated facts and remaining uncertainties.
Define the function precisely. Some systems have multiple protocols for data, methods, configuration, or administration. An active system history does not automatically imply that changes to all relevant datasets have been captured. Conversely, a deactivated subfunction may affect only a limited scope. Document the actual status, version, settings, and relevant responsibilities. Avoid premature conclusions such as either "only a checkbox" or "all data unusable."
Limit additional risks and retain traces
Notify the relevant Quality and System Owners and establish a controlled response. Depending on risk, affected activities may be restricted, additional reviews introduced, or data temporarily excluded from certain decisions. The appropriate action depends on the process, data criticality, and available controls. A blanket continuation or a full-scale halt cannot be justified solely from the technical assessment.
Ensure relevant existing records and configuration information are preserved according to an appropriate procedure. Avoid uncoordinated modifications, restore operations, or cleanup activities that could overwrite traces. In cases requiring immediate hazard mitigation, the approach shall be coordinated and documented with the relevant responsible parties. Specify who captured which information and when. A subsequent copy must be clearly identifiable as such and shall not be presented as an original record generated simultaneously.
Limit the possible time period to a verifiable extent
Look for verifiable state changes: configuration modifications, installations, upgrades, authorization events, released baselines, and prior audit evidence. A screenshot taken today does not demonstrate the state from three months ago. The absence of an entry does not, by itself, prove the date of deactivation. The reliability of each indication is assessed. A time window may initially be described as "no later than" or "not reliably prior to" when a precise boundary cannot be substantiated.
Capture affected projects, devices, customers, or data classes. Assess whether the same configuration pattern has been applied in other systems. A single finding may indicate a broader root cause, without necessarily implying that all systems are impacted. Use criteria for expansion and limitation. Historical evaluation should clearly demonstrate the rationale for including one area and excluding another. A convenient fixed retrospective period does not substitute for such justification.
Evaluate data and product traces separately from the configuration.
Review the decisions that may be based on the potentially affected data. This may include inspection reports, material decisions, and previously released batches. Assess the uncertainty introduced by the missing history and identify independent information that supports the evaluation. Other records may be helpful, but they can only substitute for the missing audit trail to the extent that they actually support the specific statement required. A mere plausibility assumption does not close the gap.
The responsible quality and, if applicable, product authorities decide on any additional required actions. Investigations, restrictions, retests, or regulatory communications are evaluated in accordance with established procedures. This general guidance does not establish a universal reporting deadline nor a automatic recall decision. Similarly, a subsequent analysis shall not implicitly replace an unresolved historical data situation. The original record and any subsequent evaluation remain distinct.
Sources: [3]
Example: three weeks with unclear history
In the exercise scenario, the current configuration shows a deactivated data history. A previously released test from three weeks ago confirms the previously active function. Between these two time points, multiple analysis sequences were performed. An upgrade occurred within this timeframe, but it does not alone establish causality. The team first prepares a list of affected data and decisions and reviews existing change and operational records.
The action plan separates immediate control, investigation, and permanent improvement. It forwards open questions rather than closing them through current activation. The following work aid is a standalone proposal. It does not define a general regulatory response and does not require that each case undergo the same technical forensic approach. What is essential is a risk-appropriate, traceable handling with clear responsibility for the professional implications.
| Question to address | Appropriate information | Expected result |
|---|---|---|
| Which function was not active? | Actual Settings and System Scope | Precise control test result |
| For how long could the gap have existed? | Reliable Baselines, Changes, and Reviews | Justified Time Window with Uncertainty |
| Which data and decisions are affected? | Data population and process relationships | Complete assessable list of affected parties |
| Which intermediate control applies now? | Risk and permitted activities | Effective action with accountability |
| Why did the gap occur? | Technical and organizational root cause evidence | Traceable root cause assessment |
| How do we recognize sustained efficacy? | Appropriate tests and operational validation evidence | Verified control instead of closed task only |
Why Reactivation Alone Is Insufficient
In the June 4, 2019 Warning Letter regarding ABC Testing, the FDA highlighted deactivated audit trails and inadequate access controls. The proposed activation and setup of user accounts constituted an insufficient response; between-checks and procedures for monitoring and verifying analytical data were missing. This historical isolated incident is not a binding template applicable to every facility. It, however, illustrates the distinction between technical remediation and a sustainable, comprehensive corrective action.
During reactivation, verify that the actually required events are now captured. An active setting may still exhibit insufficient coverage. Test relevant creation, modification, and, if necessary, additional required actions using controlled data. Ensure that the timing, actor, content, and interrelation are documented to the required extent. The evidence should also demonstrate that the routinely assigned responsible personnel can reasonably review and assess the history.
Sources: [4]
Link root causes and CAPA to the intended process objective
Investigate both technical and organizational root causes jointly. Was a configuration change introduced during the upgrade? Was the requirement ambiguous? Could too many end-users have modified the function? Was there a lack of proof following a system backup? A "staff training" measure may be insufficient if an uncontrolled technical pathway remains. Conversely, technical restrictions do not automatically resolve gaps in responsibility for review and change evaluation.
Link the CAPA to the intended future operation. A standardized configuration review, appropriate roles, and clear handovers can prevent similar errors and simultaneously alleviate daily workload. Prioritize based on risk, dependencies, and strategic objectives. Fully address the specific finding without reducing the entire project to an isolated regulatory documentation exercise. Effectiveness is demonstrated by functional controls and reliable decisions, not merely by increased documentation.
Sources: [5]
Close historical gaps and future efficacy separately
The conclusion must demonstrate what can actually be substantiated from the past. Which data were evaluated? What limitations remain? What decisions were made and by whom? A current functional test can confirm future control, but it does not reconstruct events that were never recorded. Do not create retroactively dated audit trail entries to fabricate an apparently complete history.
For future operations, state monitoring, change evaluation, review, and escalation procedures will be defined. Appropriate triggers such as upgrades, migration, or system duplication will be assessed. The frequency of ongoing monitoring activities will be justified based on risk. If the root cause affects multiple systems, the effectiveness will be evaluated to the corresponding extent. Thus, historical investigations and sustainable prevention will remain linked, without conflating their distinct objective evidence.
Project experience and collaboration
Preserve evidence of the gap and investigate its impact
If an audit trail was disabled, we establish the affected period and actual use with QA, IT and process owners. We preserve original information. Missing history must not be replaced with entries created after the event. Independent records may help explain individual transactions, subject to their limits. Where the impact cannot be adequately bounded, broader assessment and possibly immediate restrictions on use are necessary. We support investigation and verification that control has been restored. Reactivating the trail does not establish control retrospectively.
Frequently asked questions about Responding to a disabled audit trail
Must all data be rejected when an audit trail is missing?
Not as a general rule. Assess the affected data, required conclusions, independent evidence and residual uncertainty. The responsible quality and product owners decide under the applicable procedures. A general loss of confidence or unexamined continued use cannot replace that reasoned decision.
Does it suffice to disable the audit trail?
No. This can address future controls, but it does not close the historical gap. Investigate timeframes, data and product sequences, intermediate controls, and root causes. The activated function must demonstrate its required coverage within the specific system. Procedures and responsibilities are part of the ongoing action.
Sources: [4]
Can we retroactively enter missing historical entries?
Not to be considered as a presumed original system history. A subsequent reconstruction or evaluation may be required, but must clearly demonstrate origin, timing, basis, and uncertainty. Retain the original information and distinguish it from post-hoc investigations. A artificially complete history would obscure the data integrity.
Does another system log prove that nothing was changed?
Only if it reliably covers the specific necessary statement. Review the content, completeness, identities, time period, and log protection. A connection or startup protocol may address different questions compared to a record-level change history. The evidential strength is explicitly limited, rather than being broadly transferred.
Sources: [2]
How far must the investigation extend?
The boundary is justified based on robust evidence of system status, data criticality, and root cause analysis. There is no fixed, universal retrospective period established from this general case. When the start date cannot be reliably determined, uncertainty remains evident and is assessed. A conveniently selected timeframe does not substitute for professional justification.
Is a training session sufficient as a CAPA?
This depends on the established root cause. Training alone does not eliminate an ongoing technical workarounds. Review roles, configurations, change procedures, and review processes. Effectiveness assessment indicates that the relevant root cause is under control and that the required documentation is reliably generated during operations.
Sources: [5]
Sources and review date
As of October 15, 2026. The classification is based on the listed regulatory and authority sources. Investigation steps and examples are internal work aids.
- European Commission: EU GMP Annex 11, Computerised Systems (2011)Current version as per EudraLex inventory as of 15.10.2026. Application, risks, operation, and data lifecycle; specific testing proposals are derived independently.
- PIC/S: PI 041-1, Good Practices for Data Management and Integrity (2021)Inspection Guidance on Data Integrity; in particular electronic data, transmission, access, and review.
- FDA: Data Integrity and Compliance With Drug CGMP, Q&A (2018)Final Guidance: Data completeness, electronic source data, review and investigation. Guidance and binding regulations differ.
- FDA Warning Letter: ABC Testing, 04.06.2019Specific historical inspection case: Activation of audit trails alone was insufficient; intermediate checks and procedures were missing. No general action catalog applicable to every facility.
- EMA / ICH: Q10 Pharmaceutical Quality System, Step 5 (2008)Section 3.2.2: Root Cause Analysis, CAPA, and Effectiveness Evaluation. Specific action plans and exercises provided in the contribution are standalone methodological aids.
Sources checked: .
How we can support you
Daniel Herrmann Consulting combines professional assessment with hands-on work on the specific system. During the first meeting, we clarify your intended use, existing documentation, and the next most appropriate action step.
- Audit and inspection preparation →
Go into an audit or inspection with accessible evidence and a prepared team. We begin with a walkthrough, rank open issues by risk and rehearse the inspection with your team.

