EU Pharmacovigilance Inspection Findings: Recurring Patterns and What Inspectors Actually Test

Introduction

A pharmacovigilance inspection is not simply a document review. It is a regulatory test of whether the pharmacovigilance system works in practice.

Inspectors may examine the same process from several directions: written procedures, system configuration, individual cases, training, performance data, agreements, quality records, interviews and evidence of management oversight. A system can therefore appear compliant when viewed through a single document and still fail when the inspector follows a real activity from beginning to end.

This article starts a dedicated series on EU pharmacovigilance inspection findings. It focuses on recurring patterns rather than reproducing the older general discussion of finding classification and CAPA already present elsewhere in the QPPV knowledge base.

The objective is to answer a practical question:

What do EU pharmacovigilance inspectors actually test, and what makes a deficiency significant?

The distinction between an actual inspection finding and an illustrative example is important. Where this article describes findings reported by EMA's Pharmacovigilance Inspectors Working Group (PhV IWG), they are identified as such. Illustrative scenarios are labelled accordingly.

1. Who Performs EU Pharmacovigilance Inspections?

Responsibility for conducting pharmacovigilance inspections rests with national competent authorities. EMA has an important coordination and harmonisation role, including maintaining EU-level procedures and coordinating inspections relating to centrally authorised products (CAPs). citeturn0search1turn0search5

For human CAPs, the supervisory authority is the competent authority of the Member State where the PSMF is located. EU inspection programmes are risk-based and can include routine and for-cause inspections, with national programmes accounting for the majority of EU/EEA pharmacovigilance inspections. citeturn2view0

This matters for interpretation of inspection statistics. An EMA PhV IWG annual report does not represent every pharmacovigilance inspection conducted in Europe. It reports the defined inspection population covered by the programme and report.

2. What Is an Inspector Actually Trying to Establish?

At system level, an inspector is trying to establish whether the MAH has the personnel, systems and facilities necessary to meet its pharmacovigilance obligations and whether those arrangements function as described. The 2024 PhV IWG report states that CAP inspections are system inspections in which one or more products may be selected as examples so that specific information can be traced and verified through processes. citeturn2view0

A useful model is:

Regulatory requirement
        ↓
Written system
        ↓
Operational process
        ↓
Actual record / transaction
        ↓
Quality control
        ↓
Management oversight
        ↓
Evidence of effectiveness

The inspector can enter the chain at any point and move in either direction.

For example, an inspector may start with an SOP and then ask for a real ICSR processed under that SOP. Alternatively, the inspector may select a case and then ask which procedure, training, system control and quality check governed its handling.

3. The Evidence Chain

A strong inspection response does not depend on one document. It depends on consistent evidence across the system.

Typical evidence can include:

• approved procedures and work instructions;
• training records;
• safety-database configuration;
• case-processing records;
• literature-search records;
• submission acknowledgements;
• reconciliation records;
• quality-control results;
• performance indicators;
• audit reports;
• deviations and CAPAs;
• vendor oversight records;
• PSMF content;
• QPPV review and escalation evidence; and
• management-review records.

The inspector's question is often not "Do you have a procedure?" but rather "Show me evidence that the procedure is implemented and controlled."

4. Why a Local Error Can Become a System Finding

An isolated error is not automatically evidence of systemic failure.

However, an apparently local issue becomes more concerning when the inspector can demonstrate one or more of the following:

• the same failure occurred in multiple cases;
• the process allows the failure to recur;
• quality controls did not detect it;
• management information did not reveal it;
• the written system does not define the control adequately;
• training did not prepare staff to perform the activity;
• the vendor or affiliate interface is unclear;
• previous audits or CAPAs identified the same weakness; or
• the organisation cannot determine the true scope of the problem.

The important issue is therefore control architecture, not merely error counting.

5. The 2024 Human Inspection Dataset

The EMA PhV IWG annual report for 2024 provides a useful recent reference point.

For human CAP pharmacovigilance inspections in the programme, 59 inspections were conducted in 2024: 52 at QPPV/PSMF (MAH) sites, one at a global pharmacovigilance site and six at subcontractor/licensing-partner/affiliate sites. citeturn2view0

For CHMP-requested human inspections conducted in 2024, the report records 87 deficiencies: 0 critical, 29 major and 58 minor. The three most common finding areas were:

1. management and reporting of adverse reactions;
2. quality management systems; and
3. the PSMF. citeturn2view0

This is a useful empirical anchor for the series. It shows that the recurring inspection problem is not one exotic failure mode; it is concentrated in the core operating machinery of pharmacovigilance.

6. The 2023 Dataset Shows a Similar Pattern

The PhV IWG 2023 annual report recorded 94 deficiencies in CHMP-requested human pharmacovigilance inspections: one critical, 32 major and 61 minor. The three most common areas were again QMS, management/reporting of adverse reactions and the PSMF. citeturn1search15

The recurrence across consecutive reporting years is more informative than a single year's ranking. It suggests that these areas deserve particular attention when an MAH assesses inspection readiness.

It does not, however, mean that every MAH has the same deficiencies or that the ranking can be extrapolated to all national inspection programmes.

7. The Three-Layer Inspection Test

The recurring areas can be understood through three questions.

Layer 1 — Is the requirement understood?

Does the organisation know what must be done, by whom and within what timeframe?

Layer 2 — Is the process controlled?

Does the system make compliant performance likely and detect failures when they occur?

Layer 3 — Can the organisation prove it?

Can the MAH produce reliable evidence showing that the process operated over the relevant period?

A weakness at any layer can generate an inspection finding.

8. Finding Area One: Management and Reporting of Adverse Reactions

Management and reporting of adverse reactions was the leading human finding area in 2024. EMA recorded 22 findings in this area; 12 were major and 10 were minor. Among the major findings were deficiencies involving submission and follow-up, medical review and MedDRA coding, literature screening, and receipt and collation of ICSRs from all sources at a single collection point within the EU. citeturn2view0

This is important because it demonstrates that case-processing findings are broader than simple reporting-timeliness errors.

Inspectors can examine the whole case lifecycle:

Source
 ↓
Receipt
 ↓
Identification / triage
 ↓
Case validity
 ↓
Data entry
 ↓
Medical review
 ↓
Coding
 ↓
Follow-up
 ↓
Quality control
 ↓
Submission
 ↓
Reconciliation / closure

A deficiency anywhere in that chain can affect the reliability of the pharmacovigilance system.

9. Finding Area Two: Quality Management System

QMS findings are important because the quality system is the mechanism through which the MAH controls the pharmacovigilance processes themselves.

An inspector may therefore look beyond the existence of SOPs and examine:

• whether procedures reflect actual practice;
• whether deviations are identified and investigated;
• whether CAPAs address root causes;
• whether training follows procedural changes;
• whether quality controls are risk-based;
• whether performance indicators are meaningful; and
• whether management receives sufficient information to intervene.

A QMS finding can consequently be a symptom of weakness elsewhere in the PV system.

10. Finding Area Three: PSMF

The PSMF is a particularly powerful inspection lens because it describes the pharmacovigilance system while providing a route to supporting evidence.

The inspector may compare the PSMF against operational reality.

Potential discrepancies include:

• responsibilities that do not match current organisational arrangements;
• obsolete service providers;
• missing or outdated process descriptions;
• inconsistent information about affiliates;
• incomplete audit information;
• outdated quality-system descriptions; or
• evidence that the documented system does not reflect the system actually being operated.

The important principle is simple:

The PSMF should describe the pharmacovigilance system that actually exists, not the system the organisation wishes existed.

11. What the Inspector May Do Next

If an inspector identifies a discrepancy, the next question may be about scope.

For example:

"You have shown me one case with this problem. How did you determine whether other cases were affected?"

That question tests whether the organisation has a reliable way to identify the population exposed to the same process weakness.

This is where data quality, reconciliation, quality-control records and management information become important.

12. A Recurring Pattern: The Control That Exists but Does Not Control

One of the most useful ways to understand inspection findings is to distinguish between process existence and process effectiveness.

A control can exist on paper while failing operationally.

Examples of illustrative scenarios include:

• an SOP requires second-person review, but the system does not record whether the review occurred;
• a literature process requires escalation, but no evidence demonstrates that escalation thresholds are monitored;
• a vendor agreement allocates responsibilities, but the MAH has no performance evidence showing the vendor met them;
• a CAPA is closed because a procedure was revised, without evidence that the revised process prevented recurrence.

These examples are illustrative rather than reported inspection findings. Their purpose is to show the type of control-versus-evidence gap that inspectors can explore.

13. Traceability Is a Core Inspection Concept

Inspectors need to be able to move from an assertion to supporting evidence.

If an organisation states that all serious ICSRs are submitted on time, the inspector may ask for the performance data, then sample cases, then submission acknowledgements, then the underlying receipt dates and triage records.

This creates a traceability chain:

claim → metric → sample → source record → system event → control → responsible function.

Weak traceability makes a compliant process difficult to demonstrate. More importantly, it can prevent the MAH itself from knowing whether the process is functioning reliably.

14. Inspection Finding Versus Inspection Question

Not every difficult inspection question becomes a finding.

Inspectors can ask probing questions to understand the system, test consistency or determine the scope of an issue. A finding normally requires evidence of a deficiency against an applicable requirement or an ineffective system control.

This distinction matters when communicating inspection experience internally. An organisation should not convert every inspector question into a supposed regulatory requirement, and an educational article should not present hypothetical inspection questions as published findings.

15. Finding Classification Does Not Replace Root-Cause Analysis

The classification of a finding tells the organisation something about regulatory significance. It does not by itself explain why the problem occurred.

A major finding could result from different underlying causes:

• inadequate process design;
• ineffective quality control;
• system configuration;
• unclear ownership;
• insufficient training;
• vendor failure;
• poor change control;
• weak management oversight; or
• several interacting causes.

The same visible deficiency can therefore require very different CAPAs depending on the root cause.

16. Repeat Findings Are Especially Informative

A repeated deficiency deserves a different analytical response from a first occurrence.

If a previous CAPA was intended to prevent a recurrence and the same problem is found again, the inspector may reasonably examine:

1. whether the original root cause analysis was adequate;
2. whether the corrective action addressed the cause;
3. whether the action was implemented as intended;
4. whether effectiveness was actually tested; and
5. whether management responded when early warning indicators appeared.

A repeat finding can therefore reveal a failure of organisational learning rather than merely another procedural error.

17. The PSMF as a Cross-Check, Not Just a Document

The PSMF can connect many inspection domains.

An inspector may use it to understand:

• the QPPV arrangement;
• organisational responsibilities;
• key processes;
• contracted activities;
• safety databases;
• quality systems;
• audits;
• performance monitoring; and
• the relationship between global and EU pharmacovigilance activities.

The PSMF can then be compared with evidence from the actual system.

This makes discrepancies particularly useful to an inspector. A PSMF that says a process is centrally controlled while operational evidence shows material local variation raises questions about governance and oversight.

18. The QPPV Is Part of the Evidence Chain

QPPV involvement should not be reduced to a signature on documents.

Inspectors may be interested in whether the QPPV has sufficient visibility of significant pharmacovigilance issues and whether escalation mechanisms allow the QPPV to act when required.

Relevant evidence may include:

• governance meetings;
• escalation records;
• quality reviews;
• inspection preparation;
• CAPA oversight;
• significant deviation assessment;
• safety-system performance information; and
• evidence of interaction with senior management.

The precise organisational model varies, but the underlying question is whether the QPPV role is operationally meaningful.

19. Outsourcing Does Not Remove Inspection Accountability

A pharmacovigilance activity can be performed by an affiliate, service provider, licensing partner or other contractor while remaining part of the MAH's pharmacovigilance system.

The 2024 inspection programme included inspections of subcontractor/licensing-partner/affiliate sites as well as QPPV/PSMF and global pharmacovigilance sites. citeturn2view0

This reflects an important inspection principle: the regulator can follow the activity to the location where it is actually performed.

The MAH should therefore be able to demonstrate:

• defined responsibilities;
• appropriate agreements;
• qualification and training;
• oversight;
• performance monitoring;
• deviation and escalation pathways; and
• audit or inspection readiness.

20. The 2019–2022 Evidence Base

EMA's report on pharmacovigilance tasks in the EU Member States and EMA provides an additional historical view of recurring inspection issues.

For 2019–2022, EMA reported that the most common issues in CHMP-requested pharmacovigilance inspections included ICSR quality in EudraVigilance, including coding errors, insufficient information for a valid case and insufficient follow-up, as well as incomplete PSMF documentation. citeturn0search17

The report also describes other inspection triggers and examples of non-compliance, including concerns around PASS timelines and implementation of risk-minimisation measures such as registries and educational materials. citeturn0search17

This historical evidence complements the 2023 and 2024 PhV IWG reports and reinforces the importance of looking at recurring system themes rather than isolated observations.

21. Why the Same Finding Can Have Different Consequences

The regulatory significance of a deficiency depends on context.

Relevant factors can include:

• patient-safety impact;
• number of products affected;
• geographical scope;
• duration;
• recurrence;
• detectability;
• whether the issue is ongoing;
• whether the MAH knew or should have known about it;
• effectiveness of existing controls; and
• quality of the response once the issue was identified.

Consequently, two organisations can have superficially similar deficiencies that receive different regulatory attention because the surrounding risk is different.

22. Finding Themes Should Be Analysed Across Processes

A mature inspection-readiness programme should not examine findings only by department.

Consider an illustrative pattern:

Case-processing delay
        ↓
QC did not detect it
        ↓
KPI did not show the problem
        ↓
Vendor SLA did not escalate it
        ↓
Deviation was not opened
        ↓
QPPV did not receive escalation

The visible problem is a case-processing issue. The systemic problem may involve QMS, vendor oversight, performance monitoring and governance simultaneously.

This cross-functional perspective is central to the inspection series.

23. What Inspectors Can Learn From KPIs

Performance indicators can be useful inspection evidence, but only if they measure meaningful controls.

An inspector may ask:

• What does this KPI actually measure?
• What is the denominator?
• What is the target?
• How was the target selected?
• What happens when performance falls below target?
• Who reviews the result?
• Is there evidence of trend analysis?
• Have thresholds changed without justification?

A dashboard full of green indicators does not establish control if the indicators are poorly designed or exclude important failure modes.

24. Automation and Inspection Evidence

Increasing automation does not remove the need for inspection evidence.

The 2024 PhV IWG report records discussion of automation and AI in pharmacovigilance, alongside EudraVigilance/EVDAS, digital transformation and risk-based inspection planning. citeturn2view0

Where an automated process performs a safety-critical activity, the organisation should be able to explain:

• what the system does;
• what inputs it relies on;
• what controls detect failures;
• how changes are governed;
• how exceptions are handled; and
• what evidence demonstrates continued reliable operation.

Automation can reduce manual error while simultaneously creating new inspection questions around validation, change control, monitoring and accountability.

25. What a Good Pre-Inspection Self-Test Looks Like

A useful self-test is not simply a document checklist.

For each critical PV process, select a real transaction and reconstruct it from beginning to end.

For example:

ICSR: receipt → triage → validity → entry → medical review → coding → follow-up → QC → submission → acknowledgement → reconciliation.

Then ask:

• Can every transition be demonstrated?
• Are timestamps reliable?
• Are responsibilities clear?
• Did the system generate the expected controls?
• Were exceptions escalated?
• Can the organisation reproduce the evidence?

This is much closer to the way an inspector can test a system than simply confirming that SOPs exist.

26. The "Show Me" Principle

A useful internal inspection discipline is to replace statements with evidence requests.

Instead of:

"We have a robust follow-up process."

ask:

"Show me five recent cases requiring follow-up, the requests sent, responses received, escalation where necessary and the final assessment."

Instead of:

"Our vendor is monitored."

ask:

"Show me the vendor KPI trend, deviations, escalations, quality meetings, audit history and evidence of action when performance deteriorated."

This approach exposes gaps between policy and practice before an inspector does.

27. The Difference Between Readiness and Compliance

Inspection readiness is not a substitute for compliance.

An organisation can assemble an impressive inspection room, prepare presentations and retrieve documents quickly while still operating an ineffective system.

The stronger model is:

compliant system → reliable evidence → efficient inspection readiness.

Inspection preparation should therefore be treated as a verification exercise, not as a presentation exercise.

28. Turning an Inspection Finding Into a Risk Question

Once a deficiency is identified, the most useful next question is not simply "How do we fix the procedure?"

It is:

What risk did the failed control create, and what population could have been affected?

This leads to a structured assessment:

1. Identify the failed requirement or control.
2. Identify the period during which it may have failed.
3. Identify products, countries, sources or cases potentially affected.
4. Determine whether the failure was detected by another control.
5. Assess actual or potential patient-safety and regulatory impact.
6. Determine whether immediate containment is necessary.
7. Identify the true root cause.
8. Define corrective and preventive actions.

The scope assessment is particularly important. A CAPA that fixes today's visible error but leaves the affected historical population unidentified is unlikely to provide a complete response.

29. Root Cause: Avoiding the "Training Failure" Trap

Training is sometimes identified as the root cause simply because an individual performed an activity incorrectly.

That can be appropriate, but it should not be assumed.

If trained personnel repeatedly make the same error, possible underlying causes include:

• ambiguous instructions;
• poor system design;
• conflicting procedures;
• unrealistic workload;
• inadequate quality controls;
• unclear ownership;
• weak vendor interfaces; or
• an organisational process that makes the desired behaviour difficult.

A training CAPA may therefore be insufficient when the system itself creates the error.

30. CAPA Should Address the Cause, Not the Finding's Wording

Suppose an inspection identifies repeated delayed case submissions.

A weak response might be:

"Retrain case processors on reporting timelines."

A stronger analysis asks why the delays occurred.

If the root cause is an interface failure that prevents urgent cases from entering the correct workflow, retraining alone cannot prevent recurrence.

The final CAPA may need a combination of:

• immediate containment;
• system correction;
• process redesign;
• escalation controls;
• vendor remediation;
• training; and
• effectiveness monitoring.

31. Effectiveness Verification Is Part of the Remediation

An action should not be considered effective simply because it was completed.

For example:

Procedure revised ≠ process corrected

Training completed ≠ staff perform correctly

System patch deployed ≠ system performs reliably

Vendor warned ≠ vendor performance improved

The verification method should therefore correspond to the failed control.

If the problem was reporting timeliness, measure reporting timeliness. If it was missed follow-up, examine follow-up performance. If it was a documentation-control failure, test document accuracy and currency over time.

32. What Makes a CAPA Inspection-Ready?

A strong CAPA package allows an inspector to reconstruct:

Finding
  ↓
Risk assessment
  ↓
Scope assessment
  ↓
Root cause
  ↓
Immediate containment
  ↓
Corrective action
  ↓
Preventive action
  ↓
Implementation evidence
  ↓
Effectiveness verification
  ↓
Management approval / closure

The evidence should be internally consistent.

For example, if the root cause says that a vendor interface failed, the CAPA should not consist only of an internal training action. If the CAPA claims that the vendor interface was redesigned, the inspection package should contain the change evidence and subsequent performance data.

33. Inspection Follow-Up Is Itself a Controlled Process

The EU Union procedure on follow-up of pharmacovigilance inspections requires information on findings, proposed CAPA and recommendations to be communicated within a timeframe appropriate to the seriousness of the issue and, in relevant circumstances, before the inspection is closed. citeturn0search19

The procedure also recognises the value of coordination when national and supervisory-authority inspections identify related non-compliances, including the possibility of an integrated CAPA plan for the same or similar issues. citeturn0search19

This is important for large MAHs. A finding discovered at one affiliate or contractor should not automatically be treated as a local issue if the same global process operates elsewhere.

34. When Should a Finding Trigger Broader Review?

A broader review is particularly appropriate when the finding involves:

• a central safety database;
• a global SOP;
• a common vendor;
• a shared literature process;
• a global case-processing model;
• a common IT interface;
• a centrally controlled RMM;
• a global training programme; or
• a common quality-control mechanism.

The question is whether the finding identifies a local implementation problem or a global control failure.

35. Inspection Findings and the Member State Layer

EU pharmacovigilance systems operate across multiple Member States and organisational locations.

A process can therefore have several interfaces:

EU requirement → central PV → affiliate → local HCP/source → vendor → central database → regulatory submission.

The weakest interface may determine the performance of the overall process.

This is particularly relevant for activities involving local implementation, such as safety communications, risk-minimisation measures, local literature, national reporting arrangements and affiliate oversight.

The MAH should be able to distinguish legitimate national adaptation from uncontrolled process variation.

36. What the 2024 Findings Tell Us About Priorities

The 2024 human inspection data provide a practical prioritisation signal.

The leading finding areas were adverse-reaction management/reporting, QMS and PSMF. Within adverse-reaction management/reporting, the major findings included submission/follow-up, medical review and coding, literature screening and the EU collection point for ICSRs. citeturn2view0

This does not mean an MAH should prepare only those three areas. It means these areas deserve particularly careful end-to-end testing when building an inspection-readiness programme.

37. What the 2023 Findings Add

The 2023 report showed the same three leading areas: QMS, adverse-reaction management/reporting and PSMF. citeturn1search15

The value of the two-year comparison is that it supports a recurring-theme interpretation rather than a one-year anomaly.

The article series will therefore use these areas as anchors and then examine them separately in subsequent articles.

38. What the Historical Data Add

The 2019–2022 EMA report provides additional examples of recurring ICSR and PSMF weaknesses and notes inspection triggers involving data-quality problems, PASS timelines and risk-minimisation implementation. citeturn0search17

Taken together, the sources suggest that inspection risk is concentrated around a few broad principles:

• data must be complete and reliable;
• processes must operate within required timelines;
• responsibilities must be clear;
• quality systems must detect and correct failure;
• the PSMF must reflect reality; and
• the MAH must be able to demonstrate effective oversight.

39. A Practical Inspection-Readiness Matrix

40. What Inspectors May Notice When the System Is Mature

A mature system does not necessarily have zero deviations.

Instead, it can demonstrate that deviations are:

• detected;
• assessed;
• prioritised;
• escalated when appropriate;
• investigated;
• corrected;
• monitored for recurrence; and
• used to improve the system.

This is an important distinction. A regulator is not necessarily looking for an organisation that claims perfection. The stronger signal is a system that knows where it is weak and responds before weaknesses become significant compliance failures.

41. The QPPV Perspective

From a QPPV perspective, inspection readiness should answer five questions:

1. Where can the system fail?
2. How would we detect the failure?
3. How quickly would it reach the QPPV?
4. How would we determine its scope and risk?
5. How would we demonstrate that remediation worked?

This is more useful than maintaining a static list of documents to present during an inspection.

42. The Most Important Shift in Inspection Preparation

The most important shift is from document readiness to system evidence readiness.

Document readiness asks:

"Do we have the SOP?"

System evidence readiness asks:

"Can we demonstrate that the SOP-controlled process has operated correctly, consistently and effectively?"

That difference is central to the inspection series.

43. How This Series Will Use Inspection Evidence

The subsequent articles in this series will move from the recurring pattern identified here into specific inspection domains.

The planned progression is:

1. PSMF and pharmacovigilance-system governance;
2. ICSR and case processing;
3. signal management;
4. RMP and risk minimisation;
5. PSUR/PBRER and benefit-risk evaluation;
6. vendor, affiliate and outsourcing oversight;
7. QMS, CAPA, training and documentation;
8. inspection evidence, traceability and data integrity;
9. QPPV oversight, escalation and management governance;
10. cross-functional and interface findings; and
11. root-cause analysis and effective CAPA.

The purpose of splitting the series is to allow each domain to be examined using the same evidence-based framework rather than producing one long catalogue of deficiencies.

44. What Counts as a Strong Inspection Finding Example?

A strong published example should identify its evidentiary basis.

There are three useful categories:

Category A — Published regulatory finding

The finding is explicitly reported by an authoritative regulatory source, such as an EMA PhV IWG annual report or an official inspection outcome.

Category B — Regulatory expectation illustrated by a scenario

The regulatory requirement is established, but the scenario is constructed to explain how an inspector could test it. It must be labelled illustrative.

Category C — Practical QPPV hypothesis

The scenario is based on professional reasoning about how a system could fail. It should not be presented as evidence that regulators have actually identified that specific deficiency.

This distinction will be maintained throughout the series.

45. Why Published Findings Are More Valuable Than Generic Checklists

A checklist can tell a QPPV what to inspect internally.

A published finding can additionally show what regulators have actually considered important enough to record as a deficiency.

The best approach is to combine both:

published finding → regulatory requirement → operational control → inspection evidence → internal self-test.

This prevents inspection preparation from becoming either purely theoretical or overly dependent on historical anecdotes.

46. Limits of Inspection Statistics

Inspection statistics need careful interpretation.

The 2024 figures cited in this article concern CHMP-requested human pharmacovigilance inspections in the PhV IWG reporting population. They do not constitute a prevalence estimate for every pharmacovigilance inspection in the EU/EEA. National inspection programmes account for most EU/EEA pharmacovigilance inspections. citeturn1search2turn2view0

Likewise, the number of findings in a category depends on the number and scope of inspections, the sites selected and the inspection questions pursued.

The data should therefore be used to identify recurring inspection themes, not to calculate the probability that a particular MAH will receive a particular finding.

47. EU Inspection Procedures Provide the Regulatory Backbone

The practical interpretation of findings should remain anchored to the EU inspection framework.

EMA publishes Union procedures covering preparation, conduct and reporting of pharmacovigilance inspections, coordination, information sharing, record keeping and follow-up. These procedures support harmonisation and mutual recognition of pharmacovigilance inspections across the Member States. citeturn0search1

GVP Module III is the central GVP module for pharmacovigilance inspections. EMA's current GVP page identifies Module III as the applicable human pharmacovigilance inspection guidance. citeturn0search0

The practical implication is that inspection preparation should not be based solely on historical findings. It should start with the current regulatory framework and then use historical findings to understand where controls commonly fail.

48. The Inspection Finding as a System Signal

An inspection finding can be viewed at three levels:

Level 1 — The observed deficiency

Level 2 — The failed control

Level 3 — The system condition that allowed the failure

For example, a late report is the observed deficiency. The failed control may be ineffective triage or escalation. The system condition may be an inadequate workflow, unclear ownership or weak performance monitoring.

A mature response addresses all three levels.

49. Practical QPPV Checklist

Before an inspection, the QPPV should be able to answer:

• What are our highest-risk PV processes?
• Which processes have generated deviations or CAPAs recently?
• Which findings have occurred previously?
• Are any CAPAs overdue or not yet effectiveness-verified?
• Which vendors perform safety-critical activities?
• Where do global and local responsibilities intersect?
• Does the PSMF accurately describe the current system?
• Can a real ICSR be traced end-to-end?
• Can the organisation demonstrate literature-search performance?
• Can signal decisions be reconstructed?
• Can RMP/RMM implementation be demonstrated where relevant?
• Can PSUR/PBRER conclusions be traced to their evidence?
• Can management demonstrate oversight of significant PV risks?
• If a control failed yesterday, would we know today?
• If an inspector identified a failure today, could we determine its historical scope?

50. Key Takeaways

• EU pharmacovigilance inspections are system inspections, not merely document inspections.
• National competent authorities conduct inspections, while EMA coordinates and harmonises important elements of the EU inspection system. citeturn0search1turn0search5
• The 2024 PhV IWG report recorded 87 deficiencies in CHMP-requested human pharmacovigilance inspections, with management/reporting of adverse reactions, QMS and PSMF among the three most common finding areas. citeturn2view0
• The 2023 report identified the same three leading areas, supporting their treatment as recurring themes rather than a single-year anomaly. citeturn1search15
• Historical EMA reporting also identified ICSR quality and incomplete PSMF documentation among common inspection issues. citeturn0search17
• A process existing on paper does not prove that the control works.
• Inspectors can trace an activity from a requirement to an actual transaction and back through the quality system.
• A local deficiency may become systemic when the same control operates across products, countries, vendors or affiliates.
• CAPA effectiveness must be demonstrated against the failed control, not merely by completing an action.
• Published inspection findings and illustrative scenarios must be clearly distinguished.
• The strongest inspection-readiness model is compliant system → reliable evidence → efficient inspection readiness.

51. References

1. European Medicines Agency. Good pharmacovigilance practices (GVP) — current GVP modules, including Module III: Pharmacovigilance inspections. Current EMA GVP framework. urlEMA GVP guidanceturn0search0
2. European Medicines Agency. Pharmacovigilance inspection procedures: human. Union procedures on preparation, conduct, reporting, coordination, information sharing, record keeping and follow-up of pharmacovigilance inspections. urlEMA human pharmacovigilance inspection proceduresturn0search1
3. European Medicines Agency. Annual report of the Pharmacovigilance Inspectors' Working Group for 2024, EMA/INS/PhV/122716/2025. Findings and inspection activity for human and veterinary CAP programmes. urlEMA PhV IWG Annual Report 2024turn1search14
4. European Medicines Agency. Annual report of the Pharmacovigilance Inspectors' Working Group for 2023, EMA/INS/PhV/471069/2024. Human CAP inspection activity and categorisation of findings. urlEMA PhV IWG Annual Report 2023turn1search15
5. European Medicines Agency. Report on pharmacovigilance tasks from EU Member States and the European Medicines Agency (EMA), 2019–2022. Inspection activity and recurring findings. urlEMA pharmacovigilance tasks report 2019–2022turn0search17
6. European Medicines Agency. Union procedure on the follow-up of pharmacovigilance inspections, EMA/INS/PhV/327777/2018. Inspection finding follow-up, CAPA and coordination. urlEMA inspection follow-up procedureturn0search19
7. European Medicines Agency. Coordination of pharmacovigilance inspections. Role of EMA and the PhV Inspectors Working Group in EU inspection harmonisation. urlEMA coordination of pharmacovigilance inspectionsturn0search5
8. European Medicines Agency. Annual Report 2024 — Inspections and compliance. EU inspection programme context and the predominance of national pharmacovigilance inspection programmes. urlEMA Annual Report 2024 — inspections and complianceturn1search2

Regulatory Note

This article is an educational analysis of the EU pharmacovigilance inspection framework and published inspection evidence. Reported findings are attributed to the cited regulatory sources. Illustrative scenarios are explicitly identified as illustrative and must not be interpreted as published regulatory findings. Current applicable legislation, GVP guidance, Union procedures, national requirements and product-specific obligations should be checked before using this material for operational or regulatory decisions.