Article 31 Pharmacovigilance Referral: Legal Basis, Procedure and Regulatory Consequences in the EU

Introduction

An Article 31 pharmacovigilance referral is one of the mechanisms through which a pharmacovigilance concern can be brought into a formal Union-level regulatory procedure. It is important precisely because the procedure sits at the boundary between pharmacovigilance assessment and regulatory decision-making: safety information is evaluated scientifically, a regulatory recommendation or position is developed through the applicable European committees, and the resulting decision is implemented through the legal framework applicable to the affected marketing authorisations.

The procedure should not be understood simply as an "EMA safety review". Its legal basis, initiating conditions, committee responsibilities and consequences are defined by EU medicines legislation. The identity of the initiating authority, the nature of the pharmacovigilance information, the products within scope and their authorisation routes all matter to understanding what happens next.

The most useful starting distinction is between the legal trigger and the scientific subject matter. Article 31 is a legal referral mechanism. A safety concern may be the subject of the procedure, but the fact that the issue concerns safety does not by itself explain the legal route. The reader must first establish that the procedure is an Article 31 referral arising from the evaluation of pharmacovigilance data and that the statutory conditions for Union-level action are satisfied.

A second distinction is equally important: the PRAC recommendation is not the European Commission decision. PRAC performs the central pharmacovigilance assessment, but the subsequent regulatory route depends on the authorisation status of the affected products and on the applicable provisions of the legislation.

This article therefore follows the procedure as a regulatory sequence rather than as a collection of committee descriptions.

1. What Is an Article 31 Pharmacovigilance Referral?

An Article 31 pharmacovigilance referral is a Union-level procedure under Article 31 of Directive 2001/83/EC that arises from the evaluation of pharmacovigilance data concerning authorised medicinal products where the interests of the Union are involved.

The formulation matters. Article 31 is not itself synonymous with pharmacovigilance. Article 31 can also be used for a non-pharmacovigilance referral concerning matters such as quality or efficacy. The pharmacovigilance pathway is distinguished by the source and regulatory character of the information giving rise to the procedure.

For a pharmacovigilance Article 31 referral, PRAC is the principal scientific committee responsible for the pharmacovigilance assessment. The procedure may subsequently involve CMDh, CHMP and the European Commission depending on the products affected and the applicable legal route.

A useful conceptual model is:

Pharmacovigilance information
          ↓
Safety concern of Union significance
          ↓
Article 31 referral
          ↓
PRAC assessment
          ↓
PRAC recommendation
          ↓
CMDh and/or CHMP pathway
          ↓
European regulatory decision
          ↓
Implementation and continued monitoring

The exact sequence should always be confirmed against the initiating legal provision and the referral-specific documentation.

2. The Legal Basis: Article 31 of Directive 2001/83/EC

The principal legal basis is Article 31 of Directive 2001/83/EC. The general referral provisions in Articles 32 to 34 operate alongside the pharmacovigilance-specific provisions of the Directive.

The legal framework is therefore layered rather than contained in a single paragraph. Article 31 establishes the referral mechanism and its Union-level context; the subsequent provisions establish procedural elements and consequences; and the pharmacovigilance provisions govern the specific safety-related route.

This is important when reading guidance or an individual referral notice. Operational guidance may describe the procedure in practical terms, but the legal effect of the procedure derives from the legislation.

For regulatory interpretation, the correct hierarchy is:

1. identify the applicable provision of Directive 2001/83/EC;
2. identify the procedural guidance applicable to that provision;
3. identify the specific referral notice and questions;
4. follow the committee documents and final decision.

An EMA question-and-answer document can explain how the procedure operates in practice, but it should not be treated as replacing the statutory text.

3. The Meaning of "Interests of the Union"

The expression "interests of the Union" is central to the Article 31 mechanism.

It identifies a regulatory situation in which the issue is sufficiently important to warrant assessment through a Union-level procedure rather than being left solely to separate national regulatory processes.

In a pharmacovigilance referral, the issue arises from the evaluation of safety information concerning an authorised medicinal product or products. The Union-interest element provides the legal rationale for bringing that issue into the common European regulatory framework.

It should not, however, be reduced to a simple numerical threshold such as a particular number of cases, a particular reporting rate or a particular disproportionality value. The existence of a serious safety concern and the legal assessment of Union interest are related but distinct questions.

A sound regulatory reading therefore asks two separate questions:

• What does the pharmacovigilance evidence show?
• Why does the issue require assessment through the Article 31 Union procedure?

Keeping those questions separate prevents scientific evidence from being mistaken for the legal trigger itself.

4. Why Article 31 Pharmacovigilance Referrals Exist

The EU pharmacovigilance system operates across multiple Member States, multiple marketing authorisation holders and multiple authorisation routes. A safety issue may therefore extend beyond a single national authorisation or a single company.

The same active substance may be marketed through:

• nationally authorised products;
• mutual recognition procedures;
• decentralised procedures;
• centrally authorised products;
• multiple marketing authorisation holders.

A major safety concern can consequently produce a fragmented regulatory picture if every authorisation is assessed independently.

The Article 31 pharmacovigilance mechanism provides a route for a qualifying issue to be considered at Union level. Its purpose is not merely to accelerate an existing national procedure. It creates a common regulatory assessment in circumstances where the legislation provides for Union involvement.

This distinction is important because the procedure is part of the architecture of EU medicines regulation. It is not simply an administrative escalation of ordinary signal management.

5. Who Can Initiate an Article 31 Pharmacovigilance Referral?

The applicable EMA procedural guidance identifies three potential initiators for an Article 31 pharmacovigilance referral:

• a Member State competent authority;
• the European Commission;
• a marketing authorisation holder.

The initiating mechanism should be distinguished from the subsequent scientific assessment. The initiator brings the issue into the referral framework; PRAC then performs the pharmacovigilance assessment within the procedure defined by the legislation.

The referral notification identifies the safety concern and the questions to be assessed. It also explains why the matter falls within the Union-level mechanism.

The role of an MAH requires particular care in interpretation. An MAH may seek an Article 31 pharmacovigilance referral, but the practical process for establishing the Union interest and defining the question referred to PRAC follows the applicable regulatory framework. An MAH should therefore not be described as having unilateral control over the scope or legal initiation of the procedure.

The initiating authority and the scientific assessor perform different functions, and those functions should remain distinct when reconstructing the regulatory history.

6. The Referral Notification

The referral notification is the document that formally establishes the subject matter of the procedure.

From a regulatory reading perspective, it should answer at least four questions:

1. What medicinal product or products are concerned?
2. What safety issue is being referred?
3. What questions are being submitted for assessment?
4. Why is the matter being addressed through the Union procedure?

The notification should be read before the later scientific documents because it defines the scope of the questions that PRAC is being asked to address.

This distinction becomes important when the final recommendation contains a broader discussion of the product's safety profile. The existence of additional scientific discussion does not necessarily mean that every issue discussed became a formal referral question.

The notification is therefore the starting point for reconstructing the procedural scope.

7. Which Medicinal Products Can Be Within Scope?

An Article 31 pharmacovigilance referral can encompass affected medicinal products with valid marketing authorisations in the European Economic Area, subject to the scope established by the referral and the applicable legislation.

The products within scope can include nationally authorised and centrally authorised medicines. They can also include products marketed by different marketing authorisation holders.

This breadth is important for pharmacovigilance because the same safety issue may arise from evidence concerning one product but have implications for other products containing the same active substance or otherwise affected by the same risk.

The scope should therefore be established from the actual referral documentation rather than inferred solely from the product that first generated the safety concern.

For each referral, the reader should establish:

• active substance or other relevant product identifier;
• pharmaceutical forms and strengths within scope;
• marketing authorisation holders affected;
• authorisation route;
• Member States involved where relevant;
• any explicit exclusions or limitations.

The scope is a regulatory fact, not an assumption based on pharmacological similarity alone.

8. Article 31 Pharmacovigilance Versus Article 31 Non-Pharmacovigilance

Article 31 should not be treated as a single undifferentiated referral category.

The distinction between pharmacovigilance and non-pharmacovigilance referrals is fundamental because it determines the scientific committee and procedural pathway.

The practical consequence is significant. If a document is described simply as an "Article 31 referral", the reader should not immediately assume that PRAC is involved. The initiating basis must first be established.

This distinction also prevents terminology from contaminating interpretation of the other EU referral mechanisms discussed elsewhere in this series.

9. Article 31 Is a Regulatory Procedure, Not a Signal Category

An Article 31 referral should not be treated as a pharmacovigilance signal category.

A signal is an item of safety information that warrants evaluation within the pharmacovigilance system. An Article 31 referral is a statutory regulatory procedure.

The two can be connected in practice:

Safety information
      ↓
Signal detection or other PV evaluation
      ↓
Scientific assessment
      ↓
Significant regulatory concern
      ↓
Article 31 referral, where statutory conditions are met

But the existence of a signal does not automatically create an Article 31 referral.

Conversely, identifying an Article 31 referral does not by itself establish that a particular signal has been validated, that causality has been proven, or that the benefit-risk balance is negative.

The regulatory procedure and the internal pharmacovigilance classification should therefore remain conceptually separate.

10. The Role of PRAC in an Article 31 Pharmacovigilance Referral

For a pharmacovigilance Article 31 referral, the Pharmacovigilance Risk Assessment Committee (PRAC) is the principal scientific committee responsible for the safety assessment within the Agency's procedure.

That statement should be understood precisely. PRAC does not simply "approve" or "reject" a medicine. It assesses the pharmacovigilance question referred to it, evaluates the available evidence, considers the implications for the benefit-risk balance and adopts the recommendation provided for by the applicable legal framework.

The evidence considered can vary substantially between referrals. Depending on the question, the assessment may draw on spontaneous reports, epidemiological evidence, clinical studies, post-authorisation safety studies, literature, EudraVigilance data, regulatory information and other relevant sources.

The important regulatory point is that the evidence base is referral-specific. It should not be assumed that every Article 31 procedure follows an identical scientific assessment model.

PRAC's role can therefore be expressed as:

Referral question
       ↓
Evidence relevant to that question
       ↓
PRAC scientific assessment
       ↓
Benefit-risk evaluation
       ↓
PRAC recommendation

The recommendation is the bridge between the pharmacovigilance assessment and the subsequent regulatory stage.

11. What PRAC Is Actually Assessing

The central question in an Article 31 pharmacovigilance procedure is not merely whether an adverse event has been reported.

The assessment may require PRAC to determine, depending on the referral:

• whether an association between the medicinal product and the safety concern is supported by the evidence;
• the seriousness and clinical relevance of the risk;
• the frequency or magnitude of the risk where it can be estimated;
• which patient populations are affected;
• whether particular risk factors modify the risk;
• whether the risk is already adequately described;
• whether existing risk-minimisation measures are sufficient;
• whether additional measures could adequately control the risk;
• and whether the overall benefit-risk balance remains favourable under the proposed conditions of use.

The exact questions are determined by the referral.

This is why the questions referred to PRAC deserve careful attention. They provide the legal and procedural boundary within which the scientific assessment is conducted.

12. PRAC Assessment Is Not the Same as Routine Signal Management

A common source of confusion is treating an Article 31 procedure as simply a more formal version of signal management.

The two activities are related but different.

Routine pharmacovigilance may identify and assess signals throughout the product lifecycle. Many signals are handled without any Union referral. They may lead to routine monitoring, additional analysis, a variation, an update to the risk management plan or other pharmacovigilance action.

An Article 31 referral is different because a qualifying matter has entered a statutory Union-level regulatory procedure.

The distinction can be expressed as:

The existence of an Article 31 referral therefore tells the reader something about the regulatory status of the issue, not simply about the existence of a pharmacovigilance signal.

13. The Questions Referred to PRAC Matter

The referral questions are not a procedural formality.

They determine what PRAC has been asked to assess and provide the framework for interpreting the recommendation.

For example, a referral may ask PRAC to consider whether a particular adverse reaction requires changes to product information. Another may require assessment of whether the benefit-risk balance remains favourable for a defined patient population. Another may concern whether existing risk-minimisation measures adequately control an identified risk.

A reader should therefore avoid jumping directly from the referral title to the final regulatory conclusion.

A better reading sequence is:

1. read the referral notification;
2. identify the formal questions;
3. identify the evidence relevant to those questions;
4. read PRAC's reasoning;
5. identify the recommendation;
6. follow the subsequent regulatory pathway.

This prevents a broad safety discussion from being mistaken for the precise issue on which the committee was required to act.

14. Evidence in the PRAC Assessment

The evidence base in an Article 31 pharmacovigilance procedure should be read as a body of evidence rather than as a list of individual reports.

Depending on the case, the assessment can involve different evidence streams, such as:

• individual case safety reports;
• aggregate pharmacovigilance analyses;
• epidemiological studies;
• observational databases;
• clinical trials;
• post-authorisation studies;
• scientific literature;
• mechanistic or pharmacological information;
• exposure data;
• information from other regulatory authorities.

Different evidence types answer different questions.

For example, spontaneous reports can be important for detecting a potential association, while epidemiological studies may contribute more directly to estimation of relative or absolute risk. Clinical information may help characterise severity, risk factors or clinical management.

The regulatory conclusion therefore depends on the integrated assessment rather than on one pharmacovigilance metric.

15. Benefit-Risk Balance in the Referral

The regulatory endpoint of a pharmacovigilance assessment is not normally a binary question of whether a risk exists.

Medicines commonly have risks. The relevant regulatory question is whether the benefits continue to outweigh the risks under the conditions of use, taking account of the evidence and possible risk-control measures.

The assessment may therefore consider:

• magnitude of benefit;
• seriousness of the disease;
• availability of alternatives;
• nature and severity of the identified risk;
• incidence or estimated frequency;
• preventability or detectability of the risk;
• affected population;
• reversibility of harm;
• effectiveness of existing or proposed risk-minimisation measures.

A referral can consequently result in a range of outcomes. The existence of a significant safety concern does not logically imply suspension or revocation.

16. PRAC Recommendation Versus Final Regulatory Decision

One of the most important distinctions in the entire procedure is the difference between the PRAC recommendation and the final regulatory decision.

PRAC performs the pharmacovigilance assessment and adopts its recommendation. That recommendation then enters the applicable legal pathway.

For nationally authorised products, CMDh has a defined role after PRAC.

For centrally authorised products, CHMP considers the PRAC recommendation.

Where the applicable procedure leads to a European Commission decision, the Commission's decision is the legally binding regulatory endpoint.

The conceptual sequence is therefore:

PRAC recommendation
       ↓
Applicable committee / regulatory pathway
       ↓
Position or opinion
       ↓
European Commission, where applicable
       ↓
Binding regulatory outcome

It is therefore inaccurate to describe the PRAC recommendation itself as the final European Union decision.

17. Nationally Authorised and Centrally Authorised Products

The authorisation status of the affected products is one of the most important procedural variables.

A pharmacovigilance referral can involve:

1. nationally authorised products;
2. centrally authorised products;
3. both categories.

This matters because the post-PRAC route differs.

For nationally authorised products, CMDh is the relevant coordination body after the PRAC recommendation.

For centrally authorised products, CHMP considers the recommendation.

Where both categories are within scope, the procedure must be understood as containing corresponding regulatory pathways for the different authorisation categories.

The authorisation route should therefore be established early in any regulatory assessment.

18. The CMDh Role After PRAC

For nationally authorised medicinal products, the PRAC recommendation enters the CMDh process.

CMDh is the Coordination Group for Mutual Recognition and Decentralised Procedures – Human. Its role in this context is not to repeat the pharmacovigilance assessment performed by PRAC from first principles. Rather, it considers the PRAC recommendation within the regulatory framework applicable to nationally authorised products.

This distinction is important because PRAC and CMDh have different institutional functions.

A simplified model is:

PRAC
Scientific pharmacovigilance assessment
             ↓
CMDh
Regulatory coordination for NAPs

The precise legal consequences depend on the outcome reached by CMDh and the applicable provisions of the Directive.

19. CMDh Consensus and Majority Outcomes

For nationally authorised products, the CMDh outcome can depend on whether agreement is reached.

Where CMDh reaches consensus, the resulting position is implemented through the applicable national regulatory mechanisms and timetable.

Where consensus cannot be reached and the applicable majority procedure is used, the matter can proceed to the European Commission decision-making route provided by the legislation.

The distinction is therefore more than a voting detail. It can determine the subsequent institutional pathway.

The simplified model is:

PRAC recommendation
        ↓
      CMDh
     /     \
consensus  majority
   ↓           ↓
implementation  Commission pathway

The actual legal consequences should always be checked against the current legislation and the individual referral documents.

20. The CHMP Role After PRAC

For centrally authorised products, the PRAC recommendation is considered by CHMP.

CHMP is the Committee for Medicinal Products for Human Use and is responsible for the scientific opinion applicable to centrally authorised medicines within its remit.

In this context, CHMP considers the pharmacovigilance recommendation and adopts the opinion required by the applicable referral procedure.

The distinction between PRAC and CHMP should be maintained:

• PRAC: pharmacovigilance risk assessment;
• CHMP: medicinal-product scientific opinion within the central authorisation framework.

The CHMP opinion is then part of the material leading to the applicable European Commission decision.

21. When Both CMDh and CHMP Are Relevant

A referral can affect both nationally and centrally authorised products.

In that situation, it is misleading to describe the procedure as simply "PRAC to CHMP" or simply "PRAC to CMDh".

The regulatory consequences must be followed for each authorisation category.

A useful model is:

                  PRAC
                   ↓
          PRAC recommendation
             /           \
            /             \
     National products   Central products
            ↓                   ↓
          CMDh                CHMP
            ↓                   ↓
     Applicable position   CHMP opinion
            \                   /
             \                 /
              ↓               ↓
          Union regulatory outcome

The individual referral documents determine exactly how the branches are coordinated and what legal instrument follows.

22. What the PRAC Recommendation Can Contain

The recommendation can address the regulatory measures considered appropriate in light of the pharmacovigilance assessment.

Depending on the case, this can include recommendations concerning:

• maintenance of the existing authorisation;
• variation of product information;
• restrictions on use;
• contraindications;
• warnings and precautions;
• additional risk-minimisation measures;
• monitoring requirements;
• suspension;
• revocation.

These are examples of possible regulatory consequences, not a predetermined checklist applied to every Article 31 referral.

The recommendation should therefore be read in its full procedural context.

23. Why a Referral Does Not Mean Withdrawal

The announcement of an Article 31 pharmacovigilance referral is sometimes interpreted as an announcement that a medicine is going to be withdrawn.

That is incorrect.

The referral establishes a formal assessment of the issue. The purpose of the assessment is to determine the appropriate regulatory response.

Possible conclusions can include:

• no change;
• targeted product-information changes;
• additional risk-minimisation measures;
• restrictions in a particular population;
• suspension;
• revocation.

The existence of the procedure therefore tells the reader that a Union-level regulatory question is being considered. It does not, by itself, answer that question.

24. Product Information as a Regulatory Outcome

Where the evidence supports continued use with additional precautions, the regulatory outcome may include changes to product information.

Depending on the issue, changes can concern:

• therapeutic indication;
• contraindications;
• warnings and precautions;
• adverse-reaction information;
• dosing or administration;
• monitoring;
• special populations.

The exact wording and implementation route depend on the authorisation category and final regulatory instrument.

For a QPPV or regulatory-affairs professional, the practical task is therefore not simply to identify that "product information changed" but to determine:

1. what changed;
2. why it changed;
3. which products are affected;
4. what implementation action is required;
5. what pharmacovigilance or risk-management activities must follow.

25. Risk-Minimisation Measures

A referral can result in additional or strengthened risk-minimisation measures when the identified risk can be controlled without withdrawing the medicinal product.

The measure should be linked to the risk it is intended to control.

Examples can include:

• targeted warnings;
• contraindications;
• monitoring requirements;
• educational materials;
• controlled-use conditions;
• communication to healthcare professionals.

The existence of a risk-minimisation measure does not itself demonstrate that the overall benefit-risk balance is negative. In many cases, the purpose of risk minimisation is precisely to allow continued use while reducing a clinically important risk.

26. Urgent Protective Measures During the Procedure

The existence of a Union referral does not necessarily prevent a Member State from taking urgent protective action where the applicable legal conditions are met.

This reflects an important feature of the EU medicines system: immediate public-health protection and Union-level scientific assessment are not necessarily mutually exclusive.

Where urgent action is taken, the applicable notification and procedural requirements under the legislation must be followed.

For regulatory readers, the key distinction is between:

• an urgent national protective measure taken during the procedure; and
• the final Union regulatory outcome reached through the referral mechanism.

The first should not automatically be treated as evidence that the second has already been determined.

27. Re-Examination of a CHMP Opinion

Where the applicable Article 31 procedure results in a CHMP opinion, the relevant applicant or marketing authorisation holder may have a right to request re-examination under the applicable legal framework.

The procedural deadlines are critical.

EMA guidance describes a short period for notifying the Agency of the intention to request re-examination, followed by the period for submitting the detailed grounds. The exact deadlines and procedural requirements should be checked against the current applicable legislation and the referral-specific EMA documentation rather than copied from a generic summary.

This is an important example of why procedural articles should distinguish general framework from case-specific timetable.

28. Re-Examination Is a Defined Scientific Procedure

Re-examination should not be described loosely as an appeal against the Commission.

It is a mechanism within the committee procedure through which the relevant scientific opinion can be reconsidered on the grounds submitted by the applicant or MAH.

The grounds matter because they define what the committee is being asked to reconsider.

A careful regulatory reconstruction should therefore identify:

• the original opinion;
• the notification of intention to request re-examination;
• the detailed grounds;
• the committee's re-examination assessment;
• the final opinion after re-examination;
• the subsequent Commission decision, where applicable.

These are separate procedural events and should not be collapsed into one generic "appeal" stage.

29. The European Commission Decision

Where the applicable procedure leads to a Commission decision, the Commission's decision is the legally binding Union regulatory endpoint.

This distinction is especially important for centrally authorised products. The scientific committee opinion provides the scientific conclusion within the medicines regulatory system; the Commission decision gives effect to the applicable Union regulatory measure.

For nationally authorised products, the relationship between the Commission decision and national implementation must also be understood in light of the legal route applicable to the referral.

The practical sequence is therefore:

Scientific assessment
       ↓
Committee recommendation / opinion
       ↓
Commission decision where required
       ↓
Implementation through the applicable authorisation system

The final decision should always be read rather than inferred from the PRAC recommendation.

30. Regulatory Implementation Is Part of the Procedure's Consequence

A referral does not end conceptually when the Commission decision is published.

The regulatory outcome must be translated into operational action.

Depending on the outcome, implementation may involve:

• variation of marketing authorisations;
• changes to SmPC and package leaflet;
• changes to labelling;
• risk-management-plan updates;
• additional risk-minimisation measures;
• healthcare-professional communication;
• national implementation for nationally authorised products;
• internal pharmacovigilance monitoring.

The implementation requirements should be derived from the final regulatory instrument and applicable procedural documents.

This is particularly important for a QPPV because the regulatory decision can create new pharmacovigilance obligations even though the scientific assessment itself has concluded.

31. The Procedure Does Not End Pharmacovigilance

A completed Article 31 referral does not terminate pharmacovigilance surveillance.

The post-referral period can generate additional information about whether the regulatory measures are effective and whether the residual risk remains acceptable.

The resulting cycle is:

Safety evidence
      ↓
Article 31 assessment
      ↓
Regulatory action
      ↓
Implementation
      ↓
Continued pharmacovigilance
      ↓
New evidence if available

This is one reason why a referral should be understood as part of the product lifecycle rather than as an isolated regulatory event.

32. How to Read an Article 31 Procedure Correctly

For practical regulatory work, the most reliable reading method is to reconstruct the procedure chronologically and legally.

Start with the initiating document and identify:

1. the legal basis;
2. the initiating authority;
3. the medicinal products within scope;
4. the precise referral questions;
5. the evidence considered;
6. the PRAC recommendation;
7. the authorisation routes affected;
8. the subsequent CMDh or CHMP step;
9. any re-examination;
10. the final Commission decision;
11. the implementation requirements.

This approach prevents a common error: treating a later committee document as though it were the document that initiated the procedure.

The regulatory history is a sequence, and each document has a defined place within that sequence.

33. How the Referral Is Initiated in Practice

The legal mechanism and the practical initiation of an Article 31 pharmacovigilance referral should be distinguished.

A safety concern may be identified through ordinary pharmacovigilance activity, a regulatory review, new epidemiological evidence, a safety communication, or another source. Identification of the concern is not itself the formal referral.

The formal procedure begins when the competent authority or other legally entitled initiator invokes the Article 31 mechanism and the required notification is transmitted through the applicable regulatory process.

For a regulatory professional, this means that the timeline should be reconstructed from the formal referral event, not from the date on which the underlying safety signal was first detected.

A useful chronology is:

Safety concern identified
        ↓
Internal / national regulatory assessment
        ↓
Decision that Union referral is appropriate
        ↓
Formal Article 31 notification
        ↓
EMA / committee procedure

The dates of these events may differ substantially. Treating them as one date can produce an inaccurate regulatory history.

34. The Referral Notice as the Procedural Anchor

Once the referral is formally initiated, the referral notice becomes the central document for understanding the procedure.

It normally identifies the medicinal products concerned, describes the safety issue and sets out the questions to be considered. It may also specify the legal basis and procedural context.

The notice should be preserved as part of the regulatory record because later documents should be interpreted against it.

For an MAH, a useful working practice is to create a referral control record containing:

This is an operational control, not a substitute for the official regulatory documents.

35. The Importance of the Referral Scope

A referral can have a broader or narrower scope than the initial safety concern that triggered it.

For example, an issue may first be identified in one product or one Member State but subsequently be considered in relation to multiple medicinal products or authorisation holders.

The scope should therefore be determined from the referral notice and subsequent procedural documents.

A useful distinction is:

• triggering evidence — information that brought the concern to regulatory attention;
• referral scope — products and questions formally placed into the Union procedure;
• assessment evidence — information considered by PRAC in answering those questions;
• regulatory outcome — measures ultimately applied to the products within scope.

These four concepts should not be conflated.

36. The Marketing Authorisation Holder's Position During the Referral

The MAH is an important participant in the procedure, but it is not the decision-maker.

The MAH's responsibilities during a pharmacovigilance referral can include providing relevant safety information, responding to questions, reviewing the scientific assessment, preparing submissions and implementing the eventual regulatory outcome.

The response should be evidence-based and aligned with the actual referral questions.

A common mistake is to prepare a general safety narrative when the regulatory question is narrower. A stronger response maps each substantive conclusion to the question being assessed and identifies the evidence supporting that conclusion.

For internal governance, the MAH should maintain a clear distinction between:

1. the company's scientific assessment;
2. the company's regulatory position;
3. the committee's independent assessment;
4. the final regulatory decision.

These are not interchangeable.

37. Preparing the Pharmacovigilance Evidence Package

The evidence package should allow the committee to understand both the safety concern and the context in which it occurs.

Depending on the referral, this can include:

• case-series analysis;
• cumulative review of individual cases;
• exposure estimates;
• observed and expected event patterns;
• epidemiological analyses;
• clinical-trial data;
• literature review;
• biological plausibility;
• risk factors;
• dose-response information where relevant;
• dechallenge and rechallenge information where relevant;
• alternative explanations;
• background incidence;
• existing risk-minimisation measures.

The appropriate package is referral-specific. More data is not necessarily better if it does not answer the regulatory question.

The key quality criterion is traceability: the conclusion should be possible to trace back to the underlying evidence.

38. Case-Series Review in a Referral

Where individual case safety reports are central to the issue, a structured case-series review can be particularly important.

The assessment should distinguish between:

• case validity;
• clinical description;
• temporal relationship;
• alternative causes;
• concomitant medicines;
• dechallenge and rechallenge;
• outcome;
• seriousness;
• relevant risk factors;
• consistency across cases.

A simple count of cases is rarely sufficient to establish the regulatory significance of a safety concern.

The committee may need to understand whether the cases represent a coherent clinical pattern and whether the observed pattern is compatible with the known pharmacology and epidemiology of the event.

39. Epidemiological Evidence

Epidemiological evidence can materially change the interpretation of a pharmacovigilance concern.

Where an appropriate study is available, the assessment may consider measures such as relative risk, rate ratio, odds ratio, hazard ratio or absolute risk, depending on the study design.

Interpretation requires attention to:

• study population;
• exposure definition;
• outcome definition;
• comparator;
• confounding;
• selection bias;
• information bias;
• latency;
• duration of follow-up;
• statistical uncertainty;
• clinical relevance.

A statistically significant association is not automatically equivalent to a causal relationship, and the absence of statistical significance does not necessarily exclude a clinically important risk.

The regulatory interpretation depends on the totality of evidence.

40. Causality and Regulatory Decision-Making

Causality is one component of the assessment, but regulatory action does not always require mathematical certainty about causation.

The relevant question can be whether the evidence supports a sufficient level of concern to justify a change in the conditions of use or other risk-control measures.

This is particularly important where the event is serious and the available evidence is necessarily incomplete.

A regulatory assessment should therefore avoid two extremes:

• treating every temporal association as proof of causality;
• requiring absolute proof before any protective measure can be considered.

The appropriate evidentiary threshold depends on the legal and scientific context of the referral and the nature of the decision being considered.

41. Signal Strength and Clinical Significance Are Different

The statistical or epidemiological strength of a signal should not be confused with its clinical importance.

A rare but catastrophic event may warrant regulatory attention even if statistical power is limited. Conversely, a common event with a weak clinical consequence may produce a large number of reports without requiring major regulatory intervention.

The assessment should therefore consider both:

• evidence strength; and
• clinical and public-health significance.

This distinction is particularly useful when explaining why a referral can lead to action despite uncertainty in the estimated incidence.

42. Risk Factors and Vulnerable Populations

A referral may identify that the risk is concentrated in a particular population.

Relevant factors can include:

• age;
• renal or hepatic impairment;
• pregnancy;
• comorbidities;
• genetic characteristics;
• concomitant medicines;
• treatment duration;
• dose;
• previous exposure;
• other clinical characteristics.

The regulatory consequence may then be targeted rather than product-wide.

For example, a risk may be manageable through a contraindication in a defined population, enhanced monitoring, dose modification or a warning directed at prescribers.

The final measure should correspond to the population in which the evidence demonstrates a clinically meaningful risk.

43. Existing Risk-Minimisation Measures Must Be Evaluated

An important part of the referral assessment is determining whether existing measures already control the identified risk.

The question is not simply whether a warning exists. The committee may need to consider whether the measure is:

• scientifically appropriate;
• sufficiently prominent;
• understood by healthcare professionals;
• capable of changing clinical behaviour;
• supported by monitoring;
• effective in reducing the risk.

Where existing measures are inadequate, the referral may lead to additional risk minimisation.

Where evidence demonstrates that the risk is not adequately controllable, more restrictive regulatory action may be considered.

44. Risk-Minimisation Effectiveness

Risk-minimisation measures should be considered in terms of their intended effect, not merely their existence.

A warning that does not change clinical behaviour may have limited regulatory value. Similarly, an educational programme should not be treated as effective simply because it was distributed.

Where effectiveness data are available, they can form part of the assessment.

Relevant evidence may include:

• prescribing behaviour;
• monitoring compliance;
• healthcare-professional knowledge;
• patient behaviour;
• incidence before and after intervention;
• utilisation patterns;
• additional pharmacovigilance data.

This provides a bridge between the regulatory decision and the post-referral pharmacovigilance system.

45. The Role of the Risk Management Plan

The risk management plan is often relevant to implementation of the outcome of a pharmacovigilance referral.

Where the referral identifies an important risk or introduces new risk-minimisation measures, the RMP may need to be updated in accordance with the applicable regulatory requirements.

The update can affect:

• safety concerns;
• pharmacovigilance activities;
• additional pharmacovigilance activities;
• routine risk-minimisation measures;
• additional risk-minimisation measures;
• planned effectiveness evaluation.

The RMP should therefore be treated as part of the implementation system rather than as a document that is independent of the referral outcome.

46. Pharmacovigilance System Impact

A referral outcome can affect the operational pharmacovigilance system.

Depending on the decision, the MAH may need to modify:

• case-processing instructions;
• medical review criteria;
• signal detection specifications;
• targeted surveillance;
• aggregate reporting content;
• literature surveillance;
• risk-management activities;
• safety communication processes.

The exact changes depend on the final regulatory requirements.

A QPPV should therefore ask not only "What did the regulator decide?" but also:

What must change in the pharmacovigilance system because of that decision?

That question turns the regulatory outcome into an operational control.

47. Interaction With Signal Management After the Referral

The referral does not eliminate the need for routine signal management.

The safety issue remains part of the product's evolving safety profile. New evidence can modify the understanding of the risk and can reveal whether the regulatory measures remain appropriate.

Post-referral signal management should therefore distinguish between:

• the original referred issue;
• new evidence concerning that issue;
• unrelated new signals;
• evidence about the effectiveness of risk minimisation.

This distinction prevents every subsequent safety observation from being treated as though it were part of the original referral.

48. Aggregate Reports After an Article 31 Referral

The outcome of an Article 31 procedure can become relevant to subsequent aggregate safety reporting.

Depending on the product and reporting framework, the MAH may need to discuss the referral, newly characterised risk, regulatory measures or emerging evidence in relevant periodic safety evaluations.

The exact reporting treatment depends on the applicable product category, reporting obligation and timing.

The important principle is that the referral becomes part of the product's regulatory history and should be considered when evaluating subsequent safety information.

49. Documentation and Audit Trail

An Article 31 referral creates a substantial regulatory record.

For an MAH, the internal record should normally allow an inspector or auditor to reconstruct:

1. when the company became aware of the referral;
2. what the referral questions were;
3. who was assigned responsibility;
4. what evidence was reviewed;
5. what position the company adopted;
6. what submissions were made;
7. what regulatory conclusions were received;
8. what implementation actions were taken;
9. how effectiveness was monitored.

The objective is not simply document retention. It is decision traceability.

50. QPPV Oversight of an Article 31 Referral

The QPPV's role is primarily one of pharmacovigilance system oversight and assurance rather than acting as the company's sole regulatory decision-maker.

Depending on the organisation and referral, QPPV oversight can include ensuring that:

• relevant safety information is identified and escalated;
• the company's pharmacovigilance response is appropriately governed;
• submissions are medically and scientifically supported;
• regulatory commitments are incorporated into the safety system;
• risk-management activities are updated where required;
• implementation is tracked;
• post-referral monitoring is maintained.

The exact allocation of responsibilities should follow the company's governance model, SOPs and delegation arrangements.

The QPPV should nevertheless be able to explain how the referral affected the pharmacovigilance system and how the organisation assured compliance with the resulting requirements.

51. Common Misreadings of Article 31 Pharmacovigilance Referrals

Several recurring errors make referral documents harder to interpret.

Misreading 1: "Article 31 means PRAC"

Not necessarily. The legal basis must first be established, including whether the referral is pharmacovigilance-related.

Misreading 2: "PRAC recommendation is the final decision"

It is not. The applicable regulatory pathway must be followed to the legally effective outcome.

Misreading 3: "A referral means withdrawal"

A referral is an assessment procedure, not a predetermined outcome.

Misreading 4: "The triggering signal defines the final scope"

The formal referral notice determines the procedural scope.

Misreading 5: "Statistical significance determines the regulatory outcome"

Regulatory assessment integrates statistical, clinical, epidemiological and pharmacological evidence.

Misreading 6: "The procedure ends when the decision is published"

Implementation and post-referral pharmacovigilance remain relevant.

52. A Practical Article 31 Referral Reading Checklist

When reviewing a live or historical Article 31 pharmacovigilance referral, the following sequence is useful:

This checklist is particularly useful during inspection preparation or regulatory due diligence.

53. Article 31 Referral Timeline: What to Record

A referral timeline should use actual procedural dates rather than approximate intervals whenever the information is available.

At minimum, record:

• date of formal referral;
• date of PRAC assessment milestones;
• date of recommendation;
• date of CMDh or CHMP outcome;
• date of re-examination request, if any;
• date of final opinion;
• date of Commission decision, where applicable;
• implementation deadlines;
• completion dates for required actions.

The exact timetable is procedure-specific. Generic timelines should therefore be treated as explanatory rather than as substitutes for the official timetable.

This distinction is particularly important because statutory and procedural deadlines can change and can differ according to the regulatory route.

54. What Makes an Article 31 Referral Different From Routine Regulatory Variation

A routine variation generally begins with an application to change an existing marketing authorisation.

An Article 31 referral is different because the regulatory system itself initiates a Union-level assessment of the defined issue under the statutory referral mechanism.

The difference affects governance, scientific assessment, committee involvement and the legal route to the outcome.

A referral can ultimately require a variation or changes to product information, but the variation is then an implementation consequence of the regulatory decision rather than the mechanism by which the original safety question was created.

This distinction helps prevent two different regulatory processes from being described as though they were interchangeable.

55. Relationship With Other EU Referral Mechanisms

Article 31 should be interpreted within the wider EU referral architecture.

Other referral provisions address different regulatory circumstances, including procedures under Articles 107i, 107k, 29, 30 and 31 of Directive 2001/83/EC and Article 20 of Regulation (EC) No 726/2004.

The legal trigger, committee involvement, product scope and procedural consequences differ between these mechanisms.

The correct first question when encountering a referral is therefore:

Which legal referral mechanism has actually been invoked?

Only after that should the reader infer which committee, evidence standard, timetable and implementation pathway apply.

This is particularly important when terminology such as "safety referral" is used informally in secondary sources.

56. Article 31 and the QPPV's Regulatory Awareness

A QPPV does not need to become a specialist EU lawyer to oversee a referral effectively. The essential requirement is the ability to connect the legal procedure to the pharmacovigilance system.

The QPPV should be able to answer:

• What is the safety issue?
• What legal referral mechanism is being used?
• What products are affected?
• What did PRAC conclude?
• What regulatory action followed?
• What has the company implemented?
• How is the resulting risk being monitored?

This is the practical intersection between EU regulatory affairs and pharmacovigilance system oversight.

57. Reading the Final Regulatory Outcome

The final outcome of an Article 31 pharmacovigilance referral should be read as a legal and operational document, not merely as the last paragraph of the scientific assessment.

A useful approach is to separate four layers:

1. scientific conclusion — what the evidence supports;
2. regulatory conclusion — what action is required;
3. legal instrument — which document gives the outcome effect;
4. implementation requirement — what the marketing authorisation holder and competent authorities must do.

This separation is particularly important when a scientific recommendation contains several possible measures but the final legal instrument adopts only some of them.

The implementation team should therefore work from the final applicable decision and its annexes, while using the PRAC and CHMP documents to understand the reasoning behind the decision.

58. Annexes Can Be More Important Than the Narrative

Regulatory decisions and opinions frequently contain annexes that specify the practical consequences of the outcome.

For a pharmacovigilance referral, these may contain revised product information, conditions, timelines or other regulatory requirements.

A regulatory review that reads only the narrative decision can therefore miss the precise wording that must be implemented.

For implementation purposes, the working record should identify:

• the operative decision;
• affected products;
• revised texts;
• implementation deadlines;
• conditions attached to the authorisation;
• required communications;
• any required follow-up studies or monitoring.

The annexes should be treated as part of the regulatory instrument rather than supplementary reading.

59. Product-by-Product Implementation

A referral can affect several products whose regulatory histories are not identical.

Implementation should therefore be performed at product level rather than assuming that one global action is sufficient.

For each affected marketing authorisation, establish:

• authorisation number;
• product name;
• strength and pharmaceutical form;
• marketing authorisation holder;
• authorisation route;
• applicable regulatory action;
• implementation deadline;
• completion evidence.

This is particularly important where the same active substance is marketed by several MAHs or through different authorisation routes.

A referral-level conclusion can therefore require multiple implementation workstreams.

60. National Implementation for Nationally Authorised Products

For nationally authorised products, the regulatory outcome must be translated into the applicable national authorisation framework.

The exact mechanics depend on the legal route and the outcome reached through the referral.

The MAH should therefore distinguish between:

• the Union-level regulatory conclusion;
• the national administrative implementation;
• the company's internal completion of the required changes.

These can occur on different dates.

For audit purposes, the MAH should retain evidence demonstrating that the required national implementation was completed within the applicable timetable.

61. Central Authorisation Implementation

For centrally authorised products, the implementation pathway follows the central authorisation framework and the applicable Commission decision.

The MAH should map the decision to the affected marketing authorisation documentation and ensure that the required product-information and pharmacovigilance changes are incorporated through the appropriate regulatory process.

Again, the key point is to distinguish the date of the scientific opinion from the date and legal effect of the final regulatory decision.

A scientific conclusion can therefore predate the point at which the final legally binding action takes effect.

62. The Regulatory Decision as a Source of Pharmacovigilance Obligations

A referral outcome can create or modify obligations that directly affect the pharmacovigilance system.

Examples include:

• new identified risks;
• important potential risks;
• additional monitoring;
• targeted follow-up;
• additional pharmacovigilance activities;
• additional risk-minimisation measures;
• changes to routine surveillance.

The QPPV and relevant safety functions should translate these requirements into controlled activities rather than treating the regulatory decision as a document-retention exercise.

The implementation record should link each regulatory requirement to an owner, due date, evidence of completion and, where appropriate, effectiveness monitoring.

63. Updating the Safety Specification

Where the referral establishes or materially changes the characterisation of a risk, the safety specification may need to be updated within the applicable risk-management framework.

The update should reflect the final regulatory conclusion rather than an earlier hypothesis considered during the procedure.

This distinction matters because the scientific discussion can contain uncertainties and alternative hypotheses that are not ultimately incorporated into the final safety specification.

The regulatory outcome should therefore be used as the authoritative basis for the final internal classification, subject to the applicable RMP guidance and regulatory documentation.

64. Updating Signal Detection Strategy

A referral outcome can also affect how future cases and signals are detected.

For example, if the procedure establishes a clinically important risk associated with a defined population, the signal-management strategy may need to reflect that knowledge.

Possible changes can include:

• revised search terms;
• updated medical concepts;
• new subgroup analyses;
• targeted cumulative review;
• enhanced monitoring of specific outcomes;
• revised signal validation criteria.

The objective is not to create an artificial permanent "Article 31 signal" category. The objective is to ensure that the pharmacovigilance system incorporates the scientific knowledge generated by the referral.

65. Effectiveness Evaluation After Risk-Minimisation Changes

When the regulatory outcome introduces additional risk-minimisation measures, the organisation should identify how effectiveness will be assessed where required.

Effectiveness can concern two different questions:

1. process effectiveness — was the measure delivered as intended?
2. outcome effectiveness — did the measure achieve the intended reduction or control of risk?

These questions should not be conflated.

A communication may have been distributed to all intended recipients while failing to change prescribing behaviour. Conversely, a measurable change in prescribing may occur without proving that the clinical risk itself has been reduced.

The evaluation design should therefore be aligned with the purpose of the measure.

66. Medical Information and Safety Communication

Some referral outcomes require communication to healthcare professionals or patients.

The communication should accurately reflect the final regulatory wording and should not introduce claims that go beyond the regulatory decision.

Internal review should therefore confirm consistency between:

• final regulatory decision;
• product information;
• safety communication;
• medical-information responses;
• training material;
• pharmacovigilance monitoring instructions.

Inconsistent communication can create a new regulatory risk even when the underlying implementation is technically complete.

67. Monitoring Compliance With the Outcome

Implementation should be monitored until all regulatory commitments are demonstrably complete.

A useful implementation tracker can contain:

The exact fields should reflect the organisation's quality system.

The principle is simple: every regulatory requirement should have a traceable implementation path.

68. Deviations From Regulatory Implementation

If an implementation deadline cannot be met, the issue should be managed through the applicable quality and regulatory processes.

The organisation should assess:

• what requirement was missed;
• why it was missed;
• whether patients or public health could be affected;
• whether an interim control is required;
• whether the authority must be informed;
• what corrective and preventive actions are appropriate.

The existence of a referral does not create an exception to normal pharmacovigilance and quality-system expectations.

69. CAPA and Root-Cause Analysis

A regulatory implementation failure may require formal deviation management and, depending on its significance, CAPA.

Root-cause analysis should distinguish between:

• failure to identify the requirement;
• failure to assign ownership;
• inadequate procedural controls;
• inadequate system configuration;
• communication failure;
• insufficient regulatory intelligence;
• inadequate oversight.

A superficial conclusion such as "deadline missed" does not identify the systemic cause.

The objective of CAPA is to reduce recurrence risk, not merely to document the incident.

70. Inspection Readiness for an Article 31 Referral

An inspector may reasonably expect the MAH to demonstrate that it understood the referral, implemented the outcome and maintained appropriate pharmacovigilance oversight.

The inspection package should therefore make it possible to reconstruct:

Referral
   ↓
Scientific assessment
   ↓
Company response
   ↓
Regulatory decision
   ↓
Implementation
   ↓
Ongoing monitoring

Documents should be linked sufficiently clearly that an inspector does not have to reconstruct the history from unrelated systems.

71. Governance During a Referral

Because a referral crosses regulatory, medical, pharmacovigilance, quality and sometimes commercial functions, governance should be explicit.

A cross-functional governance model can assign responsibility for:

• regulatory correspondence;
• scientific assessment;
• case-series and epidemiology;
• submission preparation;
• executive escalation;
• product-information implementation;
• RMP updates;
• communication;
• post-referral monitoring.

The QPPV should have appropriate visibility of the safety aspects and the resulting pharmacovigilance obligations.

The exact organisational model will differ between companies, but accountability should not be ambiguous.

72. Decision Logs and Scientific Judgement

Major scientific judgements made during a referral should be documented in a controlled manner.

Examples include decisions concerning:

• case inclusion criteria;
• alternative causes;
• exposure denominators;
• subgroup definitions;
• interpretation of conflicting studies;
• proposed risk-minimisation measures.

The purpose is not to document every discussion. It is to preserve the reasoning behind material decisions that affect the regulatory submission or the company's interpretation of the evidence.

This becomes particularly valuable when a referral extends over many months or involves staff changes.

73. Handling Conflicting Evidence

Conflicting evidence is not unusual in pharmacovigilance.

A high-quality assessment should explain why apparently inconsistent findings do or do not change the overall conclusion.

Relevant considerations can include:

• differences in study design;
• different populations;
• different exposure definitions;
• outcome misclassification;
• statistical power;
• residual confounding;
• differences in follow-up;
• biological plausibility.

The regulatory conclusion should be based on the totality of evidence rather than selected studies that support a preferred position.

74. Uncertainty in an Article 31 Assessment

A referral can produce a regulatory action even when scientific uncertainty remains.

The appropriate response to uncertainty may be additional monitoring, restricted use, additional risk minimisation or further evidence generation rather than immediate withdrawal.

The regulatory reasoning should therefore be read for how uncertainty was handled.

Important questions include:

• What remains uncertain?
• Does the uncertainty affect the benefit-risk conclusion?
• Can the uncertainty be reduced through additional evidence?
• Can the risk be controlled while uncertainty remains?

This is often more informative than focusing only on whether causality was described as "established" or "not established".

75. Benefit-Risk Reasoning Under Uncertainty

Benefit-risk evaluation is inherently comparative.

The relevant question is not simply whether a medicine has a risk, but whether the totality of benefits and risks remains acceptable for the intended population and conditions of use.

Where uncertainty is material, the assessment can consider the consequences of both action and inaction.

For example:

• unrestricted continued use may expose patients to a serious uncertain risk;
• excessive restriction may deprive patients of important therapeutic benefit;
• targeted risk minimisation may provide an intermediate control strategy.

The chosen regulatory measure should therefore be understood in relation to both the identified risk and the therapeutic benefit.

76. Population-Level Versus Individual-Level Risk

A referral can concern a population-level safety signal while regulatory measures ultimately operate at the individual-patient level.

The evidence may describe an increased rate across a population, whereas the final measure may instruct clinicians to avoid treatment in a defined subgroup or monitor individual patients.

The translation from population evidence to individual clinical action should therefore be explicit in the regulatory reasoning and product information.

77. Regulatory Proportionality

The regulatory response should be considered in relation to the seriousness and controllability of the risk and the benefits of treatment.

A proportional response may involve:

• no change;
• additional information;
• targeted monitoring;
• contraindication in a subgroup;
• additional risk minimisation;
• restriction of use;
• suspension or revocation.

The presence of multiple possible measures is why the final decision should be read carefully rather than inferred from the existence of the referral.

78. What the Referral Does Not Establish

An Article 31 pharmacovigilance referral does not, by itself, establish:

• causality;
• a negative benefit-risk balance;
• a manufacturing defect;
• a product-quality failure;
• regulatory non-compliance by the MAH;
• that withdrawal will occur;
• that every product containing the active substance is affected.

Each of those conclusions requires its own evidence and regulatory basis.

This negative definition is useful because public summaries can sometimes compress a complex referral into a much stronger claim than the actual regulatory documents support.

79. How to Read Public EMA Materials Without Overinterpreting Them

Public EMA materials are valuable sources for understanding the referral, but different documents have different functions.

A public referral page may provide a concise description. A PRAC recommendation provides scientific reasoning. A CHMP opinion addresses the applicable central regulatory stage. A Commission decision establishes the legally binding Union outcome where applicable.

The reader should therefore identify the document type before relying on a statement as a legal conclusion.

A useful hierarchy is:

Legal text
   ↓
Formal referral / regulatory decision
   ↓
Committee opinion or recommendation
   ↓
Procedural guidance
   ↓
Public summary / secondary explanation

Lower-level summaries are useful for orientation but should not replace the primary regulatory documents when making a live regulatory determination.

80. How to Cite an Article 31 Referral in Regulatory Writing

A regulatory document should identify the referral precisely enough that another professional can locate the underlying procedure.

Where relevant, include:

• Article 31 legal basis;
• active substance or product;
• referral subject;
• date or procedural identifier;
• PRAC recommendation;
• subsequent committee opinion or position;
• Commission decision where applicable.

Avoid citing only a general EMA page when the conclusion depends on a specific referral document.

Precise citation improves reproducibility and reduces the risk of attributing a statement from one procedural stage to another.

81. Distinguishing Regulatory Fact From Scientific Interpretation

A strong referral assessment separates statements into three categories:

Regulatory fact: what the legal or procedural documents state.

Scientific finding: what the evidence supports.

Interpretation: what the reviewer concludes from those facts and findings.

For example:

Regulatory fact: PRAC recommended a specific regulatory measure.

Scientific finding: the assessment identified a clinically important risk in a defined population.

Interpretation: the measure appears proportionate to the identified risk under the conditions described.

This structure makes regulatory writing more defensible and easier to audit.

82. Version Control of the Referral Record

The regulatory record should distinguish between different versions of submissions, assessments and final documents.

This is particularly important where the scientific assessment evolves during the procedure.

The controlled record should identify:

• document version;
• submission date;
• author / owner;
• regulatory stage;
• superseded versions;
• final status.

The final regulatory conclusion should never be reconstructed from an obsolete draft when the authoritative document is available.

83. Closing the Referral Internally

Once the regulatory implementation is complete, the organisation should formally close the internal referral workstream according to its quality system.

Closure should normally confirm:

• all regulatory commitments completed;
• product-information changes implemented;
• RMP changes completed where required;
• risk-minimisation measures operational;
• pharmacovigilance systems updated;
• relevant monitoring established;
• outstanding actions transferred to routine lifecycle processes.

Closure does not mean the safety issue is forgotten. It means the exceptional referral project has been converted into controlled lifecycle activities.

84. The Article 31 Referral as a Lifecycle Event

The most useful way to understand an Article 31 pharmacovigilance referral is as a lifecycle event rather than a single committee meeting.

Pre-referral safety evidence
          ↓
Formal Union referral
          ↓
PRAC assessment
          ↓
Recommendation
          ↓
CMDh / CHMP pathway
          ↓
Commission decision where applicable
          ↓
Implementation
          ↓
Risk-management and PV follow-up
          ↓
Future evidence

Each stage has a different regulatory purpose.

Understanding those purposes is more useful than memorising committee names or procedural labels in isolation.

85. Practical Takeaways for Pharmacovigilance Professionals

For day-to-day work, the essential lessons are these:

1. Identify the legal referral mechanism first.
2. Read the formal referral notice before the later scientific documents.
3. Separate the triggering safety concern from the formal referral scope.
4. Treat PRAC as the pharmacovigilance scientific assessment stage.
5. Do not confuse a PRAC recommendation with the final legal decision.
6. Determine whether affected products are nationally or centrally authorised.
7. Track CMDh and CHMP consequences according to the authorisation route.
8. Read the final decision and annexes for the operative requirements.
9. Translate the outcome into RMP, PV-system and product-information actions.
10. Maintain post-referral monitoring and a complete audit trail.

These principles provide a reliable framework for interpreting both historical and live Article 31 pharmacovigilance referrals.

86. Final Perspective

Article 31 pharmacovigilance referrals are best understood as the point where a significant pharmacovigilance concern enters a formal Union regulatory decision-making process.

The scientific question and the legal question are connected, but they are not identical. PRAC evaluates the pharmacovigilance evidence and adopts its recommendation. The subsequent regulatory pathway depends on the products affected and the applicable legal framework. CMDh and CHMP perform their respective roles, and where the procedure requires it, the European Commission establishes the binding Union outcome.

For the QPPV and pharmacovigilance organisation, the work does not end with the committee recommendation or Commission decision. The outcome must be translated into product information, risk management, pharmacovigilance processes, communication, monitoring and documented compliance.

The strongest way to read an Article 31 referral is therefore as a complete regulatory chain:

Evidence
  → Referral
  → Scientific assessment
  → Regulatory recommendation
  → Legal decision
  → Implementation
  → Continued pharmacovigilance

That chain is the essential structure of the Article 31 pharmacovigilance system.

References

The following primary and official regulatory sources should be used when checking the legal and procedural statements in this article. The applicable consolidated legislation and procedure-specific documents should always take precedence over a general explanatory article.

1. Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Community code relating to medicinal products for human use, as amended. In particular, Article 31 and the related provisions governing referral procedures and pharmacovigilance, including Articles 32, 33, 34 and 107i–107k.
2. European Medicines Agency (EMA), Questions and answers: Article 31 pharmacovigilance referrals. This provides practical and procedural guidance on initiation, PRAC assessment, recommendation, CMDh/CHMP follow-up, re-examination and publication. urlEMA Article 31 pharmacovigilance referrals Q&Ahttps://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/referral-procedures-human-medicines/questions-answers-article-31-pharmacovigilance-referrals
3. European Medicines Agency, Referral procedures: regulatory and procedural guidance. This provides the wider collection of procedural guidance applicable to human-medicine referrals and related regulatory procedures. urlEMA referral regulatory and procedural guidancehttps://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/referral-procedures-human-medicines/referral-procedures-regulatory-procedural-guidance
4. European Medicines Agency, Guide to information on human medicines evaluated by EMA — What EMA publishes and when. This is useful for understanding the publication sequence for referral documents, opinions, assessment reports, annexes and related regulatory information. urlEMA publication guide for human medicineshttps://www.ema.europa.eu/en/documents/other/guide-information-human-medicines-evaluated-european-medicines-agency-what-agency-publishes-when_en.pdf
5. European Medicines Agency, Questions and answers: Article 31 non-pharmacovigilance referrals. This is useful for distinguishing the pharmacovigilance Article 31 pathway from the non-pharmacovigilance Article 31 pathway. urlEMA Article 31 non-pharmacovigilance referrals Q&Ahttps://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/referral-procedures-human-medicines/questions-answers-article-31-non-pharmacovigilance-referrals
6. Regulation (EC) No 726/2004 of the European Parliament and of the Council, as amended, establishing Union procedures for the authorisation, supervision and pharmacovigilance of medicinal products for human and veterinary use and establishing a European Medicines Agency. The applicable provisions should be read together with Directive 2001/83/EC where centrally authorised products are concerned.
7. EMA, PRAC-related procedural and regulatory guidance, including guidance concerning rapporteur/co-rapporteur appointment, pharmacovigilance assessment and publication of safety-related recommendations.
8. European Commission, Notice to Applicants, Volume 2A — Procedures for marketing authorisation, Chapter 3: Union Referral Procedures. The current version should be consulted for the applicable procedural framework and implementation requirements.
9. European Commission, Notice to Applicants, Volume 2A — Decision-making procedure for the adoption of Commission Decisions. The current version should be consulted when a referral proceeds to a Commission decision.
10. EMA and CMDh procedural guidance relevant to nationally authorised medicinal products, mutual-recognition and decentralised procedures, implementation of CMDh positions and related product-information changes.

Primary-document hierarchy

When there is any conflict between a general explanation and the documents governing a particular referral, use the following hierarchy:

1. applicable EU legislation;
2. the formal referral notification and procedure-specific documents;
3. the final Commission decision or applicable CMDh position;
4. the final CHMP opinion and PRAC recommendation, as applicable;
5. current EMA/CMDh procedural guidance;
6. secondary summaries and explanatory material.

The hierarchy is particularly important because referral procedures evolve, individual timetables differ and procedural guidance can be revised.

Regulatory Note

This article is an educational and regulatory-reference document. It is intended to explain the structure and practical interpretation of Article 31 pharmacovigilance referrals for medicinal products for human use in the European Union. It is not legal advice and should not be used as a substitute for the applicable legislation, official regulatory guidance or procedure-specific documents.

The article has been reviewed against official EU and EMA materials available at the date shown in the article metadata. EU medicines legislation, EMA guidance, CMDh procedures and publication practices can change. A live regulatory assessment must therefore use the current consolidated legislation and current procedure-specific documentation.

The article deliberately distinguishes general procedural principles from case-specific facts. The existence, scope, timetable, questions, committee documents, re-examination rights, implementation requirements and final outcome of an individual Article 31 referral must be established from the documents for that specific procedure.

In particular, do not infer the final legal outcome solely from a PRAC recommendation, a committee meeting highlight, a public summary or a secondary source. Where applicable, the final Commission decision or CMDh consensus position and its annexes should be checked for the operative requirements.

Where an individual regulatory procedure is underway, the procedure-specific EMA timetable, communications, formal regulatory documents and applicable legal provisions take precedence over generic descriptions in this article.

Nothing in this article should be interpreted as creating a regulatory obligation that is not contained in applicable EU legislation, a binding regulatory decision or an applicable regulatory requirement.