Safety Specification
The Safety Specification is the part of the EU Risk Management Plan that describes the medicinal product's safety profile with emphasis on the risks and uncertainties that require focused risk management. Its purpose is not to reproduce the SmPC adverse-reaction section or catalogue every theoretical concern. It identifies the safety issues that matter enough to shape the pharmacovigilance plan or risk minimisation strategy.
- Safety Specification
- Purpose and Regulatory Framework
- Safety Specification Is Broader Than the Summary of Safety Concerns
- The Three Safety-Concern Categories
- Not Every Adverse Reaction Is an RMP Safety Concern
- How Evidence Becomes a Safety Concern
- Characterising Risks, Not Merely Naming Them
- Relationship With the Pharmacovigilance Plan
- Relationship With Risk Minimisation
- Relationship With Benefit-Risk Evaluation
- Special Situations
- Lifecycle Management
- Potential Failure Modes
- Inspection Considerations
- Practical Review Checklist
- Key Takeaways
- References
- Regulatory Note
Purpose and Regulatory Framework
GVP Module V Rev. 2 sets out the structure and principles of the Safety Specification in Part II of the EU RMP. The current integrated EU RMP format, Rev. 2.0.1, translates those principles into the submission structure used by applicants and marketing-authorisation holders.
The Safety Specification supports three linked tasks:
- characterise what is already known about the product's safety profile;
- identify important uncertainties that remain; and
- determine which safety concerns require proactive pharmacovigilance or risk minimisation.
The outcome is summarised in Module SVIII, which lists the important identified risks, important potential risks and missing information that constitute the RMP's summary of safety concerns.
Safety Specification Is Broader Than the Summary of Safety Concerns
A common misunderstanding is to treat the Safety Specification as only the final three-category list. In fact, Part II contains the evidence needed to understand that list.
Depending on the product, the Safety Specification considers areas such as:
- epidemiology of the indication and target population;
- non-clinical safety findings;
- clinical-trial exposure;
- populations not studied in clinical trials;
- post-authorisation experience;
- identified and potential risks; and
- missing information.
These modules create the scientific context from which Module SVIII is derived.
The Three Safety-Concern Categories
Important identified risks
An identified risk is an undesirable clinical outcome for which there is sufficient scientific evidence that it is caused by the medicinal product. It becomes important when it could materially affect the benefit-risk balance or has implications for public health or individual patient management such that focused evaluation or risk minimisation is warranted.
The detailed classification principles are covered in [[important-identified-risks]].
Important potential risks
A potential risk is an undesirable clinical outcome for which there is scientific basis to suspect an association but insufficient evidence to establish causality. It becomes important when confirmation of the risk could materially affect benefit-risk or risk management.
See [[important-potential-risks]].
Missing information
Missing information is a relevant gap in knowledge about safety in a population or use situation where there is scientific reason to suspect that the safety profile could differ and available information is insufficient.
Absence of data alone is not enough. See [[missing-information]].
Not Every Adverse Reaction Is an RMP Safety Concern
This principle keeps the RMP focused. An adverse reaction may be established, described in product information and adequately managed through routine measures without remaining an important identified risk in the RMP.
Similarly, a theoretical mechanism does not automatically become an important potential risk, and an underrepresented population does not automatically become missing information.
The Safety Specification therefore performs selection, not transcription.
How Evidence Becomes a Safety Concern
Classification requires two separate judgements: what the evidence supports and whether the issue is important for risk management.
Evidence can arise from clinical trials, spontaneous reports, epidemiology, literature, non-clinical studies, pharmacology, class effects and regulatory assessments. No universal case count, reporting ratio, relative-risk threshold or statistical boundary determines classification.
A structured evaluation asks:
- What clinical outcome or knowledge gap is being considered?
- What evidence supports an association or a plausible difference in safety?
- How certain is the evidence?
- What is the likely clinical and public-health consequence?
- Would focused pharmacovigilance or risk minimisation improve management of the issue?
This prevents statistical signals, isolated cases or mechanistic hypotheses from being promoted automatically into RMP safety concerns.
Characterising Risks, Not Merely Naming Them
A useful Safety Specification describes enough about an important risk to explain why and how it should be managed. Relevant dimensions may include:
- seriousness and severity;
- frequency or incidence where estimable;
- preventability and reversibility;
- time to onset and duration;
- dose or exposure relationship;
- risk factors and susceptible populations;
- clinical consequences; and
- uncertainty in the available evidence.
The level of detail should be proportionate to what is known and what decisions depend on the characterisation.
Relationship With the Pharmacovigilance Plan
The Safety Specification defines the question; the pharmacovigilance plan explains how remaining uncertainty will be addressed.
For each safety concern, the organisation should determine whether routine pharmacovigilance is sufficient or whether an additional pharmacovigilance activity is justified.
An additional activity should not exist merely because an item appears in Module SVIII. It should answer a specific unresolved question that routine pharmacovigilance cannot address adequately.
For PASS, see [[pass-and-rmps]].
Relationship With Risk Minimisation
Risk minimisation asks a different question: what should be done to reduce the probability or impact of harm?
Routine measures such as product information may be sufficient for many safety concerns. Additional risk minimisation measures are justified only when routine measures are insufficient to support safe and effective use.
The scientific chain should therefore be visible:
evidence → safety concern → remaining uncertainty → pharmacovigilance need → risk minimisation need.
Relationship With Benefit-Risk Evaluation
Importance is contextual. The same adverse outcome can have different risk-management significance in a life-threatening disease with few alternatives than in a mild condition with many safe alternatives.
Assessment should therefore consider the magnitude of benefit, disease severity, alternatives, preventability of harm and the population exposed. The Safety Specification cannot be interpreted independently of the therapeutic context.
Special Situations
Class effects
A class effect can support a potential-risk hypothesis, but transferability should be examined. Mechanism, target, structure, exposure and indication may differ sufficiently that class evidence does not apply equally to every product.
New populations or indications
A new indication may change the target population and therefore the relevance of existing risks or missing information. The Safety Specification should be reassessed rather than simply copied unchanged from the existing indication.
Generic and established products
For established substances, the safety profile may already be well characterised. Proportionality remains important: a focused Safety Specification is preferable to importing historical concerns that no longer require active management.
Biological medicinal products
For biologics, immunogenicity, manufacturing-related attributes and mechanism-specific risks may require careful scientific characterisation. These are not automatically RMP safety concerns; their inclusion still depends on evidence and risk-management importance.
Lifecycle Management
The Safety Specification should evolve as evidence changes. A potential risk may become identified, missing information may be resolved, and a previously important risk may become sufficiently characterised and managed that focused RMP treatment is no longer necessary.
Reclassification or removal should follow scientific evidence, not fixed time periods or corporate thresholds. The rationale should explain what changed and why the new classification better reflects the current state of knowledge.
Potential Failure Modes
The following are illustrative failure modes, not published inspection findings.
| Failure mode | Why it is weak |
|---|---|
| Copying every SmPC adverse reaction into Module SVIII | confuses known adverse reactions with safety concerns requiring focused management |
| Treating every signal as a potential risk | signal status is not the same as RMP classification |
| Listing every trial-excluded population as missing information | absence of data alone is insufficient |
| Using numerical cut-offs for importance | GVP does not prescribe universal thresholds |
| Carrying historical concerns forward without reassessment | makes the RMP progressively less focused |
| Adding a safety concern without assessing PV/RMM consequences | breaks the purpose of the RMP |
| Removing a concern without documenting why focused management is no longer needed | weakens traceability |
Inspection Considerations
An inspector could examine whether the safety concern list is scientifically defensible and whether it aligns with the rest of the pharmacovigilance system.
Relevant questions include:
- What evidence supports each category?
- Why is the issue important enough for RMP focus?
- How are signal conclusions reflected?
- Are additional pharmacovigilance activities linked to explicit uncertainties?
- Are risk minimisation measures proportionate to the concern?
- Why were concerns added, reclassified or removed?
- Are the RMP, PSUR/PBRER and product information scientifically consistent?
The aim is traceability of reasoning, not production of a particular internal classification form.
Practical Review Checklist
This checklist is recommended operational practice.
- Is each safety concern defined as a specific clinical outcome or knowledge gap?
- Does the evidence support identified, potential or missing-information classification?
- Is importance justified in the therapeutic context?
- Have ordinary adverse reactions been excluded where routine management is sufficient?
- Have data gaps been distinguished from true missing information?
- Are class and mechanistic arguments product-relevant?
- Does each additional PV activity answer a defined uncertainty?
- Are risk minimisation measures proportionate?
- Are recent signal, study and regulatory conclusions incorporated?
- Are additions, reclassifications and removals traceable?
Key Takeaways
The Safety Specification is the scientific foundation of the RMP. It converts a broad safety evidence base into a focused set of concerns that require active risk management.
Its three formal safety-concern categories—important identified risks, important potential risks and missing information—describe different states of evidence and uncertainty.
Classification requires both evidence and importance. No universal case count, disproportionality threshold or exposure cut-off determines inclusion.
The Safety Specification is dynamic. It should become more focused as uncertainties are resolved and should change when new evidence materially alters risk management.
References
- European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) Module V — Risk management systems (Rev. 2). EMA/838713/2011 Rev. 2.
- European Medicines Agency. Guidance on the format of the risk management plan in the EU — integrated format (Rev. 2.0.1). EMA/164014/2018 Rev. 2.0.1.
- European Medicines Agency. Risk management plans (RMP) in post-authorisation phase: questions and answers. EMEA-H-19984/03 Rev. 118, updated 13 July 2026.
- International Council for Harmonisation. ICH E2E Pharmacovigilance Planning.
- European Union. Commission Implementing Regulation (EU) No 520/2012, consolidated version current at 12 February 2026.
- European Union. Directive 2001/83/EC, as amended.
- European Union. Regulation (EC) No 726/2004, as amended.
Regulatory Note
This article distinguishes binding EU requirements, GVP guidance, scientific interpretation and recommended operational practice. As of 8 September 2026, GVP Module V Rev. 2 and the integrated RMP format Rev. 2.0.1 remain the principal published EU references for the Safety Specification.