GVP Module VII: Writing the PSUR Safety Evaluation

Introduction

The scientific value of a Periodic Safety Update Report (PSUR) depends less on the volume of information collected than on the quality of the evaluation performed on that information.

GVP Module VII requires the PSUR to provide a comprehensive and critical analysis of the benefit-risk balance, considering new or emerging safety information in the context of cumulative information on risks and benefits. citeturn0search0turn0search17

This means that writing a PSUR is not primarily an exercise in summarising databases, reproducing signal reports or transferring information from previous reports. The writer must establish what the evidence means, how it changes or confirms the existing safety profile, how uncertainty affects interpretation and whether the overall benefit-risk balance requires any action.

The distinction between data presentation and scientific evaluation is therefore fundamental.

1. The Evaluation Is the Core of the PSUR

A useful conceptual model is:

Evidence
   ↓
Data quality and limitations
   ↓
Clinical interpretation
   ↓
Risk and benefit characterisation
   ↓
Integration of cumulative knowledge
   ↓
Benefit-risk evaluation
   ↓
Conclusion and action

Each step answers a different question.

A table can show that reports increased. It cannot by itself establish whether the underlying risk increased.

A study can report a statistically significant association. It does not automatically establish causality or clinical importance.

A known risk can remain important even when no new signal is identified.

The PSUR evaluation must connect the evidence to the regulatory conclusion.

2. Start With the Regulatory Question

Before drafting an evaluation, the author should understand what the PSUR is required to establish.

The central questions are:

1. What is known about the safety profile?
2. What new information became available during the reporting interval?
3. Does that information materially change the understanding of existing risks?
4. Are there new or emerging safety concerns?
5. What is the quality and uncertainty of the evidence?
6. What does the cumulative evidence indicate?
7. Does the benefit-risk balance remain favourable for the authorised indications and relevant populations?
8. Are changes to pharmacovigilance, risk minimisation, product information or other regulatory measures warranted?

These questions should drive the evaluation rather than the structure of the source databases.

3. Data Presentation Versus Critical Evaluation

The PSUR contains both information presentation and scientific assessment, but they should not be confused.

Consider the following example:

During the reporting interval, 420 cases of hepatic events were received compared with 310 in the previous interval.

That is a data observation.

A critical evaluation would ask:

• Did exposure increase?
• Did reporting behaviour change?
• Were the cases clinically comparable?
• Did seriousness or outcomes change?
• Was there an increase in a particular clinical phenotype?
• Were there changes in concomitant medicines or underlying disease?
• Was there a regulatory communication that stimulated reporting?
• Is there supporting evidence from studies or literature?
• Does the cumulative evidence indicate a change in risk?

The conclusion should follow from that analysis rather than from the raw count.

4. Characterising the Evidence

Evidence should be characterised according to its source, quality and relevance.

Important dimensions can include:

• study design;
• population;
• exposure;
• comparator;
• outcome definition;
• follow-up duration;
• completeness of information;
• statistical precision;
• consistency across sources;
• biological plausibility;
• and potential biases or confounding.

Different evidence types answer different questions.

A spontaneous case may provide valuable clinical detail but usually cannot establish incidence. A controlled study may provide a denominator but have limited generalisability. A literature review may identify a pattern across multiple populations but contain heterogeneous methods.

The evaluation should preserve these distinctions.

5. Strengths and Limitations of the Evidence

A defensible PSUR does not hide uncertainty.

For each important conclusion, the author should consider what makes the evidence persuasive and what limits its interpretation.

Strengths may include:

• consistent findings across independent sources;
• strong temporal association;
• appropriate comparator data;
• clinically coherent case characteristics;
• dose-response evidence;
• biological plausibility;
• and reproducibility.

Limitations may include:

• incomplete case information;
• small sample size;
• confounding;
• selection bias;
• reporting bias;
• exposure uncertainty;
• outcome misclassification;
• and limited follow-up.

The purpose is not to assign an artificial numerical score to every piece of evidence. It is to make the reasoning transparent.

6. New Information in the Context of Existing Knowledge

New information should be compared with the established safety profile.

For an existing risk, the evaluation should ask whether the new evidence:

• confirms the known risk;
• changes its clinical characterisation;
• changes its frequency or severity;
• identifies a new population at risk;
• identifies a new risk factor;
• changes understanding of preventability;
• or challenges the previous assessment.

For a potential new risk, the assessment should consider whether the evidence is sufficient to support a new safety concern or whether further investigation is required.

The objective is not to label every new observation as a new risk.

7. Characterisation of Risks

Risk characterisation should go beyond naming an adverse reaction.

Where appropriate, the evaluation should consider:

• clinical nature;
• seriousness;
• frequency or reporting pattern;
• severity;
• outcome;
• reversibility;
• predictability;
• preventability;
• risk factors;
• susceptible populations;
• dose relationship;
• time to onset;
• and the effectiveness of available risk-management measures.

The relevant characteristics depend on the specific risk.

A risk that is rare but fatal may require a different assessment from a common, mild and reversible adverse reaction.

8. Characterisation of Benefits

The benefit side of the assessment should be equally disciplined.

Relevant considerations can include:

• nature of the therapeutic effect;
• magnitude of effect;
• clinical relevance;
• duration;
• consistency;
• generalisability;
• patient population;
• unmet medical need;
• and available alternatives.

The PSUR should not assume that the benefits described at initial authorisation remain unchanged forever. New evidence can strengthen, weaken or refine the understanding of therapeutic benefit.

9. Baseline Versus New Benefit Information

Where there is little or no new relevant benefit information, the PSUR should still establish the baseline understanding against which the safety findings are interpreted.

Where important new efficacy or effectiveness information is available, its relevance to the benefit-risk balance should be critically assessed.

EMA's Module VII guidance specifically calls for consideration of the strength of evidence, effect size, methodological limitations, consistency, clinical relevance, generalisability and other factors when evaluating efficacy and effectiveness. citeturn0search18

The amount of benefit information included should therefore be proportionate to its relevance to the current safety and benefit-risk assessment.

10. Medical Need and Important Alternatives

Benefit-risk interpretation requires clinical context.

The PSUR should consider the medical need for the product in the authorised indications and summarise important alternatives, including no treatment where relevant.

This context matters because the clinical significance of a risk cannot always be understood without considering the disease being treated and the available therapeutic alternatives.

For example, the tolerance for a serious adverse reaction may differ substantially between treatment of a self-limiting condition and treatment of a life-threatening disease with few effective alternatives.

The assessment should explain the relevance of that context rather than use it as a reason to discount a risk.

11. Indication-Specific Evaluation

Benefit-risk balance is specific to an indication and population.

Where a medicinal product is authorised for multiple indications, the benefit-risk evaluation should therefore be considered separately for each indication where appropriate. If important differences exist among populations within an indication, those differences should also be addressed where possible. citeturn0search17

A single overall conclusion can obscure important differences.

For example, a safety concern may have little effect on benefit-risk for one indication but materially affect another because the expected therapeutic benefit, population or available alternatives differ.

12. Integrating the Evidence

The integrated evaluation should bring together the key findings rather than repeat each preceding section.

A useful writing pattern is:

Finding
  ↓
Clinical significance
  ↓
Strength of evidence
  ↓
Consistency with existing knowledge
  ↓
Impact on risk
  ↓
Impact on benefit-risk
  ↓
Action or rationale for no action

This is the point at which the author demonstrates scientific judgement.

The integrated benefit-risk section should not simply reproduce the earlier descriptions of benefits and risks. EMA's Module VII guidance specifically distinguishes characterisation from the integrated benefit-risk analysis. citeturn0search18

13. Avoiding Mechanical Conclusions

Weak conclusions often use phrases such as:

• "No new safety signal was identified, therefore the benefit-risk balance remains positive."
• "The number of cases increased but no action is required."
• "The literature did not identify any new concerns."

These statements may ultimately be correct, but they are insufficient without the reasoning that supports them.

A defensible conclusion explains:

1. what was evaluated;
2. what changed;
3. what did not change;
4. how strong the evidence is;
5. what uncertainty remains;
6. and why the resulting conclusion follows.

14. When the Evidence Is Conflicting

Conflicting evidence should be analysed rather than averaged into a vague conclusion.

Suppose spontaneous reports suggest a potential association while observational studies do not demonstrate an increased risk.

The evaluation should consider whether the difference can be explained by:

• different populations;
• different exposure definitions;
• outcome ascertainment;
• confounding;
• comparator selection;
• statistical power;
• latency;
• or differences in the underlying question being answered.

The conclusion should identify which evidence is more informative for the specific safety question and why.

15. Cumulative Evidence and Emerging Trends

A PSUR is periodic, but its conclusions are cumulative.

A pattern that is weak in one reporting interval may become more persuasive when additional evidence accumulates.

Conversely, an early concern may weaken as additional evidence fails to confirm it.

The author should therefore avoid treating each PSUR as an isolated report. The current evaluation should explicitly identify whether the cumulative evidence:

• strengthens;
• weakens;
• confirms;
• refines;
• or leaves unchanged

the previous understanding of important risks and benefits.

16. Uncertainty Is Part of the Conclusion

A scientific conclusion can be robust while still containing uncertainty.

For example:

The available evidence supports an association, but the magnitude of risk remains uncertain because exposure estimates are incomplete and an important potential confounder cannot be fully excluded.

That is more informative than either an unjustifiably definitive conclusion or a statement that "more data are needed" without explaining why.

Uncertainty should be linked to its consequence for the regulatory decision.

17. What the Writer Should Be Able to Defend

For every material conclusion, the PSUR author should be able to answer:

• What evidence supports this conclusion?
• Where did the evidence come from?
• How complete and reliable is it?
• What alternative explanations were considered?
• How does it compare with previous knowledge?
• What is the clinical significance?
• What uncertainty remains?
• Does it change the risk characterisation?
• Does it change the benefit-risk balance?
• Does it require action?

If the answer cannot be reconstructed from the PSUR and its controlled supporting records, the evaluation is difficult to defend during regulatory assessment or inspection.

18. Evaluating Individual Safety Concerns

A PSUR may contain numerous safety issues, but not every issue deserves the same depth of analysis.

The level of evaluation should be proportionate to the potential significance of the issue and the strength of the available evidence.

A useful prioritisation approach is to consider:

• seriousness of the potential outcome;
• clinical relevance;
• novelty;
• strength of evidence;
• magnitude or potential magnitude of risk;
• affected population;
• preventability;
• consistency across sources;
• and potential effect on the benefit-risk balance.

This does not mean that less prominent information should be ignored. It means that the scientific effort should be directed toward the questions most capable of changing the regulatory conclusion.

19. Signals Within the PSUR Evaluation

Signal management and PSUR preparation are separate processes, but their outputs intersect.

A signal identified during the reporting interval should be evaluated through the appropriate signal-management process. The PSUR then places relevant conclusions from that process into the periodic safety context.

For a signal that remains under evaluation, the PSUR should explain its status and what is known about its potential significance without presenting an unresolved hypothesis as an established risk.

For a closed signal, the PSUR should reflect the conclusion where it is relevant to the overall safety evaluation.

The PSUR should therefore answer the regulatory question rather than recreate the complete signal-management workflow.

20. Known Risks Without New Signals

An important known risk may remain central to the benefit-risk assessment even if no new signal is generated.

The author should consider whether the reporting interval provides evidence that:

• the risk remains appropriately characterised;
• its clinical features have changed;
• its frequency or severity appears different;
• new risk factors have emerged;
• existing risk-minimisation measures remain appropriate;
• or the cumulative evidence changes the previous assessment.

"No new signal" is therefore not equivalent to "nothing happened."

21. New Safety Information That Does Not Establish a New Risk

Not every concerning observation becomes a new risk.

A cluster of cases may require investigation without providing sufficient evidence to establish an association. A study may identify an imbalance that requires contextual interpretation. A literature report may raise a hypothesis without establishing causality.

The PSUR should preserve this distinction.

Appropriate scientific language can include:

• "the available evidence does not currently establish an association";
• "the finding warrants continued monitoring";
• "the evidence is insufficient to characterise the event as a new risk";
• or "the cumulative evidence supports the existing risk characterisation."

The wording should reflect the actual strength of the evidence.

22. From Safety Finding to Regulatory Implication

A strong PSUR evaluation connects scientific findings to regulatory consequences.

The reasoning can be expressed as:

Safety finding
      ↓
Clinical interpretation
      ↓
Risk characterisation
      ↓
Benefit-risk impact
      ↓
Regulatory implication
      ↓
Action / continued monitoring / no action

The final step should not be predetermined.

Possible consequences may include further investigation, changes to product information, changes to risk-minimisation measures, additional pharmacovigilance activity or other regulatory action, depending on the evidence and applicable requirements.

23. Product Information Changes

Where the PSUR evaluation supports a change to product information, the report should make the relationship between the evidence and proposed change clear.

EMA's current procedural advice states that the MAH should consider the impact of the cumulative safety data and risk-benefit analysis on the marketing authorisation and draw conclusions on the need for changes to product information. For EU single-assessment procedures, proposed changes based on PSUR data are addressed through the relevant PSUR process rather than being duplicated in a separate variation procedure. citeturn0search0

The PSUR conclusion should therefore not merely state that a label change is proposed. It should explain the safety finding, its significance and the rationale for the proposed regulatory action.

24. RMP Implications

The PSUR and RMP serve different functions but should remain consistent.

A PSUR conclusion can identify information that has implications for:

• the safety specification;
• identified or potential risks;
• missing information;
• pharmacovigilance activities;
• or risk-minimisation measures.

The appropriate RMP action depends on the evidence and regulatory context. A PSUR should not automatically trigger an RMP update merely because a safety issue was discussed.

EMA's current procedural advice permits an RMP update to accompany a PSUR for certain centrally authorised product situations when the RMP changes are a direct result of data presented in the PSUR. citeturn0search0

The important control is traceability between the PSUR conclusion and any resulting RMP decision.

25. Risk-Minimisation Implications

Where evidence indicates that an existing risk-minimisation measure may be inadequate, the PSUR evaluation should make the issue visible.

Questions can include:

• Is the risk still appropriately characterised?
• Is the target population correctly defined?
• Is the measure reaching the intended audience?
• Is the measure understood and followed?
• Is there evidence of effectiveness?
• Has the risk changed in a way that makes the existing measure insufficient?

Detailed effectiveness methodology belongs in the relevant risk-management framework, but the PSUR should integrate important conclusions that affect the benefit-risk assessment.

26. Medical Review and Scientific Writing

The PSUR should demonstrate clinical reasoning, not merely statistical reporting.

Medical review should consider:

• plausibility of the reported event;
• clinical phenotype;
• alternative causes;
• disease-related background risk;
• concomitant medicines;
• temporal relationship;
• dechallenge and rechallenge where relevant;
• biological mechanism;
• and consistency with other evidence.

The extent of clinical assessment should be proportionate to the issue.

For important safety concerns, the reasoning should be sufficiently explicit that another qualified reviewer can understand why the conclusion was reached.

27. Case Narratives: Use Evidence, Do Not Flood the PSUR

The PSUR is not intended to reproduce all individual case narratives.

EMA's current procedural advice states that systematic inclusion of individual case listings and case narratives is not required, although clinical evaluations of individual cases should be included where integral to the scientific analysis of a signal or safety concern. Line listings may be requested for adverse reactions of special interest during assessment. citeturn0search0

The practical principle is therefore selective clinical evidence.

A case narrative belongs in the evaluation when it materially helps explain a safety concern, not simply because the case exists.

28. Quantitative Analysis

Quantitative methods can strengthen a PSUR evaluation, but numbers do not replace clinical interpretation.

Depending on the question, quantitative analyses may include:

• reporting-rate comparisons;
• observed-versus-expected analyses;
• epidemiological estimates;
• incidence or event-rate comparisons;
• subgroup analyses;
• or other appropriate statistical evaluations.

The method should match the question and data quality.

A statistical difference may be clinically unimportant, while a clinically important rare event may not produce a statistically compelling result.

29. Disproportionality and Reporting Rates

Disproportionality measures can help identify patterns within spontaneous-reporting data, but they should not be interpreted as direct measures of incidence or causality.

Similarly, changes in reporting rates may reflect changes in exposure, awareness, reporting behaviour or case ascertainment.

The PSUR should therefore describe what the analysis can support and avoid implying conclusions that exceed the method's limitations.

30. Subgroup Analysis

Important differences may be hidden by aggregate results.

Where clinically justified, the evaluation should consider relevant subgroups such as:

• age;
• sex;
• indication;
• dose;
• treatment duration;
• renal or hepatic impairment;
• concomitant treatment;
• pregnancy;
• and other clinically meaningful characteristics.

Subgroup analyses should be hypothesis-driven and interpreted cautiously, particularly where sample sizes are small or multiple comparisons are made.

31. Temporal Patterns

Time can be clinically informative.

The evaluation may need to consider:

• latency from exposure to event;
• duration of treatment;
• changes after discontinuation;
• recurrence on re-exposure;
• and whether the observed pattern is consistent with the proposed mechanism.

Temporal association is evidence, but temporal association alone does not establish causality.

32. Background Incidence and Confounding

Some events occur commonly in the underlying population or in the disease being treated.

A PSUR should therefore consider background incidence where it is relevant to interpreting an observed pattern.

Confounding can arise from:

• the underlying disease;
• age or comorbidity;
• concomitant medicines;
• healthcare utilisation;
• indication-related treatment selection;
• and other differences between exposed and comparison populations.

Ignoring these factors can exaggerate or underestimate the apparent contribution of the medicinal product.

33. Benefit-Risk Analysis Is Not a Numerical Score

A benefit-risk evaluation may use quantitative or semi-quantitative methods where appropriate, but GVP Module VII requires the assumptions, considerations, judgement or weighting supporting the conclusions to be clear when such methods are used. Economic considerations such as cost-effectiveness are not part of the PSUR benefit-risk evaluation. citeturn0search17

A numerical score should therefore never substitute for the clinical and regulatory reasoning underlying the conclusion.

34. Indication and Population Context

A medicinal product may have a different benefit-risk balance across indications or populations.

The evaluation should therefore avoid statements such as "the product remains favourable" when the evidence actually supports a more qualified conclusion.

Where appropriate, the writer should state which indication and population are being evaluated and explain important differences.

This is particularly important when:

• therapeutic benefit differs substantially;
• risk factors differ;
• exposure differs;
• alternatives differ;
• or the evidence base is not equally developed across populations.

35. When There Is Little New Information

A PSUR can contain little genuinely new information during a particular interval.

That does not justify a superficial conclusion.

The author should establish that the relevant evidence was considered and explain why the cumulative assessment remains unchanged.

A concise evaluation can be scientifically strong if it identifies the relevant evidence and clearly explains why no material change in benefit-risk has occurred.

36. When There Is a Large Amount of New Information

The opposite problem is an excessive amount of information.

A PSUR should not become unreadable because every available analysis is reproduced.

The writer should distinguish:

• information that changes the assessment;
• information that confirms existing knowledge;
• information that is supportive but non-decisive;
• and information that does not materially affect the conclusion.

The most important evidence should be visible in the integrated evaluation.

37. Writing the Conclusion

A strong conclusion normally contains four elements:

1. What was found.
2. What it means.
3. How certain the organisation is.
4. What should happen next.

For example, rather than writing:

No change in the benefit-risk balance was identified.

A stronger conclusion would explain which important new information was evaluated, how it affected the existing risk characterisation, what cumulative evidence was considered and why the overall balance remains unchanged.

The precise wording must of course reflect the actual evidence.

38. Avoiding Overstatement

Scientific writing should not claim more than the evidence supports.

Avoid converting:

• association into causality;
• absence of reports into absence of risk;
• lack of statistical significance into proof of no effect;
• lack of a validated signal into proof that no safety concern exists;
• or a positive benefit-risk conclusion into proof that no future change is possible.

The PSUR is a periodic assessment precisely because knowledge evolves.

39. Avoiding Understatement

The opposite error is excessive caution that prevents a meaningful conclusion.

Statements such as "the evidence is inconclusive and further monitoring is required" should explain what is inconclusive, why it matters and what monitoring is intended to resolve.

Regulatory assessors need to understand the decision logic, not simply the existence of uncertainty.

40. The Reviewer Test

Before approval, a PSUR safety evaluation should pass a simple test:

Could an independent qualified reviewer reconstruct the scientific reasoning from the evidence presented and the controlled supporting records?

If not, the problem may not be the amount of data. It may be the connection between the data and the conclusion.

41. Inspection Perspective

An inspection may examine whether the MAH can demonstrate that its PSUR conclusions were supported by controlled evidence.

Relevant evidence may include:

• source-data extracts;
• reconciliation records;
• signal assessments;
• medical review records;
• statistical analyses;
• literature assessments;
• study reports;
• exposure estimates;
• benefit-risk working documents;
• review comments;
• and documented responses to material challenges.

The final PSUR should not contain conclusions that cannot be traced back to an underlying evidence base.

42. Common Writing Failures

Common weaknesses include:

• describing data without evaluating them;
• copying previous conclusions without reassessment;
• treating case counts as incidence;
• ignoring exposure changes;
• failing to discuss conflicting evidence;
• treating every signal as a confirmed risk;
• providing generic benefit statements;
• failing to evaluate differences between indications;
• hiding important uncertainty;
• and proposing actions without explaining their evidence basis.

These are practical failure modes and should not be presented as universal regulatory findings unless supported by an authoritative inspection source.

43. Practical Writing Checklist

Before finalising the safety evaluation, ask:

44. QPPV Perspective

The QPPV should be able to challenge a PSUR conclusion rather than merely confirm that the report has been completed.

Useful questions include:

• What changed since the previous PSUR?
• Which finding could most plausibly change the benefit-risk balance?
• What evidence argues against that interpretation?
• What uncertainty remains?
• Are the conclusions consistent with signal-management activity?
• Are they consistent with the RMP?
• Are proposed actions proportionate?
• Can every material conclusion be defended from the evidence?

This is part of effective PV oversight because the PSUR is an integrated assessment of the PV system's accumulated knowledge.

45. Key Takeaways

• A PSUR is an assessment, not a data dump.
• Critical evaluation connects evidence to clinical meaning and regulatory consequence.
• New information must be interpreted against cumulative knowledge.
• Risks and benefits should be characterised separately before being integrated.
• Benefit-risk evaluation is indication- and population-specific.
• Evidence strength, limitations and uncertainty should be explicit.
• Conflicting evidence should be analysed rather than averaged away.
• Quantitative methods support, but do not replace, clinical judgement.
• Product-information and RMP implications should be traceable to the PSUR evidence.
• Conclusions should state what was found, what it means, how certain the assessment is and what action follows.
• The final reasoning should be reconstructable from controlled supporting evidence.

References

1. European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) Module VII — Periodic safety update report (Rev. 1).
2. European Medicines Agency. Periodic safety update reports (PSURs), current preparation and assessment guidance.
3. European Medicines Agency. GVP Module VII Explanatory Note, current revision.
4. Commission Implementing Regulation (EU) No 520/2012 on the performance of pharmacovigilance activities, as amended.
5. European Parliament and Council. Directive 2001/83/EC, as amended.
6. Regulation (EC) No 726/2004, as amended.
7. ICH E2C(R2). Periodic Benefit-Risk Evaluation Report (PBRER).

Regulatory Note

This article is an educational explanation of the scientific evaluation required in an EU PSUR. It does not replace current GVP Module VII, applicable EU legislation, EMA procedural guidance, the EURD framework, PSUR/PSUSA procedural documents or an organisation's approved procedures.

Regulatory requirements and guidance may change. Before preparing or assessing a PSUR, the current applicable regulatory documents and procedural requirements should be verified.

Practical failure modes and inspection considerations in this article are illustrative unless an authoritative inspection source is specifically identified.

46. Practical Scenario: A Known Risk With Increasing Reports

An established adverse reaction shows a substantial increase in reports during the reporting interval.

The first conclusion should not be that the risk has increased by the same proportion.

The evaluation should examine:

• changes in exposure;
• reporting behaviour;
• changes in awareness;
• case completeness;
• seriousness and outcomes;
• indication or population changes;
• relevant concomitant medicines;
• literature and study evidence;
• and whether the clinical phenotype has changed.

The conclusion may ultimately be that the increase represents increased reporting rather than increased underlying risk. Alternatively, the cumulative evidence may support a genuine change in risk characterisation.

The important point is that the conclusion follows the analysis.

47. Practical Scenario: A Rare Serious Event

A small number of cases describe a rare but potentially fatal event.

The numerical count may be insufficient to establish frequency, but the clinical seriousness can justify detailed evaluation.

The author should examine the clinical characteristics, alternative explanations, temporal relationship, dechallenge or rechallenge where available, biological plausibility and supporting evidence from other sources.

A low number of reports should not automatically result in a conclusion that the issue is unimportant.

48. Practical Scenario: Positive Study With Limited Generalisability

A new study demonstrates a clinically meaningful benefit, but the study population differs substantially from the population treated in routine practice.

The evaluation should describe the strength of the finding while considering whether it can reasonably be generalised.

A statistically persuasive result may have limited relevance if the studied population, treatment duration or clinical setting differs materially from the authorised use.

The benefit assessment should therefore distinguish internal evidence strength from applicability to the relevant patient population.

49. Practical Scenario: Safety Signal Without Immediate Regulatory Action

A potential signal has been identified and remains under evaluation.

The PSUR should not prematurely characterise the signal as an established risk. It should describe the available evidence, the limitations and the current assessment status.

If continued evaluation is appropriate, the report should make clear what is known and what remains uncertain.

The absence of immediate regulatory action does not mean the issue has been dismissed.

50. Practical Scenario: Evidence Weakens a Previous Concern

A concern identified in an earlier reporting period is evaluated using additional cases, studies and literature.

The cumulative evidence may fail to confirm the original hypothesis.

A strong PSUR should say so explicitly and explain why the newer evidence changes the previous interpretation.

The report should not preserve an obsolete concern merely because it appeared in an earlier PSUR.

51. Practical Scenario: Different Benefit-Risk Balances Across Indications

A medicinal product has two authorised indications.

In the first indication, the product provides substantial benefit where few alternatives exist. In the second, several effective alternatives are available and the therapeutic benefit is smaller.

The same safety risk can therefore have different implications for the two indications.

The benefit-risk evaluation should make the distinction visible rather than applying a single undifferentiated conclusion to both indications. GVP Module VII specifically recognises that benefit-risk balance is specific to an indication and population. citeturn0search17

52. Practical Scenario: No Material New Information

Suppose the reporting interval produces no new safety signal, no important new efficacy information and no meaningful change in exposure or risk-management measures.

The conclusion may be concise, but it should still demonstrate that the relevant evidence was evaluated.

A defensible conclusion is not simply:

No change.

It should explain why the cumulative evidence remains consistent with the existing assessment.

53. Practical Scenario: Important Information Arrives After the DLP

A significant safety finding becomes available shortly after the DLP.

The organisation should first determine whether the information requires action through another immediate or regulatory pathway. The fact that the DLP has passed does not make important safety information irrelevant.

The information should then be handled in accordance with the applicable PSUR requirements and controlled procedures.

The PSUR timeline must never become a reason to delay necessary safety action.

54. Practical Scenario: A Proposed Label Change

The PSUR evaluation identifies evidence supporting a change to the product information.

The report should connect:

New evidence
     ↓
Risk interpretation
     ↓
Benefit-risk consequence
     ↓
Proposed wording / action
     ↓
Regulatory pathway

The proposed change should be traceable to the evidence and should be consistent with the applicable EU PSUR procedure.

55. Inspection Finding Pattern: Conclusions Not Supported by Evidence

One of the most important practical weaknesses is a disconnect between the evidence and the conclusion.

Examples include:

• a conclusion that the benefit-risk balance remains favourable without analysing important new safety information;
• a statement that reporting has increased without considering exposure;
• a claim that a signal is absent because no formal signal was validated;
• or a proposed action without a documented scientific rationale.

These examples illustrate inspection risk; they should not be represented as quotations or specific regulatory findings unless an authoritative inspection report is cited.

56. Inspection Finding Pattern: Copy-Forward Evaluation

Copying the previous PSUR's conclusions into a new reporting cycle without reassessing the cumulative evidence creates a significant governance weakness.

A previous conclusion can remain valid, but the new report should demonstrate that it has been reconsidered in light of the current reporting interval and cumulative evidence.

Evidence of meaningful reassessment can include:

• updated data tables;
• changed exposure estimates;
• current signal-management conclusions;
• new study findings;
• literature review;
• and documented scientific review.

57. Inspection Finding Pattern: Unclear Benefit-Risk Rationale

A conclusion that simply states "benefit-risk remains positive" may be inadequate if important new information was presented elsewhere in the report.

The integrated analysis should explain why those findings do or do not change the balance.

The rationale should be specific to the relevant indication and population rather than relying on a generic statement about the product.

58. Inspection Finding Pattern: Unsupported Statistical Interpretation

A PSUR may contain sophisticated statistical analyses that are not interpreted appropriately.

Potential weaknesses include:

• treating disproportionality as incidence;
• treating statistical significance as clinical importance;
• ignoring multiple comparisons;
• failing to address confounding;
• or presenting a confidence interval without explaining its clinical meaning.

The solution is not to avoid quantitative analysis. It is to connect the method and its limitations to the clinical question.

59. Inspection Finding Pattern: Evidence From Different Processes Does Not Reconcile

A PSUR may draw on multiple PV processes whose conclusions are inconsistent.

For example:

• signal management may describe a concern as ongoing;
• the RMP may not reflect the concern;
• the PSUR may describe no change;
• and safety communication records may show escalation.

Such discrepancies do not necessarily mean one process is wrong, but the organisation should be able to explain the differences.

Cross-process reconciliation is therefore an important governance control.

60. Inspection Finding Pattern: Important Uncertainty Is Hidden

A report can be technically accurate while still presenting an overly confident conclusion.

If exposure estimates are highly uncertain, a study has major methodological limitations or important cases lack follow-up, the limitation should be visible where it affects interpretation.

The objective is not to weaken every conclusion. It is to ensure that the confidence expressed matches the evidence available.

61. Inspection-Ready Evidence Trail

A defensible PSUR process should allow reconstruction of the major scientific decisions.

A simplified evidence chain is:

Source data
    ↓
Controlled analysis
    ↓
Medical / scientific interpretation
    ↓
Draft conclusion
    ↓
Cross-functional review
    ↓
Final conclusion
    ↓
Regulatory action / rationale for no action

The organisation should know where the supporting evidence for each material conclusion resides and who reviewed it.

62. QPPV Oversight of the Safety Evaluation

The QPPV's role is not to rewrite every paragraph of the PSUR, but to provide effective oversight of the pharmacovigilance system and its conclusions.

Questions appropriate for QPPV review include:

• Does the report reflect the current safety profile?
• Are important new findings appropriately evaluated?
• Are the conclusions consistent with signal management?
• Are important risks and uncertainties visible?
• Is the benefit-risk conclusion clinically defensible?
• Are proposed actions proportionate?
• Are there discrepancies with the RMP or safety communications that require explanation?
• Can the organisation demonstrate the evidence underlying the conclusion?

63. Governance of Material Changes During Review

The scientific evaluation may change during internal or regulatory review.

Material changes should be controlled so that the final conclusion remains traceable to the evidence available and the rationale for the change is understood.

A controlled review process should distinguish editorial corrections from substantive scientific changes.

Substantive changes may require reassessment of linked sections, conclusions, proposed actions and regulatory implications.

64. Final Review Questions

Before the PSUR is approved, the responsible team should be able to answer:

Evidence

• Have all relevant evidence sources been considered?
• Are important limitations visible?
• Has cumulative evidence been assessed?

Risk

• Are important risks appropriately characterised?
• Have changes from the previous assessment been identified?
• Are important new signals or concerns appropriately addressed?

Benefit

• Are the key benefits still appropriately characterised?
• Is their clinical relevance clear?
• Are indication and population differences addressed?

Integration

• Does the benefit-risk analysis actually integrate the evidence?
• Is the conclusion supported by the preceding assessment?
• Are uncertainties and alternative explanations addressed?

Action

• Are proposed changes to product information justified?
• Are RMP implications considered where relevant?
• Are further investigations or risk-minimisation actions justified?
• If no action is proposed, is the rationale explicit?

Governance

• Has appropriate medical and cross-functional review occurred?
• Can the evidence trail be reconstructed?
• Can the QPPV defend the final conclusion?

65. What a High-Quality PSUR Safety Evaluation Looks Like

A high-quality safety evaluation has several characteristics.

It is critical, because it questions what the evidence actually demonstrates.

It is cumulative, because the current assessment is not isolated from previous knowledge.

It is clinical, because statistical findings are interpreted in terms of patients and treatment context.

It is transparent, because strengths, limitations and uncertainty are visible.

It is integrated, because risks and benefits are brought together rather than evaluated in isolation.

It is proportionate, because effort is concentrated on findings capable of changing the regulatory conclusion.

It is traceable, because material conclusions can be linked to controlled evidence.

And it is action-oriented, because the assessment makes clear what should happen as a consequence of the evidence or why no additional action is warranted.

66. Key Takeaways

Writing the safety evaluation is the point at which the PSUR becomes a scientific regulatory assessment rather than a collection of pharmacovigilance outputs.

The author should move systematically from evidence to interpretation, from interpretation to risk and benefit characterisation, and from those assessments to an integrated benefit-risk conclusion.

The strongest evaluations do not avoid uncertainty. They explain it. They do not treat every observation as a new risk. They distinguish hypotheses from established findings. They do not merely repeat previous conclusions. They demonstrate current reassessment.

Most importantly, the final conclusion should be proportionate to the evidence and traceable to the underlying scientific record.

References

1. European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) Module VII — Periodic safety update report (Rev. 1).
2. European Medicines Agency. Periodic safety update reports (PSURs), current preparation and assessment guidance.
3. European Medicines Agency. GVP Module VII Explanatory Note, current revision.
4. Commission Implementing Regulation (EU) No 520/2012 on the performance of pharmacovigilance activities, as amended.
5. European Parliament and Council. Directive 2001/83/EC, as amended.
6. Regulation (EC) No 726/2004, as amended.
7. ICH E2C(R2). Periodic Benefit-Risk Evaluation Report (PBRER).

Regulatory Note

This article is an educational explanation of writing the safety evaluation of an EU PSUR. It does not replace current GVP Module VII, applicable EU legislation, EMA procedural guidance, PSUR/PSUSA guidance or an organisation's approved procedures.

Regulatory requirements and guidance may change. Before preparing or assessing a PSUR, the current applicable regulatory documents and procedural requirements should be verified.

The inspection-finding patterns and practical scenarios in this article are illustrative unless an authoritative inspection source is specifically identified.