Writing the Signal Evaluation Section in a PBRER

Introduction

Signal evaluation is the scientific centre of the Periodic Benefit-Risk Evaluation Report (PBRER). Every important safety observation identified during the reporting interval ultimately converges within this section, where available evidence is critically evaluated to determine whether the medicinal product's recognised safety profile should change.

Unlike summary tabulations or case listings, signal evaluation is not primarily concerned with describing data. Instead, it explains how diverse sources of pharmacovigilance information have been assessed, integrated and interpreted to reach scientifically defensible conclusions regarding potential safety concerns.

For many reviewers, this section provides the clearest demonstration of the quality of an organisation's pharmacovigilance system because it illustrates how emerging observations progress from initial detection through structured scientific evaluation and ultimately influence the ongoing benefit-risk assessment.

Accordingly, this section should present well-reasoned medical assessments rather than collections of case summaries.

Regulatory Basis

Signal evaluation within the PBRER is described principally in ICH E2C(R2).

Within the European Union, signal management is governed by Good Pharmacovigilance Practices (GVP) Module IX, while presentation of completed signal evaluations within the PBRER is described in GVP Module VII.

Although the PBRER should summarise important completed and ongoing signal evaluations relevant to the reporting interval, it is not intended to replace the organisation's signal management system.

Instead, it provides a structured regulatory summary of the scientific evaluations that contribute to the evolving understanding of the medicinal product's safety profile.

Why This Section Exists

Every medicinal product generates numerous safety observations throughout its lifecycle.

Only a small proportion ultimately influence the recognised benefit-risk profile.

The purpose of signal evaluation is to distinguish observations requiring regulatory attention from those adequately explained by existing knowledge, alternative causes or insufficient supporting evidence.

This section therefore demonstrates the scientific reasoning underpinning important safety decisions rather than merely documenting that signal management activities occurred.

Regulatory reviewers use this section to understand not only the conclusions reached by the Marketing Authorisation Holder but also the quality of the scientific reasoning supporting those conclusions.

Relationship to Other Sections of the PBRER

Signal evaluation occupies a central position within the PBRER.

Information presented earlier in the report provides the evidence upon which signal evaluations depend.

For example:

• changes to the Reference Safety Information establish the recognised safety profile;

• estimated patient exposure provides essential clinical context;

• summary tabulations identify reporting patterns requiring review;

• findings from clinical trials, non-interventional studies and scientific literature contribute additional evidence.

The conclusions reached within the Signal Evaluation section subsequently influence:

• characterisation of important identified risks;

• characterisation of important potential risks;

• benefit evaluation;

• integrated benefit-risk analysis;

• recommendations regarding risk minimisation measures;

• regulatory conclusions presented later within the report.

Signal evaluation should therefore be viewed as the scientific bridge linking safety observations with regulatory decision-making.

Writing Tip

Readers should understand not only the conclusion reached for each signal but also the reasoning that led to that conclusion.

Principles of Signal Evaluation

Signal evaluation is not a search for evidence supporting a predetermined conclusion. It is a structured scientific assessment that considers all available evidence, including findings that support, weaken or refute a potential association between a medicinal product and an observed safety issue.

The objective is to determine whether the totality of available evidence justifies a change in the current understanding of the medicinal product's safety profile.

Consequently, every signal evaluation should remain objective, balanced and scientifically reproducible.

Authors should avoid selectively presenting evidence that supports only one conclusion.

Instead, the evaluation should demonstrate that alternative explanations have been considered and that the final conclusion reflects the overall weight of the available evidence.

Defining the Safety Question

Every signal evaluation should begin with a clearly defined medical question.

Examples include:

• Does the medicinal product increase the risk of acute pancreatitis?

• Is the observed hepatic injury causally related to treatment?

• Does long-term exposure increase the incidence of malignancy?

• Is the reported arrhythmia consistent with the known pharmacology of the medicinal product?

Clearly defining the question focuses the subsequent evaluation and prevents unnecessary discussion of unrelated findings.

Well-defined questions also facilitate consistent interpretation by regulatory reviewers.

Sources of Evidence

Signal evaluation rarely depends upon a single source of information.

Instead, aggregate physicians integrate evidence obtained from multiple complementary sources.

These may include:

• individual case safety reports;

• cumulative case review;

• interval case review;

• summary tabulations;

• disproportionality analyses;

• clinical trials;

• post-authorisation safety studies (PASS);

• observational studies;

• patient registries;

• scientific literature;

• regulatory authority assessments;

• class effects;

• non-clinical studies where relevant;

• pharmacological and mechanistic evidence.

Each evidence source contributes differently to the overall evaluation.

The strength of the assessment depends not upon the number of sources consulted but upon the quality, consistency and relevance of the available evidence.

Scientific Principle

Signal evaluation should integrate all relevant evidence. No individual evidence source should be regarded as definitive in isolation.

Assessing the Quality of the Evidence

Before interpreting the available data, reviewers should critically evaluate the quality of the evidence itself.

Important considerations include:

• completeness of case documentation;

• diagnostic certainty;

• temporal relationship;

• biological plausibility;

• alternative medical explanations;

• concomitant medications;

• underlying disease;

• dechallenge information;

• rechallenge information;

• dose-response relationship;

• consistency across independent reports.

Incomplete or poorly documented cases generally contribute less to the overall assessment than well-documented reports supported by objective clinical evidence.

Similarly, findings reproduced consistently across several independent evidence sources generally provide stronger support than isolated observations.

Integrating Multiple Evidence Sources

One of the defining characteristics of expert signal evaluation is integration of evidence.

For example, an apparent increase in spontaneous reports may initially suggest a potential safety concern.

However, the interpretation may change substantially after considering:

• patient exposure;

• published epidemiological studies;

• findings from ongoing PASS;

• clinical trial safety data;

• regulatory assessments;

• biological plausibility;

• class effects.

Conversely, absence of a statistical signal does not necessarily exclude an important clinical association when strong mechanistic evidence and well-documented individual cases exist.

The aggregate physician should therefore evaluate the totality of available evidence rather than assigning excessive importance to any single information source.

Considering Alternative Explanations

Scientific evaluation requires active consideration of explanations other than a causal relationship with the medicinal product.

Alternative explanations may include:

• the underlying disease;

• concomitant medications;

• medical history;

• age-related risk;

• confounding by indication;

• reporting bias;

• stimulated reporting;

• diagnostic bias;

• random variation;

• coding artefacts.

Failure to consider plausible alternative explanations may lead to overinterpretation of weak safety observations.

Conversely, dismissal of genuine signals without adequate evaluation may delay important regulatory action.

Balanced consideration of competing explanations is therefore central to robust pharmacovigilance practice.

Medical Review Consideration

Ask not only "What evidence supports this signal?" but also "What evidence argues against it?" and "Could another explanation account for the observed findings?" Robust signal evaluations address all three questions explicitly.

Applying Causality Principles During Signal Evaluation

Signal evaluation requires assessment of whether the available evidence supports a causal association between the medicinal product and the observed safety concern. Although no single framework is mandated for every evaluation, experienced aggregate physicians routinely apply well-established scientific principles when interpreting the totality of available evidence.

These principles should not be regarded as rigid decision rules. Instead, they provide a structured approach for evaluating the strength, consistency and biological credibility of the available data.

The importance of each consideration varies according to the nature of the medicinal product, the adverse event under evaluation and the quality of the available evidence.

Biological Plausibility

One of the first questions during signal evaluation is whether the proposed association is biologically plausible.

Reviewers should consider:

• the known pharmacology of the medicinal product;
• mechanism of action;
• pharmacokinetic properties;
• pharmacodynamic effects;
• known class effects;
• receptor distribution;
• target organ toxicity.

A biologically plausible mechanism strengthens the credibility of a proposed association but is not essential for identifying a genuine adverse reaction.

Conversely, lack of an immediately obvious mechanism should not automatically lead to rejection of a signal.

Many recognised adverse drug reactions were identified before their underlying biological mechanisms became fully understood.

Temporal Relationship

The timing of exposure in relation to the adverse event is fundamental to causal assessment.

Important questions include:

• Did exposure precede the adverse event?

• Is the time to onset compatible with current pharmacological knowledge?

• Does the temporal pattern remain consistent across multiple independent reports?

Events occurring before treatment initiation or after biologically implausible intervals are generally less supportive of causality, although exceptions exist for delayed toxicities and immunological mechanisms.

Consistency of temporal relationships across independent reports frequently strengthens the overall evaluation.

Dechallenge and Rechallenge

Information regarding dechallenge and rechallenge may provide valuable supporting evidence.

Reviewers should evaluate:

• whether the adverse event improved following discontinuation;

• whether recurrence occurred after re-exposure;

• whether alternative explanations exist.

Positive rechallenge findings may substantially strengthen a causal hypothesis.

However, absence of rechallenge information is common in routine pharmacovigilance practice and should not be interpreted as evidence against causality.

Similarly, intentional rechallenge is often unethical following serious adverse reactions and should therefore not be expected.

Dose-Response Relationship

Where appropriate, reviewers should consider whether the available evidence suggests a relationship between exposure and the occurrence or severity of the adverse event.

Examples include:

• increasing frequency with higher doses;

• earlier onset following higher exposure;

• cumulative toxicity during prolonged treatment;

• improvement following dose reduction.

Not every adverse reaction demonstrates a dose-response relationship.

Idiosyncratic reactions, hypersensitivity reactions and many immune-mediated adverse events may occur independently of dose.

Consequently, absence of a dose-response relationship should be interpreted within the clinical context rather than regarded as evidence against causality.

Consistency Across Evidence Sources

Confidence in a signal generally increases when similar observations are identified independently across multiple sources of evidence.

Examples include:

• spontaneous reports;

• clinical trials;

• observational studies;

• patient registries;

• published literature;

• regulatory assessments;

• independent databases.

Consistency does not require identical findings.

Rather, reviewers should consider whether different evidence sources describe observations that are compatible with the same underlying safety concern.

Alternative Explanations

Every signal evaluation should actively explore competing explanations.

Examples include:

• progression of the underlying disease;

• concomitant medicinal products;

• pre-existing medical conditions;

• diagnostic uncertainty;

• confounding by indication;

• reporting bias;

• surveillance bias;

• random variation.

The objective is not simply to identify alternative explanations but to determine whether they provide a more convincing explanation than a causal association with the medicinal product.

Failure to consider plausible alternatives may weaken the scientific credibility of the evaluation.

Contradictory Evidence

Signal evaluation should present all relevant evidence, including findings that do not support the proposed association.

Examples include:

• large epidemiological studies showing no increased risk;

• negative clinical trial findings;

• mechanistic evidence arguing against causality;

• well-designed observational studies with conflicting conclusions.

Discussion of contradictory evidence demonstrates scientific objectivity and allows regulators to understand how competing information was considered before reaching the final conclusion.

Balanced evaluations generally carry greater scientific credibility than reports presenting only supportive evidence.

The Totality of the Evidence

No single observation establishes causality.

Similarly, absence of any individual consideration does not exclude a causal association.

Expert signal evaluation therefore depends upon integration of all available evidence.

Reviewers should weigh the strengths and limitations of individual evidence sources, assess their consistency and determine whether the overall body of evidence justifies modification of the recognised safety profile.

The conclusion should reflect the quality, consistency and biological credibility of the complete evidence base rather than any single study, statistical analysis or individual case report.

Scientific Principle

Signal evaluation is an exercise in scientific judgement rather than a mechanical scoring process. The strength of the conclusion depends upon the totality of the available evidence, interpreted objectively, transparently and in the context of current medical and pharmacological knowledge.

Writing Individual Signal Evaluations

Each signal evaluation presented within the PBRER should tell a coherent scientific story. The objective is not to document every activity undertaken during signal management but to explain how the available evidence was evaluated and how the final scientific conclusion was reached.

A well-written signal evaluation allows the reader to understand the question being investigated, the evidence considered, the strengths and limitations of that evidence, the reasoning applied by the reviewer and the impact of the findings upon the medicinal product's benefit-risk profile.

The level of detail should remain proportionate to the importance of the signal. Signals with significant regulatory or clinical implications generally require more comprehensive discussion than signals that are rapidly excluded following review of the available evidence.

Begin with the Safety Question

Every signal evaluation should begin by clearly defining the medical question.

For example:

• Is acute pancreatitis causally associated with the medicinal product?

• Does long-term treatment increase the risk of interstitial lung disease?

• Is the observed increase in hepatic injury consistent with the recognised safety profile?

A clearly stated question establishes the scope of the evaluation and helps readers understand the purpose of the subsequent discussion.

Avoid beginning immediately with case summaries or statistical outputs without first explaining what is being evaluated.

Explain Why the Signal Was Evaluated

Readers should understand why the signal entered formal evaluation.

The source of the signal may include:

• routine signal detection activities;

• review of cumulative case reports;

• disproportionate reporting analyses;

• scientific literature;

• regulatory authority requests;

• findings from clinical trials;

• post-authorisation safety studies (PASS);

• observational studies;

• class-wide safety concerns;

• internal medical review.

The origin of the signal provides important context but should not be interpreted as evidence supporting or refuting causality.

Signals originating from different sources should be evaluated using the same scientific standards.

Summarise the Evidence Reviewed

Rather than describing every document individually, authors should provide a concise overview of the evidence considered during the evaluation.

This may include:

• cumulative case review;

• interval cases;

• medically confirmed cases;

• literature publications;

• clinical trial safety analyses;

• epidemiological studies;

• PASS findings;

• non-clinical evidence where relevant;

• class effects;

• previous regulatory assessments.

The objective is to demonstrate that the evaluation considered all relevant evidence rather than relying upon a single source of information.

Present the Medical Assessment

The medical assessment forms the scientific core of the signal evaluation.

This discussion should integrate the available evidence into a balanced clinical interpretation.

Where appropriate, the reviewer should discuss:

• clinical characteristics of reported cases;

• diagnostic certainty;

• temporal relationships;

• biological plausibility;

• dechallenge and rechallenge information;

• dose-response observations;

• consistency across evidence sources;

• important strengths of the evidence;

• important limitations and uncertainties.

The emphasis should remain on scientific interpretation rather than repetition of case narratives or statistical summaries.

Discuss Alternative Explanations

Robust signal evaluations actively consider explanations other than a causal association with the medicinal product.

Potential alternatives include:

• progression of the underlying disease;

• concomitant therapies;

• recognised complications of treatment;

• confounding by indication;

• demographic characteristics;

• co-morbidities;

• reporting bias;

• diagnostic uncertainty.

Alternative explanations should be analysed objectively rather than dismissed superficially.

Where uncertainty remains, it should be acknowledged explicitly.

Explain the Medical Conclusion

Every evaluation should conclude with a clear scientific judgement.

Possible conclusions include:

• the available evidence supports a causal association;

• the available evidence does not currently support a causal association;

• additional evidence is required before reaching a conclusion;

• the recognised safety profile remains unchanged;

• modification of the recognised safety profile is justified.

The conclusion should arise naturally from the preceding discussion rather than appearing as an unsupported opinion.

Importantly, the medical conclusion should distinguish between absence of evidence and evidence of absence. Failure to demonstrate a causal relationship during the current reporting interval does not necessarily exclude the possibility that future evidence may alter the assessment.

Describe Regulatory and Pharmacovigilance Consequences

Where the evaluation results in action, the consequences should be summarised briefly.

Examples include:

• closure of the signal;

• continuation of signal monitoring;

• revision of the Company Core Safety Information (CCSI);

• updates to the Reference Safety Information (RSI);

• revisions to product information;

• implementation of additional pharmacovigilance activities;

• introduction or modification of risk minimisation measures;

• requests for additional clinical or epidemiological studies;

• referral to regulatory authorities.

Where no regulatory action is considered necessary, this should also be stated clearly together with the scientific rationale.

Writing Tip

Every signal evaluation should answer five questions:

What was the concern?

Why was it evaluated?

What evidence was reviewed?

What conclusion was reached?

What changed as a result?

If a reviewer can answer those five questions after reading the evaluation, the section has usually achieved its objective.

Writing Different Types of Signal Evaluations

Not every signal evaluation should be written in the same manner.

The structure, level of detail and scientific emphasis should reflect the maturity of the available evidence, the regulatory status of the signal and its potential impact upon the benefit-risk profile of the medicinal product.

Applying the same template to every signal frequently results in unnecessary repetition for straightforward evaluations and insufficient discussion for complex safety issues.

Instead, authors should adapt the narrative to the scientific question being addressed while maintaining a consistent overall structure.

Newly Validated Signals

Newly validated signals generally require the most comprehensive discussion because regulators may be reviewing the issue for the first time within the context of the PBRER.

The evaluation should explain:

• why the signal was validated;
• the evidence available at the Data Lock Point;
• important strengths and limitations of the evidence;
• alternative explanations considered;
• remaining uncertainties;
• whether the recognised safety profile should change.

Where the evaluation remains incomplete, the discussion should explain what additional information is required rather than speculating about future conclusions.

Ongoing Signal Evaluations

Signals under active evaluation require a balanced discussion.

The report should summarise:

• the current evidence;
• important findings obtained during the reporting interval;
• activities completed;
• outstanding questions;
• planned pharmacovigilance activities where relevant.

Authors should avoid implying that a conclusion has already been reached when scientific evaluation remains ongoing.

Equally, ongoing evaluation should not be presented as evidence supporting causality.

The purpose is to inform regulators of scientific progress while accurately reflecting the current state of knowledge.

Refuted Signals

Refuted signals deserve careful scientific explanation.

Simply stating that a signal has been closed provides little educational or regulatory value.

Instead, the evaluation should explain why the available evidence no longer supports the proposed association.

Examples include:

• alternative medical explanations identified;

• inconsistent clinical findings;

• absence of biological plausibility;

• negative epidemiological studies;

• contradictory clinical trial evidence;

• diagnostic uncertainty;

• reporting artefacts;

• duplicate cases.

Importantly, closure of a signal does not imply that the underlying adverse event can never become a signal again if new evidence emerges.

The conclusion should therefore remain proportionate to the currently available evidence.

Confirmed Signals

Where evaluation supports a causal association, the report should explain how the recognised safety profile has changed.

The discussion should address:

• evidence supporting the conclusion;

• impact upon benefit-risk evaluation;

• changes to the Company Core Safety Information (CCSI) or Reference Safety Information (RSI), where applicable;

• product information updates;

• additional pharmacovigilance activities;

• risk minimisation measures;

• remaining uncertainties.

The emphasis should remain on the scientific reasoning supporting the conclusion rather than merely documenting the resulting regulatory actions.

Signals Arising from Regulatory Authority Requests

Some signal evaluations are initiated following requests from regulatory authorities rather than routine internal signal detection.

Examples include requests from PRAC, national competent authorities, the FDA, MHRA or other regulatory agencies.

These evaluations should be written using the same scientific principles as internally generated evaluations.

The origin of the request should provide context but should not influence the objectivity of the medical assessment.

The scientific conclusion should remain based upon the totality of the available evidence rather than the source of the request.

Class Effect Evaluations

Evaluation of potential class effects presents additional challenges.

Evidence relating to structurally or pharmacologically related medicinal products may strengthen or weaken the plausibility of a proposed association.

However, observations affecting one member of a therapeutic class should not automatically be assumed to apply to all related products.

Authors should discuss:

• similarities in pharmacology;

• differences in molecular structure where relevant;

• available evidence from related medicinal products;

• biological plausibility;

• product-specific evidence.

The emphasis should remain on evaluating the medicinal product covered by the PBRER while considering relevant class information objectively.

Signals Supported by Multiple Evidence Sources

Some evaluations involve substantial evidence from spontaneous reports, clinical trials, epidemiological studies, published literature, PASS and regulatory assessments.

In these situations, the discussion should integrate the evidence into a coherent scientific narrative rather than presenting each source independently.

The objective is to explain how the various evidence sources complement or contradict one another and how they collectively influence the medical conclusion.

Readers should understand the overall scientific reasoning rather than simply reviewing separate summaries of individual studies.

Writing Tip

Adapt the depth of discussion to the scientific importance of the signal.

A minor signal closed following review of several poorly documented case reports should not receive the same level of discussion as a signal that may alter the recognised benefit-risk profile of the medicinal product.

Avoiding Cognitive Bias During Signal Evaluation

Signal evaluation is fundamentally a process of scientific judgement. Although structured methodologies, standard operating procedures and multidisciplinary review reduce subjectivity, every medical reviewer remains vulnerable to cognitive biases that may influence interpretation of the available evidence.

Experienced aggregate physicians recognise these biases and actively attempt to minimise their influence by evaluating all relevant evidence objectively, documenting alternative explanations and ensuring that conclusions remain proportionate to the available data.

The purpose of this section is not to apply formal psychological theory but to highlight common reasoning errors that may affect pharmacovigilance decision-making.

Confirmation Bias

Confirmation bias occurs when reviewers preferentially seek or emphasise evidence supporting an existing hypothesis while giving insufficient attention to contradictory findings.

For example, after identifying several convincing cases of hepatotoxicity, a reviewer may unconsciously focus on additional supportive cases while overlooking well-conducted epidemiological studies demonstrating no increased risk.

Balanced signal evaluation requires equal consideration of evidence supporting and challenging the proposed association.

Authors should therefore discuss contradictory evidence explicitly and explain why the overall conclusion remains scientifically justified.

Anchoring Bias

Early impressions may influence subsequent interpretation of evidence.

For example, if an initial case series strongly suggests a causal association, reviewers may continue interpreting later evidence through that perspective even when more robust data become available.

Conversely, early dismissal of a signal may delay recognition of an emerging safety concern.

Signal evaluations should therefore be reviewed periodically as new evidence becomes available rather than relying excessively upon earlier conclusions.

Availability Bias

Recent, memorable or clinically dramatic cases naturally attract attention.

Fatal events, unusual clinical presentations or cases receiving widespread publicity may appear more important than they truly are when considered within the complete cumulative dataset.

Reviewers should ensure that highly memorable cases do not receive disproportionate emphasis compared with the totality of available evidence.

Authority Bias

Scientific conclusions should not depend solely upon the source of the opinion.

Recommendations from regulatory authorities, external experts or senior colleagues deserve careful consideration but should still be evaluated critically against the available evidence.

Similarly, internally generated evaluations should be supported by objective scientific reasoning rather than organisational preference.

The credibility of a signal evaluation depends upon the quality of its evidence rather than the authority of the individual proposing the conclusion.

Publication Bias

Published literature represents an important component of signal evaluation.

However, studies demonstrating statistically significant associations are generally more likely to be published than studies reporting negative findings.

Consequently, reviewers should interpret published evidence within the broader context of spontaneous reports, clinical trials, observational studies and regulatory assessments rather than relying exclusively upon published literature.

Confounding and Alternative Explanations

Many apparent safety associations are influenced by factors unrelated to the medicinal product itself.

Examples include:

• confounding by indication;

• concomitant therapies;

• disease severity;

• age;

• co-morbidities;

• lifestyle factors.

Reviewers should actively evaluate these alternative explanations before concluding that the medicinal product is responsible for the observed event.

Failure to consider confounding appropriately may lead to incorrect attribution of risk.

Premature Closure

One of the most important errors during signal evaluation is reaching a conclusion before sufficient evidence has been reviewed.

Premature closure may result in:

• inappropriate dismissal of genuine signals;

• unnecessary continuation of weak signals;

• inadequate exploration of contradictory evidence;

• incomplete assessment of biological plausibility.

Experienced reviewers remain willing to revise their conclusions as new information becomes available.

Scientific conclusions should evolve with the evidence rather than remain fixed following the earliest stages of evaluation.

Maintaining Scientific Objectivity

The purpose of signal evaluation is not to prove or disprove causality.

Instead, it is to determine whether the totality of available evidence supports modification of the recognised safety profile of the medicinal product.

High-quality evaluations therefore demonstrate:

• balanced consideration of all relevant evidence;

• explicit discussion of uncertainties;

• acknowledgement of important limitations;

• objective assessment of competing explanations;

• conclusions proportionate to the available evidence.

These principles strengthen both the scientific credibility of the evaluation and the confidence of regulatory reviewers.

Medical Review Consideration

Throughout the evaluation, continually ask:

"If the available evidence suggested the opposite conclusion, would I recognise it?"

This simple question encourages reviewers to reassess assumptions, examine contradictory evidence objectively and minimise the influence of unconscious bias.

Presenting Signal Evaluations to Regulators

The objective of a signal evaluation is not simply to record that an assessment has been performed. Rather, it is to communicate the scientific reasoning supporting the conclusions reached by the Marketing Authorisation Holder in a manner that is transparent, balanced and reproducible.

Regulatory assessors should be able to understand what question was evaluated, which evidence was considered, how conflicting findings were interpreted and why the final conclusion was reached. A well-written evaluation allows the reviewer to follow the scientific reasoning without needing to reconstruct the assessment independently.

Consequently, clarity of presentation is as important as scientific accuracy.

Write for the Reviewer

Signal evaluations should be written with the regulatory reviewer in mind.

The reviewer should never have to guess:

• why the signal was evaluated;

• what evidence was reviewed;

• which evidence carried the greatest weight;

• why contradictory findings were discounted;

• what uncertainty remains;

• whether further pharmacovigilance activities are planned.

Each evaluation should answer these questions naturally as the discussion progresses.

Scientific reasoning should therefore be explicit rather than implied.

Present Evidence in a Logical Sequence

Although different organisations use different templates, effective signal evaluations generally follow a consistent scientific progression.

Begin by defining the medical question.

Then summarise the evidence reviewed.

Next, discuss the strengths and limitations of that evidence.

Consider competing explanations.

Finally, explain the medical conclusion and any resulting pharmacovigilance or regulatory actions.

This logical progression allows regulators to understand how the conclusion emerged from the available evidence.

Present Supporting and Contradictory Evidence Together

Balanced scientific writing requires discussion of evidence supporting and challenging the proposed association.

Presenting only favourable evidence weakens the credibility of the evaluation and may raise regulatory concerns regarding objectivity.

Where contradictory evidence exists, it should be described openly together with the reasons why the overall conclusion nevertheless remains appropriate.

Likewise, where the evidence remains inconclusive, uncertainty should be acknowledged explicitly rather than masked by overly definitive language.

Transparency generally strengthens regulatory confidence.

Use Tables Judiciously

Tables can improve clarity when summarising complex information.

Examples include:

• summary of key case characteristics;

• comparison of major epidemiological studies;

• overview of clinical trial findings;

• chronology of important regulatory milestones;

• comparison of evidence supporting and opposing a causal association.

However, tables should complement rather than replace scientific discussion.

The interpretation of the evidence should always remain within the narrative.

Readers should never be expected to infer the medical conclusion solely from numerical tables.

Summarising Individual Cases

Most signal evaluations do not require detailed presentation of every Individual Case Safety Report (ICSR).

Instead, authors should identify clinically important patterns observed across the available cases.

Where individual cases are particularly informative—for example because of positive rechallenge, compelling temporal association or unusual clinical features—they may be described briefly to illustrate important aspects of the evaluation.

The emphasis should remain on explaining why the cases contribute to the overall assessment rather than reproducing case narratives that have already been reviewed elsewhere.

Expressing Scientific Uncertainty

Uncertainty is an inherent feature of pharmacovigilance.

Accordingly, signal evaluations should communicate uncertainty honestly and precisely.

Appropriate expressions include:

• the available evidence suggests;

• the current data support;

• the evidence remains inconclusive;

• a causal association cannot currently be excluded;

• additional evidence is required.

Avoid unnecessarily definitive statements where important uncertainties remain.

Conversely, avoid excessive qualification when the available evidence clearly supports a conclusion.

The degree of certainty expressed should reflect the quality and consistency of the available evidence.

Avoid Overstating Statistical Findings

Statistical analyses contribute valuable evidence but should not be interpreted in isolation.

Disproportionality analyses, reporting ratios and other statistical methods identify observations that warrant medical review. They do not establish causality.

Similarly, statistically non-significant findings do not necessarily exclude clinically important associations, particularly for rare adverse reactions or products with limited exposure.

Authors should therefore integrate statistical findings with clinical, pharmacological and epidemiological evidence before reaching scientific conclusions.

Presenting Regulatory Actions

Where a signal evaluation results in regulatory or pharmacovigilance action, the discussion should distinguish clearly between the scientific conclusion and the actions that followed.

For example, revision of the Company Core Safety Information, implementation of additional risk minimisation measures or initiation of further studies should be presented as consequences of the evaluation rather than evidence supporting the evaluation itself.

This distinction helps reviewers understand the sequence of scientific assessment followed by regulatory decision-making.

Writing Tip

A regulatory reviewer should be able to explain your reasoning after reading the evaluation—even if they ultimately disagree with your conclusion. The purpose of the narrative is to make the scientific reasoning transparent, not to persuade the reader through selective presentation of evidence.

Common Mistakes

Preparation of the Signal Evaluation section requires scientific judgement, balanced interpretation and clear communication. Deficiencies identified during regulatory review are often related less to the scientific conclusion itself than to inadequate explanation of how that conclusion was reached.

The following mistakes are encountered repeatedly during internal quality review, regulatory assessment and pharmacovigilance inspections.

Beginning with the Conclusion

One of the most common weaknesses is presenting the final conclusion before explaining the evidence.

Statements such as:

"The signal was refuted."

or

"No causal relationship was identified."

provide little value unless the reader understands how those conclusions were reached.

The discussion should first present the scientific question, summarise the available evidence, consider alternative explanations and then explain why the final conclusion is justified.

Scientific reasoning should always precede scientific conclusions.

Treating All Evidence Equally

Not all evidence contributes equally to signal evaluation.

For example, a well-documented case demonstrating a positive rechallenge may provide greater support than several poorly documented spontaneous reports.

Similarly, a well-designed epidemiological study may carry greater weight than isolated anecdotal observations.

The evaluation should therefore discuss the quality of evidence, not merely the quantity.

Ignoring Contradictory Evidence

Balanced signal evaluations acknowledge evidence that does not support the proposed association.

Failure to discuss contradictory findings may create the impression of selective reporting.

Examples include:

• negative epidemiological studies;

• inconsistent clinical trial findings;

• alternative pharmacological explanations;

• conflicting regulatory assessments.

Readers should understand why the overall conclusion remains appropriate despite conflicting evidence.

Confusing Association with Causation

Detection of an association does not establish causality.

Likewise, statistical disproportionality alone does not confirm that the medicinal product caused the observed adverse event.

Authors should distinguish clearly between:

• observed reporting patterns;

• statistical observations;

• biological plausibility;

• causal assessment;

• regulatory conclusions.

Failure to separate these concepts frequently results in overstated conclusions.

Overinterpreting Small Numbers

Rare adverse events often involve very small numbers of reports.

Small numerical differences should be interpreted cautiously, particularly when exposure is limited or case quality is poor.

Conversely, small numbers should not automatically be dismissed when the events are clinically important, biologically plausible or consistently observed across several independent evidence sources.

The emphasis should remain on clinical significance rather than numerical magnitude.

Failing to Explain Remaining Uncertainty

Many signal evaluations remain inconclusive.

This is a scientifically acceptable outcome.

Authors should clearly explain:

• what uncertainties remain;

• why the available evidence is insufficient;

• what additional information would strengthen the evaluation;

• whether continued monitoring is appropriate.

Attempting to present uncertain evidence as definitive weakens scientific credibility.

Inconsistent Conclusions Across the PBRER

The conclusions presented within the Signal Evaluation section should align with subsequent discussions throughout the report.

For example:

• confirmed signals should be reflected within Characterisation of Risks where appropriate;

• important findings influencing benefit-risk should appear within the Integrated Benefit-Risk Evaluation;

• product information changes should correspond to discussions of the Reference Safety Information;

• additional pharmacovigilance activities should be reflected consistently elsewhere in the report.

Internal consistency strengthens regulatory confidence in the scientific integrity of the PBRER.

Inspection and Regulatory Assessment Considerations

Signal evaluation represents one of the principal areas examined during regulatory assessment because it demonstrates how the Marketing Authorisation Holder identifies, evaluates and manages emerging safety concerns.

Regulatory assessors commonly evaluate whether:

• important signals have been included;

• evaluations reflect current scientific knowledge;

• conclusions are supported by the available evidence;

• contradictory evidence has been considered appropriately;

• uncertainty has been described transparently;

• conclusions remain consistent with the recognised safety profile;

• regulatory actions correspond to the scientific conclusions.

During pharmacovigilance inspections, inspectors may additionally examine the governance surrounding signal evaluation.

Examples include:

• traceability between the signal management system and the PBRER;

• documentation supporting scientific conclusions;

• multidisciplinary review and approval processes;

• consistency between internal signal assessments and regulatory submissions;

• implementation of resulting pharmacovigilance activities and risk minimisation measures;

• communication of important conclusions to relevant organisational functions.

Inspectors may also compare signal evaluations presented in successive PBRERs to determine whether evolving evidence has been interpreted consistently and whether previous regulatory commitments have been fulfilled.

Inspection Insight

Inspectors rarely focus only on whether the "correct" conclusion was reached. They are equally interested in whether the organisation can demonstrate a robust, reproducible and scientifically defensible process for evaluating emerging safety information.

Key Takeaways

Signal evaluation represents the scientific core of the PBRER.

Its purpose is not to catalogue signals or reproduce case summaries but to explain how available evidence has been integrated to determine whether the recognised safety profile of the medicinal product should change.

High-quality signal evaluations define the medical question clearly, consider all relevant evidence objectively, acknowledge uncertainty, evaluate alternative explanations and present conclusions that are proportionate to the strength of the available evidence.

Ultimately, the value of a signal evaluation lies not in the conclusion itself but in the transparency, consistency and scientific rigour of the reasoning that supports it.

How a Senior Aggregate Physician Thinks

Experienced aggregate physicians rarely ask:

"Does this signal appear convincing?"

Instead, they ask:

"If another experienced physician reviewed exactly the same evidence, would they understand why I reached this conclusion?"

Their objective is not to prove a hypothesis but to challenge it rigorously.

They actively search for evidence that weakens their preferred explanation, evaluate competing hypotheses fairly and revise their conclusions whenever new data justify doing so.

They recognise that pharmacovigilance decisions are seldom based on certainty. Instead, they are based on careful interpretation of incomplete evidence, guided by scientific reasoning, clinical judgement and regulatory principles.

This disciplined approach produces evaluations that remain robust even when future evidence ultimately changes the scientific understanding of the medicinal product.

Continue Reading

• [[signal-evaluation-methodology]]
• [[signal-detection]]
• [[signal-management]]
• [[writing-the-characterisation-of-risks-section-in-a-pbrer]]
• [[writing-the-integrated-benefit-risk-evaluation-section-in-a-pbrer]]
• [[benefit-risk-evaluation]]
• [[scientific-writing-in-pharmacovigilance]]

References

Primary Regulatory References

1. ICH E2C(R2): Periodic Benefit-Risk Evaluation Report.

2. EMA. Good Pharmacovigilance Practices (GVP) Module VII – Periodic Safety Update Report.

3. EMA. Good Pharmacovigilance Practices (GVP) Module IX – Signal Management.

4. Commission Implementing Regulation (EU) No 520/2012.

5. Directive 2001/83/EC.

6. Regulation (EC) No 726/2004.

Signal Management Guidance

1. CIOMS Working Group VIII. Practical Aspects of Signal Detection in Pharmacovigilance.

2. CIOMS publications relating to signal management and benefit-risk evaluation.

3. EMA guidance and procedural advice relating to signal management and PRAC signal procedures.

Supporting Scientific Literature

1. Hauben M, Aronson JK. Defining 'signal' and its subtypes in pharmacovigilance.

2. Council for International Organizations of Medical Sciences (CIOMS). Relevant publications on causality assessment and signal evaluation.

3. Selected peer-reviewed publications relating to pharmacoepidemiology, causal inference and post-authorisation signal evaluation.