Writing the Data in Summary Tabulations Section in a PBRER

Introduction

Summary tabulations provide a structured overview of the cumulative adverse event data available for a medicinal product. They allow regulators and Marketing Authorisation Holders to review large volumes of safety information systematically and identify patterns that may require further scientific evaluation elsewhere within the PBRER.

Although summary tabulations contain numerical data, they should not be regarded as evidence of causality or proof of emerging safety concerns. Instead, they provide an organised presentation of accumulated safety information that supports medical review, signal detection and the overall benefit-risk evaluation.

This section therefore serves as a bridge between individual case reports and the detailed scientific assessment presented later in the PBRER.

Regulatory Basis

Preparation of summary tabulations is described within ICH E2C(R2) and reflected in European Union Good Pharmacovigilance Practices (GVP) Module VII.

These tabulations summarise adverse reaction data accumulated since international birth and, where appropriate, during the reporting interval.

The objective is to facilitate scientific review rather than simply satisfy a reporting requirement.

Authors should therefore understand both how the tabulations were generated and how they should be interpreted.

Why This Section Exists

Individual case reports provide detailed clinical information but become increasingly difficult to review as cumulative experience with a medicinal product grows.

Summary tabulations organise these data into a structured format that allows reviewers to identify patterns requiring further assessment.

The tabulations themselves do not determine whether a safety signal exists.

Rather, they provide one source of information that contributes to subsequent signal evaluation, risk characterisation and benefit-risk assessment.

Accordingly, this section should be regarded as a descriptive presentation of available data rather than a scientific interpretation of those data.

What Are Summary Tabulations?

Summary tabulations are structured presentations of adverse event data derived from the global pharmacovigilance safety database.

Depending upon the reporting requirements and characteristics of the medicinal product, tabulations may include cumulative data, interval data or both.

Cases are commonly grouped using the Medical Dictionary for Regulatory Activities (MedDRA), allowing reviewers to examine reporting patterns according to System Organ Class (SOC), Preferred Term (PT) and other standardised coding levels.

The precise format of the tabulations may differ between organisations and products. However, the underlying objective remains the same: to present adverse event data in a consistent and reviewable manner that supports subsequent medical assessment.

Writing Tip

This section explains what data are presented and how they were organised. Detailed interpretation of important findings belongs later within the Signal Evaluation and Characterisation of Risks sections.

Preparing the Summary Tabulations

Although the final summary tabulations are usually generated using validated pharmacovigilance database systems, preparation of this section involves considerably more than producing a series of tables.

Before the tabulations are generated, authors should understand which cases have been included, the applicable Data Lock Point, the coding standards applied, the version of MedDRA used, and any case reconciliation activities completed during the reporting interval.

The objective is to ensure that the tabulations accurately represent the cumulative global safety database at the time of report preparation.

Accordingly, aggregate report authors should understand the provenance of the data rather than simply reproducing system-generated output.

Sources of Data

Summary tabulations are typically generated from the organisation's validated global safety database.

Depending upon the pharmacovigilance system and product portfolio, the database may contain cases originating from:

• spontaneous reports;
• solicited reports;
• interventional clinical trials where applicable;
• non-interventional studies;
• scientific literature;
• patient support programmes;
• market research programmes where reportable;
• licensing partners;
• affiliated Marketing Authorisation Holders;
• regulatory authorities;
• published case reports.

The exact sources included should be consistent with the scope of the PBRER and organisational procedures governing aggregate reporting.

Where multiple databases contribute to the cumulative safety dataset, reconciliation activities should be completed before generation of the summary tabulations.

Defining the Dataset

Before reviewing the tabulations, authors should confirm that the dataset corresponds to the intended reporting scope.

Important considerations include:

• the Data Lock Point used for extraction;
• products included within the report;
• formulations and strengths included;
• approved indications where relevant;
• geographical scope;
• inclusion of partner and affiliate cases;
• management of duplicate reports;
• handling of follow-up information;
• inclusion of literature cases;
• reconciliation with clinical trial databases where applicable.

Errors at this stage may affect every subsequent section of the PBRER.

For this reason, understanding the dataset is as important as reviewing the tabulated results themselves.

Case Counting Principles

One of the most important concepts in summary tabulations is understanding exactly what is being counted.

Depending upon the tabulation, counts may represent:

• individual case safety reports (ICSRs);
• individual patients;
• adverse events;
• Preferred Terms (PTs);
• reactions;
• fatal outcomes;
• serious reports.

These measures are not interchangeable.

For example, a single patient may contribute several adverse reactions coded to different Preferred Terms, while multiple follow-up reports may relate to the same underlying case.

Aggregate report authors should therefore understand the counting methodology used within their validated safety database before interpreting numerical differences between reporting periods.

Detailed case-counting methodology is discussed in [[summary-tabulations-in-pharmacovigilance]].

Data Quality Before Medical Review

Summary tabulations should never be interpreted without considering data quality.

Medical reviewers should understand whether important data management activities have already been completed before preparation of the report.

Examples include:

• duplicate detection and reconciliation;
• MedDRA recoding;
• incorporation of follow-up information;
• reconciliation with partner databases;
• quality control of seriousness classification;
• verification of fatal outcomes;
• correction of coding inconsistencies;
• database quality review prior to Data Lock Point.

The purpose of this review is not to repeat database quality control but to ensure that important deficiencies are recognised before scientific interpretation begins.

Medical Review Consideration

Before interpreting apparent trends within the summary tabulations, confirm that important changes are not explained by database reconciliation activities, MedDRA version updates, coding corrections or incorporation of large volumes of follow-up information shortly before the Data Lock Point.

Reviewing the Summary Tabulations

Medical review should begin with an overall assessment of the dataset rather than detailed examination of individual Preferred Terms.

Questions that should be considered include:

• Does the overall distribution of reports appear consistent with previous reporting periods?

• Have any unexpected increases occurred within particular System Organ Classes?

• Are reporting patterns consistent with changes in product utilisation?

• Have new indications or expanded geographical marketing influenced the dataset?

• Have important coding changes occurred since the previous PBRER?

Only after this initial review should attention turn to individual Preferred Terms or specific safety topics requiring detailed scientific assessment.

The summary tabulations should guide subsequent review rather than determine scientific conclusions independently.

How an Aggregate Physician Reviews Summary Tabulations

Review of summary tabulations is a scientific exercise rather than a numerical exercise. Experienced aggregate physicians do not simply compare case counts between reporting periods. Instead, they evaluate whether the observed reporting patterns are medically plausible, whether they are consistent with current knowledge of the medicinal product and whether any findings require further scientific investigation.

The objective of the review is not to identify every numerical difference but to determine which observations may influence the overall benefit-risk evaluation.

Accordingly, review should proceed systematically from the overall dataset towards individual safety topics rather than beginning immediately with isolated Preferred Terms.

Step 1 – Review the Dataset Before Reviewing the Data

Before interpreting any apparent trends, the reviewer should confirm that the dataset itself is suitable for comparison.

Questions that should be considered include:

• Has the Data Lock Point changed as expected?

• Has worldwide exposure changed substantially?

• Have important new markets been launched?

• Have additional indications been approved?

• Has the reporting interval differed from previous reports?

• Have large numbers of historical partner cases been migrated into the safety database?

• Has the MedDRA version changed?

• Have major database reconciliation activities occurred?

Apparent increases in reporting may reflect changes in the dataset rather than genuine changes in the safety profile.

The reviewer should therefore understand the context before interpreting the numbers.

Step 2 – Perform a High-Level Medical Review

The initial review should focus upon the overall characteristics of the dataset.

Rather than examining individual Preferred Terms immediately, the reviewer should consider broader questions.

Does the overall pattern of reporting appear consistent with previous reporting intervals?

Have particular System Organ Classes increased disproportionately?

Has the proportion of serious reports changed?

Has the proportion of fatal reports changed?

Have reporting patterns altered following expansion into new patient populations?

At this stage, the objective is to identify areas requiring more detailed review rather than reaching scientific conclusions.

Step 3 – Identify Clinically Meaningful Changes

Not every increase in case numbers requires further investigation.

Reviewers should instead focus upon changes that may have medical or regulatory significance.

Examples include:

• previously unrecognised adverse reactions;

• important increases involving known serious adverse reactions;

• unexpected fatal outcomes;

• unusual reporting patterns;

• medically important events;

• events affecting special populations;

• clusters involving new indications;

• unexpected demographic patterns.

These observations should subsequently be evaluated using all available sources of evidence rather than relying solely upon summary tabulations.

Step 4 – Review Important Risks

Known important identified risks and important potential risks deserve particular attention during medical review.

The reviewer should consider whether reporting patterns remain consistent with the recognised safety profile and whether cumulative experience suggests important changes requiring further evaluation.

Review should also consider whether implemented risk minimisation measures appear consistent with observed reporting patterns.

The purpose is not to measure effectiveness directly but to identify observations requiring further assessment within the benefit-risk evaluation.

Step 5 – Review New and Unusual Findings

Experienced reviewers often spend considerable time examining observations that do not fit the expected safety profile.

Examples include:

• clinically unusual Preferred Terms;

• unexpected combinations of adverse reactions;

• unexpected age distributions;

• unusual geographical clustering;

• disproportionate reporting following introduction of a new formulation;

• unexpected reporting following manufacturing or formulation changes;

• reports associated with newly approved indications.

These observations do not necessarily represent emerging safety signals.

However, they frequently justify additional review using individual case narratives, literature, signal detection outputs, epidemiological evidence or clinical expert opinion.

Step 6 – Place Numerical Findings into Clinical Context

Summary tabulations should never be interpreted independently of the wider benefit-risk evaluation.

Changes in reporting frequency should always be considered alongside:

• estimated patient exposure;

• changes in utilisation;

• newly approved indications;

• expansion into additional countries;

• changes in reporting obligations;

• implementation of additional pharmacovigilance activities;

• stimulated reporting;

• regulatory communications;

• important published literature.

Only after considering these factors should the reviewer determine whether further scientific assessment is required.

Medical Review Consideration

A numerical increase is an observation, not a conclusion. The role of the aggregate physician is to determine whether that observation can be explained by changes in exposure, reporting behaviour, coding practices or clinical use before considering the possibility of an evolving safety issue.

Summary Tabulations Are Only One Source of Evidence

Summary tabulations should never be interpreted in isolation.

They represent one component of a comprehensive pharmacovigilance assessment that also includes:

• individual case review;

• interval case review;

• cumulative case review;

• signal detection activities;

• clinical trial findings;

• non-interventional studies;

• scientific literature;

• regulatory authority communications;

• findings from ongoing PASS;

• information relating to class effects;

• benefit-risk assessments.

Scientific conclusions should always integrate these complementary sources of evidence rather than relying solely upon numerical tabulations.

Writing Tip

Avoid statements such as "the increase in reports demonstrates an increased risk."

Instead, describe the observation objectively and explain how it was further evaluated using the broader pharmacovigilance evidence base.

Interpreting Trends Without Being Misled

One of the greatest challenges in aggregate report preparation is distinguishing meaningful changes in the safety profile from changes that simply reflect differences in reporting behaviour, product utilisation or data management.

Summary tabulations present reported adverse events. They do not directly measure the incidence of adverse reactions, establish causality or quantify clinical risk. Consequently, changes observed within the tabulations should always be interpreted within their broader scientific and regulatory context.

Experienced reviewers therefore approach apparent trends with scientific curiosity rather than immediate concern.

The appropriate question is not:

"Why did the numbers increase?"

Instead, it is:

"What factors could explain the observed change?"

Only after alternative explanations have been considered should conclusions regarding emerging safety concerns be explored.

Changes in Patient Exposure

Perhaps the most important consideration is whether changes in reporting simply reflect changes in the number of patients receiving the medicinal product.

Expansion into new countries, approval of additional indications, increased prescribing or improved patient access may substantially increase the number of adverse event reports without altering the underlying safety profile.

Conversely, declining utilisation following loss of patent protection, commercial withdrawal or changing clinical practice may reduce the number of reported cases despite no change in the intrinsic safety characteristics of the product.

Summary tabulations should therefore always be interpreted together with the estimated patient exposure described elsewhere within the PBRER.

Stimulated Reporting

Reporting behaviour frequently changes following important safety communications.

Distribution of a Direct Healthcare Professional Communication (DHPC), publication of regulatory safety alerts, revisions to product information, media attention or publication of influential scientific literature may all increase awareness of particular adverse reactions.

Such increases often represent improved recognition and reporting rather than an actual increase in the frequency of the adverse reaction.

Authors should therefore consider whether important communications or regulatory activities occurred during the reporting interval before interpreting increases within the summary tabulations.

Expansion into New Patient Populations

Approval of new indications or extension of use into different patient populations may substantially influence reporting patterns.

For example, introduction of a medicinal product into paediatric practice, oncology, elderly populations or patients with significant comorbidities may alter the types of adverse reactions reported.

Similarly, expansion into regions with different disease epidemiology, prescribing practices or pharmacovigilance systems may influence the composition of the safety database.

Changes observed under these circumstances should therefore be interpreted in relation to the characteristics of the newly exposed population.

Changes in Data Quality and Database Content

Apparent trends may also result from changes within the pharmacovigilance database itself.

Examples include:

• incorporation of large numbers of follow-up reports;

• migration of historical partner databases;

• reconciliation of duplicate cases;

• implementation of new MedDRA versions;

• correction of coding inconsistencies;

• receipt of delayed literature cases;

• integration of newly acquired product portfolios.

These activities may substantially alter summary tabulations without representing any change in the medicinal product's safety profile.

Medical reviewers should therefore understand important database activities completed during the reporting interval before interpreting numerical differences.

The Weber Effect and Product Lifecycle

Several observational studies have described changes in spontaneous reporting patterns during the lifecycle of medicinal products.

For some products, spontaneous reporting is highest during the early post-marketing period before gradually stabilising as clinical experience accumulates. This phenomenon has historically been referred to as the Weber effect.

Although not all medicines demonstrate this pattern, reviewers should recognise that spontaneous reporting behaviour may change over time for reasons unrelated to changes in intrinsic product safety.

Similarly, renewed interest following approval of additional indications, introduction of new formulations or significant regulatory communications may temporarily alter reporting behaviour.

The product lifecycle should therefore be considered when interpreting cumulative reporting trends.

Rare Events and Random Variation

For medicinal products with relatively small exposure or rare adverse reactions, modest numerical fluctuations may occur simply through random variation.

An increase from two reports to four reports, although representing a 100% relative increase, may have little clinical significance when considered in the context of worldwide utilisation.

Conversely, a small numerical increase involving a previously unrecognised fatal adverse reaction may warrant immediate detailed review.

The clinical importance of an observation should therefore never be judged solely by the magnitude of the numerical change.

When Trends Warrant Further Investigation

Certain observations should prompt additional medical assessment.

Examples include:

• unexpected clusters of clinically similar reports;

• previously unrecognised serious adverse reactions;

• unusual demographic patterns;

• unexpected fatal outcomes;

• changes involving important identified or potential risks;

• unexpected reporting within special populations;

• findings consistent across multiple independent data sources.

These observations should not automatically be interpreted as confirmed safety signals.

Instead, they indicate areas requiring further scientific evaluation using individual case review, signal detection activities, epidemiological evidence, published literature and other available pharmacovigilance information.

Medical Review Consideration

Numerical trends should initiate scientific review rather than determine scientific conclusions. Apparent increases should always be interpreted in conjunction with exposure estimates, product utilisation, reporting behaviour, coding practices, regulatory activities and the totality of available evidence.

The Importance of Scientific Judgement

Summary tabulations provide valuable descriptive information but cannot replace expert medical assessment.

Experienced aggregate physicians recognise that identical numerical findings may require different interpretations depending upon the medicinal product, patient population, treatment duration, exposure estimates and current understanding of the recognised safety profile.

Scientific judgement therefore remains central to interpretation of aggregate safety data.

The role of the physician is not merely to identify changes within the tables but to determine whether those changes alter the overall understanding of the medicinal product's benefit-risk balance.

Questions Every Aggregate Physician Should Ask

Summary tabulations should stimulate scientific enquiry rather than encourage rapid conclusions. Before interpreting apparent trends or identifying potential safety concerns, reviewers should systematically consider whether alternative explanations exist.

The following questions provide a practical framework for medical review.

Does the Dataset Explain the Observation?

Before examining individual adverse reactions, confirm that the observed pattern is not explained by changes in the dataset itself.

Consider:

• Has the Data Lock Point changed appropriately?

• Have historical partner cases been migrated?

• Have newly acquired product portfolios been incorporated?

• Have duplicate reconciliation activities been completed?

• Has a new MedDRA version been implemented?

• Have important coding corrections occurred?

• Has the geographical scope changed?

If the dataset has changed substantially, numerical comparisons with previous reporting intervals should be interpreted cautiously.

Could Exposure Explain the Change?

An increase in reported cases should always be interpreted together with exposure information.

Ask:

• Has patient exposure increased?

• Has a new indication been approved?

• Has the product been launched in additional countries?

• Has utilisation changed substantially?

• Has the treated population changed?

An increase proportional to patient exposure may not indicate any change in the underlying safety profile.

Could Reporting Behaviour Have Changed?

Reporting patterns frequently change independently of product safety.

Consider whether the reporting interval included:

• Direct Healthcare Professional Communications;

• important regulatory safety communications;

• media attention;

• publication of influential scientific literature;

• patient awareness campaigns;

• litigation;

• enhanced pharmacovigilance activities;

• organised patient support programmes.

Each may influence spontaneous reporting independently of any biological change in product safety.

Are the Cases Clinically Similar?

Numerical increases alone provide limited information.

Reviewers should determine whether the reported cases describe:

• similar clinical presentations;

• similar time to onset;

• similar patient characteristics;

• similar outcomes;

• similar risk factors;

• similar routes of administration;

• similar dose relationships.

Clusters of medically similar cases generally warrant greater attention than unrelated reports sharing only the same Preferred Term.

Are Particular Populations Affected?

Review whether the observation is concentrated within a specific population.

Examples include:

• paediatric patients;

• elderly patients;

• pregnant women;

• patients with renal impairment;

• patients with hepatic impairment;

• oncology populations;

• patients receiving combination therapy.

Restriction of an observation to one population may provide important clues regarding the underlying mechanism.

Does the Observation Affect a Recognised Risk?

Determine whether the observation relates to:

• an important identified risk;

• an important potential risk;

• missing information;

• an already labelled adverse reaction;

• a previously evaluated signal.

Existing knowledge frequently provides the context required to interpret apparent changes appropriately.

Is the Observation Supported by Other Evidence?

Summary tabulations represent only one component of the aggregate safety evaluation.

Review whether similar findings are supported by:

• individual case review;

• signal detection outputs;

• interval case review;

• cumulative case review;

• clinical trials;

• non-interventional studies;

• PASS;

• scientific literature;

• regulatory authority communications;

• class effects.

Consistency across multiple independent evidence sources generally strengthens confidence in the medical interpretation.

Does the Observation Require Further Evaluation?

Not every numerical change requires additional investigation.

However, further assessment should be considered when observations are:

• medically important;

• unexpected;

• clinically consistent;

• serious;

• fatal;

• increasing disproportionately after accounting for exposure;

• supported by multiple evidence sources;

• inconsistent with the recognised safety profile.

Where additional evaluation is required, the summary tabulations should direct the reviewer towards the appropriate scientific assessment rather than provide the final conclusion.

Medical Review Checklist

Before concluding that a trend represents an emerging safety issue, confirm that alternative explanations involving exposure, reporting behaviour, database changes, coding practices, product utilisation and recognised risks have been considered systematically.

Common Mistakes

Several recurring deficiencies are encountered during preparation and review of this section.

One of the most frequent errors is interpreting increases in case numbers as evidence of increased clinical risk without considering patient exposure or changes in reporting behaviour.

Another common mistake is reviewing Preferred Terms in isolation without considering related MedDRA groupings, clinical context or individual case characteristics.

Authors should also avoid drawing conclusions directly from summary tabulations. Numerical observations should instead guide subsequent scientific evaluation using the totality of available evidence.

Failure to recognise important database changes, including MedDRA version updates, partner database migrations or duplicate reconciliation activities, may lead to incorrect interpretation of apparent trends.

Finally, reviewers should avoid treating all numerical increases equally. Clinical importance depends upon the nature of the event, the affected population, exposure, biological plausibility and supporting evidence rather than numerical magnitude alone.

Inspection and Regulatory Assessment Considerations

Regulatory assessors expect summary tabulations to provide an accurate and reproducible summary of the cumulative safety database.

During assessment, reviewers may examine whether:

• the tabulations are consistent with the reported Data Lock Point;

• the appropriate MedDRA version has been used;

• cumulative and interval data have been presented correctly;

• important observations identified within the tabulations are discussed appropriately elsewhere in the PBRER;

• numerical findings are interpreted in conjunction with patient exposure and other supporting evidence;

• important changes in reporting patterns have been evaluated scientifically rather than described without interpretation.

During pharmacovigilance inspections, organisations may additionally be expected to demonstrate:

• procedures governing generation of summary tabulations;

• quality control of database extracts;

• reconciliation of partner and affiliate data;

• management of duplicate reports;

• governance of MedDRA version implementation;

• traceability between the validated safety database and the aggregate report.

The objective is to demonstrate that summary tabulations accurately represent the underlying safety database and support a scientifically robust benefit-risk evaluation.

Relationship to Other Sections of the PBRER

The Data in Summary Tabulations section should not be interpreted independently of the remainder of the PBRER. Its principal purpose is to provide a structured summary of the cumulative safety database that supports subsequent scientific evaluation.

Observations identified during review of the summary tabulations frequently influence the depth of discussion required within later sections of the report. For example, an apparent increase in reports relating to an important identified risk may justify a more detailed discussion within the Characterisation of Risks section. Likewise, unusual reporting patterns involving previously unrecognised adverse reactions may require formal signal evaluation.

Conversely, absence of important changes within the summary tabulations does not exclude the possibility of new safety information arising from clinical trials, post-authorisation safety studies (PASS), scientific literature, regulatory authority communications or other pharmacovigilance activities.

The tabulations therefore represent one component of an integrated safety assessment rather than an independent assessment of the medicinal product's benefit-risk profile.

Writing Style Recommendations

This section should remain descriptive.

Authors should avoid presenting medical conclusions that belong within subsequent sections of the PBRER.

Narratives accompanying the summary tabulations should:

• describe objectively what has been reviewed;
• avoid speculation regarding causality;
• distinguish observations from conclusions;
• avoid unnecessary statistical interpretation;
• remain consistent with subsequent signal evaluation and benefit-risk discussions.

The emphasis should remain on providing a reliable foundation for later scientific interpretation.

Key Takeaways

Summary tabulations organise cumulative adverse event data into a structured format that supports systematic medical review.

They do not establish causality, quantify clinical risk or confirm the existence of safety signals. Instead, they provide a descriptive overview of the available safety database and help identify observations requiring further scientific evaluation.

High-quality aggregate physicians interpret summary tabulations within the broader context of patient exposure, product utilisation, data quality, reporting behaviour, published evidence and the recognised safety profile of the medicinal product.

The most valuable contribution of the reviewer is therefore not identifying numerical differences but determining whether those differences require further medical assessment.

How a Senior Aggregate Physician Thinks

Experienced aggregate physicians rarely begin by asking whether the number of reported cases has increased.

Instead, they first seek to understand the context surrounding the data.

They ask whether changes in patient exposure, reporting behaviour, regulatory activity, MedDRA coding, database reconciliation or clinical practice could explain the observed findings before considering whether the data suggest an evolving safety issue.

They recognise that summary tabulations are descriptive tools rather than decision-making tools.

Scientific conclusions are reached only after integrating evidence from individual case review, signal detection activities, clinical studies, epidemiological investigations, published literature, regulatory assessments and the overall benefit-risk evaluation.

This disciplined approach reduces the likelihood of both overlooking genuine safety concerns and overinterpreting random fluctuations within spontaneous reporting data.

Continue Reading

The concepts discussed in this article are closely related to the following guides:

• [[summary-tabulations-in-pharmacovigilance]]
• [[signal-detection]]
• [[signal-evaluation]]
• [[benefit-risk-evaluation]]
• [[estimating-patient-exposure-in-pharmacovigilance]]
• [[medical-literature-monitoring]]
• [[writing-the-signal-evaluation-section-in-a-pbrer]]
• [[writing-the-characterisation-of-risks-section-in-a-pbrer]]

References

Primary Regulatory References

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

2. European Medicines Agency. Good Pharmacovigilance Practices (GVP) Module VII – Periodic Safety Update Report.

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

4. Directive 2001/83/EC.

5. Regulation (EC) No 726/2004.

MedDRA and Coding Standards

1. MedDRA® Introductory Guide (current version).

2. MedDRA® Term Selection: Points to Consider.

3. MedDRA® Data Retrieval and Presentation: Points to Consider.

Aggregate Reporting and Pharmacovigilance

1. CIOMS Working Group reports relevant to aggregate safety reporting.

2. ENCePP Guide on Methodological Standards in Pharmacoepidemiology.

3. ISPE Good Pharmacoepidemiology Practices.

Supporting Scientific Literature

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

2. Edwards IR, Aronson JK. Adverse drug reactions: definitions, diagnosis and management.

3. Selected peer-reviewed publications relating to spontaneous reporting systems, pharmacovigilance methodology and aggregate safety evaluation.