GVP Module VI: Difficult ICSR Validity and Assessment Scenarios

Introduction

Some of the most difficult ICSR decisions arise when a report appears clinically important but does not clearly contain the information needed to establish a valid individual case safety report.

Examples include statements such as:

• "the patient was hospitalised";
• "the patient died";
• "the medicine did not work";
• "adverse event NOS";
• "the patient was exposed";
• or "there was an abnormal laboratory result".

These statements can be highly relevant to pharmacovigilance. They do not, however, automatically answer the question of whether a valid ICSR exists.

A useful principle is:

Safety-relevant information
          ≠
Automatically a valid ICSR

The assessment should begin with the applicable regulatory criteria and the information actually contained in the source. The processor should not manufacture missing information merely because the report appears important.

1. The Four Minimum Elements

Under the EU pharmacovigilance framework, an ICSR requires four minimum elements:

1. an identifiable reporter;
2. an identifiable patient;
3. one or more suspected medicinal products; and
4. one or more suspected adverse reactions.

If one of these minimum elements is absent, the report does not meet the minimum criteria for a valid ICSR.

The distinction between a report being important and being valid is central to this article.

2. Do Not Infer a Missing Adverse Reaction

A frequent processing error is to infer an adverse reaction from a circumstance or outcome.

For example:

Hospitalisation
     ↓
What happened clinically?
     ↓
Was an adverse reaction reported?

The processor should not invent a reaction simply because hospitalisation occurred.

Similarly:

Death
     ↓
What was the clinical event or suspected reaction?

The outcome of death is not itself a substitute for an adverse reaction when the applicable validity criterion has not been met.

3. Hospitalisation-Only Reports

A report stating that a patient was hospitalised may be clinically important, but hospitalisation is a seriousness criterion rather than, by itself, an adverse reaction.

The organisation should determine whether the source provides a reported adverse reaction or other clinical event that satisfies the applicable ICSR requirements.

If the source contains only hospitalisation with no reportable adverse reaction, the organisation should not create a fictional reaction merely to make the case valid.

4. Death-Only Reports

A report may state that a patient died while receiving a medicinal product without identifying a suspected adverse reaction or cause of death.

The organisation should distinguish:

• the patient's death;
• the cause of death, if known;
• any clinical event preceding death;
• and any suspected relationship with the medicinal product.

Death is an outcome. It should not automatically be converted into an adverse reaction.

5. Fatal Outcome With Unknown Cause

A case can initially contain a fatal outcome while the cause remains unknown.

The organisation should record the information actually available and determine whether the minimum ICSR criteria are satisfied. Follow-up may subsequently establish the cause or a relevant clinical event.

The absence of a known cause should not be replaced with an unsupported diagnosis.

6. Lack of Efficacy

A statement that a medicinal product was ineffective does not automatically establish a valid adverse reaction.

The organisation should assess the specific regulatory context and the information provided.

Relevant clinical information can include:

• indication;
• disease severity;
• treatment duration;
• dose;
• adherence;
• disease progression;
• and clinical outcome.

Lack of efficacy also has product- and context-specific reporting considerations and should not be processed using an indiscriminate universal rule.

7. Adverse Event NOS

A source may contain a vague statement such as "adverse event", "AE NOS" or an equivalent nonspecific expression.

The organisation should assess whether the information actually identifies an adverse reaction sufficiently to satisfy the applicable minimum criteria.

A vague label should not automatically be expanded into a specific medical event that the source did not report.

8. Exposure Without a Clinical Event

Exposure to a medicinal product can be pharmacovigilance-relevant without being an adverse reaction.

Examples include:

• accidental exposure;
• overdose;
• pregnancy exposure;
• occupational exposure;
• or medication error.

The organisation should assess the applicable special-situation requirements separately from the question of whether an adverse reaction occurred.

The next chunk will cover product complaints, laboratory abnormalities, incomplete diagnoses, evolving cases, validity reassessment and difficult borderline examples.

9. Product Complaints Without a Clinical Event

A product complaint may contain information about a defect, packaging problem, device issue or other quality concern without describing an adverse reaction in a patient.

The organisation should determine whether there is an associated patient and clinical event that meets the ICSR criteria. A quality complaint and an ICSR can be related, but they are not interchangeable concepts.

Where a product complaint includes an adverse reaction, the relevant pharmacovigilance assessment should proceed alongside the applicable quality investigation.

10. Abnormal Laboratory Results

An isolated abnormal laboratory value does not automatically establish a valid ICSR.

The assessment should consider whether the source describes a suspected adverse reaction or clinically meaningful event and whether the minimum criteria are otherwise satisfied.

The organisation should avoid inventing a diagnosis from a laboratory result unless the source or appropriate medical assessment supports it.

11. Diagnosis Without a Suspected Product

A clinical diagnosis can be important information but does not by itself establish a suspected relationship with a medicinal product.

The organisation should determine whether a medicinal product is identified as suspected and whether the remaining minimum criteria are satisfied.

12. Exposure Without an Adverse Reaction

Exposure events can require pharmacovigilance assessment even where no adverse reaction is reported.

Examples include medication error, overdose, pregnancy exposure and occupational exposure.

The organisation should apply the specific regulatory requirements for the situation without converting exposure into an adverse reaction merely to create an ICSR.

13. Incomplete Patient Information

A report may contain a product and an event but insufficient information to identify the patient.

The organisation should assess whether the available information meets the applicable patient-identifiability requirement. A vague reference to a patient population does not automatically establish an identifiable individual patient.

Follow-up may be appropriate when the missing information can reasonably be obtained.

14. Incomplete Reporter Information

The reporter must be identifiable under the applicable validity framework.

A report that merely states that "someone reported this" without sufficient information to identify the source may require further assessment or follow-up.

The organisation should not manufacture reporter details from assumptions about the source channel.

15. Evolving Reports

A report can begin as incomplete and later become a valid ICSR.

For example:

Initial information
      ↓
Missing minimum element
      ↓
Follow-up / additional source
      ↓
Minimum criteria satisfied
      ↓
Valid ICSR

The organisation should have processes for recognising when additional information changes the validity status.

16. A Valid ICSR Can Become More Complete

Validity is not the end of assessment.

Once the minimum criteria are satisfied, additional information may materially improve the case, including seriousness, outcome, diagnosis, causality or clinical chronology.

The organisation should not confuse "valid" with "complete".

17. A Report Can Change During Follow-Up

Follow-up can also reveal that the initial interpretation was incorrect.

For example, a report initially understood as an isolated symptom may later reveal a different diagnosis, a duplicate case or a different medicinal product exposure.

The case should be reassessed rather than simply appended with new text.

18. Duplicate Information

An apparently incomplete report may contain information that matches another existing case.

Duplicate assessment should be performed according to the controlled duplicate process. The organisation should avoid creating a new case merely because a new source has provided additional information about an existing patient/event.

19. Hospitalisation Plus an Unspecified Event

A source may say that a patient was hospitalised following treatment but provide no identifiable clinical event beyond the statement that an "adverse event" occurred.

The organisation should determine whether the source actually provides a reportable adverse reaction or sufficiently specific clinical information.

Hospitalisation should not be used to fill the missing reaction element.

20. Death Plus an Unspecified Event

Similarly, a source may report that a patient died after exposure but provide no adverse reaction or clinical event.

The organisation should distinguish the outcome from the reaction and assess the minimum criteria using the information actually available.

Where appropriate, follow-up may seek the cause of death, preceding clinical events and other relevant information.

21. Lack of Efficacy Plus Clinical Deterioration

A report may initially state only that a medicine did not work. Follow-up may reveal disease deterioration, hospitalisation or another clinical event.

The later information can change the regulatory and clinical assessment. The organisation should reassess the report based on the complete information rather than applying the initial lack-of-efficacy label indefinitely.

22. Inspection Perspective

These scenarios are important because inspection findings can arise when organisations demonstrate weak control over case validity, data quality or follow-up.

An inspector may ask:

• How do you determine whether a report is a valid ICSR?
• How are borderline reports handled?
• How do you prevent processors from inventing missing information?
• How do you identify cases that become valid after follow-up?
• How are invalid reports documented and retained where required?
• How are product complaints and pharmacovigilance cases reconciled?
• How are death-only and lack-of-efficacy reports assessed?

The organisation should be able to demonstrate a controlled, reproducible decision process.

23. Common Failure: Creating a Reaction From the Outcome

A processor sees "hospitalised" or "death" and enters an adverse reaction that was never reported.

This creates artificial data and can distort safety analysis.

The correct approach is to preserve what the source actually says and obtain further information where appropriate.

24. Common Failure: Treating Every Safety-Relevant Record as an ICSR

Pharmacovigilance systems can receive safety-relevant information that belongs to different processes.

Examples include product complaints, medication errors without clinical consequences, exposure reports and study information.

The organisation should apply the correct regulatory pathway rather than treating every record as identical.

25. Common Failure: No Documented Validity Decision

An organisation may have a database case but be unable to demonstrate why the minimum criteria were considered satisfied.

Inspection readiness requires traceability of the assessment and appropriate evidence supporting material processing decisions.

The next chunk will cover regulatory decision trees, difficult borderline examples, reassessment, inspection findings, actionable controls and References + Regulatory Note.

26. A Practical Validity Decision Framework

A consistent assessment can be structured as follows:

1. Is there an identifiable patient?
          ↓ yes
2. Is there an identifiable reporter?
          ↓ yes
3. Is a medicinal product suspected?
          ↓ yes
4. Is an adverse reaction or other reportable clinical event described?
          ↓ yes
        VALID ICSR

If an element is missing, the organisation should determine whether the available information can reasonably be clarified through follow-up or another reliable source.

The framework should be applied using the current regulatory definitions and controlled procedures rather than processor intuition.

27. What Happens When a Minimum Element Is Missing?

A missing element should trigger the appropriate process rather than an automatic database workaround.

The organisation may need to:

• seek clarification;
• retain the information in an appropriate safety repository;
• link it to an existing case;
• monitor for additional information;
• or document why the minimum criteria are not met.

The exact treatment depends on the source and applicable requirements.

28. Do Not Manufacture Identifiability

Identifiability should be based on information actually available from the source.

For example, a statement that "a number of patients experienced events" does not automatically establish an identifiable individual patient for each event.

Similarly, a generic reference to "the doctor" does not necessarily establish an identifiable reporter if the source contains no sufficient information to identify that person.

29. Difficult Scenario: Death Reported by a Known Reporter

Consider:

"Dr Smith reports that her patient died while taking Product X. No cause of death is known."

The information contains an identifiable reporter, an identifiable patient, a suspected medicinal product and a fatal outcome. The critical question is whether a suspected adverse reaction has actually been reported.

The organisation should not invent a cause of death merely because the outcome is fatal. It should assess the available information against the current ICSR validity requirements and determine whether follow-up is appropriate.

30. Difficult Scenario: Hospitalisation With No Reaction

Consider:

"Patient was admitted to hospital while taking Product X. No further information is available."

Hospitalisation is clinically significant but is not itself a substitute for a reported adverse reaction.

The organisation should not create a reaction such as "hospitalisation" merely to satisfy a database field. Appropriate clarification should be sought where possible.

31. Difficult Scenario: Lack of Efficacy With No Clinical Event

Consider:

"Product X did not work."

This may be relevant pharmacovigilance information, but the organisation should assess whether the applicable reporting requirements are met in the particular product and clinical context.

If follow-up establishes a clinically significant deterioration or other reportable event, the assessment may change.

32. Difficult Scenario: Adverse Event NOS

Consider:

"Patient experienced an adverse event."

The organisation should determine whether this statement itself identifies a suspected adverse reaction sufficiently to satisfy the applicable criteria. It should not replace the nonspecific term with an invented diagnosis.

Where clarification is possible, follow-up should seek the nature of the event and relevant clinical details.

33. Difficult Scenario: Laboratory Abnormality

Consider:

"ALT increased after starting Product X."

The organisation should determine whether the laboratory abnormality constitutes the reported adverse reaction in the source and whether the remaining validity criteria are satisfied.

If the source provides no indication that the abnormality is associated with the product, the organisation should not assume causality merely from temporal sequence.

34. Difficult Scenario: Product Complaint Plus Injury

A quality complaint can contain a patient injury or adverse reaction.

The organisation should ensure that the quality and pharmacovigilance processes communicate appropriately while maintaining the distinction between:

• the quality defect;
• the exposure;
• the clinical event;
• and the ICSR assessment.

35. Reassessment and Audit Trail

When information changes the validity status or material assessment of a case, the organisation should preserve an appropriate audit trail.

A reviewer should be able to understand why the case was initially considered incomplete and what information subsequently changed the assessment.

36. Inspection Findings and Evidence

Inspection findings relating to ICSR quality commonly expose weaknesses in minimum-information assessment, data quality, follow-up and case processing.

An inspection-ready organisation should therefore be able to demonstrate:

• the applicable validity criteria;
• controlled procedures;
• trained personnel;
• documented follow-up;
• quality-control checks;
• appropriate case status decisions;
• and traceability of material changes.

Where an organisation rejects or does not process information as a valid ICSR, it should be able to explain the regulatory basis for that decision.

37. Controls for Difficult Cases

Useful controls include:

• validity decision trees;
• processor training using borderline examples;
• medical escalation criteria;
• controlled terminology;
• follow-up triggers;
• duplicate checks;
• quality-control sampling of invalid and borderline reports;
• and periodic review of recurring validity errors.

Quality control should examine both false positives and false negatives.

38. What Good Looks Like

A mature case-validity process:

• applies the current minimum criteria consistently;
• does not infer missing clinical information;
• distinguishes outcomes from reactions;
• recognises that safety-relevant information may follow different regulatory pathways;
• actively identifies cases that become valid after follow-up;
• reassesses evolving cases;
• maintains traceability;
• and can explain decisions during inspection.

39. Final Principles

1. A clinically important report is not necessarily a valid ICSR.
2. Hospitalisation is a seriousness concept, not automatically an adverse reaction.
3. Death is an outcome and should not be converted into an invented cause or reaction.
4. Lack of efficacy requires context-specific assessment.
5. Vague terms such as "adverse event NOS" should not be expanded without evidence.
6. Exposure and medication-use circumstances should not automatically be converted into reactions.
7. Product complaints and pharmacovigilance cases may overlap but remain distinct processes.
8. Validity can change when additional information becomes available.
9. The distinction between invalid, incomplete and valid cases should be controlled and documented.
10. Inspection readiness requires evidence that validity decisions are systematic, reproducible and traceable.

Key Takeaways

• Start with the information actually reported.
• Apply the four minimum ICSR criteria without inventing missing facts.
• Do not confuse seriousness or outcome with the adverse reaction element.
• Use follow-up to resolve uncertainty where appropriate.
• Reassess cases when new information changes their status.
• Treat borderline validity decisions as a quality-system issue, not merely a processor judgement.

References

1. European Medicines Agency. Good Pharmacovigilance Practices (GVP), Module VI — Collection, management and submission of reports of suspected adverse reactions to medicinal products.
2. European Medicines Agency. ICH E2D(R1) — Post-Approval Safety Data: Definitions and Standards for Management and Reporting, including EU implementation material.
3. European Medicines Agency. GVP Annex I — Definitions.
4. European Commission. Commission Implementing Regulation (EU) No 520/2012, as amended.
5. European Parliament and Council. Directive 2001/83/EC, as amended.
6. European Medicines Agency. Pharmacovigilance inspection reports and inspection-related pharmacovigilance quality findings, as applicable.

Regulatory Note

This article is an educational explanation of difficult ICSR validity and assessment scenarios under the EU pharmacovigilance framework. It does not replace current GVP Module VI, applicable EU legislation, ICH E2D(R1) implementation material, product-specific requirements or organisational procedures.

Regulatory guidance and definitions can change. The current EMA source documents and applicable legislation should be checked before using an interpretation to change a live pharmacovigilance process.

The examples are deliberately illustrative. They are not presented as descriptions of specific regulatory inspection cases unless the relevant source is explicitly identified.