Common Signal Management Failures

Table of contents

• Introduction
• Failure modes across the signal lifecycle
• Treating detection as management
• No documented detection rationale
• Poor signal validation
• Treating a threshold as a decision
• Incomplete signal assessment
• Confusing cases with rates
• Ignoring negative or contradictory evidence
• Weak escalation
• Inconsistent documents
• No follow-up after closure
• Controls and inspection evidence
• Traceability
• Quality and competence
• QPPV and governance
• Metrics used carefully
• Practical review checklist
• References
• Regulatory Note

Introduction

Signal management includes detection, validation, assessment, decision-making, communication and follow-up. A statistical output is only one input. The system must show what was reviewed, how decisions were reached, who was accountable and how unresolved uncertainty was managed.

The patterns below are potential failure modes, not a claim that they are findings from a specific authority. They are useful prompts for procedures, audits, governance and inspection preparation.

Failure modes across the signal lifecycle

Treating detection as management

A disproportionality output or literature alert is not the completed process. Define validation, clinical review, prioritisation, assessment, action and monitoring.

No documented detection rationale

Record data sources, scope, method, frequency, product coverage, exclusions, thresholds if used and the rationale for changes. A method can be flexible, but the process must be explainable.

Poor signal validation

Validation should distinguish a statistical or clinical observation from a signal requiring assessment. Review case quality, duplicates, reporting patterns, chronology, alternative explanations, biological plausibility, class information and exposure.

Treating a threshold as a decision

A threshold can trigger review; it should not replace medical judgement. Conversely, a signal below a numerical threshold may still matter if the event is serious, novel, specific or clinically important.

Incomplete signal assessment

An assessment should define the question, evidence reviewed, uncertainty, conclusion, action and follow-up. “No signal” is not an adequate rationale when the record does not show what was examined.

Confusing cases with rates

Spontaneous reports support detection and characterisation but usually cannot provide an unbiased incidence rate. Avoid presenting case counts as frequency without appropriate exposure and denominator data.

Ignoring negative or contradictory evidence

A balanced assessment includes trials, epidemiology, literature, class data, exposure patterns, alternative causes and information that weakens the hypothesis. Selective evidence makes the conclusion fragile.

Weak escalation

Define when a signal goes to the QPPV, safety governance, aggregate reporting, risk management, medical review or regulatory contact. Escalation should be timely, documented and linked to the decision.

Inconsistent documents

Signal records, PSUR/PBRER, RMP, product information and safety communications should not contradict one another. Differences may be justified by scope or timing, but the rationale should be visible.

No follow-up after closure

A decision to close or monitor a signal is not the end of the evidence lifecycle. Track planned searches, studies, label review, regulatory outcomes and new cases.

Controls and inspection evidence

Traceability

GVP Module IX expects an audit trail for signal-management activities. Retain dates, analyses, decisions, rationale, roles, review and actions in a retrievable record.

Quality and competence

Define role-based training, quality control and review. Critical processes should include appropriate audit coverage, including relevant service providers and contractors.

QPPV and governance

The QPPV should have access to significant signals, decisions, unresolved risks and cross-document implications. Governance should be able to see workload, overdue assessments, repeat themes and actions.

Metrics used carefully

Possible metrics include timeliness of review, overdue actions, closure rationale, completion of planned follow-up and consistency across products. Metrics are indicators; they are not substitutes for case-level and scientific review.

Practical review checklist

For each signal, ask:

• What observation triggered review?
• Was the signal validated and duplicates addressed?
• What data sources and periods were searched?
• What is the clinical phenotype and plausible latency?
• What alternative explanations were considered?
• What evidence supports and weakens the association?
• What is known about absolute risk and benefit?
• Who reviewed and approved the conclusion?
• What action and follow-up were assigned?
• Can the record be retrieved and reconstructed?

References

• EMA, GVP Module IX: Signal management
• EMA, Signal management
• EMA, GVP Module II: Pharmacovigilance system master file
• EMA, GVP Module VII: Periodic safety update report

Regulatory Note

GVP Module IX provides the EU framework for signal management, but methods, thresholds, roles and escalation routes should be tailored to the product and system. A documented rationale is more important than a nominally sophisticated algorithm.