Lesson 7- Signal Detection, Risk Assessment and management

Question 1

Define signal according to WHO.

Answer 1

Reported information suggesting a possible causal relationship between a drug and adverse event, previously unknown or incompletely documented.

Question 2

What does CIOMS define as a signal?

Answer 2

Information from multiple sources suggesting a new causal association (adverse/beneficial) requiring verificatory action.

Question 3

How does EMA define a safety signal?

Answer 3

Information on a new or known adverse event potentially caused by a medicine, warranting further investigation.

Question 4

What do all signal definitions have in common?

Answer 4

Focus on information, potential harm, evaluation, and hypothesis formation.

Question 5

List qualitative methods for signal detection

Answer 5

• Spontaneous reports
• Case series
• Literature
• Social media
• Health authority alerts
• ICSRs (Individual Case Safety Reports).

Question 6

Name quantitative methods for signal detection.

Answer 6

• PRR (Proportional Reporting Ratio)
• ROR (Reporting Odds Ratio)
• MGPS (Multi-item Gamma Poisson Shrinker)
• BCPNN (Bayesian Confidence Propagation Neural Network).

Question 7

What is PRR? Provide an example.

Answer 7

Proportional Reporting Ratio compares ADR frequency for a drug vs. all others. Example: Rifabutin and uveitis (PRR = 556).

Question 8

What is a limitation of data mining in signal detection?

Answer 8

Disproportionality measures association, not causation; requires validation with traditional methods.

Question 9

What is signal validation?

Answer 9

Evaluating if sufficient evidence exists to justify further analysis (e.g., using spontaneous reports, literature, regulatory documents).

Question 10

What sources are used in signal validation?

Answer 10

• Spontaneous reports,
• Product information
• Scientific literature
• Pharmacoepidemiologic studies
• MAH data.

Question 11

Define signal confirmation.

Answer 11

Procedural step to identify signals requiring discussion in pharmacovigilance risk assessment committees.

Question 12

Why is signal prioritization necessary?

Answer 12

To focus resources on high-likelihood/high-impact signals and discard false positives.

Answer 13

Seriousness

# Public health impact
# Strength of evidence
# Novelty
# Disproportionality score.

Question 14

What is the evidence score in UK MHRA’s impact analysis?

Answer 14

Based on disproportionality (e.g., PRR), biological plausibility, and evidence strength.

Question 15

What are the prioritization categories?

Answer 15

High priority (urgent action), gather more info, low priority, no action.

Question 16

What is assessed during signal evaluation?

Answer 16

Causality, frequency, seriousness, clinical implications, preventability.

Question 17

How is causality determined?

Answer 17

Balance of evidence supporting cause-effect (e.g., rechallenge response, exclusion of confounders).

Question 18

What defines ADR frequency?

Answer 18

Categorized as very common (>10%)
Common (1–10%)
Uncommon (0.1–1%)
Rare (0.01–0.1%)
Very rare (<0.01%).

Question 19

List possible regulatory actions post-signal confirmation.

Answer 19

Update product labels, restrict indications, issue safety alerts, recall products, mandate risk minimization activities.

Question 20

How might the SPC (Summary of Product Characteristics) be updated?

Answer 20

Modify dosing instructions, add contraindications, update warnings, include new ADRs, adjust pregnancy/lactation advice.

Question 21

What is the goal of risk minimization?

Answer 21

Prevent ADRs through HCP education, monitoring, dose adjustments, or restricted use.

Question 22

What is included in a drug safety communication?

Answer 22

Nature of the problem, evidence summary, actions taken, implications for HCPs/patients, risk-benefit balance, contact details.

Question 23

Outline the steps in signal management.

Answer 23

1. Detection: Use qualitative/quantitative methods.
2. Validation: Assess evidence sufficiency.
3. Prioritization: Rank signals by impact.
4. Assessment: Evaluate causality, frequency, and risk.
5. Action: Implement regulatory/risk mitigation measures.
6. Communication: Inform stakeholders.

Question 24

What SPC section might be updated for dosing instructions?

Answer 24

Example: Reduce dose for elderly patients or limit treatment duration to mitigate cumulative toxicity.

Question 25

How does **MGPS** differ from **PRR**?

Answer 25

MGPS uses Bayesian shrinkage to reduce noise in rare events; PRR compares observed vs. expected proportions.

Question 26

What is the **public health score** in prioritization?

Answer 26

Based on case numbers, health consequences, and reporting rate relative to drug exposure.

Question 27

How does **ROR (Reporting Odds Ratio)** differ from **PRR**?

Answer 27

- **ROR**: Compares odds of an event occurring with a drug vs. without it. - **PRR**: Compares proportion of a specific ADR for a drug vs. all other drugs.

Question 28

What is the **BCPNN (Bayesian Confidence Propagation Neural Network)**?

Answer 28

A data mining method using Bayesian statistics to identify drug-event associations with disproportional reporting in databases like VigiBase.

Question 29

What procedural steps follow **signal confirmation**?

Answer 29

Escalation to pharmacovigilance risk assessment committees, initiation of regulatory reviews, or design of additional studies (e.g., pharmacoepidemiologic).

Question 30

What are the **Bradford Hill criteria** used for?

Answer 30

To assess causality in pharmacovigilance, including strength of association, consistency, specificity, temporality, and biological plausibility.

Question 31

Give an example of updating the **Contraindications** section of the SPC.

Answer 31

Adding a contraindication for patients with a specific genetic mutation (e.g., HLA-B*5701 for abacavir hypersensitivity).

Question 32

What is **EudraVigilance**?

Answer 32

The EU’s database for managing and analyzing ADR reports, maintained by the EMA.

Question 33

How do **national regulatory databases** contribute to signal detection?

Answer 33

They provide localized ADR data, enabling early identification of region-specific safety concerns (e.g., traditional medicine interactions).

Question 34

What is included in the **basic model for drug safety communication**?

Answer 34

1. Nature of the problem (drug, hazard, risk factors). 2. Evidence summary. 3. Actions taken (label updates, studies). 4. Implications for HCPs/patients. 5. Risk-benefit balance. 6. Contact details for further info.

Question 35

Why is data mining insufficient alone for signal detection?

Answer 35

It identifies statistical associations, not causation. Requires clinical validation (e.g., case reviews, biological plausibility).

Question 36

Break down the **final assessment phase** of signal management.

Answer 36

1. **Causality determination**: Is the drug-event link plausible? 2. **Risk quantification**: Frequency, severity, and preventability. 3. **Decision-making**: Regulatory action, risk minimization, or no action.