PVC Signal Detection & Analysis 2 — Questions and Answers
Question 1: Which statistical measure is used in Empirical Bayes Geometric Mean (EBGM) signal detection to shrink estimates toward the prior mean?
- Bayesian shrinkage toward the global reporting rate (Correct answer)
- Maximum likelihood estimation without prior correction
- Poisson regression with covariate adjustment
- Chi-square test with Yates correction
Correct answer: Bayesian shrinkage toward the global reporting rate
EBGM applies Bayesian shrinkage that pulls individual drug-event estimates toward the overall reporting rate when observed counts are low, reducing noise.
Question 2: In the context of signal detection, what is the primary risk of using a very low minimum threshold (e.g., N≥1) for reporting count?
- Excessive false-positive signals from sparse data instability (Correct answer)
- Reduced sensitivity for detecting true drug-event associations
- Overestimation of the Reporting Odds Ratio confidence interval
- Underrepresentation of serious adverse events
Correct answer: Excessive false-positive signals from sparse data instability
Very low count thresholds produce unstable, noisy signals dominated by chance associations rather than true pharmacovigilance signals.
Question 3: When a signal is detected in spontaneous reporting data, what is the MOST important next step before escalating to a formal safety review?
- Signal validation to confirm the association is not an artifact (Correct answer)
- Immediate regulatory notification to the FDA
- Withdrawal of the product from the market
- Initiation of a Phase IV clinical trial
Correct answer: Signal validation to confirm the association is not an artifact
Signal validation involves assessing whether a detected disproportionality reflects a true biological association or is due to confounding, data quality issues, or statistical artifact.
Question 4: Which of the following is a recognized limitation of spontaneous reporting systems (SRS) that directly impacts signal detection sensitivity?
- Unknown denominator and underreporting of adverse events (Correct answer)
- Mandatory reporting requirements for all healthcare providers
- Duplicate suppression algorithms that remove valid reports
- Overrepresentation of mild, non-serious adverse events
Correct answer: Unknown denominator and underreporting of adverse events
SRS lack a defined denominator (exposed population) and suffer from significant underreporting, making true incidence rates impossible to calculate directly.
Question 5: What is 'masking' (or 'competition bias') in the context of pharmacovigilance signal detection?
- A true signal for one drug-event pair is suppressed by a stronger signal from another drug for the same event (Correct answer)
- Intentional suppression of adverse event reports by manufacturers
- A coding error that assigns reports to the wrong MedDRA term
- Underreporting that selectively hides serious events
Correct answer: A true signal for one drug-event pair is suppressed by a stronger signal from another drug for the same event
Masking occurs when a strong signal for one drug-event combination reduces the apparent disproportionality for another drug-event pair in the same database.
Question 6: In MedDRA hierarchy, which level is most commonly used as the unit of analysis in disproportionality signal detection?
- Preferred Term (PT) (Correct answer)
- System Organ Class (SOC)
- High Level Term (HLT)
- Lowest Level Term (LLT)
Correct answer: Preferred Term (PT)
Preferred Terms are the primary analysis level for signal detection because they represent distinct medical concepts with sufficient granularity and coding consistency.
Question 7: Which approach helps address the problem of time-to-onset variability when analyzing signals for drug-induced liver injury (DILI)?
- Restricting the analysis window to reports with onset within a clinically plausible latency period (Correct answer)
- Excluding all reports that lack liver enzyme data
- Using only fatal reports to increase signal specificity
- Applying a fixed 30-day post-exposure window for all DILI signals
Correct answer: Restricting the analysis window to reports with onset within a clinically plausible latency period
Restricting to biologically plausible latency windows improves signal specificity by excluding temporally implausible associations for DILI.
PVC Signal Detection & Analysis 2
Which statistical measure is used in Empirical Bayes Geometric Mean (EBGM) signal detection to shrink estimates toward the prior mean?