Signal detection and statistical methods in safety data MCQs With Answer

Introduction: Signal detection and statistical methods are vital components of pharmacovigilance, enabling identification of potential safety issues from spontaneous reports, clinical data, and electronic health records. This set of MCQs focuses on core concepts such as disproportionality measures (PRR, ROR, IC, EBGM), Bayesian approaches, observed-versus-expected calculations, sequential testing, and common pitfalls like confounding and masking. Designed for M.Pharm students, the questions reinforce both theoretical understanding and practical interpretation of safety metrics, thresholds for signaling, and steps in signal validation and management to support robust drug safety surveillance.

Q1. Which disproportionality measure compares the proportion of reports for a specific adverse event for a given drug to the proportion for all other drugs?

• Reporting Odds Ratio (ROR)
• Proportional Reporting Ratio (PRR)
• Information Component (IC)
• Empirical Bayes Geometric Mean (EBGM)

Correct Answer: Proportional Reporting Ratio (PRR)

Q2. In a 2×2 contingency table used for PRR calculation, which cell represents the number of reports of the specific adverse event for the drug of interest?

• a (drug, event)
• b (drug, other events)
• c (other drugs, event)
• d (other drugs, other events)

Correct Answer: a (drug, event)

Q3. The Information Component (IC) used in Bayesian signal detection is primarily derived from which database system?

• FAERS
• VigiBase (WHO)
• EudraVigilance
• MedWatch

Correct Answer: VigiBase (WHO)

Q4. Which of the following statements best describes the Empirical Bayes Geometric Mean (EBGM)?

• It is a frequentist measure equivalent to chi-square statistic
• It shrinks disproportionality estimates toward the null to reduce false positives
• It is used exclusively for time-to-onset analysis
• It requires individual-level denominator data from population registries

Correct Answer: It shrinks disproportionality estimates toward the null to reduce false positives

Q5. Which threshold rule is commonly used to flag a PRR signal according to Evans et al. (2001)?

• PRR ≥ 2, chi-square ≥ 4, and at least 3 cases
• PRR ≥ 1.5, p-value < 0.05, and at least 5 cases
• PRR ≥ 3, IC025 > 0, and at least 2 cases
• PRR ≥ 2.5, EBGM > 2, and at least 10 cases

Correct Answer: PRR ≥ 2, chi-square ≥ 4, and at least 3 cases

Q6. The Reporting Odds Ratio (ROR) is interpreted as:

• The ratio of observed to expected events in exposed vs unexposed person-time
• The odds of reporting a specific event for the drug versus all other drugs
• The Bayesian posterior probability of causality
• The absolute incidence rate per 1,000 patients

Correct Answer: The odds of reporting a specific event for the drug versus all other drugs

Q7. Which problem occurs when a very commonly reported adverse event for one drug reduces the apparent disproportionality of a true association for another drug?

• Confounding by indication
• Masking
• Simpson’s paradox
• Channeling bias

Correct Answer: Masking

Q8. In observed-versus-expected (O/E) analyses for safety, what is required to calculate the expected number of events?

• Spontaneous report counts only
• Background incidence rates and exposed person-time
• Only the total number of reports in the database
• Case narratives and causality assessments

Correct Answer: Background incidence rates and exposed person-time

Q9. Which statistical issue arises when conducting many disproportionality tests across thousands of drug-event pairs?

• Survivor bias
• Multiple testing leading to inflated false positive rate
• Non-differential misclassification
• Time-varying confounding

Correct Answer: Multiple testing leading to inflated false positive rate

Q10. The Information Component lower 95% credibility limit is denoted as:

• ICupper
• IC025
• IC95
• IClower

Correct Answer: IC025

Q11. Which method is most appropriate to reduce spurious signals from small counts in disproportionality analysis?

• Require a minimum case count threshold and apply shrinkage estimators
• Exclude all events with less than 100 reports
• Use only reporting rates without adjustment
• Apply Bonferroni correction to raw PRR values

Correct Answer: Require a minimum case count threshold and apply shrinkage estimators

Q12. Sequential probability ratio testing (SPRT) in pharmacovigilance is used for:

• Adjusting for confounding in cohort studies
• Continuous monitoring to detect safety signals early
• Calculating background incidence rates
• Transforming MedDRA terms into SOCs

Correct Answer: Continuous monitoring to detect safety signals early

Q13. Which metric compares observed spontaneous reports to those expected from background rates while accounting for age and sex stratification?

• Crude Reporting Rate
• Standardized Morbidity Ratio (SMR)-like O/E
• PRR without stratification
• Kaplan-Meier estimate

Correct Answer: Standardized Morbidity Ratio (SMR)-like O/E

Q14. In Bayesian data mining, shrinkage is applied to:

• Increase variance of rare event estimates
• Bring extreme disproportionality estimates closer to the overall mean
• Remove confounding by indication
• Convert spontaneous reports into incidence rates

Correct Answer: Bring extreme disproportionality estimates closer to the overall mean

Q15. Which is a limitation of disproportionality analyses using spontaneous reporting databases?

• They provide precise incidence rates for adverse events
• They cannot detect signals for rare events
• They are subject to reporting bias and lack of accurate denominators
• They eliminate the need for further pharmacoepidemiologic studies

Correct Answer: They are subject to reporting bias and lack of accurate denominators

Q16. The Reporting Odds Ratio (ROR) is calculated as (a/c) / (b/d). What does ‘d’ represent in this formula?

• Number of reports of the event with the drug
• Number of reports of other events with the drug
• Number of reports of other events with other drugs
• Number of reports of the event with other drugs

Correct Answer: Number of reports of other events with other drugs

Q17. Which approach helps assess temporal association between drug exposure and event onset in spontaneous reports?

• Disproportionality ratios without dates
• Time-to-onset (TTO) analysis and plotting distribution
• Using only aggregated monthly counts
• Applying Bonferroni correction to PRR values

Correct Answer: Time-to-onset (TTO) analysis and plotting distribution

Q18. Which analytic improvement does stratification by age, sex, or region provide in disproportionality analysis?

• Eliminates all reporting bias
• Controls for confounding by these variables and uncovers masked signals
• Converts PRR directly into incidence rates
• Replaces need for shrinkage estimators

Correct Answer: Controls for confounding by these variables and uncovers masked signals

Q19. What is the primary purpose of signal validation after initial statistical detection?

• To immediately update product labeling without review
• To confirm clinical plausibility, exclude artefacts, and prioritize investigations
• To compute new disproportionality metrics only
• To publish the signal in the literature automatically

Correct Answer: To confirm clinical plausibility, exclude artefacts, and prioritize investigations

Q20. Which best practice reduces false positive signals when mining large safety databases?

• Rely solely on PRR thresholds without clinical review
• Combine statistical thresholds with clinical assessment and evidence synthesis
• Ignore stratification and aggregate all reports
• Report every statistical alert as an official safety signal

Correct Answer: Combine statistical thresholds with clinical assessment and evidence synthesis

Download