Meta-analysis models and pooled data interpretation MCQs With Answer

Meta-analysis models and pooled data interpretation MCQs With Answer

This set of 20 multiple-choice questions is crafted for M.Pharm students to reinforce advanced understanding of meta-analysis models and pooled-data interpretation. The questions cover fixed-effect and random-effects models, heterogeneity measures (I², Tau², Cochran’s Q), choice and transformation of effect measures (OR, RR, MD, SMD), weighting and pooling methods, publication bias and funnel-plot interpretation, meta-regression, subgroup and sensitivity analyses, prediction intervals, and robust estimators such as DerSimonian–Laird and Hartung–Knapp. Each item contains plausible distractors and a clear correct answer to support exam preparation and critical appraisal in pharmacoepidemiology and pharmacoeconomics.

Q1. Which assumption best describes the fixed-effect model in meta-analysis?

• All studies estimate different but related true effects
• There is a single true effect size shared by all studies
• Study-level covariates explain all between-study variance
• Effects vary randomly with no common distribution

Correct Answer: There is a single true effect size shared by all studies

Q2. Which statement about the random-effects model is true?

• It assumes no between-study heterogeneity
• It pools study estimates giving larger studies exclusive weight
• It models a distribution of true effects across studies
• It uses study sample size only to determine weights

Correct Answer: It models a distribution of true effects across studies

Q3. Which metric quantifies the proportion of variability due to heterogeneity rather than chance?

• Cochran’s Q statistic
• Tau-squared (Tau²)
• I² statistic
• Standard error of pooled estimate

Correct Answer: I² statistic

Q4. What does Tau² represent in a random-effects meta-analysis?

• Proportion of total variation due to sampling error
• Estimated between-study variance of true effects
• Chi-squared value for heterogeneity test
• Average within-study variance

Correct Answer: Estimated between-study variance of true effects

Q5. Which method is a commonly used estimator for between-study variance?

• Inverse-variance fixed estimator
• DerSimonian–Laird method
• Kaplan–Meier estimator
• Bonferroni correction

Correct Answer: DerSimonian–Laird method

Q6. In meta-analysis of odds ratios, why are log transformations commonly used?

• To force the OR to lie between 0 and 1
• To stabilize variances and normalize the sampling distribution
• To eliminate the need for confidence intervals
• To convert OR into absolute risk differences

Correct Answer: To stabilize variances and normalize the sampling distribution

Q7. What is the purpose of a funnel plot in meta-analysis?

• To display time-to-event data across trials
• To visually assess potential publication bias or small-study effects
• To compute the pooled effect estimate directly
• To estimate between-study variance numerically

Correct Answer: To visually assess potential publication bias or small-study effects

Q8. Cochran’s Q test primarily assesses which aspect of meta-analysis?

• Presence of publication bias
• Statistical significance of pooled effect size
• Presence of heterogeneity beyond chance
• Appropriateness of effect measure (OR vs RR)

Correct Answer: Presence of heterogeneity beyond chance

Q9. When would you prefer to report a prediction interval instead of only a confidence interval?

• When you need the pooled estimate precision across included studies
• When predicting the range of true effects in a new individual study
• When sample sizes are identical across studies
• When there is no heterogeneity detected

Correct Answer: When predicting the range of true effects in a new individual study

Q10. Which pooling metric is most appropriate for continuous outcomes measured on the same scale across studies?

• Odds ratio (OR)
• Risk ratio (RR)
• Mean difference (MD)
• Standardized mean difference (SMD)

Correct Answer: Mean difference (MD)

Q11. The standardized mean difference (SMD) is used primarily when:

• Outcomes are dichotomous and rare
• Continuous outcomes use different measurement scales across studies
• All studies report identical scale units
• Time-to-event outcomes are reported

Correct Answer: Continuous outcomes use different measurement scales across studies

Q12. Which weighting method is most common in inverse-variance meta-analysis?

• Weight proportional to study sample size only
• Weight proportional to inverse of within-study variance
• Equal weighting of all studies
• Weight based on study quality score

Correct Answer: Weight proportional to inverse of within-study variance

Q13. Which of the following describes the Hartung–Knapp adjustment?

• An approach to remove small-study effects from funnel plots
• A method to calculate heterogeneity using I²
• An alternative method to derive more accurate confidence intervals in random-effects meta-analysis
• A technique to impute missing outcome data in trials

Correct Answer: An alternative method to derive more accurate confidence intervals in random-effects meta-analysis

Q14. Meta-regression is primarily used to:

• Estimate the pooled effect when there is no heterogeneity
• Assess whether study-level covariates explain heterogeneity in effect sizes
• Detect publication bias using regression intercepts
• Replace subgroup analysis in all circumstances

Correct Answer: Assess whether study-level covariates explain heterogeneity in effect sizes

Q15. What is a common limitation of doing multiple subgroup analyses in a meta-analysis?

• They always increase statistical power
• They can lead to spurious findings due to multiple comparisons
• They eliminate between-study heterogeneity completely
• They are a substitute for sensitivity analyses

Correct Answer: They can lead to spurious findings due to multiple comparisons

Q16. In presence of zero events in one arm of a trial, which approach is commonly applied for pooling binary outcomes?

• Exclude the study entirely without adjustment
• Apply a continuity correction (e.g., add 0.5 to cells)
• Convert the outcome to a continuous measure
• Use Kaplan–Meier estimates instead

Correct Answer: Apply a continuity correction (e.g., add 0.5 to cells)

Q17. Which procedure helps evaluate the robustness of meta-analysis results to analytic choices?

• Sensitivity analysis
• Forest plot only
• Study-level randomization
• Descriptive table without statistics

Correct Answer: Sensitivity analysis

Q18. Small-study effects refer to:

• The phenomenon where smaller studies tend to show different, often larger, effect sizes than larger studies
• The reduction of bias by including only small trials
• The statistical technique to weight studies equally
• The exclusive presence of heterogeneity in large trials

Correct Answer: The phenomenon where smaller studies tend to show different, often larger, effect sizes than larger studies

Q19. In a forest plot, a pooled effect whose confidence interval crosses the line of no effect implies:

• The pooled estimate is statistically significant at conventional alpha
• The pooled estimate is not statistically significant at conventional alpha
• The meta-analysis has no heterogeneity
• All individual studies are non-significant

Correct Answer: The pooled estimate is not statistically significant at conventional alpha

Q20. Which guideline is most relevant when reporting systematic reviews and meta-analyses?

• CONSORT
• STROBE
• PRISMA
• ARRIVE

Correct Answer: PRISMA

Download