Dropouts in clinical studies MCQs With Answer

Introduction: Dropouts in clinical studies are participants who discontinue or become lost to follow-up before a study’s planned endpoint. For M.Pharm students, understanding dropouts is essential because they influence internal validity, statistical power, and regulatory acceptance of trial results. This blog offers focused multiple-choice questions that explore reasons for dropout, classifications (e.g., MCAR, MAR, MNAR), analytical methods to handle missing data (intention-to-treat, per-protocol, LOCF, multiple imputation, mixed models), impact on bias and precision, design strategies to minimize dropout, and regulatory/reporting expectations. These MCQs aim to deepen conceptual and applied knowledge for robust trial design, analysis, and interpretation in pharmaceutical research.

Q1. What classification describes missing outcome data that are related only to observed data but not to the unobserved outcome itself?

• Missing Completely at Random (MCAR)
• Missing at Random (MAR)
• Missing Not at Random (MNAR)
• Informative Censoring

Correct Answer: Missing at Random (MAR)

Q2. Which analysis population includes all randomized participants in the groups to which they were assigned, regardless of adherence or dropout?

• Per-Protocol Analysis
• As-Treated Analysis
• Intention-to-Treat (ITT) Analysis
• Complier Average Causal Effect (CACE) Analysis

Correct Answer: Intention-to-Treat (ITT) Analysis

Q3. Which approach to handling missing data potentially biases estimates by carrying the last observed measurement forward to later time points?

• Multiple Imputation
• Last Observation Carried Forward (LOCF)
• Mixed-Effects Model for Repeated Measures (MMRM)
• Inverse Probability Weighting

Correct Answer: Last Observation Carried Forward (LOCF)

Q4. Which type of dropout is most likely to introduce bias because the reason for missingness depends on the unobserved outcome itself?

• Missing Completely at Random (MCAR)
• Missing at Random (MAR)
• Missing Not at Random (MNAR)
• Non-informative Dropout

Correct Answer: Missing Not at Random (MNAR)

Q5. Which statistical method models all available repeated measures and makes valid inferences under MAR without imputing missing values explicitly?

• Complete-Case Analysis
• Mixed-Effects Model for Repeated Measures (MMRM)
• Last Observation Carried Forward (LOCF)
• Simple Mean Imputation

Correct Answer: Mixed-Effects Model for Repeated Measures (MMRM)

Q6. In sample size calculation, how are anticipated dropouts typically handled?

• Ignore dropouts and calculate sample size solely on effect size
• Increase the sample size by inflating for expected dropout rate
• Decrease sample size because dropouts reduce variance
• Use LOCF to adjust sample size downwards

Correct Answer: Increase the sample size by inflating for expected dropout rate

Q7. Which regulatory reporting guideline explicitly requires a flow diagram showing numbers of participants assessed for eligibility, randomized, and who completed or discontinued treatment?

• STROBE
• PRISMA
• CONSORT
• ARRIVE

Correct Answer: CONSORT

Q8. Differential dropout between treatment arms primarily threatens which aspect of a randomized trial?

• External validity only
• Internal validity by introducing selection bias
• Blinding but not bias
• Randomization sequence generation

Correct Answer: Internal validity by introducing selection bias

Q9. Which missing-data mechanism allows unbiased estimates from complete-case analysis provided the mechanism is true?

• Missing at Random (MAR)
• Missing Not at Random (MNAR)
• Missing Completely at Random (MCAR)
• Informative Censoring

Correct Answer: Missing Completely at Random (MCAR)

Q10. Which method uses multiple plausible datasets to reflect uncertainty about missing values and combines results using Rubin’s rules?

• Single Imputation with Mean
• Multiple Imputation
• Complete-Case Analysis
• Last Observation Carried Forward (LOCF)

Correct Answer: Multiple Imputation

Q11. In survival analysis, dropout due to unrelated reasons that stops follow-up but is independent of the future event hazard is termed:

• Informative Censoring
• Non-informative Censoring
• Left Censoring
• Interval Censoring

Correct Answer: Non-informative Censoring

Q12. Which sensitivity analysis approach evaluates how conclusions change when assuming different missing-data mechanisms, including MNAR?

• Complete-Case Sensitivity
• Pattern-Mixture Models and Selection Models
• Standard Per-Protocol Analysis only
• Single Best-Imputation Check

Correct Answer: Pattern-Mixture Models and Selection Models

Q13. Which retention strategy is least likely to reduce dropout in a long-term clinical trial?

• Regular participant reminders and flexible visit schedules
• Financial compensation and reimbursement for travel
• Keeping follow-up visits strictly at inconvenient times without alternatives
• Establishing strong participant–investigator communication

Correct Answer: Keeping follow-up visits strictly at inconvenient times without alternatives

Q14. Which analytic approach weights participants by the inverse probability of being observed to handle dropout-related bias?

• Multiple Imputation
• Inverse Probability Weighting (IPW)
• Last Observation Carried Forward (LOCF)
• Hot-Deck Imputation

Correct Answer: Inverse Probability Weighting (IPW)

Q15. For non-inferiority trials, why are dropouts a special concern when using per-protocol analyses?

• Per-protocol analyses ignore efficacy entirely
• Dropouts can bias results toward non-inferiority if non-adherent participants are excluded
• Per-protocol analyses always increase type I error regardless of dropouts
• Dropouts have no impact on non-inferiority conclusions

Correct Answer: Dropouts can bias results toward non-inferiority if non-adherent participants are excluded

Q16. Which missing-data handling method incorrectly assumes that the missing value equals the overall sample mean and can understate variability?

• Regression Imputation
• Mean Imputation
• Multiple Imputation
• Mixed-Effects Modeling

Correct Answer: Mean Imputation

Q17. Which statement is true regarding Last Observation Carried Forward (LOCF)?

• LOCF always provides unbiased estimates under MNAR
• LOCF preserves within-subject trajectories perfectly
• LOCF can underestimate variability and bias treatment effects
• LOCF is the gold standard recommended by regulators for handling missing data

Correct Answer: LOCF can underestimate variability and bias treatment effects

Q18. What is the primary reason to perform a sensitivity analysis related to dropouts?

• To confirm that the primary analysis code runs correctly
• To assess how robust study conclusions are to different assumptions about missing data
• To replace CONSORT reporting requirements
• To increase the nominal sample size post hoc

Correct Answer: To assess how robust study conclusions are to different assumptions about missing data

Q19. Which outcome of high dropout rates is most likely to reduce statistical power?

• Increased internal validity
• Reduced effective sample size
• Perfect balance between groups
• Lower type II error

Correct Answer: Reduced effective sample size

Q20. Which practice is recommended in trial protocols to handle anticipated dropouts and missing data?

• No description is necessary; handle missing data after study completion
• Pre-specify primary analysis population, missing-data assumptions, and sensitivity analyses
• Plan to exclude all participants with any missing data from all analyses
• Guarantee that there will be zero dropouts so no plan is needed

Correct Answer: Pre-specify primary analysis population, missing-data assumptions, and sensitivity analyses

Download