Introduction:
This blog presents a focused set of multiple-choice questions on crossover study designs, tailored for M.Pharm students studying Research Methodology & Biostatistics. Crossover trials are widely used in pharmacology and bioequivalence studies because each participant receives multiple treatments, allowing within-subject comparisons and greater statistical efficiency. The questions below cover design types (2×2, Latin square, Williams), washout and carryover effects, appropriate statistical models (Grizzle, mixed-effects ANOVA), sample size and analysis strategies, and practical issues such as dropout handling and bioequivalence criteria. Answers are provided to aid self-assessment and to deepen understanding of methodological choices critical for clinical pharmacology research.
Q1. In a standard 2×2 crossover design (AB/BA), what is the primary advantage compared to a parallel-group design?
- Each subject receives only one treatment, reducing complexity
- It eliminates all period and sequence effects
- Within-subject comparison reduces variability and required sample size
- Treatment carryover is never a concern
Correct Answer: Within-subject comparison reduces variability and required sample size
Q2. Which component is explicitly modeled in the classic Grizzle ANOVA for a 2×2 crossover?
- Carryover, period, treatment and subject-within-sequence effects
- Only treatment and period effects, ignoring carryover
- Random slopes for treatment over time
- Time-varying covariates as fixed effects
Correct Answer: Carryover, period, treatment and subject-within-sequence effects
Q3. What is the usual recommended approach if a significant carryover effect is detected in a 2×2 crossover?
- Ignore it and proceed with pooled analysis
- Analyze only first-period data as a parallel-group comparison
- Combine periods and adjust by increasing sample size
- Switch to nonparametric tests for both periods
Correct Answer: Analyze only first-period data as a parallel-group comparison
Q4. In bioequivalence studies comparing AUC and Cmax, why are pharmacokinetic measures often log-transformed before analysis?
- Log transformation makes the data binary
- It stabilizes variance and makes ratios symmetric for confidence intervals
- Log transform eliminates carryover effects
- It always produces normally distributed data regardless of original distribution
Correct Answer: It stabilizes variance and makes ratios symmetric for confidence intervals
Q5. Which design is specifically used to control first-order carryover across multiple treatments with minimal subjects?
- Parallel-group randomized design
- Latin square or Williams design
- Factorial crossover design
- Open-label sequential design
Correct Answer: Latin square or Williams design
Q6. When calculating sample size for a 2×2 crossover aimed at demonstrating bioequivalence, which parameter is most critical?
- Between-subject variance only
- Within-subject variance of the log-transformed PK measure
- The expected dropout rate only
- Number of sequences regardless of variance estimates
Correct Answer: Within-subject variance of the log-transformed PK measure
Q7. Which assumption underlies the increased efficiency of crossover designs?
- Treatment effects are identical across all subjects and times
- Within-subject correlation is negligible
- Within-subject variability is smaller than between-subject variability
- No period effects exist
Correct Answer: Within-subject variability is smaller than between-subject variability
Q8. In a 2×2 crossover, what does a sequence effect refer to?
- An interaction between subject demographics and treatment
- Difference attributable to the order in which treatments are given
- Random measurement error within a period
- Technical drift in laboratory assays over time
Correct Answer: Difference attributable to the order in which treatments are given
Q9. Which analytic approach is preferred for modern crossover analysis when missing data and unequal variances are possible?
- Paired t-test on change scores only
- Mixed-effects model with fixed effects for sequence, period and treatment and random subject effects
- Nonparametric Wilcoxon signed-rank test always
- Simple ANOVA ignoring subject-level random effects
Correct Answer: Mixed-effects model with fixed effects for sequence, period and treatment and random subject effects
Q10. How is washout period length primarily determined in crossover pharmacokinetic studies?
- By regulatory guidance irrespective of drug pharmacology
- Using a fixed 24-hour rule for all drugs
- Based on drug half-life, typically several half-lives to minimize carryover
- By the longest expected treatment period in the study
Correct Answer: Based on drug half-life, typically several half-lives to minimize carryover
Q11. What is the principal disadvantage of crossover designs?
- They always require larger sample sizes than parallel trials
- Risk of carryover and longer study duration per subject
- They cannot control within-subject variability
- They are unsuitable for chronic stable conditions
Correct Answer: Risk of carryover and longer study duration per subject
Q12. In a Williams design for four treatments, the primary goal is to:
- Maximize number of periods while ignoring balance
- Balance first-order carryover and sequence effects across treatments
- Ensure each subject receives only one treatment
- Assess dose-response within each subject
Correct Answer: Balance first-order carryover and sequence effects across treatments
Q13. Which statement about testing for carryover in crossover trials is correct?
- Significant carryover testing is mandatory and always reliable
- Testing for carryover can be underpowered and is controversial; prevention is preferable
- If carryover test is non-significant, no washout is needed
- Carryover tests remove the need for randomization
Correct Answer: Testing for carryover can be underpowered and is controversial; prevention is preferable
Q14. For bioequivalence, the common acceptance criterion for the 90% confidence interval of the geometric mean ratio is:
- 0.5 to 2.0
- 0.8 to 1.25
- 0.9 to 1.1
- 0.7 to 1.3
Correct Answer: 0.8 to 1.25
Q15. Which effect is tested by including period as a fixed effect in the crossover ANOVA?
- Carryover from previous treatments
- Systematic differences between study periods (e.g., time trends)
- Within-subject random variability
- Subject-by-treatment interaction
Correct Answer: Systematic differences between study periods (e.g., time trends)
Q16. Incomplete or two-stage crossover designs are employed primarily when:
- All subjects can tolerate all treatments without dropout
- There are too many treatments to administer to every subject in a feasible timeframe
- One-period designs are always preferred
- Carryover effects are desired to be maximized
Correct Answer: There are too many treatments to administer to every subject in a feasible timeframe
Q17. What does intra-subject correlation (or within-subject correlation) in crossover trials indicate?
- Correlation between measurements across different subjects
- Degree to which repeated measures on the same subject are related
- Correlation between randomization sequence and outcome
- Association between period number and dropout rate
Correct Answer: Degree to which repeated measures on the same subject are related
Q18. If a two-treatment crossover shows a significant treatment-by-period interaction, the appropriate interpretation is:
- No treatment effect exists
- Treatment effects vary by period, suggesting potential carryover or time-dependent effects
- The study is automatically bioequivalent
- Sequence assignment was successful and no adjustment is needed
Correct Answer: Treatment effects vary by period, suggesting potential carryover or time-dependent effects
Q19. When handling missing data in crossover trials, which strategy is generally recommended?
- Exclude all subjects with any missing data from analysis
- Use modern mixed-model methods that can handle missing-at-random data without biased complete-case removal
- Impute missing data deterministically using the grand mean only
- Replace missing values with zeros to maintain sample size
Correct Answer: Use modern mixed-model methods that can handle missing-at-random data without biased complete-case removal
Q20. Which of the following best describes an advantage of Latin square crossover designs?
- They require no randomization
- They can control for two blocking factors (period and sequence) while balancing treatments
- They are suitable only for two treatments
- They eliminate within-subject variability completely
Correct Answer: They can control for two blocking factors (period and sequence) while balancing treatments
