Help
What is a Crossover Design?
A crossover design is a repeated measures design where each subject receives multiple treatments in a specific sequence. Key features:
- Each subject serves as their own control
- Reduces between-subject variability
- Requires fewer subjects than parallel-group designs
- Treatments separated by washout periods
Design Types Explained
- Latin Square: Each treatment appears once in each period, and each subject receives each treatment once. Balanced for period effects.
- Williams Square: Extension of Latin Square that also balances first-order carryover effects. Each treatment follows every other treatment equally often.
- All Sequences: Complete counterbalancing with all possible orderings. Uses more sequences but provides maximum balance.
Washout Period
The washout period allows the effects of the previous treatment to dissipate before starting the next treatment. Calculate based on:
- Drug half-life: Time for concentration to reduce by 50%
- 5 half-lives: Standard washout (96.875% elimination)
- Physiological effects: May need longer if biological effects persist
- Disease stability: Ensure condition returns to baseline
When to Use Crossover Designs
- Good for: Chronic stable conditions, symptomatic treatments, within-subject comparisons
- Not suitable for: Curative treatments, progressive diseases, long-lasting effects
- Examples: Pain medications, asthma treatments, antihypertensive drugs, cognitive enhancers
Example: 2×2 Crossover (AB/BA)
The classic two-treatment, two-period crossover:
- Sequence 1: Drug A → Washout → Drug B
- Sequence 2: Drug B → Washout → Drug A
- Balances order effects
- Half the subjects needed vs parallel design