Crossover trial
A crossover trial is a randomized study in which each participant receives two or more interventions in turn, such as a supplement and then a placebo, so that each person serves as their own comparison.
What it means in practice
Parallel-group trials give each participant one intervention, while crossover trials give each participant a sequence of interventions [3]. The simplest version is the AB/BA design: participants are randomized to start on A or B, then switch to the other [3].
Because each participant acts as their own control, a crossover trial needs fewer participants than a parallel-group trial to reach the same statistical power [1]. The design brings its own problems. Carry-over is when the effect of the first intervention persists into the second period, so the result depends on the order in which treatments were received [1]. A period effect is when the outcome changes over time regardless of treatment [1]. A washout is a gap between periods with no treatment, meant to let the first treatment wear off [1].
Cochrane's handbook describes the design as suited to interventions with a temporary effect in stable, chronic conditions, and warns that carry-over is likely when a substance has a very long elimination half-life [3].
Why it matters when reading nootropics research
A trial might report that the same volunteers completed attention tests after a supplement on one visit and after a placebo on another. That is a crossover design, and it is only as good as its handling of order effects.
Reporting has often been thin. In a review of 116 randomized crossover trials indexed in MEDLINE in December 2000, the median sample size was 15, carry-over was addressed in the methods of 29% of trials, and 70% reported and defended a washout period [2]. Only 29% gave confidence intervals or standard errors that would let the data enter a meta-analysis [2]. A CONSORT extension now sets out what crossover reports should include, with a modified flowchart and baseline table [1].
Things to check: how long the washout was, how the design and analysis addressed carry-over and period effects, and whether the analysis accounted for repeated measurements from the same people.
References
- Dwan K, Li T, Altman DG, et al. CONSORT 2010 statement: extension to randomised crossover trials. BMJ. 2019.
- Mills EJ, Chan AW, Wu P, et al. Design, analysis, and presentation of crossover trials. Trials. 2009.
- Cochrane. Cochrane Handbook for Systematic Reviews of Interventions, Chapter 23: Including variants on randomized trials. October 2019 (chapter last updated).