Ask Mallard

Crossover trial

A crossover trial gives each participant both treatments in a randomly assigned order, so each person acts as their own control.
NHMRC EVIDENCE LEVEL · 2009Level II · InterventionProvisional design classification. Study quality and certainty of findings require separate appraisal.
Why this level?Participants or clusters are randomly allocated to concurrent comparison groups. NHMRC 2009 source

Each participant receives both conditions in a randomised order, so each is their own comparison. Carryover between periods is the threat this design has to answer, and the washout is how it answers it.

What question does this help answer?

The same question a parallel trial answers, but with the between-person variation removed. Because each comparison is within a person, far fewer participants are needed for the same precision.

When to use it

Use it for a stable chronic condition, an outcome that responds quickly and then reverses, and a treatment whose effect washes out. Symptom control in asthma, pain, or reflux are the classic settings.

When not to use it

Do not use it when the condition progresses during the trial, when the outcome is permanent (death, cure, a first event), or when the treatment can change the disease rather than just its symptoms. In those settings the two periods are not comparable and the design is invalid rather than merely inefficient.

How it works in plain language

Participants are randomized to a sequence rather than a treatment: some receive A then B, others B then A. A washout period separates them. The analysis compares treatments within participants and must account for the PERIOD as well, because outcomes often drift over time regardless of treatment.

What the two sequences look like

A two-period crossover. Participants are randomized to a sequence: one sequence receives A, then a washout, then B; the other receives B, then a washout, then A. Each participant receives both treatments once.Crossover: randomized to a sequence, with a washout between periodsFirst periodWashoutSecond periodSequence A then BTreatment AWashoutTreatment BSequence B then ATreatment BWashoutTreatment ATreatment AWashoutTreatment BEach participant is their own comparison, so the period must be modelled too. Carryover that the washoutdoes not remove breaks the design, and testing for it afterwards is underpowered.
Conceptual diagram. It shows the structure of the idea, not data from any study.

Why Mallard may recommend it

Mallard may recommend a crossover when the condition described is chronic and stable, the outcome is a symptom or measurement rather than an event, and recruitment is the binding constraint.

Key assumptions and design requirements

  • No carryover: each participant returns to baseline before the next period begins.
  • The washout is long enough for the treatment's effect to disappear, and its length is justified.
  • The analysis adjusts for period as well as treatment.
  • Dropout is analysed carefully, since a participant lost after one period contributes no within-person comparison.
  • Sample size accounts for the WITHIN-person standard deviation, which is what the design exploits.

What to watch out for

The design assumes each participant returns to baseline between periods, and a carryover effect breaks that assumption in a way the data cannot reliably diagnose: testing for carryover after the fact is underpowered and biases the main comparison. It suits stable chronic conditions and outcomes that reverse, and it is unusable when the condition progresses or the outcome is permanent.

Clinical example

Question: Does a new inhaler improve morning peak flow more than the standard one? Design: Randomize stable asthma patients to inhaler A for four weeks then B, or B then A, with a two-week washout, and compare peak flow within participants while adjusting for period.

How to report it

CONSORT extension for randomised crossover trials is the guideline for this design. It asks for the order of interventions each participant was randomised to, the number and duration of the periods and of the washout, how carryover was considered, and an analysis based on comparisons within each participant. A reporting guideline is not a design manual: use it to make the study transparent, not as a substitute for careful design. Check the target journal's own requirements as well, which this page cannot know.

Code for this method
Loading…
Recent studies that used this
Searching PubMed…
Related Learn pages
Sources & further reading
  1. Overview of clinical study designs (opens in a new tab) · Clin Exp Emerg Med · doi 10.15441/ceem.23.036
  2. Clinical research study designs: The essentials (opens in a new tab) · Pediatr Investig · doi 10.1002/ped4.12166
  3. CONSORT 2010 statement: extension to randomised crossover trials (opens in a new tab) · BMJ · doi 10.1136/bmj.l4378
THINLY SOURCED, and flagged as such: both references are general design overviews rather than crossover methodology. A crossover-specific methods reference should replace one of them before this page is published. Neither covers carryover testing, which is the design's central difficulty.
Last source review: 2026-09-06
AI-generated educational content
Mallard uses reputable methodological and reporting sources, but this material may contain errors or omit important context. Verify important methodological decisions with the cited sources and, when appropriate, a qualified research methodologist or statistician.