Home / Blog / Clinical Trial Designs: Parallel, Crossover & Beyond
Clinical Research

Clinical Trial Designs: Parallel, Crossover & Beyond

14 Jun 2026 2 min read By admin

A clinical trial’s design decides what its result can prove. You can classify every design two ways — by how subjects are grouped and by what hypothesis it tests. The map below sorts them all.

Clinical trial designs By structure By hypothesis Parallel groupseparate groups (most common) Crossoverown control + washout Factorial (2×2)two interventions at once Cluster-randomisedgroups, not individuals Adaptive / sequentialadjusts on interim data Superioritybetter than control Non-inferioritynot worse by a pre-set margin Equivalenceneither better nor worse Controls: placebo · active · historical Blinding: open · single · double · triple
Trial designs classified by structure (how subjects are grouped) and by hypothesis (what they aim to show).

Designs by structure

DesignFeature
Parallel groupSubjects randomised to one treatment for the whole study (most common).
CrossoverEach subject receives both treatments in sequence (own control); needs a washout.
Factorial (2×2)Tests two or more interventions simultaneously.
Cluster-randomisedGroups/clusters (hospitals, villages) randomised, not individuals.
Adaptive / sequentialDesign adapts based on accumulating interim data.

Designs by hypothesis

  • Superiority — show the test treatment is better than control.
  • Non-inferiority — show the test is not worse than control by a pre-set margin.
  • Equivalence — show the test is neither better nor worse, within margins.

Controls & blinding

  • Controls: placebo, active comparator, historical, or no treatment.
  • Blinding: open-label, single, double or triple blind.

Crossover design — pros & cons

AdvantagesDisadvantages
Fewer subjects (within-subject comparison)Carryover effect (needs washout)
Removes inter-subject variabilityPeriod effect; longer duration
Each subject is their own controlOnly for chronic, stable conditions (not acute/curable)

Other key concepts

  • Endpoints — primary, secondary and surrogate (defined a priori).
  • Analysis sets — intention-to-treat (ITT) vs per-protocol; plus run-in and enrichment designs.
  • The RCT is the gold standard, reported per CONSORT.
Exam tip: parallel (separate groups) vs crossover (own control + washout) vs factorial (two interventions). Crossover suits chronic, stable disease but risks carryover. Know superiority vs non-inferiority vs equivalence, and that the RCT is the gold standard.

Choose the design to fit the question: parallel for most, crossover to economise on stable disease, factorial to test two things at once — and frame the hypothesis as superiority, non-inferiority or equivalence. Get that right and the trial can actually answer what you asked.

Test yourself

0 / 2

A quick check on this topic — tap an answer for instant feedback.

  1. Q1. A crossover trial requires a ___ between the two treatments:

  2. Q2. A non-inferiority trial aims to show the test treatment is:

Question 1 of 2
Share
Clinical Research
← All articles