Ask Mallard

N-of-1 trial

An N-of-1 trial is a randomized, usually blinded, multiple-crossover trial run within a single patient to answer a question about that patient.
NHMRC EVIDENCE LEVEL · 2009Not yet classified · InterventionProvisional design classification. Study quality and certainty of findings require separate appraisal.
Why this level?The design family and its defining features have not been specified for this hierarchy. NHMRC 2009 source

The unit is one patient, and the hierarchy rates what a design can support for a population. A randomised multiple-crossover in one person can be strong evidence for that person and says nothing on its own about anyone else -- a distinction the ladder does not make.

What question does this help answer?

It answers "does this treatment work for THIS person", which is often the question in front of a clinician and is not the question a parallel-group trial answers.

When to use it

Use it for a chronic stable condition, a treatment with rapid onset and offset, and an outcome that can be measured repeatedly and reverses. It is well suited to deciding whether to continue a medication whose benefit for an individual is genuinely unclear.

When not to use it

It cannot be used for conditions that progress, for outcomes that occur once, or for treatments that cure or permanently modify disease. It also cannot, on its own, answer a population question.

How it works in plain language

The patient receives treatment and comparator in multiple randomized, blinded pairs of periods, with washouts between. Outcomes are recorded in each period and compared within the patient. Repeating the pairs is what distinguishes signal from day-to-day fluctuation.

What the treatment periods look like

An N-of-1 trial in one person. The treatments are given in repeated pairs of periods with a washout between them, and the order within each pair is randomized, so it differs from pair to pair.N-of-1 trial: one person, repeated pairs, randomized order within each pairFirst periodWashoutSecond periodFirst pairTreatment AWashoutTreatment BSecond pairTreatment BWashoutTreatment AThird pairTreatment AWashoutTreatment BTreatment AWashoutTreatment BThe answer applies to the person studied. It needs a stable condition, an outcome that responds quicklyand reverses, and a short washout.
Conceptual diagram. It shows the structure of the idea, not data from any study.

Why Mallard may recommend it

Mallard may recommend an N-of-1 design when the question is explicitly about an individual patient's response and the condition and treatment meet the stability and reversibility requirements.

Key assumptions and design requirements

  • The condition is stable over the trial and the treatment effect reverses on withdrawal.
  • Periods are randomized and, where possible, blinded to patient and assessor.
  • Enough pairs of periods are run to distinguish effect from variability.
  • The outcome is measured repeatedly with an instrument sensitive enough for within-person change.
  • Aggregating a SERIES of N-of-1 trials is the route to a population estimate, and needs its own analysis plan.

What to watch out for

The answer applies to the patient studied, and generalizing it requires aggregating a series of such trials. It needs a condition that is stable, an outcome that responds quickly and reverses, and a treatment with a short washout: most questions do not meet all three.

Clinical example

Question: Does this patient's chronic pain actually respond to the drug they have been taking for a year? Design: Four randomized blinded pairs of two-week periods on drug and placebo with washouts, with daily pain scores, compared within the patient.

How to report it

CENT, the CONSORT extension for N-of-1 trials, is the guideline for this design. Report the number and length of treatment periods, how the order of treatments was randomised and whether it was blinded, any washout between periods, and how the analysis handled carryover and the repeated measurements within the one patient. 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.

Recent studies that used this
Searching PubMed…
Related Learn pages
Sources & further reading
  1. CONSORT extension for reporting N-of-1 trials (CENT) 2015 Statement (opens in a new tab) · BMJ · doi 10.1136/bmj.h1738
  2. Overview of clinical study designs (opens in a new tab) · Clin Exp Emerg Med · doi 10.15441/ceem.23.036
CENT covers reporting rather than analysis. Methods for aggregating a series of N-of-1 trials into a population estimate are outside it and need a separate reference.
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.