Ask Mallard
Study design

Basket trial

4 min read

Test one targeted treatment across disease groups with a shared marker, while keeping each group's results visible.

NHMRC EVIDENCE LEVEL · 2009Level IV · InterventionProvisional design classification. Study quality and certainty of findings require separate appraisal.
Why this level?A single case series has post-test or pre-test/post-test outcomes, without a comparison group. NHMRC 2009 source

The single-arm version taught here has no concurrent control and matches an NHMRC intervention case series, level IV. A randomized basket would be classified by that allocation instead.

Fit

When to use it, and when not to

Use it when

  • You have a treatment aimed at a marker found across tumor types.
  • You can recruit and report each disease cohort with a clear marker assay.

Not when

  • You need a definitive comparison against usual care. Use randomization within eligible cohorts. Read: Randomized controlled trial
  • You need several marker-matched treatments within one tumor type. Use an umbrella trial.

Before you start

What you need

Decide these first
  1. The shared marker and assayDefine positivity, assay performance and specimen handling.
  2. The disease groupsSet eligibility and recruitment targets separately for each group.
  3. The response definition and timingoutcomeTypeUse common assessment rules and record duration and harms.
  4. Whether cohorts share statistical informationSpecify pooling or borrowing rules, stopping rules and sensitivity to different treatment effects.
  5. Assumptions and methods for missing dataCompare missing completely at random, missing at random and missing not at random. Describe which data are absent.
It rests on these
  1. The target is relevant in each tumor context.How to check: Assess biology and prior evidence by disease group.
  2. Response is measured consistently.How to check: Use prespecified assessment times and independent review where feasible.
  3. Any borrowing allows plausible differences between cohorts.How to check: Simulate scenarios with benefit in some groups and no benefit in others.
  4. Missing at random may be plausible after recorded predictors are included.How to check: Consider multiple imputation for missing covariates under missing at random. Multiple imputation is not inherently better.
  5. Missing not at random remains possible.How to check: Use a delta or tipping-point sensitivity analysis for missing not at random.

Plain language

How it works

Enroll patients with different tumor types that share a defined molecular feature. Give the targeted treatment under one protocol and assess each disease cohort. The version taught here is single-arm.

What to preserve

Some basket trials instead randomize within cohorts. A marker can act differently across tissue types, so a pooled response rate may mislead. If the design borrows information across cohorts, state and test the assumptions that make sharing credible. CONSORT applies to randomized variants. For a single-arm basket, report cohort flow, response rules and uncertainty without implying randomization.

Basket trial in outline
One targeted treatment spans tumor types. Keep cohort results separate; the shared marker does not prove shared benefit.Basket trialTumor type A with shared markerTumor type B with shared markerSame targeted treatmentReport type A resultsReport type B resultsOne targeted treatment spans tumor types. Keep cohort results separate; the shared marker does not proveshared benefit.
Conceptual diagram. It shows the structure of the idea, not data from any study.

Worked example

A worked example

Invented for illustration
Setting
Two tumor cohorts share a marker and receive the same drug.
Question
Is there a signal of activity in each group?
Approach
Report each cohort separately before considering any pooling.
Result
Cohort A has 8 responses among 20 patients, or 40 in 100. Cohort B has 2 among 20, or 10 in 100.
What it means
Together that is 10 of 40, or 25 in 100. The pooled rate hides a marked difference and proves no comparative treatment benefit.

In practice

Recent studies PubMed search · not vetted

Searching PubMed…

Watch out

Common mistakes

The one to remember

A shared marker does not guarantee a shared response across tumor types. Uncontrolled response rates describe activity; they do not establish benefit against usual care.

  • Assuming the same marker guarantees the same effect.Instead: Keep cohort results visible. Let differences in tumor biology inform the borrowing plan.
  • Using an uncontrolled response rate as a causal treatment effect.Instead: Treat it as activity evidence. Comparative benefit needs a suitable control design.
  • Dropping patients whose exposure or predictors could not be measured.Instead: This can cause selection bias. Retain people with multiple imputation when missing at random is plausible; assess departures.

Inside Mallard

How Mallard uses this

Plans do not link to this design yet. Mallard has no dedicated sizing form for basket trial.

No Mallard code example points to this page.
Atlas node design.basket_trial, the id a plan cue records.

Related Learn pages

Related designs and methods

Evidence

Sources & further reading

  1. Vemurafenib in Multiple Nonmelanoma Cancers with BRAF V600 Mutations (opens in a new tab) · N Engl J Med · doi 10.1056/NEJMoa1502309
  2. Efficacy of Larotrectinib in TRK Fusion-Positive Cancers in Adults and Children (opens in a new tab) · N Engl J Med · doi 10.1056/NEJMoa1714448
  3. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials (opens in a new tab) · BMJ · doi 10.1136/bmj.c332
  4. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls (opens in a new tab) · BMJ · doi 10.1136/bmj.b2393
Prepared with AI assistance; source titles, authors, years and DOIs checked against PubMed on 2026-11-06. No independent professional approval is claimed. The vemurafenib basket supports reporting responses by cancer type. The larotrectinib analysis pools marker-positive cancers across three protocols; it supports cross-tumor activity, not randomized benefit.
Last source review: 2026-11-06
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.