A basket trial tests one targeted treatment in patients grouped by a shared biomarker, whatever their tumour type. The data problem is small cohorts, strict biomarker eligibility and a stop-or-continue decision for each basket. This page shows how to set that up in an EDC, what the numbers look like, and what stays with your statistician.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Baskets open
6
Enrolled
54
Responders (demo)
17
Baskets under review
1
14 of 20
12 of 20
9 of 20
8 of 20
7 of 20, under review
4 of 20
The short answer
The design
In a basket trial, a single investigational therapy is tested in patients who share a molecular feature, for example a particular gene alteration, across different tumour histologies. Each histology forms a basket. Typical primary endpoints are objective response rate or disease control measured against pre-specified criteria, and a common structure is a small first stage per basket followed by expansion if enough responses are seen. Statistical designs range from independent baskets, in which each is analysed alone, to hierarchical or Bayesian models that borrow information between baskets. The choice is the statistician's, and the background is in the Bayesian clinical trial design guide.
Do not confuse this with a platform trial. A platform trial is about time: treatment arms enter and leave under a master protocol with a shared control, as described in adaptive platform trial software. A basket trial is about population: one treatment, many histologies, usually no concurrent control, and a decision per basket. The software needs differ accordingly. A platform trial stresses amendments and shared-control dates; a basket trial stresses eligibility documentation, small-cohort data quality and per-basket reporting.
Basket designs appear most often in oncology and rare disease, where targeted therapies and sparse populations make histology-independent enrolment attractive. For the therapeutic area view see EDC for oncology clinical trials and EDC for rare disease clinical trials. Capture is suitable for Phase 1 to 3; most basket trials are Phase 2, covered by EDC for Phase 2 trials.
Data model
There are two sensible patterns. Choose with your statistician before the build, because it affects every export.
| Question | Baskets as arms | Baskets as a coded variable |
|---|---|---|
| What it means | Each histology is an arm in the study design | One treatment arm; histology is a required, coded eligibility field |
| Visit schedule | Arm-specific visits where baskets differ | One schedule for all baskets |
| Best when | Baskets have different assessments or follow-up | Baskets share assessments and follow-up |
| Reporting per basket | Filter by arm | Filter or group on the basket variable in the export |
| Adding a basket | Add an arm and its visits through an amendment | Add a code to the list, through an amendment |
| Risk | More structure to maintain | Easier to mis-code at entry, so add an edit check |
Capture has epochs, arms and arm-specific visits, and dropdown or single-choice fields with edit checks. Neither pattern is a dedicated basket feature, so confirm the chosen pattern on a demo.
Eligibility
The biomarker result is the entry ticket for every patient, so it deserves its own section on the eligibility form, with the assay, the result, the date and the reference to the source report. The form below is a demo.
Histology and biomarker
Histology (basket)
BASKETBiomarker result
BMRSLTAuto-query: indeterminate result does not meet inclusion criterion 4
Assay reference
BMASSAYTest date
BMDTCUpload the protocol to the free sandbox and review the AI-drafted eligibility and baseline forms. Nothing is saved without approval. No credit card, no time limit.
Enrolment tracking
A basket tracker answers three questions: who is in, who has been assessed, and which baskets are near a decision point. The columns below are examples from a demo study.
| Basket | Screened | Enrolled | Evaluable | Responses | Status |
|---|---|---|---|---|---|
| Colorectal | 19 | 14 | 12 | 4 | Open |
| Lung | 17 | 12 | 10 | 3 | Open |
| Biliary | 12 | 9 | 7 | 3 | Open |
| Thyroid | 10 | 8 | 6 | 3 | Open |
| Bladder | 11 | 7 | 7 | 1 | Under review |
| Other solid | 6 | 4 | 2 | 0 | Open |
Demo data. Evaluable and Responses come from the response assessment form; the Status column is a human decision recorded by the committee, not calculated by the EDC.
Where the counts come from
Capture does not have a dedicated basket dashboard, so the per-basket view is built from what the system gives you. Screening and enrolment come from the participant records and the basket field, the response category from the response assessment form, and the adverse event rows from the safety forms. Export by date range and by site, group on the basket variable, and publish the tracker as a simple table on a schedule. The screening and enrolment log template is a useful starting layout and the enrolment forecast calculator projects when a basket might reach its target.
Small baskets reward discipline. If ten patients decide a stage, one missing assessment is ten percent of the evidence, so run a weekly query listing for the evaluable population and chase it. Edit checks that raise an auto-query when a response assessment is missing at the scheduled visit, or the date falls outside the window, do the first pass; a data manager does the rest. The tools for that are in query management software and clinical data cleaning.
Because oncology basket trials usually run across several centres, keep the forms identical at all sites and review the exports by site; cross-site data harmonisation covers the differences that creep in. Site coordinators see names; researchers see coded IDs, so statisticians and committees work on coded data.
Setup
Baskets as arms or as a coded variable. Write the decision into the data management plan so exports are reproducible.
Upload the protocol; the AI builder drafts the visit schedule and forms, and you review every field. Use templates for adverse events, concomitant medications and laboratory data.
Required histology and biomarker fields, with edit checks tied to each inclusion criterion and an eligibility confirmation signed by the investigator.
A response assessment form that mirrors your protocol criteria, plus adverse event forms. The CTCAE grade lookup, ECOG performance status template and adverse event template save time.
Researcher, site coordinator and admin roles, with any unblinded roles limited. Run scripted tests in the sandbox for each edit check and skip rule.
Move forms from draft to approved, because only approved forms appear in the casebook, then enrol site by site.
Decide who sees the per-basket table, how often it is refreshed and who records the committee decision for each basket.
Division of labour
| Task | Capture | Statistician or committee |
|---|---|---|
| Capture biomarker, histology and basket | Yes: required fields, edit checks, audit trail | Defines the criteria |
| Count participants per basket | Exports and tracking tables from the basket variable | Confirms the evaluable population definition |
| Response classification | Records the assessment against your form | Applies the protocol criteria and any central review |
| Stage-one stopping for a basket | Not applied by the EDC | Pre-specified rule applied by the statistician |
| Borrowing between baskets or Bayesian models | Supplies cleaned data and SDTM exports | Runs the model in a validated environment |
| Safety review | Adverse event forms and queries, SAE listings | Committee review of safety signals |
| Audit | Field-level, append-only audit trail | Inspection strategy |
Safety reporting timelines are set by the protocol and regulations, not by the EDC.
Cost
Basket trials tend to be modest in size, which suits the published price list. For example, a Phase 2 basket study with 120 participants across six baskets, enrolling over 12 months and followed for 6 months each, is live for 18 months with 720 participant-months. On Core at list rates that is 18 x $1,000 + 720 x $50 = $18,000 + $36,000 = $54,000, or $450 per participant. This is a ceiling arithmetic example only, because rates above 100 participants are lower but not published. ePRO and eConsent are included in Core; Full Platform at $5,000 per month and $30 per participant adds stratified and blinded randomisation, drug supply, eTMF and CTMS. Study setup by the Capture team is $0 and the study fee starts when the first participant is screened. See pricing, and use the free sandbox before committing.
For sample size planning, the pilot study sample size calculator and survival sample size calculator are free tools on this site; a basket stage size comes from your statistician's design.
A basket trial tests one targeted therapy in patients who share a molecular feature, across different tumour types or conditions. Each group, usually defined by histology, is a basket analysed on its own response.
A basket trial is one treatment across many populations, with a decision per basket. A platform trial runs under a master protocol while arms are added or dropped over time, usually with a shared control.
No dedicated basket module is claimed. You model baskets as arms or as a coded eligibility variable, and the per-basket tracking view is built from exports. Confirm the pattern on a demo.
No. The EDC records and exports the data. Stopping rules, borrowing models and committee decisions belong to your statistician and committee.
Use required fields for histology, result, assay, test date and a reference to the source report, with an edit check against the inclusion criterion. The report itself stays in the source documents.
Yes. Export as CSV or Excel with a data dictionary, filtered by date range or site, and group on the basket variable. SDTM datasets are available as SAS XPT with Define-XML.
At published Core list rates, 120 participants enrolled over 12 months and followed for 6 come to about $54,000, or $450 per participant. Rates above 100 participants are lower but not published.
Yes. The free sandbox has every feature, no credit card and no time limit. You pay only once you go live with real participants.
Keep exploring
EDC for oncology clinical trials
The therapeutic area view.
Adaptive platform trial software
A different master-protocol design.
EDC for rare disease trials
Small populations.
Eligibility screening eCRF
A starting form.
Enrolment forecast calculator
Project when baskets fill.
Pricing
Core, Full Platform and Enterprise.
Upload the protocol to the free sandbox, review the drafted eligibility and response forms, and test the edit checks. Every feature, no credit card, and you pay only once you go live.