A platform trial keeps one protocol and one control group running while treatment arms are added, dropped or graduated over years. The data capture has to survive that. This page separates what an EDC does for you from what your trial statistician and independent committee own, and shows how Capture fits.
Free sandbox · No credit card · 21 CFR Part 11 aligned
| Assessment | Q1 | Q2 | Q3 | Q4 | Q5 | Q6 |
|---|---|---|---|---|---|---|
| Shared control | ||||||
| Arm A (from start) | S | |||||
| Arm B (added Q2) | ||||||
| Arm C (added Q4) |
x = enrolling this quarter · S = stopped enrolling, follow-up continues
The short answer
Definition
A conventional multi-arm trial fixes its arms at the start and runs to a single final analysis. A platform trial runs under a master protocol with a standing infrastructure: arms enter and leave as evidence accrues, the control arm is shared, and decisions at interim looks, such as dropping an arm for futility or graduating one for efficacy, follow rules written in advance. The design is adaptive in that those decisions change what happens next, and the regulatory background is covered in the ICH E20 adaptive designs guide; the statistical side in the Bayesian clinical trial design guide.
Operationally, that creates problems an ordinary EDC build never faces. The protocol will be amended many times, perhaps once or twice a year. Sites activate and close against specific arms. A participant can be eligible for some arms and not others, so the eligibility form branches. Two participants enrolled a year apart may belong to different protocol versions and still be in the same analysis. And the statistician needs reproducible snapshots on demand without unblinding anyone.
None of this requires an exotic data system. It requires a configurable one with a trustworthy audit trail and exports that carry the right dates and arm labels. That is the role Capture plays. It is suitable for Phase 1 to 3, and the Phase 2 and 3 pages (EDC for Phase 2, EDC for Phase 3) describe the common use cases.
Division of labour
The clearest way to scope a platform trial is to say who owns each function. The right column is not a feature list of Capture; it names the system or person that typically carries the work.
| Function | In Capture (the EDC) | Owned elsewhere |
|---|---|---|
| Arms, epochs and arm-specific visits | Defined in the study; each arm sees only its own schedule | The protocol and master protocol appendices |
| Randomisation | Simple in Core; stratified and blinded randomisation in Full Platform, set at study setup | Allocation ratios and any response-adaptive rule, specified by the trial statistician |
| Eligibility by arm | Eligibility forms with skip logic and edit checks | The inclusion and exclusion criteria in the protocol |
| Blinding | Blinded roles never receive treatment-arm values; enforced by masked views | Unblinded statistician and pharmacy roles named in the plan |
| Interim data for a decision | Cleaned exports by date range and by site, plus SDTM datasets | The independent statistician's analysis and the data monitoring committee's recommendation |
| Stopping and graduation rules | None; the EDC does not apply them | The statistical analysis plan and committee charter |
| Amendments | Forms and visits edited in draft, tested, approved; every change audited | Protocol amendment approvals by sponsor, ethics and regulators |
| Safety | Adverse event forms, edit checks and query handling | Safety monitoring and committee review |
| Audit | Field-level, append-only audit trail, exportable | Inspection strategy and quality system |
If a vendor claims an EDC handles adaptive randomisation or a monitoring committee end to end, ask for the validation evidence and the owner of each decision.
Structure
Group visits into epochs such as screening, treatment, follow-up and early termination, and limit visits to the arms that need them. When Arm C is added, its schedule is a new set of arm-specific visits; the shared control and existing arms keep theirs.
Day 28 is an Arm B only visit, so Arm A participants never see it.
Upload the protocol to the free sandbox and let the AI draft visits and forms for review. Nothing is saved without approval. Every feature, no credit card.
Adding and dropping arms
Treat a new arm as a mini-project with its own checklist. First, the amendment is approved by the people who approve amendments. Second, the study team drafts the new forms and the arm-specific visits in the draft state and tests them in the sandbox, including every edit check. Third, the randomisation design is reviewed: the Capture facts are that arms, block size and stratification are set at study setup, so the randomisation owner must confirm what a new arm needs and agree it with the Capture team before the date sites are told to start. Do not assume an arm can be silently added to live allocation without that step. Fourth, forms are approved, because only approved forms appear in the casebook; fifth, sites are trained.
Dropping an arm is simpler for the database and harder for the people. Enrolment to that arm stops, follow-up visits continue for participants already in it, and the database keeps every record. Mark the date the arm closed to new participants, keep the arm-specific follow-up visits active, and make sure the closing decision is documented outside the EDC, in the committee minutes, so the audit trail and the decision record agree. The process for amending a live build is in protocol amendments without change orders and eCRF versioning for mid-study changes.
Amendment approved
Sponsor, ethics, regulators as required
Draft Arm C forms and visits
Draft state, AI-drafted then reviewed
Test in the sandbox
Edit checks, skip logic, roles
Confirm randomisation change
Statistician and Capture team
Approve forms and train sites
Only approved forms reach the casebook
Open arm at activated sites
Record the arm opening date
Interim looks
| Stage | Data task | Who | Where it lives |
|---|---|---|---|
| Cut-off agreed | Fix the data cut date and the arms included | Statistician with the committee | Statistical analysis plan, charter |
| Cleaning to the cut | Resolve open queries, review edit-check listings | Data manager | Capture query management |
| Snapshot taken | Export by date range, with data dictionary; SDTM if required | Data manager | Capture exports |
| Analysis | Run the pre-specified model and boundaries | Independent statistician | The statistician's validated system |
| Recommendation | Continue, drop, graduate or modify an arm | Data monitoring committee | Committee minutes (charter template) |
| Implementation | Amend schedule and randomisation if needed | Sponsor, study team | Protocol amendment, then the study build |
Capture supplies the cleaned snapshot and the audit trail. The analysis and the decision happen elsewhere, by design.
Fit check
Ask the vendor to show it, including what happens to randomisation, and how the change appears in the audit trail.
Vendor change orders on a design that amends every few months will dominate the budget. See the amendments page.
In the database through masked views, or only in the interface?
CSV with a data dictionary, SDTM with Define-XML, filtering by date range and site.
Treat any claim about adaptive randomisation or committee modules as needing validation evidence.
Platform trials run for years; model months live x fee plus participant-months x rate, as in the pricing page and the cost worksheets.
Where Capture fits
Capture suits the data capture and operations layer of a platform trial run by an academic network, a biotech or an investigator-led consortium. Your team builds and amends the study itself in the same tool you run it in, tests changes free in the sandbox, and every edit is recorded. The field-level audit trail is append-only and exportable, which supports the audit trail export inspectors ask for. Multi-site operations are covered in multi-site clinical trial management, and the analysis handoff in exporting data to R, SAS and SPSS and CDISC SDTM export software.
For allocation, Core includes simple randomisation and Full Platform adds stratified and blinded randomisation and drug supply; see RTSM software. Whatever response-adaptive rules your protocol specifies live with your statistician, who feeds allocation instructions back according to the plan, and the arrangement should be written in the data management plan before the first participant is randomised. Pricing is on the pricing page: Core is $1,000 per month including one live study plus $50 per participant per month, Full Platform is $5,000 per month per live study plus $30 per participant per month, and study setup by the Capture team is $0.
A trial run under a master protocol in which treatment arms can be added or dropped over time, usually with a shared control group and pre-specified decision rules at interim analyses.
Capture supports simple randomisation in Core and stratified and blinded randomisation in Full Platform, set at study setup. It does not contain a built-in response-adaptive randomisation algorithm; that logic belongs in your statistician's validated system.
Arms, epochs and arm-specific visits are part of the study design, and your team edits forms and visits in draft, tests and approves them. Treat an arm addition as an amendment and confirm the randomisation change with the Capture team before sites start. Test it in the sandbox first.
No. Capture supplies cleaned exports and an audit trail for the statistician and committee. There is a DSMB charter template on this site, but the committee's analysis and recommendation happen outside the EDC.
Capture enrolment, randomisation and protocol-version dates as required fields, standardise endpoint forms across arms and keep arm-specific forms arm-specific. The statistician then defines concurrent controls at analysis.
Blinded roles never receive treatment-arm values; this is enforced by masked database views, not only the interface. Name unblinded roles in your plan.
Core is $1,000 per month including one live study plus $50 per participant per month; Full Platform is $5,000 per month per live study plus $30. Study setup by the Capture team is $0. Rates above 100 participants are lower but not published.
Yes. The free sandbox has every feature, with no credit card and no time limit. You pay only once you go live.
Keep exploring
ICH E20 adaptive designs guide
The regulatory background.
Bayesian clinical trial design guide
The statistical side.
Protocol amendments without change orders
Editing a live build.
Blinded and unblinded role management
Protecting the blind.
Basket trial software
A different master-protocol design.
Pricing
Core, Full Platform and Enterprise.
Upload the master protocol to the free sandbox, draft the visits and test an arm addition. Every feature, no credit card, and you pay only once you go live.