Guide · EDC implementationUpdated September 29, 2026

EDC implementation timeline: what takes the time

Getting an EDC study live involves choosing a vendor, agreeing a contract, specifying and building the database, testing it and training sites. Here is how long each phase tends to take, what drives it, and where you can save weeks.

  • Phase by phase
  • What drives each step
  • Where to save time

Free sandbox · No credit card · 21 CFR Part 11 aligned

Typical EDC implementation
1234567891011121314151617181920

Weeks

Vendor selection
Contract
Specifications
Build
Testing (UAT)
Training and go-live
Illustrative; varies widely by vendor and study

The short version

  • The main phases are vendor selection, contracting, specifications, build, user acceptance testing, and training and go-live.
  • In vendor-built models, specifications and build are the longest phases, with rounds of review between sponsor and vendor.
  • Contracting often takes longer than expected and can run in parallel with specification work if planned.
  • Testing needs time for fixes and retesting, not just one pass.
  • Self-serve building removes the specification hand-off: the team builds, tests and approves directly, and can start before the contract is signed in a free sandbox.

Phases

Where the weeks go

In a traditional model, the sponsor writes CRF and edit check specifications, the vendor builds from them, the sponsor reviews, the vendor revises, and the cycle repeats. Each round trip takes days to weeks, and every protocol change during build restarts part of it. After the build comes user acceptance testing, where the sponsor checks forms, logic and edit checks against the specifications and raises findings for the vendor to fix.

The critical path usually runs through the protocol: a database cannot be finished before the schedule of assessments and endpoints are final. Starting the build from a near-final protocol, rather than waiting for sign-off, saves time as long as late changes are managed.

Self-serve build

In Capture, the AI Study Generator reads the protocol and drafts visits, windows and forms for your review, and your team refines, tests and approves them directly in a free sandbox. There is no specification hand-off, and testing can start as soon as the first forms are drafted. See the fastest way to launch a clinical trial.

AI Study Generator
phase2_protocol_v4.pdfStep 2 of 3
  • Reading protocol
  • Building visits and windows
  • Drafting forms

11

visits

9

draft forms

2

diaries

Phase by phase

What drives each phase

PhaseMain driversHow to shorten it
Vendor selectionNumber of vendors, demos, security reviewClear requirements; hands-on sandbox testing
ContractingLegal review, pricing negotiation, data processing termsStart early; run in parallel with build
SpecificationsProtocol maturity, review roundsBuild directly instead of writing specifications
BuildNumber of forms, edit checks, integrationsAI drafting from the protocol; template library
TestingTest scripts, findings, retestingTest as you build; risk-based focus on critical data
Training and go-liveSite availability, account set-upShort role-based training; site activation checklist

In Capture

Build and test before anything is signed

Upload the protocol and review a draft study. Refine forms and edit checks, test them with sample data in the free sandbox, and export a blank eCRF PDF for ethics submission. When you go live, sites pass an activation checklist and are ready to enrol.

  • AI-drafted visits and forms.
  • Free sandbox with every feature.
  • Blank eCRF PDF for ethics.
Clinical trial study start-up software
app.capture.study · New study
Capture AI Study Generator: describe the study or upload a protocol to generate it

Upload a protocol and see the draft

Start your build in the free sandbox today.

Start building free

Testing

Testing without the bottleneck

User acceptance testing is often compressed at the end because the build ran late. A risk-based approach helps: test the forms and logic that support primary endpoints, eligibility, safety and randomisation thoroughly, and sample the rest. Record what was tested and the results, because that evidence supports your validation.

Testing is faster when the people who designed the study can make fixes themselves. In a self-serve system, a finding is fixed and retested in the same session rather than logged for a vendor. See computer system validation.

Planning

EDC implementation checklist

Protocol maturity

Schedule of assessments and endpoints near final.

Contract in parallel

Legal and data processing terms started early.

Build owner

Who builds and approves forms.

Risk-based testing

Critical data and logic tested thoroughly.

Ethics documents

Blank eCRF and ePRO PDFs ready for submission.

Site readiness

Training and activation planned before first enrolment.

FAQ

Questions teams ask before they switch

Something not covered here? Ask us directly.

Who should be involved in the build?

The data manager, a statistician, a clinical lead and someone who knows the sites. Involving them early prevents late changes.

When should sites be trained?

Close to go-live, with short role-based training after the build is approved.

How long does EDC implementation take?

It varies widely. Vendor-built studies commonly take several months from selection to go-live; self-serve builds can be much faster, especially when the build starts from the protocol.

What is the longest phase?

In vendor-built models, usually specifications and build, because of review rounds between sponsor and vendor.

Can we build before the contract is signed?

In Capture, yes. The free sandbox has every feature, and you pay only when you go live.

Does the AI build the study automatically?

It drafts visits, windows and forms from the protocol for your review. Nothing is saved without human review.

How can testing be faster?

Test as you build, focus on critical data and logic, and let the people who designed the study fix findings directly.

Can I try it free?

Yes, in the free sandbox.

Take the database off the critical path

Draft your study from the protocol in the free sandbox.

Start building free