UAT proves the build works. Go-live proves the whole study is ready to use it: people with access, sites trained, consent in place, safety routes agreed and a plan for the first two weeks. Use this checklist as a written gate with an owner and evidence for each line.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Study
DEMO-LIVE-07
Target
T-3 days
Decision
Pending
What go-live means here
Why a gate
A clean UAT report tells you the database does what the specification says. It does not tell you that the fourth site has an account, that the investigator knows where to sign, that the ethics-approved consent text matches what is in the system or that someone will read the first queries before they age. Those are the things that go wrong in the first week of a study, and they are cheap to check the week before.
The gate has two parts. A checklist confirms each area is ready and records the evidence. A short meeting then makes a go, conditional go or no-go decision. For a small team the meeting can be fifteen minutes and the evidence a shared folder. The discipline is in writing the criteria down in advance, so the decision is not made under launch-day pressure. This page is deliberately distinct from the EDC implementation timeline, which shows how long phases take, and from common EDC implementation mistakes, which lists failure modes.
Area 1
Everything that proves the database is fit for use.
Scripts run, defects logged and closed, deviations accepted, sign-off by study lead and data manager.
No form left in draft that sites will need. In Capture only approved forms appear in the casebook.
Compare the system schedule with the current protocol version, line by line.
Boundary values, missing data and date-order errors confirmed, with priorities agreed.
Spot-check derived values (BMI, eGFR, QTc and others) against a manual calculation.
The statistician has opened a test export and agreed the structure.
Version-labelled copy for the ethics file, the investigator site file and the archive.
Area 2
The first-week failures you most often see.
Each role tested with a real account: what it sees, edits, signs and exports.
No shared logins. Delegation log matches who has data entry or signing rights.
Staff can turn on authenticator-app two-factor authentication; make it part of onboarding.
A blinded test user confirmed not to see treatment-arm values in any view or export.
Site coordinators see participant names; researchers see coded IDs. See role-based access control.
Who removes access when someone leaves, and how quickly.
Area 3
One test participant entered end to end, including a query response.
One page each with definitions, units and what to do when a value is unknown.
Site-level numbering (for example 001-0042) agreed, and enrolment links or QR codes issued per site where used.
Sites know how to reach support. Capture offers email support, typically under 24 hours, plus documentation and video tutorials.
Approvals, contracts, delegation log and training records are filed before the site is allowed to enrol.
Build the study, create test sites and users, enter a participant from consent to export and answer a query on purpose. If something is missing from your checklist, you will see it in an afternoon, with no credit card.
Areas 4 and 5
Approvals match the protocol version and the consent version going live.
If you use eConsent, the on-screen text matches the approved document, and the signature flow was tested including countersignature.
Participant signs with the email one-time code; staff countersign; the signed PDF exports with timestamps.
An example adverse event moves from entry to the person who must assess it. See adverse event reporting software.
Everyone knows how to start the expedited clock and where the sponsor's receipt is recorded. See the SAE reporting deadline calculator.
If the study is blinded, the procedure and who may use it are written and tested.
Areas 6 to 8
Both match the final build. See templates for a data management plan and data validation plan.
Who answers, who closes, and the time targets. See query management.
Source data verification fields chosen, monitoring visit plan and remote review schedule agreed.
Decide in advance how sandbox data is kept apart from live data, and confirm the handling with support. Do not assume test data carries over.
EU (Frankfurt) or USA (N. Virginia) matches the privacy notice and contracts.
Close-out exports and retention owner decided. See data archiving and retention.
Someone reviews entries, queries and sign-in problems every working day for the first two weeks.
The decision
Adapt the thresholds to your study and quality system. The point is to write them down before you need them.
| Criterion | Evidence to look at | Blocks go-live? |
|---|---|---|
| Any open critical or major UAT defect | Defect log with retest results | Yes |
| Consent version differs from the approved one | Approval letter against system text | Yes |
| No tested route for serious adverse events | End-to-end test record | Yes |
| A site has no trained, delegated user | Delegation log and training record | Yes for that site only |
| Minor wording or label issue on a non-critical form | Defect log | No; fix with a controlled change |
| Statistician has not reviewed the export | Review note | Conditional: review within the first two weeks |
Record the decision, date, attendees and any conditions with the checklist as part of the study file.
The sequence
Forms approved, schedule locked, role matrix final. Further changes need change control.
Each site completes a sandbox journey and gives feedback while it is cheap to act.
Create users, enable two-factor authentication, check delegation logs, issue enrolment links or QR codes.
Collect the checklist evidence and resolve open items. Prepare the blank eCRF PDF for the study files.
Decide, record, and tell sites. No build changes after the decision except through change control.
Support is on hand, the hypercare owner watches the first entries and the first signed consent is checked.
Review queries, entry patterns, sign-in issues and open items. Close hypercare or extend it, and record the outcome.
After launch
The first fortnight is where assumptions meet reality. Watch the volume and type of auto-queries: a flood from one rule suggests a range is wrong, not that sites are careless. Watch entry timeliness against the visit window. Check that every signed consent has a countersignature and a matching version. Review user sign-in problems and fix them the same day, since a blocked site coordinator quickly resorts to paper.
Keep a simple log of every issue, the cause and the fix, and decide whether each needs a controlled change to the build. Capture's forms move through draft and approved states, and the field-level audit trail records changes with timestamp, user, old value, new value and reason, so a change made in week one is explained in the record. For a deeper look at preparing for inspection, see the FDA inspection checklist for clinical trial data.
Capture provides 21 CFR Part 11-aligned controls. Validated use, your SOPs and the go/no-go decision belong to the sponsor. This checklist is a general aid, not a regulatory document.
UAT tests that the database meets its specification. The go-live gate also checks people, sites, consent, safety routes, documents and support, and records a decision. See the EDC UAT guide for testing.
Usually the study lead or sponsor representative together with the data manager, with input from the medical monitor and QA where you have them. Name the approvers in advance.
Start two weeks ahead so there is time for site rehearsals, then hold the decision meeting about a day before. Adjust to the size of your study.
Yes, through change control: request, assess, test in the sandbox, approve and publish a new approved version. In Capture forms move from draft to approved with an audit trail.
Yes. The Capture sandbox has every feature, no credit card and no time limit. You pay when you go live with real participants.
It is good practice. A named person reviewing entries, queries and access problems daily for two weeks catches issues while they are small.
No. It is a general aid. Follow your quality system and regulatory requirements.
Keep exploring
EDC UAT and sandbox environment
Testing before the gate.
Common EDC implementation mistakes
What to avoid.
EDC implementation timeline
Phase durations.
Fastest way to launch a clinical trial
Speed-focused guide.
Query management software
Run the query process.
Data archiving and retention
Plan the close-out.
Screening and enrollment log template
Track the first participants.
Free sandbox with every feature. No credit card, and you pay only when you go live with real participants.