Guide · Vendor transitionsUpdated September 28, 2026

How to switch EDC vendors mid-study, and when not to

Sometimes a vendor change cannot wait for the next study: a contract ends, a vendor is acquired, costs balloon or the system cannot handle an amendment. Here is how to do it without losing data integrity, and the alternatives worth considering first.

  • Honest risk assessment
  • Step-by-step plan
  • Keep the audit trail intact

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

EDC transition plan · 14 weeks
1234567891011121314

Weeks

Assess and decide
Build and validate new study
Legacy export and archive
Site training
Cutover

New data entered only in the new system from this date

Reconcile and close legacy
Cutover at a defined date, never gradually

Key points

  • Switching EDC mid-study is possible but carries real risk to data integrity, timelines and site goodwill. Treat it as a controlled change with a documented plan, not an IT project.
  • First ask whether you can avoid it: finish the current study on the old system and start the next study, extension or new cohort on the new one.
  • If you do switch, cut over on one defined date, preserve the complete legacy dataset with its audit trail, and document every decision for inspectors.
  • Most teams do not import historical data into the new system; they archive the legacy database and merge datasets at analysis, with a clear mapping.
  • Plan for site retraining, updated CRF completion guidelines, and notifying ethics committees or regulators if your procedures require it.

Deciding

Why teams switch, and why it is rarely the first option

The reasons are usually commercial or practical rather than technical: a renewal quote far above the original contract, a vendor acquired and its product sunset, a CRO relationship ending and taking the database with it, or a system that cannot implement an amendment in a reasonable time or for a reasonable change-order fee. Each of these can make staying more expensive than moving.

If you are still choosing a platform, our Medidata Rave vs Veeva Vault EDC comparison lists the questions to ask enterprise vendors, and REDCap vs Castor covers the common academic choice. But a mid-study switch has costs that do not show up in the vendor comparison. Sites must learn a new system in the middle of recruitment. Monitoring plans, edit checks and data management procedures must be rebuilt and revalidated. And the study ends up with data in two systems, which has to be explained in the clinical study report and to any inspector.

Alternatives to consider first

Can the current study finish where it is? If most participants have completed and the remaining work is follow-up and cleaning, staying often costs less. Is there a natural break point, such as the end of Part A of a two-part study, a new cohort, an open-label extension or the next protocol? Starting the new phase in a new system gives a clean boundary. Our REDCap migration guide makes the same recommendation: use the new system for a new study or arm where possible.

Switch now or at the next break point?
Switch mid-studyNext phase or study
Data in one system
Site retraining mid-recruitment
Stops a runaway renewal cost now
Needed if vendor is shutting down
Inspection explanation required

Step by step

A mid-study EDC switch in four stages

  1. 1

    Assess and document

    Record the reason for the change, the risk assessment and the cutover plan in a change control document. Agree it with the sponsor, CRO and data management lead.

  2. 2

    Build and validate

    Rebuild the current protocol version in the new system, including edit checks and the schedule, and run user acceptance testing against the old system's specifications.

  3. 3

    Cut over

    Choose a date after which all new data goes into the new system. Freeze and lock the legacy database for anything entered before it.

  4. 4

    Archive and reconcile

    Export the complete legacy dataset, audit trail and metadata, archive them, and document how the two datasets are combined for analysis.

The legacy database

What you must take out of the old system

Regulators expect you to be able to reconstruct the study from its records, which means the legacy system's data is not just the clinical values. Before access ends, secure the full dataset in an open format, the audit trail showing who changed what and why, electronic signature records, query history, the form and edit check definitions that applied at each point in time, user access records, and the export and validation documentation.

Many contracts restrict or charge for bulk exports after termination, so negotiate the exit package before you give notice. Keep the archive for the full retention period your regulations require, and make sure someone can still open it in ten years.

Import or merge at analysis?

Importing historical data into the new system looks tidy but raises hard questions: the new audit trail starts at import, not at original entry, and any mapping error becomes part of the clinical record. Most teams keep the legacy dataset as its own locked, archived source and merge the two at analysis using a documented mapping. Whichever you choose, write the rationale down.

Legacy exit package

Items

9

Secured

6

At risk

1

  • Clinical data, all forms, all subjectsCSV + XPT
  • Field-level audit trail
  • E-signature records
  • Query history with responses
  • Form and edit check versions
  • User access and role history
  • Data dictionary and annotated CRF
  • Validation documentation
  • Export after contract endFee in contract

Rebuilding fast

Rebuild the study in days, not months

The longest part of most transitions is rebuilding the study. In Capture, the AI Study Generator drafts visits, windows and forms from the protocol, and the AI form builder drafts questions from exported data dictionaries or CRF files, for your team to review. Everything can be tested with bulk test subjects before a single site switches.

  • Protocol to draft study with the AI Study Generator.
  • Forms drafted from an exported data dictionary (CSV or XLSX) or CRF (PDF or DOCX).
  • Bulk test subjects for user acceptance testing.
  • Enterprise plans include implementation assistance and support for your validation.
Protocol to study setup software
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

Test the rebuild before you give notice

Upload your protocol to a free sandbox and see how much of the study is drafted for you.

Plan your next study free

Before cutover

EDC transition checklist

Change control approved

Reason, risk assessment, cutover date and responsibilities documented and signed off.

New build validated

User acceptance testing against the current protocol version and edit check specification.

Legacy exit package secured

Data, audit trail, signatures, queries, metadata and documentation exported and archived.

Sites trained

Training completed and acknowledged before cutover; updated completion guidelines issued.

Monitoring plan updated

SDV scope and review procedures reflect the new system.

Regulatory and ethics notifications checked

Confirm whether your procedures or authorities require notification of the change.

FAQ

Questions teams ask before they switch

Something not covered here? Ask us directly.

Can you change EDC systems during a clinical trial?

Yes, with a documented plan, a validated new build, a defined cutover date and a complete archive of the legacy data and audit trail. It carries risk, so consider switching at a natural break point first.

Do I need to migrate historical data into the new EDC?

Not necessarily. Many teams archive the legacy dataset and merge it with new data at analysis using a documented mapping, which keeps the original audit trail intact.

How long does an EDC switch take?

It depends on study complexity and the number of sites. Rebuilding and validating the study usually takes the longest; AI-assisted builds can shorten that considerably.

What should I get from the old vendor?

The complete clinical dataset, audit trail, e-signature records, query history, form and edit check versions, user access history, data dictionary and validation documentation.

Do participants need to re-consent?

Not usually just because the data system changes, but check whether your consent form names the system or data processor, and follow your ethics committee's requirements.

Does Capture help with transitions?

Capture drafts studies from protocols and forms from exported files, and Enterprise plans include implementation assistance. For a mid-study switch, contact us to discuss your situation.

Make the next switch the last one

No setup fees, no change orders for amendments, and your data exportable any time.

Plan your next study free