Show information for
UK-hosted Coming soon
Castor Catalyst · For US sitesFor EU sites · UK-hosted studies coming soon

Where patient data goes, and who can open it

Version 2.1 · Updated 16 October 2026

How study data moves from the site's electronic health record (EHR) to the electronic data capture (EDC) system, and who is responsible at each step.

How study data moves from the hospital record to the electronic data capture (EDC) system, and who is responsible at each step.

Before redactionFictional patientLab report before redaction, showing the patient's name, IDs, address, phone, physician and dates, and the first metabolic panel results.
After redactionThe same lab report after redaction: the identifiers are blacked out, dates are reduced to month and year, and age, sex and all results are kept.
Castor Catalyst redaction engine on a lab report: identifiers removed, clinical values kept. In this example, dates are reduced to month and year.
  • No large language models in redaction

    The redaction engine finds identifiers with rules and small purpose-built recognisers trained only on synthetic documents; no large language models and no external services.

  • The site approves the redactions

    No document a site uploads goes to the Catalyst pipeline until the site approves the redactions, also when Castor's engine redacted.

  • No identifiers in the EDC

    The EDC holds study data only.

Where the data is

From the hospital record to the EDC, step by step

Two Castor services, kept apart. The Catalyst Source Gateway holds and redacts source for the site, with no large language models. Hosted on Microsoft Azure; no other subprocessors. The Catalyst pipeline works for the sponsor, on redacted copies only.

Data flow for US sites. Patient and site: the patient consents to the study and the original record stays at the site. The site redacts a document before upload with a site redaction tool (path A) or in the browser (path B, coming soon), or uploads it unredacted and the redaction engine redacts it in the Catalyst Source Gateway (path C). The Catalyst Source Gateway is a separate service with no large language models, hosted on Microsoft Azure with no other subprocessors, study-isolated, encrypted, in the site's country. The site approves the redactions before any upload goes to the pipeline. US only: the patient can authorise retrieval of their own health records; these participant-consented records go through the same redaction engine and skip site approval. In the Catalyst pipeline, the sponsor's service, AI extracts values from redacted copies only, then review by Castor under the review policy set by the sponsor or CRO. AI processing runs in the US. The EDC holds no personal identifiers.
Data flow for EU sites. Patient and site: the patient consents to the study and the original record stays at the site. The site redacts a document before upload with a site redaction tool (path A) or in the browser (path B, coming soon), or uploads it unredacted and the redaction engine redacts it in the Catalyst Source Gateway (path C). The Catalyst Source Gateway is a separate service with no large language models, hosted on Microsoft Azure with no other subprocessors, study-isolated, encrypted, in the site's country. The site approves the redactions before any upload goes to the pipeline. In the Catalyst pipeline, the sponsor's service, AI extracts values from redacted copies only, then review by Castor under the review policy set by the sponsor or CRO. AI processing runs in the EU; UK-hosted studies are coming soon, ask your Castor contact. The EDC holds no personal identifiers.

No document a site uploads goes to the Catalyst pipeline until the site approves the redactions, also when Castor's engine redacted. Each extracted value links to its redacted source. Today every extracted value goes to a Castor reviewer; risk-based review is coming soon.

Three ways a site sends a document

  1. Site redaction tool

    The site redacts the document before upload.

  2. In the browser Coming soon

    The site redacts on its own computer before anything is uploaded.

  3. Castor redacts

    The site uploads unredacted; Castor's redaction engine removes its identifiers in the Source Gateway. For French sites, Castor redacts (C) only once health data hosting (HDS) is confirmed for the study.

Can participants reach the hospital's records?

Records a participant authorises are their own choice and need nothing from the site.

Two public web pages list connected organisations: Epic's endpoint directory, and the map of the Trusted Exchange Framework and Common Agreement (TEFCA).

A redacted copy is still personal data

Redaction removes identifiers, but under the General Data Protection Regulation (GDPR) the copy stays personal data in Castor's hands.

Access

Who can open unredacted source

Only the site can open unredacted source. Castor reviewers, and every AI step, work on redacted copies, which are still personal data.

WhoUnredacted source?What they work with
Site teamYesThe site's own records and uploads. Approves every redaction.
Castor Catalyst redaction engineAutomated, no large language modelsReads path C uploads only to redact them.
Castor's AI pipelineNoRedacted copies only.Redacted copies only, in the EU.
Castor reviewersNoRedacted copies only. US-based for US studies.Based in the European Economic Area.
Sponsor and its contract research organisationNoStudy data in the EDC, with no personal identifiers.
Agreements

Agreements and approvals for the study

For the study, Castor works under the contract with the sponsor or CRO. If Castor redacts uploads for the site (path C), a separate agreement with the site may be needed, such as a business associate agreement (BAA). Ask your Castor contact.

For the study itself, the sponsor is controller and Castor its processor, under an Article 28 data processing agreement (DPA) with the sponsor. If Castor redacts uploads for the site (path C), a separate agreement with the site may be needed, such as a data processing agreement (DPA). Ask your Castor contact.

The clinical trial agreement (CTA) allocates responsibilities; it is not a GDPR ground. Site-controlled activities, co-sponsorship and investigator-initiated studies are assessed and documented separately.

Before the first upload

  1. Institutional review board (IRB) approval

    The study protocol is approved.

  2. Informed consent

    Uploads are for consented participants only.

  3. Authorisation or waiver

    Under the Health Insurance Portability and Accountability Act (HIPAA): an authorisation, a documented IRB or Privacy Board waiver, or another permitted pathway.1

  4. Within scope

    Upload only what that permission covers.

  5. Ethics approval

    The study protocol is approved.

  6. Lawful basis

    The controller confirms its Article 6 basis, Article 9(2) condition and any national rules.1

  7. Participant information

    Covers the processing, recipients, retention and transfers. Running trials may need an update first.2

  8. Consent

    Uploads are for consented participants only.

National rules: Germany, the Netherlands, France
  • Medical secrecy in Germany and the Netherlands

    Redaction supplements medical secrecy; it does not replace it. The site establishes the permission for disclosing unredacted records to Castor and for monitor access, under the German Criminal Code (Strafgesetzbuch, StGB) § 203 or the Dutch Medical Treatment Contracts Act (WGBO).a

  • France: health data hosting

    Health data hosting certification (hébergeur de données de santé, HDS) applies to hosted health records of French patients, including redacted copies. For French sites, Castor redacts (C) only once health data hosting (HDS) is confirmed for the study.b

Regions

Where data is stored and processed

Catalyst Source Gateway No large language models · No other subprocessors Study-isolated, encrypted, in the site's country.
  • US

    Source Gateway in the site's country; AI processing in the US. Castor reviewers and support staff for US studies are based in the US.

  • EU

    Source Gateway in the site's country; AI processing in the EU. Reviewers and support staff for EU studies are based in the European Economic Area.

  • UK

    Coming soon

    UK-hosted studies: coming soon. Ask your Castor contact.

Access by a separate entity outside the European Economic Area (EEA) is a transfer and needs cover under the EU-US Data Privacy Framework or standard contractual clauses.3 The providers used in a study, such as Google for AI extraction, are in Castor's sub-processor list.

Security

Encrypted at rest and in transit, and isolated per study. Company certifications:

  • ISO/IEC 27001

    Information security management, certified since 2017

  • ISO 9001

    Quality management, certified since 2018

  • NEN 7510

    Dutch healthcare information security standard, since 2017

On request

Documents sites can ask for

A site's privacy officer, IT security lead or IRB can ask the site's Castor contact for any of these before the first upload.

A site's data protection officer, IT security lead or ethics committee can ask the site's Castor contact for any of these before the first upload.

  • Sub-processor list

    The hosting and AI providers used in the study, with their locations.

  • Validation package

    Risk assessment, requirements and validation plan from Castor quality assurance.

  • EU AI Act assessment

    Castor's documented classification assessment, under a non-disclosure agreement.