Skip to content
2026-06-25

Cohort Discovery Service

Get Started

Overview

The Cohort Discovery Service enables researchers to query pseudonymised, OMOP-mapped health data held by Data Custodians across the UK — without the data ever leaving a Custodian's environment. It is integrated into the Health Data Research Gateway and operates on a federated architecture designed to balance research access with strong data governance.

  • Getting Started

    Understand the prerequisites and get an overview of the three onboarding workstreams.

    Getting Started

  • Architecture

    Learn how the federated platform works, how data flows, and how security is enforced.

    Architecture

  • OMOP Requirements

    Understand which OMOP CDM tables and fields must be populated for your data to be discoverable.

    OMOP Requirements

  • Connecting Bunny

    Step-by-step guide to connecting the Bunny query tool to the Cohort Discovery Service.

    Bunny Setup

  • Workstreams

    Detailed guidance for the Governance, Data, and Infrastructure onboarding workstreams.

    Workstreams

  • Reference & Support

    Glossary of terms, query tool comparison, and how to get support.

    Reference

  • Developer Guide

    Set up the full stack locally, generate synthetic test data, and contribute to the platform.

    Developer Guide

What is the Cohort Discovery Service?

The Cohort Discovery Service allows researchers to run feasibility queries against pseudonymised patient data held by Data Custodians — NHS Trusts, universities, and other healthcare organisations — without the underlying data leaving the Custodian's infrastructure. Only aggregated, disclosure-controlled counts are returned to the researcher.

The service was initially developed through the CO-CONNECT project and is now an integrated part of the Health Data Research Gateway.

HDR UK

This documentation is maintained by the Health Data Research UK Technology Team. For general enquiries, email gateway@hdruk.ac.uk or use the Need Support? button on any page of the Gateway.

How it works

graph TD
    R[Researcher] -->|Runs query via Gateway| CDS[Cohort Discovery Service]
    CDS -->|Sends task| B[Bunny / Query Tool]
    B -->|Queries locally| OMOP[(OMOP Database)]
    B -->|Returns aggregated count| CDS
    CDS -->|Displays result| R

    subgraph DC[Data Custodian Infrastructure]
        B
        OMOP
    end

Data never leaves the Custodian. The query tool (e.g. Bunny) makes only outbound requests and returns only aggregate counts — no row-level data is ever transmitted.