Transformation Directorate

Discovery data service

Owner

The Endeavour Health Charitable Trust funds the Discovery Data Service: the code for which is developed and updated by Voror Health Technologies working in partnership with health and care organisations across North East, South East and North West London. Voror is engaged through a support and maintenance contract, and then resourced for individual development projects.

The Discovery Data Service (DDS) is supported by 80 open source repositories. It is licensed under Apache Licence 2.0, which allows for free use of the software for modification, distribution, patent use, private use and commercial use.

Discovery Data Service is being done open source so that as many people as possible can benefit from it. Originally the ambition was to gift it to the UK. I want to gift it to the wider world

Sean Tickell, Founder, and CEO of Temple Black

Background 

A huge proportion of the world’s data is generated in health and care systems, and it is only increasing in volume. In London, three regional Integrated Care Systems (ICS’) alone support the records of 7.5 million patients. Those records are then called billions of times a day to deliver health and social care services or improvement. Unfortunately, a lot of historical health data is inconsistent or difficult to compare: in different formats, from different systems, and of varying quality. Multiple data standards have evolved for health and social care but none of them cover all of the relevant data. Efforts are underway to promote reliable standards and create new ones where required, but this is a massive challenge requiring sector wide collaboration. Structured, comprehensive, and precise clinical vocabularies do exist, and are fit for use in electronic health records: a large part of the problem is ensuring that pre-existing data conforms to them.

Situation

In order to build a single picture of an individual for both direct care and real time health research, large amounts of data need to be securely brought together from multiple sources. OneLondon have conducted the world’s largest public deliberation to help inform how this data is collected and used in the capital, and which uses patients are happy with (or not), but the ability to actually transform and utilise disparate information is still being developed.

Aspiration

  • Securely and safely provide data for real-time care to provide better and faster outcomes with fewer mistakes, less waste, and better use of vital clinical resources.
  • Help staff across the NHS contribute to data driven service improvement.
  • Allow health data to be deployed for proactive, preventative and joined-up care instead of reactive use.
  • Not have to recreate or reinvent anything that’s already been done for data interoperability.
  • Support the data needs of Integrated Care System formation and the transformation vision of the NHS as a whole.

Solution and impact

The Discovery Data Service is a healthcare platform built specifically to make health and care data easier to use. Rather than wait for broad agreement of standards at the point of data collection or processing, the platform brings together real-time information from Acute, Primary Care, Mental Health, and Community Health sources (among others) to support the London Care Record (LCR). Patient consent for this work is either explicit or aligned to the national opt-out.

With local, regional, or supra-regional data-sharing agreements in place (inclusive of consent handling), the DDS normalises and maps input to the Snomed-CT standard where possible, or a Discovery Code Library otherwise. This single, live, data service can then be used through the LCR for the direct care of patients, for population health, and for cross care-setting innovation.

As an example, London Ambulance Service are using the DDS to see Frailty Flags in certain parts of London. When someone rings 111, their call can be handled appropriately, and the patient can be treated effectively on the correct care pathway. This flag was used over 340,000 times in February alone.

The architecture that hosts, develops and supports the service is open source, empowering healthcare professionals to provide input on software improvements for more effective care. The information and relational models (the “ontology”) of how data handled by the DDS fits together is also open source, helping organisations within and without the NHS understand how to structure their data, and learn from the way the DDS functions.

Functionality

The Discovery Data Service is cloud-based and designed for handling data across the UK’s whole population. It’s been deployed for over 60 organisational use cases: including Shared Care Records, Frailty Flagging, Diabetes support, and proactive identification of heart problems. It:

  • Is capable of rapid and accurate data coding and assurance
  • Understands a variety of pre-existing standards, including HL7 FHIR.
  • Data providers and controlles have a right of return and deletion of their data
  • Provides open source and modular components to allow:
    • users to create accounts,
    • patients to opt-out,
    • GPs and other service providers to run rapid queries and access helpful dashboards,
    • service managers to run audits,
    • and analysts to access tool kits.

Capabilities

  • Securely collects, cleanses, protects and presents data, making it available for direct care, analysis and decision making.
  • Can be accessed on any clinician desktop across health and care. It can be used in GP surgeries, ambulances, hospitals, emergency departments and social care settings.
  • Allows for patient interaction via Patient Knows Best.
  • Allows for modules and functions to be progressively added or extended

Scope

The DDS has been in development for eight years, and has now found multiple applications across London, but in terms of potential, Sean says that the project is “still in early days… most of the benefits have yet to be delivered”.

  • Data can be absorbed at great speed, and much faster than through the use of pre-existing approaches.
  • Initial uses could be deployed across London or scaled nationwide
  • In abstract and anonymised formats it can help authorities to plan and manage their resources and finances.
  • There are very likely to be many more uses, including the ability to investigate social determinants of health.

Key learning points

  • Electronic patient records require coherent and properly structured data architecture to fully deliver their benefits.
  • A variety of processes are needed before any information is taken in or shared, and setting these up can be slow, but is vital. Agreeing on data sources and models, correct governance and use, and protocols for safety and security cannot be ignored.
  • Ongoing support of this work is a key area, and mission critical. The current team can design, develop, support, and release specific features on demand.
  • Change can be intimidating or invite scepticism but open source and its community focus helps users to see the individual and wider benefit of shared knowledge, new ideas, better solutions, and more effective ways to provide care.

Sean says:

I think there's a lot of education to take place. We welcome contributions, there's no doubt - the whole reason we're doing this is that we want to advance things as quickly as possible, by working with other experts. And I think that's where things will really accelerate when we build a bigger community with better knowledge.

Digital equalities

  • More coherent data means that more effective care can be provided when patients experience multiple or life-long conditions, allowing impairements to be accounted for in treatment pathways or emergency care.
  • Better data quality means that data driven population health assessment can more accurately identify particular at-risk groups, especially for community wide issues, such as vaccine deployment.

Give us feedback

Open Source Digital Playbook feedback survey

Page last updated: September 2022