My current work is focused on building PatientCare v4.

This is a distributed, scalable, cross-organisational, patient-centric electronic health and care record system that makes use of core foundational modular building blocks and a focus on meaningful structured standards-based data.

The front-end software is lightweight, standards-based and focused on delivering the best user experience.

Back-end software focuses on creating a seamless single patient record by making resolution and mapping of identifiers and properties first-class, with different data models for reading and writing data. As an example, that means a client can ask for a FHIR Practitioner ‘view’ of data sourced from a X.500 / LDAP record.

  • pc4: Front-end and server-side components creating an event-driven electronic health and care record. Leverages the following components:
  • hermes: SNOMED CT terminology server and embeddable library with support for inference, fast free-text search, crossmaps to and from other code systems, compositional grammar and the expression constraint language.
  • hades: A HL7 FHIR SNOMED terminology server; a thin wrapper around hermes.
  • clods: Geographical data support including NHS organisational data services and UK postcode data. Library, REST server and HL7 FHIR server.
  • concierge: Integration engine for common NHS backend services providing a set of APIs for access to live NHS systems with identifier resolution and mapping as a first-class abstraction.
  • nhspd: Library and microservice for the NHS postcode directory linking all UK postal codes to geographical health, care and administrative data.
  • dmd: A library and microservice for the UK dictionary of medicines and devices.
  • trud: Software library to automate downloads of UK reference data.

Archived projects

  • PatientCare EPR v3: An electronic health record used across Wales with integrations across multiple NHS systems, used for specialist services such as multiple sclerosis, Parkinson’s disease, motor neurone disease, neuro-rehabilitation and others. It’s used across Wales within the NHS network at It combines clinical and research cohorts using finely-grained access controls and had used my own terminology server to provide SNOMED CT functionality since 2009. Over 500 users, over 30,000 patients and over 400,000 clinical encounters recorded with structured data and SNOMED CT. It underpins much of my published research papers. It supports many services, including:
    • Multiple sclerosis; it is the largest multiple sclerosis registry in the United Kingdom
    • Motor neurone disease; all patients in South Wales with motor neurone disease are registered together with longitudinal outcomes
    • Neuromuscular disease; all patients in Wales under the neuromuscular network
    • Parkinson’s disease; all patients in Cardiff and Bridgend with Parkinson’s disease and related disorders
    • Ataxia and movement disorders registry
    • All-Wales epilepsy registry as well as supporting epilepsy surgery programme
    • Metabolic medicine
    • Botulinum toxin clinics
    • Rare disease registries and genetics
    • Welsh Neurosciences Research Tissue Bank (WNRTB)
  • go-terminology: A comprehensive terminology server. Running live and in production in NHS Wales.
  • HealthContact: iPad and iPhone application suite providing patient access to the PatientCare EPR for recording patient reported outcome measures. Replaced by hermes.
  • concierge: A set of APIs implemented in golang providing access to live NHS Wales systems with identifier resolution and mapping.
  • patientcare-ward: Proof-of-concept electronic observations and dynamically created National Early Warning Score (NEWS) chart, written in Clojurescript.
  • HealthContact web portal and iPad and iPhone applications for research recruitment and recording patient outcomes in clinic. Open source. Integrates automatically with PatientCare EPR. The first Apple ResearchKit applications deployed in the United Kingdom (2015) and trialled in the Parkinson’s clinic.
  • A version of the PatientCare EPR with a modern bespoke front-end designed only for multiple sclerosis. Uses pseudonymisation and tracks patient outcome to whole cohort in real-time (just like the full version, but hosted on Internet). Permits clinicians and patients to have web-based and iOS-based remote access and data entry into the NHS-hosted PatientCare system. Accessed by multiple NHS services across UK.
  • rsterminology2 An open-source SNOMED-CT terminology server with fast free-text autocompletion functionality built as a microservice, currently live within PatientCare and deployed independently within NHS Wales providing comprehensive search and browse functionality as well as understanding of semantic meaning and subsumption. Replaced by hermes.