Open Health Informatics

Our paper on “Linking Health Records with Knowledge Sources using OWL and RDF” was presented at the Information Technology & Communications in Health conference (ITCH) in Victoria, BC, Canada.

The paper describes a method by which the Web Ontology Language (OWL) can be used to specify a highly structured health record, following internationally recognised standards such as ISO 13606 and HL7 CDA. The structured record is coded using schemes such as SNOMED, ICD or LOINC, with the coding applied statically, on the basis of the predefined structure, or dynamically, on the basis of data values entered in the health record. The highly structured, coded record can then be linked with external knowledge sources which are themselves coded using the Resource Description Framework. These methods have been used to implement dynamic decision support in the open source cityEHR health records system. The effectiveness of the decision support depends on the scope and quality of the clinical coding and the sophistication of the algorithm used to match the structured record with knowledge sources.

Read the full paper here.

The 16th Scandinavian Conference on Health Informatics was held in Aalborg, Denmark on August 28–29. We took a long bus ride from London to get to Copenhagen and then a train back up to Aalborg, which was enough to reveal that the highest point in Denmark, Ejer Baunehøj, is only 170m above sea level.

Our paper on “A Method for Reporting of Variance in Informal Care Pathways” was a follow up to previous work on the representation of care pathways as HL7 CDA documents, conducted with colleagues from the Norwegian Centre for e-Health.

Care pathways are commonly used as a means to ensure the implementation of proven best practice in clinical care. They can be defined in an Electronic Health Record system using formal representations of template pathways which are then instantiated for specific patients, to guide clinical users through a sequence of tasks and actions. The open source cityEHR system uses HL7 CDA documents for the formal pathway definitions, but in practice we have found that some care pathways, and clinical protocols, can be implemented effectively using less formal mechanisms.

In addition to the shared patient record itself, we find that the most important of these mechanisms are notification on recording of patient events, support for care teams, annotation of the record, and the clinician-centred ‘In-Tray’. One important feature lost in this approach is the ability to report on the variance of the care provided to the patient from the best practice encapsulated in a more formally defined pathway.

To address this, we describe a method for transforming a formally defined care pathway to a finite state machine representation and querying the patient record to generate a retrospective report on the variance of states recorded in the patient record.

Read the full paper here.

Version 1.6 of cityEHR was released through sourceforge on 31st March. This version includes a number of new features and some bug fixes – it is the same version that is running in NHS hospitals as of March 2018. A summary of new features is reproduced below from the release notes.

V1.6 – New Features

System Level

1. User passwords encrypted in database (if configured)
2. Configurable password policy
3. Check for expired password on sign-on

Information Models

1. Support evaluation context in calculations
2. New option DistinctValues in entry initialValue
3. New option staticValue for elements

Clinical Data Entry (Forms, Letters, etc)

1. Initial values of entry can be all ‘distinct values’

Patient Search

1. Search time displayed only in debug mode

Cohort Search

1. Search for ‘any value’
2. Search data (i.e saved but not published) and archive for authorised users
3. Include saved cohorts as search terms
4. Search scope now included maximum, minimum
5. Cohort analysis feature finds maximum, minimum, sum, average values for cohort

In-Tray

1. No new or enhanced features

System Administration

1. Manage password – uses password policy
2. Manage records – improved merge/unmerge
3. Manage records – archive/restore records
4. Run XQuery – delete query prompts for confirmation

The specialist XML conference Balisage is held each year in Rockville, MA, just north of Washington DC. The event is a haven for XML geeks and this year we made our first appearance there, presenting a paper on “An XSLT Translator for the openEHR Archetype Definition Language”.

The cityEHR Electronic Health Records system is a pure XML application for managing patient health records, using open standards. The structure of the health record follows the definition in the ISO 13606 standard, which is used in cityEHR as a basis for clinicians to develop specific information models for the patient data they gather for clinical and research purposes. In cityEHR these models are represented as OWL/XML ontologies. The most widely adopted approach to modelling patient data in accordance with ISO 13606 is openEHR, which uses its own Archetype Definition Language to specify the information models used in compliant health records systems. This paper describes a translator for the Archetype Definition Language, implemented using XSLT and XML pipeline processing, which generates OWL/XML suitable for use in cityEHR.

Read the full paper here.

The public release of cityEHR 1.5 has been made today on sourceforge. This is the version of cityEHR installed at NHS hospitals in Q1 2017 and follows the first public release made in March 2015.

The 1.5 release is made as a single war file which can be deployed directly into Apache Tomcat and used ‘out of the box’ with the bundled demonstration application.

New features include upgraded cohort search, reporting and export facilities, notifcations, In-Tray, improved patient search and a wealth of new administration features.

Our paper on “Storing and Querying Longitudinal Data Sets in an Open Source EHR” was presented at the Information Technology & Communications in Health conference (ITCH) in Victoria, BC, Canada.

The cityEHR is an example of an open source EHR system which stores clinical data as collections of XML documents. The records gathered in routine clinical care are a rich source of longitudinal data for use in clinical studies. We describe how the standard language XQuery can be used to identify cohorts of patients, matching specified criteria. We discuss methods for ensuring good data quality and the issues in implementing XML queries on longitudinal data sets.

Read the full paper here.

Naveed Dogar has started research for her PhD at the University of Oxford, with her project on “Internationalisation of Electronic Health Records: Implementation and Evaluation of Specialised Musculoskeletal Systems”. Her research will focus on the implementation of cityEHR for Fracture Liaison Services and for club foot clinics and will seek to develop a set of key performance indicators for measuring the effectiveness of those systems. Using the KPIs, she then aims to measure the effectiveness of systems deployed in different international settings to determine whether a single information model can be made sufficiently adaptable to local clinical and cultural practices so that the same system remains effective across all deployments, whilst generating a standard set of clinical data.

Naveed is based at the Botnar Research Centre at the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences. She is a member of St Hilda’s College.

Results of a study into the scalability of the cityEHR database have been published in the proceedings of XML London, 2016. This conference is an annual two-day gathering of prominent XML practitioners and is the perfect forum to present and discuss methods for implementing large scale XML database systems, such as cityEHR.

In his paper entitled “Scalability of an Open Source XML Database for Big Data”, John Chelsom described experiments that evaluated the performance of the eXist database (the open source database used by cityEHR) when loaded with records for up to 20,000 patients and described the implementation of a federated search across multiple database nodes, allowing cityEHR to scale up potentially to millions of patient records.

The Elfin system, which manages the process and clinical data for Fracture Liaison Service has started live operation at the Oxfordshire FLS, based at the Nuffield Orthopaedic Centre. Elfin is now being used by two administrators and up to eight nurses, following a successful period of pilot running which started in December 2014.

Elfin uses cityEHR with an information model developed by Consultant Rheumatologist, Dr Kassim Javaid, with input from nurses and practitioners from Fracture Liaison services across the south of England. Elfin gathers clinical information about the patients seen by the service, tracks the process of treatment and monitoring and can generate reports required for national audit. Elfin generates letters for patients and GPs and can be used to query and extract data for research studies.

Prof John Chelsom presented the principles of Open Health Informatics and the cityEHR health records system at the NHS Open Source Open Day on 26th November in Newcastle, UK. (See his presentation slides here)

The event was attended by over 120 people with an interest in promoting the use of open source software in the NHS. A variety of approaches and applications were presented and it looks as if the drive for open source in the NHS may now be reaching critical mass.

John picked up on some interesting themes introduced by Dr Marcus Baw who is a keen advocate of all things open in health informatics. One theme challenged by John was the suggestion that open source / open standards can be used to prevent wasting time on ‘re-inventing the wheel’.

Whilst this is a laudable goal in many cases, there are also many examples where clinicians should be encouraged and enabled to re-invent wheels – open standards provide the opportunity for clinicians to make exactly the wheels they need for their own work and then share them with others who have similar, but not identical, requirements and goals.

Open standards can be used to enable diversity and interoperability between diverse systems – too often they are used to enforce uniformity which can then go on to stifle innovation