The client

Oxford University Hospitals (OUH) is a world-renowned centre of clinical excellence and one of the largest NHS teaching trusts in the UK. It became a Foundation Trust on 1 October 2015 in order to work more effectively in partnership with its patients and local community to provide high quality healthcare.

The ask

Oxford University Hospitals NHS FT (OUH) has been building the Thames Valley and Surrey Secure Data Environment (SDE) and needed to ensure that data were available in a structure that was readily consumable by researchers with standard clinical research tooling. Alongside the national initiative for SDEs, OUH have been implementing the Observational Medical Outcomes Partnership (OMOP) data model, to support the ultimate goal of ensuring researchers have seamless access to essential information, at a low cost.

What we did

To understand what the most appropriate solution would be, the Answer team conducted an options appraisal, ranking off the shelf and bespoke options against the validated business requirements. The decision made with OUH was to go with the custom solution because of the specific need for tight coupling of documentation. This was deemed to add more value than the cost associated with it.

The Answer team then simultaneously built and improved the bespoke solution, whilst conducting mappings from four source data sets to the OMOP Common Data Model v5.4 standard. In addition, a GitHub site was developed to increase adoption within the research community and invite feedback for improvements to the mappings.

A robust process for analysing and mapping source data to OMOP was developed.

The key steps are:
Defining the scope: for the MVP this involved using the OMOP minimum field set developed by HDR UK, then capturing the requirements in the Jira backlog
Conducting analysis: creating mapping recommendations and initial SQL queries for the source data
Transformation: Reviewing the mappings from the analyst(s), writing the code for the tool
Validation: Internal and external validation

The tooling is capable of outputting human readable mappings which are generated 100% from code. Critically for OUH, the tool also outputs in a machine-readable JSON format, meaning that the OUH team can leverage the tool to populate its own tool, Mauro Data Mapper.

The result

Mappings impact

Answer have developed a solution which: Implements the OMOP Data Model for a scoped set of data, including oncology specific data sets. Includes a scalable, documentable process for defining and documenting mappings and transforming data into the OMOP model. This will be used for future phases of the SDE development, ensuring efficiency and consistency in subsequent initiatives.

Provides value for OUH & researchers, by ensuring that the implemented data model and associated processes facilitate easy and efficient access to essential data for researchers, aiding in the advancement of healthcare research and the improvement of patient outcomes.