Making secondary care medicines data research ready
Launched in October 2026, the Med Map project will develop a national approach to preparing hospital medicines data for research within NHS England’s Secure Data Environment. It brings together NHS England, PRSB, the University of Liverpool, Understanding Patient Data and King’s College Hospital NHS Foundation Trust.
The project is funded by the Health Data Research Service (HDRS) as part of its technical transformation programme.
Why this work is needed
Primary care medicines data are already used widely in research. Community pharmacy dispensing records are coded in dm+d, returned to the NHS Business Services Authority for reimbursement and can be linked to other health records in all four UK nations.
Hospital medicines data cannot be used in the same way. National electronic Prescribing and Medicines Administration (ePMA) data cover around 27% of activity, and completeness and quality vary between trusts. High cost drugs data are collected in a multitude of different information systems and data collections , and are not available as research datasets through DARS.
Over the last 15 years a growing share of new medicines has been started and prescribed in hospital. Without these data, NHS England, NICE, the MHRA and researchers cannot readily tell whether a medicine reaches the patients it is intended for, how uptake varies across the country, or what outcomes it achieves in routine care. Where secure data environments do use hospital medicines data, each one builds its own mappings, pipelines and validation.
What the project will do
The project will produce the standards and tooling needed to prepare hospital medicines data consistently in any secure data environment, and will test them on ePMA and high cost drugs data within the NHS England SDE. It builds on dm+d, SNOMED CT, FHIR UK Core and the information architecture that are already developing for the Single National Formulary.
At its centre is a Therapeutic Episode Model: a computable record of a medicine together with its indication, dose, route, duration and the outcomes that follow, so that analysis does not rest on the dm+d code alone.
By 31 March 2027 the project will deliver:
- an inventory and readiness assessment of national secondary care medicines datasets, with an improvement roadmap
- the Therapeutic Episode Model, machine readable information standards and implementation specifications
- mappings between dm+d, SNOMED CT, FHIR UK Core, OMOP and the NHS England Canonical Data Model, with validation rules and data quality tools
- software assisted derivation of prescribing indication from NICE guidance and the BNF
- a reference implementation on deidentified data in an NHS England SDE workspace, linked to outcomes, with analytical code and deployment guidance published under an open licence
Who is involved
- NHS England Secure Data Environment (lead): hosts the reference implementation and provides data and workspace
- Professional Record Standards Body: programme direction, standards development and clinical assurance
- University of Liverpool: clinical pharmacology, data engineering and linkage of high cost drugs data to outcomes
- Understanding Patient Data: patient and public involvement throughout the project, with a patient representative on the project group and public materials published as project outputs.
- King’s College Hospital NHS Foundation Trust: project management and alignment with existing programmes
NICE, UK Biobank and Our Future Health have written in support of the project.
Project timeline
What the data will be used for
The use cases were developed with NHS England, NICE, UK Biobank and the academic community. Two come from NHS England National Clinical Directors. One has asked whether teplizumab is reaching eligible patients, how uptake varies across England, what outcomes are being achieved and whether the investment is good value. Another wants to review high cost monoclonal therapies for severe asthma, but the severe asthma registry is not currently linked to high cost drugs data, so there is no way to compare patients who respond with those who do not.
Linked hospital medicines data would also give NICE evidence for technology appraisal and reappraisal after a medicine is in routine use, give the MHRA real world data for safety monitoring, and let research cohorts such as UK Biobank, Our Future Health and Genomics England study treatment response alongside genomic and biomarker data.
