MHRA outlines data requirements for future UK Health Data Research Service

A new MHRA paper outlines the regulator's data requirements for the future UK Health Data Research Service (HDRS), describing the data capabilities it says the service should provide to support regulatory decision-making across medicines and medical devices.

The Baseline

• The MHRA has outlined the data capabilities it says the future UK Health Data Research Service (HDRS) should provide to support regulatory decision-making across medicines and medical devices.
• The paper identifies gaps in the current UK health data landscape and argues that the MHRA requires access to regulatory-grade, linked patient-level data to support safety monitoring and evidence generation.
• The agency recommends that the HDRS provide unified data access, comprehensive data linkage, and near real-time availability to support regulatory-grade real-world evidence (RWE).

With the UK's Health Data Research Service continuing to take shape, the Medicines and Healthcare products Regulatory Agency (MHRA) has outlined its own vision for the data infrastructure needed to support regulatory decision-making. In a newly published paper, the agency sets out the characteristics of the real-world data (RWD) ecosystem it believes will be required to underpin future regulatory activities, arguing that the HDRS should not simply be a research resource, but a regulatory infrastructure capable of generating regulatory-grade evidence.

The MHRA states that effective regulation increasingly depends on access to RWD capable of capturing the safety and performance of medicines and medical devices across the UK population. However, it emphasizes that its use of RWD differs from much academic research because it is intended to support regulatory decisions that directly affect patient safety and public health. As a result, the agency says it requires regulatory-grade data that are fit-for-purpose, representative of the UK population, and of sufficient quality, completeness, and integrity to support decision-making. The paper concludes that the MHRA "currently lacks access to sufficient regulatory-grade data to meet all its needs," limiting its ability to fully support both current regulatory activities and emerging data-driven approaches. The paper identifies several limitations in the current UK health data landscape that it says prevent the agency from meeting these requirements. While resources such as the Clinical Practice Research Datalink provide linked primary care data, it highlights the absence of comprehensive patient-level prescribing information from secondary care as a particular limitation. According to the MHRA, this restricts surveillance of hospital-prescribed medicines, including oncology therapies, personalized medicines, and high-cost drugs. The paper also highlights incomplete linkage between healthcare datasets, inconsistent recording of medical device exposure, limited availability of patient-reported outcomes and some long-term clinical outcomes, and governance processes that can delay access to data. It states that the MHRA "does not have default access to UK public health datasets, nor a fast-track governance pathway recognizing its legal remit and public health mandate."

The paper explains how these limitations affect regulatory activities that depend on access to high-quality linked data. It identifies post-market surveillance as one of the regulator's principal uses of RWD, alongside monitoring the impact of regulatory interventions and supporting evidence generation for licensing and authorization decisions. For post-market surveillance, the MHRA requires longitudinal patient-level data linking treatment exposures, health outcomes, and relevant clinical and demographic characteristics to investigate potential safety signals and assess whether observed associations are likely to be causal. The paper cites sodium valproate as an example where incomplete linkage between maternal and child records, together with limited information on developmental outcomes, restricts assessment of pregnancy risk minimization measures. It also notes that inconsistent recording of medical device exposure limits the ability to link specific devices with patient outcomes for safety monitoring.

To address these challenges, the paper argues that the HDRS should provide a single point of access to UK health data supported by trusted user status for the national regulator, a unified governance process, comprehensive linkage across primary, secondary, and specialist care, near real-time access to patient-level datasets, and adoption of a federated common data model.  The latter would support analysis while retaining the diversity of existing data sources and facilitating collaboration with international regulatory partners.

The paper argues that these capabilities would support multiple regulatory activities across the product lifecycle. It states that richer, routinely linked datasets could support regulatory-grade RWE generation for licensing and authorization decisions where randomized controlled trials are impractical or unethical. It concludes that, if designed to meet the needs of regulators alongside those of researchers and clinicians, the HDRS could provide "a foundation for public trust, effective regulation, and a vibrant life sciences ecosystem that delivers better outcomes for patients across the UK."