Regulatory Intelligence

MHRA 2030 strategy sets direction for AI device oversight

MHRA plans an adaptive framework for AI medical devices, a Centre for AI and Digital Health, and stronger lifecycle data and traceability. The 14 October strategy sets future priorities; planned UDI measures remain subject to the parliamentary process.

A strategy for AI oversight beyond initial approval

The UK Medicines and Healthcare products Regulatory Agency (MHRA) published its Strategy 2030 on 14 October 2026. For medical device developers, its commitments connect three areas: oversight of changing AI systems, specialist regulatory capability, and better data for lifecycle safety and traceability.

The strategy sets a direction for future work. Its publication does not itself bring a new package of mandatory device requirements into force. Measures requiring legislation remain subject to the relevant parliamentary process, while proposed regulatory tools and organisational changes should not be treated as already operational.

A central issue is how evidence continues to support an AI medical device after deployment. The strategy recognises that performance can change over time, between settings and across patient populations. Foundation models add further complexity. MHRA proposes an adaptive approach combining continuous monitoring of real-world performance with clearer accountability for developers, deployers and health system partners.

For manufacturers, this makes the relationship between development evidence and deployment data particularly relevant. For example, an imaging model deployed in a different hospital may encounter different equipment, workflows and patient populations. Connecting the model version with deployment conditions and monitoring results can help explain changes in performance.

Planned tools for model changes and shared evidence

MHRA plans to expand Predetermined Change Control Plans (PCCPs), allowing pre-agreed model updates to proceed without unnecessary delay. The reference to pre-agreed changes is important: the commitment does not give developers blanket permission to implement every future update without regulatory consideration.

The agency will also explore Master Files for foundation models. These would allow manufacturers and developers to share essential safety and performance information about the underlying systems used in AI-enabled medical devices. This remains an area for exploration, rather than a newly available submission mechanism.

Together, the proposals indicate an effort to make changes and their supporting evidence easier to assess while reducing unnecessary duplication. Manufacturers can usefully examine whether their change records connect the modification, its validation, the conditions covered and the evidence supporting earlier versions. That is a preparation step inferred from the strategy; the document does not establish a uniform new record format or immediate compliance deadline.

A proposed centre for clearer regulatory expectations

The strategy commits to establishing an MHRA Centre for AI and Digital Health. It would bring together specialist leadership and expertise for digital and AI-driven products, including both devices and medicines. Its intended role includes strengthening capability throughout the product lifecycle and making regulatory expectations clearer and more consistent for developers and healthcare partners.

The centre remains to be established, with details of its services and pathways to follow. Developers can monitor how it clarifies evidence expectations and engagement arrangements for different products and stages of development.

Device data and traceability remain implementation priorities

The strategy also links device safety to improved collection and use of data. MHRA plans to upgrade or replace key registration, reporting and public-access systems to support post-market surveillance and improve transparency.

Subject to the parliamentary process, new pre-market regulations would make unique device identifiers (UDIs), or an equivalent, compulsory for devices other than non-implantable custom-made devices and non-implantable investigational devices. The stated purpose is precise identification and traceability throughout the device lifecycle.

A useful practical question is whether device identifiers, software versions, clinical or performance evidence, complaints and post-market information can be reliably connected. Individually complete records can still be difficult to use if an investigation cannot associate them with the same product and version.

These preparation suggestions are analysis of the strategic direction. Manufacturers should follow subsequent legislation and implementation documents for binding obligations, dates and submission arrangements.

Sources

  1. Primary official sourceMHRA Strategy 2030 (opens in a new tab)Medicines and Healthcare products Regulatory Agency
  2. Supporting sourceMHRA Strategy 2030 - Executive summary (opens in a new tab)Medicines and Healthcare products Regulatory Agency
  3. Supporting sourceMHRA 2030 Strategy to modernise patient-centred regulation for the era of AI and personalised medicines (opens in a new tab)Medicines and Healthcare products Regulatory Agency