Guidance

Software and Artificial Intelligence (AI) as a Medical Device

Updated 25 October 2023

Overview

Software (including AI) plays an essential part in health and social care. In the UK, many of these products are regulated as medical devices (or as in vitro diagnostic medical devices (IVDs). This guidance provides access to important Software Group outputs that might be of assistance.

Innovative Devices – Software Group

Software Group are responsible for taking all reasonable steps to assure the safety of SaMD and ensure the UK public have access to technology that meets a clinical need. The Group works across the MHRA to achieve this aim for SaMD and AIaMD through:

  • Assisting with pre-market and post-market enquiries from manufacturers
  • Conducting technical file reviews and post-market surveillance activities
  • Reviewing of technical and clinical aspects of clinical investigations and exceptional use authorisations
  • Ensuring medical device regulation is fit for purpose, meets the needs of software as well as AI, and is supported by robust guidance
  • Engaging with stakeholders including industry, healthcare organisations and professionals, as well as patients and public to support the functions noted above

Classify your software as a general medical device or an IVD

See our guidance on when software applications are considered to be a medical device and how they are regulated. As Software Group works through its programme of work described in the Software and AI as a Medical Device Change Programme Roadmap (“Change Programme”) this guidance, along with other existing documents, will be updated as required.

Manufacturers as well as researchers may also wish to consult In-house manufacture of medical devices in Great Britain or our guidance on the Health Institution Exemption (HIE) – IVDR and MDR (Northern Ireland) as this may also have implications for the regulatory status of software.

Software Group also produced specific advice regarding software used during the COVID-19 pandemic.

UK Regulatory Framework for Software as a Medical Device

The MHRA announced plans for an extensive Change Programme and supporting Roadmap to drive regulatory changes including key reforms across the software as a medical device lifecycle, from qualification and classification, to requirements that apply both pre and post-market. This Programme also considers the challenges and opportunities posed by AIaMD, ensuring that these devices are appropriately evidenced, as well as address wider issues of transparency of AI (both explainability and interpretability), and adaptivity (retraining of AI models).

Software and AI as a Medical Device Change Programme Roadmap

On 17 October 2022, the MHRA published the Software and AI as a Medical Device Change Programme Roadmap that sets out further information on each work package of the Change Programme, including deliverables to meet each set of broad objectives, and further information on how the broad Change Programme will be implemented. This page will be updated as new documents are published.

Qualification and Classification

Creating an intended use statement provides guidance to help manufacturers on defining intended purpose for Software as a Medical Device (SaMD)

Post-market and vigilance

There is Information for manufacturers of medical devices about reporting adverse incidents and field safety corrective actions to the MHRA. This includes specific guidance on the vigilance reporting requirements for manufacturers of software as a medical device as part of a suite of device-specific guidance. The SaMD vigilance guidance outlines the types of events that may cause direct and indirect harm and which are considered to be reportable under the device vigilance system

If you have a problem with software as a medical or indeed any medical device, report it to MHRA using the Yellow Card scheme or via the Yellow card app.

There are specific arrangements for healthcare professionals to follow in each of the devolved administrations.

Healthcare professionals should report incidents to:

in England and Wales to the Yellow Card scheme or via the Yellow Card app

in Scotland to NHS National Services Scotland online incident reporting system and their local incident recording system

in Northern Ireland to the Northern Ireland Adverse Incident Centre and their local incident recording system

Artificial Intelligence

A Regulatory Pioneers Fund award enabled the MHRA to undertake research on methodologies for determining significant change in the way adaptive AI algorithm medical device are working and how such change should be regulated. This work was done with Brunel University as an academic partner. The work included a series of workshops between October 2021 and March 2022 which have been written up as part of the final report. This is research work and is NOT a statement of MHRA policy.

The MHRA with the U.S. Food and Drug Administration (FDA) and Health Canada have jointly identified 10 guiding principles that can inform the development of Good Machine Learning Practice (GMLP).

The MHRA, FDA and Health Canada have also jointly identified 5 guiding principles for the use of Predetermined Change Control Plans (PCCPs). These principles will help to ensure that our future guidance on PCCPs aligns internationally with these jurisdictions.

Digital mental health technology

The Wellcome Trust have funded a 3-year project led by Medicines and Healthcare product Regulatory Agency (MHRA).  In this project we will be focusing on effective regulation and evaluation of digital mental health technology. By enabling access to safe and effective digital mental health technologies, the project will improve outcomes for people with mental health conditions. This will be achieved by ensuring that digital mental health technologies are regulated to minimise risk through continuous evaluation.​

We are leading on this project in partnership with NICE, and will work with many different industry, medical and public experts such as:

  • Lived experience advisors[CH1]
  • Patients and the public
  • Clinical experts
  • Academia
  • Industry experts
  • UK and international policy makers,
  • Regulatory bodies
  • Evaluation agencies
  • Charities

The project will incorporate discussions around lived experiences which are unique when compared to discussions that focus on ‘patients’ and ‘service-users’.

Identifying as having lived a experience of mental health challenges  is a viable option to a wider range of people, whose help-seeking behaviours and self-identification are not tied to accessing clinical or medical support. 

The results of this project will include various published guidance documents to make clear the regulation and evaluation of these technologies for patients, the public, healthcare professionals and developers.

Collaborations

Software Group collaborates with several international regulators, academic groups, and UK health organisations to advance the work that is being undertaken on software and AI medical device regulation. We will highlight major programmes of partnership working with: