Reg-Fr-AIMs (Regulatory Compliance Framework for Trustworthy AI Medical Device Software) is a Research Ireland funded Frontiers for the Future Partnership award between Dundalk IT’s Regulated Software Research Centre and UCD’s Natural Computing Research & Applications Group.
Would you trust an adaptive software algorithm within a medical device that you rely upon to keep you alive, if it automatically adapts and changes its behaviour based upon changing indicators in your body? This adaptive software brings potential benefits to personalise devices to individual’s needs. We are developing a framework to assist in making this type of Artificial Intelligence technology in medical devices to be trustworthy with safe, reliable behaviour and performance. The framework can be used by regulators to create new standards, and by medical device companies to ensure they produce safe, trustworthy and better medical devices.
Medical devices, including software used as, or part of, a medical device, are a highly regulated technology with increasing regulatory requirements for devices with greater criticality. For example, the levels of safety and security that must be considered for a wellness device such as a wrist worn activity tracker are considerably less than required for a life-maintaining ventilator. There is increasing potential benefit to move towards medical devices with more autonomy that can benefit from the latest advances in Artificial Intelligence research to learn and adapt, for example, in a personalised way to the individual user. However, the current medical device regulatory environment faces challenges in terms of how to handle the potential for increasing autonomy presented by Artificial Intelligence technology. This project brings together the team from Dundalk Institute of Technology who have been instrumental in defining the existing international medical device software regulatory environment with a leading team in Artificial Intelligence research from University College Dublin, alongside two industry test-bed collaborators STATSports and BlueBridge Technologies, to develop an AI Medical Device Software framework and associated AI technologies.
Reg-Fr-AIMs project highlights to date include:
- Employment of 2 post doctoral fellow researchers and training of 5 PhD students
- Reg-Fr-AIMs team members are on the working groups of the following standards: IEC 62304, IEC 63450 and IEEEP7003
- A list of publications arising is provided below. A number of journal and conference articles are currently under active review.
- Prof Michael O’Neill was interviewed for the UCD Business Impact Podcast (June 2023) AI: Hype vs Reality.
- Prof Michael O’Neill contributed an invited article to the Business & Finance Annual 2023 on “Augmenting Individuals, Organisations and Society through Artificial Intelligence: Some Risks & Rewards”.
- Dr Róisín Loughran was invited, on behalf of Prof. Fergal McCaffery and the Regulated Software Research Centre, as keynote speaker at the 31st Irish Conference on Artificial Intelligence and Cognitive Science in Letterkenny in December 2023
- Prof. Fergal McCaffery attended the HPRA Advisory Group for medical devices to discuss AI for medical device software, on 14th June 2023 and 27th November 2023.
- PhD candidates Asifa Mahmood and Niamh St John Lynch presented at the Research in Three event in DkIT in December 2023
- Presentation on ‘Challenges of AI in Medical Device Software’ at AICon
- Presentation of paper at EuroSPI conference: Artificial Intelligence-Enabled Medical Device Standards: A Multidisciplinary Literature Review
- Membership of the Engage programme. Engage is a nine-week virtual training programme tailored for women eager to explore the commercial potential of their research, run by InterTradeIreland, Shared Island Initiative, and Government of Ireland.
- Presentation at a number of industry facing events including being an invited speaker at the 6th Annual Cybersecurity Conference, medtech industry conference with the TT group
- Panel member at the ‘Innovative Synergies: Integrating Data and Collaboration in Healthcare’ conference hosted in DkIT on 27th June 2024
- During the Summer of 2024 Reg-Fr-AIMs PhD students organised a Pint of Science event in Dundalk and presented their research
- Reg-Fr-AIMs researchers have been engaged in the design and delivery of a new UCD Professional Diploma in Artificial Intelligence and Business Analytics as part of UCD Smurfit School Executive Development.
Publications Arising:
- Hameed, M.A.S., Qureshi, A.M. and Kaushik, A., 2024. Bias Mitigation via Synthetic Data Generation: A Review. Electronics, 13(19), p.3909.