Medicines and Medical Devices Bill Debate
Full Debate: Read Full DebateLord Clement-Jones
Main Page: Lord Clement-Jones (Liberal Democrat - Life peer)Department Debates - View all Lord Clement-Jones's debates with the Department of Health and Social Care
(4 years ago)
Grand CommitteeMy Lords, I want to strongly support the amendments in the name of the noble Lord, Lord Freyberg, and myself, and to commend him on an extraordinarily well-informed speech in support of them. I absolutely support the amendments, in particular Amendment 83, which calls for the comprehensive technical review.
The noble Lord made all the points about dynamic algorithms, deep neural networks, the whole question of self-alteration by algorithm and the absolute need to avoid black-box AI, which chime strongly with everything that I know about artificial intelligence. I welcome, right at the beginning, the Minister’s agreement to meet the noble Lord, Lord Freyberg, and me to discuss these issues further. So I am taking this debate not as the end of the road but as the beginning of a dialogue on these issues.
The big question is: does the Minister really believe that we should be reviewing the definition of “medical device”? Does he believe that it is out of date but is simply reluctant to enshrine it in legislation? All Ministers hate their legislation being tampered with, but I am afraid that I am quite sure that this Bill will suffer some treatment on Report.
Even more important than the definitions is the question of how regulation takes effect on these medical devices, in particular on the new medical devices that incorporate algorithms in this way, and whether the regulation is fit for purpose. Without going into the paper at too great a length, I commend to the Minister the discussion paper from the US Food and Drug Administration, Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning-Based Software as a Medical Device. I have searched but I could not find that the MHRA had done anything of an equivalent quality. I will take a few quotes from the document that illustrate that this is exactly the kind of work that our own regulator needs to do.
The document states:
“Artificial intelligence- and machine-learning-based technologies have the potential to transform healthcare by deriving new and important insights from the vast amount of data generated during the delivery of healthcare every day … The traditional paradigm of medical device regulation was not designed for adaptive AI/ML technologies, which have the potential to adapt and optimize device performance in real-time to continuously improve healthcare for patients.”
That is positive, but it also puts on regulators a particular duty to continuously assess and monitor, as the nature of the medical device and the algorithm within it changes.
The paper sensibly goes on to suggest a completely different way of regulating this kind of dynamic algorithm and type of artificial intelligence incorporated in medical devices. It distinguishes between locked—static—algorithms and dynamic algorithms and states:
“In contrast to a locked algorithm, an adaptive algorithm (e.g., a continuous learning algorithm) changes its behavior using a defined learning process. The algorithm adaptation or changes are implemented such that for a given set of inputs, the output may be different before and after the changes are implemented.”
Those are exactly the points that the noble Lord, Lord Freyberg, made.
The FDA, of course, is one of the great regulators globally. In a sense, it carries on regardless of whatever Administration are in place. So it has a very high reputation. I hope that we are heeding the wise words of the FDA and I very much hope that we are taking on board all the suggestions that the FDA has made for better regulation of algorithms. They are different. I hope that the Minister is not being Canute-like in holding back the thought that new technologies need new treatments in regulatory terms and I am looking forward to what he has to say.