Future of UK Regulation Conference 2023: Novel technologies
This is the last of the blog series reporting from the BIA conference, ‘The Future of UK Regulation: Driving innovation in the life sciences’, which examines measures to remove roadblocks to the uptake of novel technologies and looks at what a risk proportionate approach to securing the benefits, while ensuring safety, should look like.
Pro-innovation regulation of technologies review on life sciences
The regulatory response to COVID-19 demonstrated the speed and agility with which the UK system can act. While it is unrealistic to think it can always operate at pandemic intensity, the experience points the way to how the UK can build on its new regulatory flexibility and expertise in regulatory science to bring novel technologies to market.
“What we need to find are sustainable new ways of working that take those COVID-19 lessons forward,” Dr Camilla Fleetcroft, Strategy and Innovation Director, BSI, told the conference.
Dr Fleetcroft and Professor John Bell supported the former Government Chief Science Adviser, Sir Patrick Vallance, during the review of existing rules to develop a pro-innovation regulatory framework that is less costly and more responsive to novel technologies. “We’re trying to attract innovation and investment in the UK, support UK businesses, build on the Life Sciences Strategy and really seek the opportunities in the life sciences sector,” Dr Fleetcroft said.
The review, published in May 2023, found that over and above acknowledged problems of inconsistencies and fragmentation of the regulatory system, a pressing challenge is attracting skills and talent. A key recommendation was to establish centres of excellence in regulatory science and innovation, creating a supportive environment for the Medicines and Healthcare products Regulatory Agency (MHRA). “This is about dynamic regulation with a network that can offer support to both the industry and regulators to build capacity,” Dr Fleetcroft said.
A final and overarching recommendation was to enable companies to consult the regulators as early as possible in development, to help map the route to regulatory approval. “This will give them clarity and hopefully avoid some of the roadblocks later on,” said Dr Fleetcroft. The Innovative Licensing and Access Pathway was established to enable this, and work is in hand to create a similar innovation route for medical devices.
The potential returns: Artificial Intelligence (AI) for drug discovery and clinical development
It is now possible to use DeepMind’s AlphaFold AI system to predict the 3D structure of a protein from its amino acid sequence and design drug hits against pockets in these structures, in silico.
Dr Beatrice Panico, Head of Clinical and Principal Consultant, Scendea, told delegates that AI can be used to repurpose existing drugs. That led BenevolentAI to identify baricitinib, which was predicted to reduce the ability of COVID-19 to infect lung cells. This was a huge success in a short timeframe since the process took only 90 minutes of compute time and less than three days of human work to line up the treatment for testing in hospitalised COVID-19 patients.
Another potent example of the power of AI in healthcare comes from recent research using machine learning to locate the primary tumour in the 3 – 5% of cancers where standard diagnostics fail to identify the primary site, limiting treatment options. “This is a great way to look at AI to increase the predictive and also the diagnostic power,” Dr Panico noted.
In clinical trials, it is now possible to create a digital twin of an individual patient that can then act as that patient’s control. In analysing data collected by wearable devices, AI can spot patterns and associations that humans cannot.
All of this points to a future model in which AI is applied across the full product lifecycle, encompassing target identification, to discerning the mechanism of action, combing through the literature, interrogating biobank data and testing drugs in silico before going into animal models. “That will then inform trial designs that will answer the questions we want to address,” said Dr Panico.
AI in healthcare: how’s it going?
“You may not have directly experienced AI-enabled healthcare, but I can pretty much guarantee in the next couple of years, each person in the room will experience it in one form or another”, said Professor Alastair Denniston, Consultant Ophthalmologist, University Hospitals Birmingham NHS Foundation Trust. He gave examples of AI and machine learning he is seeing in use in his local hospital – these range from AI-enabled decision support to AI-assisted assessment of images in diagnosing skin cancer, to specialist software for recording and distilling meetings and the updating of electronic health records.
“I’m excited about the potential for AI health technology because if we get it right, it can solve some of the major capacity mismatch,” said Professor Denniston. Apart from bridging the capacity gap, AI could improve the quality of healthcare, supporting more consistent 24/7 services.
But there is an alternative scenario, which is that AI leads to highly variable services because of the inbuilt biases on which to date many systems have been trained. Given this, “it is important not to exceptionalise AI, but rather to treat it like any other technology,” Professor Denniston suggested.
“Essentially, our AI systems are trained on a tiny fraction of humanity,” Professor Denniston said. And often there is no supporting data about the age, ethnicity or socioeconomic group of the participants. To address this, Standing Together has involved more than 350 stakeholders in 58 countries. Draft recommendations for healthcare dataset standards were published in April 2023.
The Regulatory Horizons Council of which Professor Denniston is a member, has published an independent report to government about the regulation of AI as a medical device.
Ensuring the safety of medical products into the future
The MHRA recognises there is a risk-benefit balance to be struck between encouraging innovation and providing access to emerging technologies, whilst ensuring patient safety. The agency’s new status as a sovereign regulator can help in achieving this, Dr Janine Jolly, Deputy Director, Benefit Risk Evaluation, MHRA, told the conference.
“We have to be able to embrace innovations in medtech and in medicine. In essence, this is very much about positive regulation, being proportionate, being responsive, being agile and making a valuable contribution to the wider health agenda,” Dr Jolly said.
The MHRA is taking steps to improve the quality of post-marketing data it receives to improve oversight of authorised products. Linked to this is the work to improve pharmacovigilance and monitoring for adverse drug reactions, from the patients’ perspective. “We’re putting patients front and centre in the work that we do,” said Dr Jolly.
As part of the MHRA’s transformation, joint benefit-risk evaluation teams are now working on both medicines and medical devices. This is to address convergence, which raises challenges not only about how products are regulated, but also how safety data is captured. “We’re making the most of AI technology within the agency, so that we can make signal detection quicker and more efficient,” said Dr Jolly.
Underlying these initiatives is Safety Connect, an integrated platform that provides an end-to-end system from initial safety alert to action taken to protect patients.
For more information, please contact Dr Christiane Abouzeid, BIA’s Head of Regulatory Affairs
More news and updates