Artificial Intelligence in Healthcare

Ian Paisley Excerpts

Thursday 5th September 2019 Westminster Hall

[The Parliamentary Under-Secretary of State for Health and Social Care (Ms Nadine Dorries)]

It is a great pleasure to serve under your chairmanship, Mr Paisley. The debate has been really interesting, particularly in the light of the number of Members who stood up and admitted that, although they could not send emails, only recently learned how to text and do not use contactless payment cards, they were very much in support of the potential of AI technology and what it can offer patients, healthcare settings and the public at large.

AI is exciting and innovative. I have been in my Department only a few days and I have learned some more from this debate. I hope to have some answers for hon. Members, every one of whom gave an example of the exciting breakthroughs and areas of application of AI, as well as of what it can deliver for patients. That is incredibly exciting.

I thank my hon. Friend the Member for Crawley (Henry Smith) for securing this debate. He made the point about misinformation and fake news. We need more of these debates because Westminster Hall, and this place as a whole, is a good forum to knock down those myths, get rid of fake news and stop fearmongering about the use of AI, because journalists who are interested in AI will follow these debates and quote what hon. Members say. We should have more debates on this subject in future.

Artificial intelligence has the potential to make a massive difference to health and care. There are significant opportunities to save money, improve care and save lives. AI technology could help personalise NHS screening and treatments for cancer, eye diseases and a range of other conditions, as well as free up staff time.

Almost all health and care services can benefit from AI in some way, but realising its potential for our health and care system depends on the involvement of patients. We are committed to working with patients to ensure that they understand and are involved in the decision making about how we use AI to deliver the impact that we both want and need.

I will give a few examples of how AI is working. Some patients have already benefited from it, as hon. Members have highlighted. John Radcliffe Hospital in Oxford has developed a system that uses AI to improve detection of heart disease and lung cancer, as the shadow Minister mentioned. Currently, 20% of heart scans result in a false positive diagnosis, and the subsequent 12,000 unnecessary operations cost the NHS about £600 million a year. The potential financial savings are huge.

Another fantastic example of the use of AI is that of Moorfields Eye Hospital’s implementation of the DeepMind AI algorithm for retina scans. The AI can correctly recommend patient treatment referrals, to the same or better standard as world-leading doctors, for more than 50 sight-threatening eye diseases. Tens of thousands of scans were taken of people with both healthy and diseased retinas, and DeepMind developed software that could detect—long before a doctor could—sight-threatening diseases and the patterns that lead to them. That is just one example.

The use of AI goes further than just diagnostics. NHS 111 online, once fully implemented, will automatically triage patients by using AI technology. The system sends patients to the most appropriate care setting and reduces unnecessary A&E visits, meaning that patients can access the care that they need faster.

We must make best use of the available resources within the NHS to harness the full potential of AI, which relies heavily on enormous amounts of data to learn and become effective at its task. That data must be shared safely, however. Health data that is shared fairly, ethically and transparently has the potential to improve outcomes for patients, improve the efficiency and efficacy of the NHS, and underpin the next wave of innovative research taking place in the UK.

To help the NHS and researchers share health data in a safe, secure and lawful way, the Government have committed to developing a policy framework that sets out our expectations for how the NHS should engage with researchers and innovators when entering data-sharing partnerships. That builds on the work of the code of conduct for data-driven health and care technology. We are committed to involving patients and the public in the development of that policy. That is key and comes back to the point made by the hon. Member for Cambridge (Daniel Zeichner). Patients must be at the heart of and engaged in projects, understanding how their data will be used in future and reassured of its safety.

To support the NHS in embedding the framework in practice, we will also set up a national centre of expertise. The centre will sit in NHSX and provide hands-on commercial and legal expertise to NHS organisations to support them in reaching fair, ethical and transparent agreements for data. Although AI has been the subject of much speculative reporting, on both benefits and risks, we know that it will bring big changes to the way in which care is developed and experienced.

While we promote the latest data-driven scientific advances in healthcare, we must always ensure that patient data is respected and properly protected. Data is vital to the delivery of safe and high-quality care, but we need to ensure that an understandable and trusted system is in place, which patients can be confident will protect their data. The Government are clear that patient data will only ever be used and/or shared when anonymised, or with the consent of the individual, unless for direct patient care. That is an important point and one that almost everyone made.

We have therefore put in place several safeguards, including legislation such as the Data Protection Act 2018, enacting GDPR; data and cyber-security standards applicable across the health and care system; and legislation that is under way to put the National Data Guardian on a statutory footing to provide an independent and authoritative voice on how data is used across the health and care system. We have also launched the national data opt-out, which gives individuals choice of how their data is used beyond their individual care. That gives patients choice, which is important.

In some instances, it will be appropriate for patient data to be shared for secondary purposes, such as when consent has been given on behalf of the patient, or there is an overwhelming public interest in sharing. The National Data Guardian is supporting work with NHSX to clarify and update guidance on the lawful use of patient data to support the understanding of the public, clinicians and industry. We do not want to hinder the progression of innovations, but all patient data should be handled with the respect and care that the public rightly expect.

We are also very aware of the ethical issues that can be raised by artificial intelligence at a personal, group and system level. Bias is a current common issue with the use of AI, and we must curtail any bias within algorithms by ensuring that the data feeding them reflects our diverse population and range of health economies. Initiatives such as DeepMind’s ethics and society research group and the Partnership on AI, which counts IBM, Microsoft, Facebook and Amazon among its members, show that industry is alive to the issues. We are already taking steps to ensure the safe development, deployment and use of AI, and the published code of conduct for data-driven health and care technology that I mentioned earlier encourages technology companies to meet a gold-standard set of principles to protect patient data to the highest standards.

NHSX announced that it would set up an “AI lab” to bring together the industry’s best academics, specialists and technology companies to build groundbreaking diagnostic tools and treatments in line with the NHS’s priorities. NHSX is delivering the Prime Minister’s grand challenge mission to use data, artificial intelligence and innovation to transform the prevention, early diagnosis and treatment of chronic diseases by 2030.

The NHS AI lab will harness the power of data science and AI to continue the UK’s great tradition of using evidence-based decisions in health, public health and social care, and to position the NHS as a world leader in artificial intelligence and machine learning. It will collaborate widely to identify impactful ways to improve the NHS through more sophisticated use of its data. Once identified, the lab will develop, test and deploy early-stage software solutions to be handed over to the NHS to implement at scale.

The operations of the AI lab will align to the core values of the NHS. Most relevant to this debate, the AI lab will protect patient privacy—to go back to the substantive concern expressed by the hon. Member for Cambridge in his speech. The AI lab will sit within the NHS and will protect patient data. It will also guarantee that the value of the healthcare data is retained by the UK public.

As well as ensuring that the technology meets the highest standards and sufficiently stringent regulation, we must ensure that the public are aware of that technology. The public must understand the principles well enough to be confident in a particular technology’s capabilities, irrespective of the statistical evidence supporting it. For the NHS to maintain the confidence that the UK public place in its brand, it must ensure that the apps and data-driven technologies that it recommends are examples of the best practice, not simply in transparency but in what they do and where the personal data goes.

There is now an opportunity for the UK to do that well, making the UK’s standards for MedTech an international benchmark, strengthening the position of digital health in the UK and enabling it to make great leaps forward. As I mentioned, the National Data Guardian and NHSX will work together to produce clarifications on the circumstances in which it is appropriate to share data. We recognise the findings of the “Putting patients at the heart of Artificial Intelligence” report produced by the all-party parliamentary group on heart and circulatory diseases and its calls for greater public engagement to avoid a souring of opinion on AI. We will continue to engage patients in the design and development of AI, where appropriate, and to raise the profile of the effectiveness and efficacy of using AI to provide health and care.

I will now go on to the points made by Members and their requests for reassurance. My hon. Friend the Member for Crawley asked how an NHS organisation investing in the new technologies would be rewarded. We are investigating how best to do that by engaging with commissioners, clinicians, business and academics. We will announce more detail in due course.

The hon. Member for Cambridge asked for an assurance that the additional NHS funding that has been announced will go ahead. Yes, the additional funding will go ahead, but we are still investigating how best to distribute it. My assurance to him is that, yes, the funding will be distributed. He himself highlighted the complexity of ensuring the fair distribution of such funding.

My hon. Friend the Member for North East Derbyshire (Lee Rowley) mentioned mitigating the risks. I hope that I covered that in my speech. A huge amount is going into mitigating such risks. For example, the Information Commissioner provides anonymisation guidance. I also refer to the points I have already made about NHSX.

The Scottish National party spokesman, the hon. Member for East Kilbride, Strathaven and Lesmahagow (Dr Cameron) talked about mental health and patients. This morning, I heard about a great example of AI helping a patient suffering with dementia. It is being used to track normal movement and behaviours. When something different or unusual happens in the home to cause concern, an alert is sent out to a first carer who can be on the scene immediately. That is another great use.

The hon. Lady also asked what we were doing about 5G. I will not try to wing this one, but will simply repeat the answer that my officials gave me word for word: we are working closely with the Department for Digital, Culture, Media and Sport, which is leading test beds—is that right?—for 5G in Liverpool and Birmingham, showing how it can improve access to services and exchange of information between patients and clinicians.

The hon. Lady also asked about international collaboration. NHSX will engage with the World Health Organisation through the Global Digital Health Partnership, and the Medicines and Healthcare Products Regulatory Agency has a strong tradition of international engagement with both the US’s Food and Drug Administration and the European Union, which is key to solving difficult regulatory questions.

In conclusion, I reiterate that AI’s potential to transform the way in which we deliver health and care in the UK is huge. Advancements in diagnosis, treatments and prevention facilitated by AI will provide frontline NHS staff with more time to spend providing care to those who need it most. Through our involvement in the Prime Minister’s grand challenge, the AI lab and our work with the National Data Guardian, we will raise the profile of AI as a health and care project, and ensure that the public are fully aware of both its benefits and the expectations they should place on the NHS.

In the last few seconds, would my hon. Friend the Member for Crawley like to wind up?

[Ian Paisley (in the Chair)]

Order. I decide who is called to wind up.

[Ms Dorries]

Sorry, Mr Paisley.

[Ian Paisley (in the Chair)]

You have about 30 seconds to wind up, Mr Smith.