Artificial Intelligence in Healthcare

Henry Smith Excerpts

Thursday 5th September 2019 Westminster Hall

[Henry Smith (Crawley) (Con)]

I beg to move,

That this House has considered involvement of patients in the use of artificial intelligence in healthcare.

It is a pleasure to serve under your chairmanship once again, Mr Paisley. I am delighted to have been granted this important debate, and I am pleased to see a number—particularly for a Thursday—of Members from all parts of the House present to take part. I declare at the outset that I am the chair of the all-party parliamentary group on heart and circulatory diseases. Earlier this year, the APPG conducted an inquiry, with the support of the British Heart Foundation, to better understand patient perspectives on artificial intelligence. It found huge potential for AI to transform the lives of those living with heart and circulatory diseases and a greater need for those affected to be included in discussions about the development and adoption of new technologies.

Before I go further, I want to speak briefly about what AI is. Put simply, AI is the term given to a set of computer actions that mimic human intelligence. Our report outlines that what separates modern AI, such as machine learning, from other types of computer program is that it can learn and improve at tasks. AI is particularly strong at finding patterns and trends in data that are not obvious through human analysis. I have mentioned machine learning, which is one type of AI. It is where algorithms—a set of rules that a computer uses to make a calculation—are used to look for patterns in data, and the computer then uses those patterns to make decisions. It looks for patterns in many different types of data, from scrutinising images to analysing genomic data.

Every day, we interact with something that uses AI. Whether it is entertainment, online shopping, wearable devices, virtual assistants, chat bots or advertising, the use of AI is ubiquitous. Whether it is through faster or more accurate diagnosis, more personalised treatment, better targeting of demand, improvements in service planning and delivery or better predictions, AI has the potential to touch all aspects of healthcare delivery and management.

Our APPG’s report, “Putting patients at the heart of artificial intelligence”, was launched in May this year. It warns that the spread of misinformation risks undermining public confidence in the use of AI in healthcare. The APPG has therefore recommended that policy makers, parliamentarians, the NHS, charities, healthcare professionals and the health technology industry should seek to engage and involve patients in the design, development and diffusion of AI. If they do not, developments in AI might not reflect the needs of the very people who could benefit from it.

It is important to ensure that fake news and the desire for a quick headline do not undermine the public’s trust and confidence in this important area of research and clinical practice. In a survey conducted for the inquiry, 91% of people with heart and circulatory diseases said that the public should be well-informed about how AI is used in healthcare. Some 90% believe it to be the responsibility of the NHS to inform the public about current and potential uses of AI in healthcare, and 48% of patients surveyed strongly support doctors using artificial intelligence technologies to assist them in diagnosing and treating heart and circulatory diseases.

Heart and circulatory diseases, including coronary heart disease, stroke and vascular dementia, affect millions of families across the UK. The halving of deaths from heart and circulatory diseases since the 1970s has been a major health success for the UK. However, such conditions still cause a quarter of all deaths in the UK and are the largest cause of premature mortality, particularly in deprived areas. Together, they make up the single biggest driver of health inequalities and cost the NHS in England at least £7.4 billion a year. As outlined in the long-term plan, it is the single biggest area where the NHS can save lives over the next 10 years.

In assessing the potential for AI, it is important to note the scale of heart and circulatory diseases in this country. The British Heart Foundation, which provides secretariat support to the APPG, reports that heart and circulatory diseases still cause a quarter of all deaths in the UK, on average killing one person every three minutes. The number of people living with heart and circulatory diseases also remains high, at 5.9 million in England. There are more than 42,000 premature deaths from cardiovascular disease each year in the UK. We must therefore utilise the enormous potential of AI across all areas to transform the way we prevent, diagnose, treat and support those living with or at risk from heart and circulatory diseases.

In my constituency of Crawley, 11,000 people were living with a heart and circulatory condition in 2017-18. Of those, 3,679 had coronary heart disease and 1,865 were living with stroke, 774 were living with heart failure and 1,985 were living with atrial fibrillation. In addition, 16,682 constituents have been diagnosed with high blood pressure, including me, and 7,555 with diabetes. While those numbers may seem high, the British Heart Foundation tells me that according to the quality outcomes framework data, Crawley is ranked 548th out of the 650 UK parliamentary constituencies for the prevalence of cardiovascular disease.

In communities around the country, including Crawley, one of the challenges of introducing AI into everyday practice in healthcare is its potential to exacerbate health inequalities. Age, ethnicity, and socioeconomic demographic factors can influence access to the best technologies. Access to new technologies is relevant because AI is currently being implemented in consumer-facing technologies, such as smartphones, which can help manage adherence to blood pressure medication, smart watches, which can track and analyse heart rates, and voice-activated assistants such as Alexa or Siri, which can act as useful reminders to take medications.

As I mentioned, the APPG on heart and circulatory diseases launched its report on AI earlier this year. Our group was grateful for the involvement and enthusiasm of the Secretary of State for Health and Social Care, who also took the time to speak at the report’s launch. Given the number of people in Crawley who have heart conditions, I wanted to keep local residents updated about my work chairing the group. Shortly after the report’s launch, I wrote in the Crawley & Horley Observer about the importance of tackling such conditions and reiterated the salience of the Department of Health and Social Care ensuring that some of this Government’s increased funding for our NHS is used to address the use of AI and its potential in the health service.

It was very much with that call in mind that, almost a month ago, I welcomed the Secretary of State’s announcement that £250 million is to be spent on the new national artificial intelligence lab to improve the health and lives of patients. The Department of Health and Social Care has said that the AI lab will bring together the industry’s best academics, specialists and technology companies. They will be working on some of the biggest challenges in health and care, identifying the patients most at risk of conditions such as heart disease. That will allow for earlier diagnosis and cheaper, more focused and personalised prevention.

The new national artificial intelligence lab will sit within NHSX, the new organisation that will oversee the digitisation of the health and care system in partnership with the accelerated access collaborative. One of the key recommendations of the APPG report is that NHSX should set up discussions with charities and the public to explore the views and concerns of patients about the use of AI in healthcare, and I would be grateful for the Minister’s assurances that through the development of the new lab, NHSX will be exploring the opinions of patients and thoroughly engaging them throughout that ongoing process.

In the past five years, we have seen AI go from struggling to identify images of cats to being able to identify skin cancer in histological sections of biopsies just as well as a team of specialist doctors with decades of combined experience. In debates on this topic, it is easy to discuss issues in what seem like abstract terms, but when patients go to see their GP, they want to see their GP. In such cases, AI could be used to create automatically the GP’s notes about their patient, reducing the time that the doctor will spend looking at their screen, for example.

There is also the issue of self-management. From dedicated apps that people use while going out for a run to the most basic step counters, more and more people use their own devices, on some level, to keep an eye on their health. AI can be used more and more in this area. Patients could use wearable devices and sensors to manage their condition at home and in the community instead of in hospital. AI systems could then monitor for unusual patient-specific patterns, such as a deterioration in a heart failure patient, and relay that information to a clinical team for further intervention. That also presents an opportunity to put patients in much better control of their care.

Our inquiry heard from experts from the University of Cambridge and the University of Oxford, who told us that NHS health checks could be better at distinguishing the risk of different types of heart condition, to ensure that the most suitable treatment can be received by the patient. On 16 August, the Department of Health and Social Care announced a review of the NHS health check service, which is offered to everyone between the ages of 40 and 74 to spot the early signs of major conditions that cause early death, including stroke, kidney disease, heart disease and type 2 diabetes.

Although the NHS health check programme has identified more than 700,000 people at high risk of cardiovascular disease over the last five years and has saved an estimated 500 lives each year, the Department of Health is right that there is potential for people to benefit even more from an enhanced tailored service. The APPG’s survey of patients with heart and circulatory diseases found that 64% had at least some awareness of the potential future uses of AI to diagnose and treat heart and circulatory diseases. However, only 17% of respondents were aware of any current uses. That represents a huge opportunity to inform patients about the opportunities of AI.

People are becoming more and more wary about the use of their personal data. From cold calls to unsubscribing from mass emails, there is increased caution from people about giving up personal information. When it came to the APPG’s inquiry, however, 86% of respondents were comfortable with their personal health data being used to help better to diagnose medical conditions. Policy makers should feel confident that patients support the use of AI in healthcare if it is done to improve health outcomes.

Trust works both ways of course, and it is important that those implementing policy and programmes are open with the public about how their information will be used. That is why patients, and the wider public, should feel involved with not only the details of what their data will be used for but the wider work of the NHS to use artificial intelligence to improve our health service. In June, when speaking on the use of AI, NHS England chief executive Simon Stevens said that

“from April next year we propose to change the way we fund care so that NHS organisations who invest in this world-leading technology will be properly rewarded for doing so.”

I would be grateful for an update from the Minister on what form that is due to take. I am sure that such an update would be welcome if colleagues are to make representations with their own health authorities and trusts.

Our report raised the issue of what patients need to know. Transparency is welcome, and it is important to specify what type of transparency, as well as its intended outcome, in addition to being clear about for whom the transparency is intended. Transparency can include outlining why an algorithm was developed, what types of data were used, and how the development was funded. Some experts have argued that the black box of AI—the difficulty in understanding how AI models reach their decisions—is not really a problem at all, as humans are equally opaque in how they arrive at decisions.

However, the ability to scrutinise, conduct quality assurance, and undertake due diligence are important parts of regulating the health system and ensuring patient safety. In November 2017, the national data guardian for health and care, Dame Fiona Caldicott, told the House of Lords Artificial Intelligence Committee about the challenges of using patient data in technology, saying:

“What we have not done is take the public with us in these discussions, and we really need their views.”

That needs to be addressed. If patients are to trust the use of AI in healthcare, they need to know they are a vital part of the journey.

Our report also looked at the regulatory framework, and how the development of such technological innovations means that health systems are becoming more complex environments to regulate. At the same time, it is important that the regulatory burden is not added to, so that the spread and adoption of new innovations is not stifled. Our inquiry found that a

“balancing act between managing expectations and encouraging hope and enthusiasm is always challenging but nevertheless important. When we say patients should be informed and clear on what AI can do for the NHS, it is not a tick-box measure. It is to provide the clarity that is needed for better diffusion of AI.”

NHS England and NICE, the National Institute for Health and Care Excellence, should encourage the development and use of reporting standards for AI research, in order to provide best practice for artificial intelligence researchers. That could also lead to greater recognition of quality in AI research, particularly among the media, policy makers, clinicians and the public.

With regard to my constituents, I mentioned the importance of Government, policy makers and NHS staff, all of whom have an important role to play in supporting patients. I am also grateful for the secretariat support provided to the APPG by the British Heart Foundation, and I pay tribute to the charity’s hardworking volunteers, including those whom I have been pleased to meet throughout Crawley, and those at the British Heart Foundation shops located on Queensway and on the Broadway in my constituency.

There is much to welcome in the NHS long-term plan. Indeed, NHS funding will grow on average by 3.4% in real terms each year from 2019-20 to 2023-24, which is of course welcome. The current funding increase will mean that the NHS can lay further foundations for service improvements. Thanks to our NHS staff, millions more people are being treated every year. Although services return to Crawley Hospital—and I continue to call for even greater provision—it remains the case that the worst decision in the history of Crawley as a new town was the removal of A&E in 2005. Our constituents expect to see improved GP provision, reduced waiting times and enhanced frontline services.

The APPG on heart and circulatory diseases welcomes the great strides made in recent years to speed up the development and diffusion of AI in the NHS. The Office for Artificial Intelligence and the AI Council have huge potential to bolster the UK’s position as a world leader in AI as part of the Government’s AI sector deal. The Centre for Data Ethics and Innovation can also cement the UK’s leadership in ethical AI and ensure that society can shape the direction of travel and reap the benefits of AI, and we hope that those initiatives will continue to be taken forward.

The chief executive of NHS England has called for this country to become a world leader in the use of AI and machine learning, stating that exploiting the boom in AI technology can help meet the target in the NHS long-term plan of making up to 30 million outpatient appointments unnecessary, in addition to saving more than £1 billion in what would have been increasing outpatient visits. The money can be reinvested in frontline care and save patients unnecessary journeys to hospitals. That reminds us that patients must be at the heart of today’s debate, and hopefully future debates in Parliament on this issue.

--- Later in debate ---

[Henry Smith]

Thank you very much again for your chairmanship, Mr Paisley.

I sincerely thank the hon. Members for Cambridge (Daniel Zeichner), for Strangford (Jim Shannon), and for East Kilbride, Strathaven and Lesmahagow (Dr Cameron), my hon. Friend the Member for North East Derbyshire (Lee Rowley), and the Opposition spokesman, the hon. Member for Burnley (Julie Cooper), for their contributions to this important debate. I congratulate the Minister, and I welcome her to her well-deserved position.

The key word I heard was “trust”, and as we go forward with AI, we need to instil that for patients.

Motion lapsed (Standing Order No. 10(6)).