Reflections on Artificial Intelligence in Medicine (AIMED22) – Is it ‘time to get real’ about AI?

Read about Dr. Hatim Abdulhussein's experience at the AIMED global summit.

Between 24 and 26 May I had the privilege of being invited to join international experts at the AIMED global summit, to share the learning and experiences from the NHS and particularly our work at Health Education England (HEE) on impact of artificial intelligence (AI) on workforce, education, and training. The conference was bravely titled ‘time to get real’ with the aim to make up for lost time as the summit had been on a hiatus during the pandemic. On the second day of the conference, I joined colleagues in education across the United States and Hong Kong, to discuss the role that AI literacy has in educational programmes. It was a great opportunity to share examples of work we are doing within our Digital, AI and Robotics in Education programme at HEE, as well as my experiences from implementing a Technologies for Health module for undergraduate medical students at Brunel Medical School. In this blog I will share reflections from the conference, and a perspective on where the NHS lies in its adoption, spread and challenges around the uptake of AI and data driven technologies.

My first reflection, having been years since I last visited San Francisco, was the sheer increase in health inequality evident just by walking through the tenderloin district in the heart of one of the world’s wealthiest cities. We can and must do better in caring for our most vulnerable, and I have always been a firm believer that if we build, implement, and evaluate technology appropriately, we will get the best for the communities we serve. The use of AI and data driven technologies must be aligned to the Core20PLUS5 approach to reducing health inequalities, ensuring solutions support our most deprived populations, and are truly inclusive and are aligned to key clinical areas of focus. In January we published the AI Roadmap, which demonstrates the current landscape of AI and data driven technologies in the NHS, highlighting the current spread, clinical areas, and types of technologies available based on current data sets. We should dig deeper to understand whether these technologies widen or narrow health inequalities, and how the technologies used can best support our rural and coastal communities, which as highlighted in the Chief Medical Officers annual report in 2021, ‘have some of the worst health outcomes in England, with low life expectancy and high rates of many major diseases’.

The conference was opened by the eminent Dr. Anthony Chang, Chair of the American Board of Artificial Intelligence in Medicine and a true clinician data scientist. He mentioned how San Francisco is the worlds leader in AI start-ups, but also sharing that London, the city I live and work as a GP in, has the second largest amount of AI start ups in the world. He also highlighted that there is a new medical publication every 5 minutes, and this exponential growth of knowledge means it is hard to keep up anymore. I was interested in the concept of sick care, that we mostly treat disease rather than prevent it. If we want to truly move to health care, where we both treat and prevent disease, we need to use other methods to understand and use this growing knowledge base. The opening session ended with the story of a patient, a boy Matea with a cardiac illness, who had developed an acute kidney injury. The father who by chance was interested in AI approached the hospital and worked with them to better understand predictive factors for the development of kidney injury so that other children in the future could benefit from earlier diagnosis and treatment. For me it was clear, if we work with our patients and citizens to develop solutions that address their needs, we will make a difference.

So, what about our workforce. Are they ready for the use of AI and data driven technologies? Well firstly, its important to remember, change is difficult. But we have all seen our lives change rapidly over the last few years, and technology, in particular devices like our phones, tablets, and laptops, are synonymous with our daily lives for most of us. In January 2018, the number of broadband connections compared to population was 98% in Western Europe, and in Northern Europe this was above 100% meaning that there were more broadband connections than people, demonstrating the way we now use multiple connected devices as part of life. So, if this is part of life, should it not be part of the way we deliver health and care? We have had our challenges with technology in the NHS, I myself have painstakingly battled with multiple NHS computers as a doctor in the last 7 years and lost a few times along the way, but if the output is not what we need or expect, we should embrace the opportunity to improve and iterate. For AI and data driven technologies, explainability will be important for this, being able to understand the relevant parts of the tool. Seeing examples in practice and directly working with AI will also help build our trust and confidence. We explore this further in our recently published report on ‘Understanding healthcare workers confidence in AI’.

I left the conference feeling truly energised and optimistic about the use of AI. In the week following I carried out my first clinic as a GP from my home computer on a day when my practice needed urgent support, and despite some slowness of the system, this was a battle I was successful in winning. Managing to deliver patient care from home is a credit to how far we’ve come, and we can go further. From the conversations I had along the way, and the feedback I received on our work at HEE and the wider NHS, I can proudly say that we are world leading on issues like AI in healthcare education, workforce impact, transformation and regulation and standards of AI in healthcare. We do all of this because we want to deliver the highest quality access and outcomes for our citizens, and I couldn’t help but leave with a sense that it is now time to get real about the potential of AI to support us on that journey.