Roundtable: debating the benefits of the NHS’s newly launched AI lab
British Prime Minister Boris Johnson has pledged £250m investment into an NHS AI lab, which will exist within NHSX, the service’s digital-focused segment. Chloe Kent spoke to a range of experts across the industry about this development, and whether it will be a worthwhile investment for the NHS.
We’ve heard from:
- Clinithink executive vice-president and general manager of life sciences Sarah Beeby
- Human+ automation consultant Oliver Cook
- EBO.ai CEO Dr Gege Gatt
- Medidata general manager and senior vice president Christian Hebenstreit
- O’Reilly chief data scientist Ben Lorica
Chloe Kent: Could the NHS AI lab pose any risk to patient's personal data?
Oliver Cook: For most services nowadays, the cost of an efficient and high-level service is the surrender of a certain amount of data. We shouldn’t be afraid of a technology that is ‘taught’ by analysing our personal information, if the correct safeguards are in place.
Gege Gatt: AI has no place in the NHS if data privacy isn’t guaranteed. In 2017, the Information Commission (ICO) ruled that an NHS trust had not done enough to safeguard patient data in an agreement it concluded with DeepMind. Since 2017 and the subsequent introduction of General Data Protection Regulation (GDPR), data security has been prioritised and is also recognised as part of the ethical framework required to build the necessary trust by patients.
Christian Hebenstreit: Our industry and the NHS is governed by the most rigid rules and regulations around protecting patient data and the more digital our industry becomes, the more systems are put in place to enhance cyber-security and ensure patient confidentiality. The NHS has always placed a significant importance on this and the AI Lab won’t change this.
Ben Lorica: The NHS holds personal information about almost every person living in the country, which means preserving privacy, collecting and cleaning data and sharing it in a secure way is crucial. So, the issue of privacy is extremely pertinent in AI’s adoption. Patients need to be assured that their privacy won’t be abused when their medical history is handed over to a computer, instead of traditionally handwritten records.
What new innovations could the AI lab lead to?
Sarah Beeby: We will see healthcare burdens such as rare disease being diagnosed more efficiently, better prediction of risk for mental health issues such as suicide, a reduction in the costs and time associated with healthcare provision, improved clinical diagnosis and supported clinical decision making.
OC: I think the potential benefits to the NHS and wider healthcare industry are dramatic. AI can be a preventative cure, by analysing data from thousands of patients to understand what might be wrong at a much earlier stage than was previously possible. Technology apps, underpinned by AI algorithms, will encourage healthier behaviour in their users. AI can also increase the speed of research into new medicines and treatments, by crunching numbers at a pace impossible for us humans.
BL: With the NHS potentially losing up to 350,000 staff by 2030, using AI will be the only way to scale services to match the mounting demand that is hitting the UK with a shrinking workforce.
What challenges could it mean for the health service?
SB: The additional work that comes with the initial adoption of the resulting products and technology within the health service does raise challenges. The introduction of innovations requires support, training, new skills sets and a refocus of current workflows, with emphasis on education and understanding around why these changes are taking place. Without an engaged workforce truly embracing change, it will be difficult to achieve the potential benefits that are available. We can only make it work for us if we work at it together.
OC: There are ethical questions around artificial intelligence in healthcare. Will AI provide more fair and objective decisions than human doctors, who, despite high levels of training, are still slightly limited by their own personal experience and biases? Or will it simply amplify human prejudices, embedding discrimination within the NHS? Artificial intelligence requires a wide variety of data from a range of socio-economic groups to be properly effective – the duty around this lies with human clinicians.
CH: While it’s great to see this financial commitment and initiative from the NHS and the UK Government, there will be challenges to overcome. To realise its full potential, the AI lab will need to keep pace with the industry and involve all relevant stakeholders within the life sciences ecosystem. Collaboration is key and there needs to be a joint agenda and common objectives across the country’s different healthcare organisations. Above all else, the success of the AI lab will require patient buy-in when it comes to engagement, sharing of data and, ultimately, trust.
BL: As they implement AI systems, practitioners will face issues that are common in every industry: collecting and cleaning data, building networks that enable appropriate data sharing, preserving privacy, explaining decisions, and more. These are all exciting challenges for AI developers work with the NHS to tackle.
What do you think are the most exciting developments in AI right now?
SB: Advances in disease knowledge, better understanding of real patient interactions with treatments and real world information on the true impact of care pathways, combined with phenotype and genomic insights, will enable a complete transformation in how diseases are prevented, when they are diagnosed and how quickly they are treated.
GG: From robot-assisted surgery to radiographic imaging analysis, AI has become critical in healthcare. We are very excited about the potential that it brings to the creation of positive patient experiences. This is not simply a basic interaction, but an effective method to connect patients to the right type of treatment and the right place. AI can augment human capabilities and resources, and this is truly transformative.
OC: The use of optical character recognition and machine learning trained models can help clinical trials with choosing the right subjects, removing the subjective noise from clinical trial medical directors that have a vested interest in including their patients in the study. This can subsequently reduce time to findings, time to treatment, time to approval and ultimately time to market.
CH: Using AI to support value discovery to make better go/no-go decisions across research and drug development is huge. Thanks to integrated systems, connected devices and AI, regulatory submissions have improved massively as well, demonstrating the value to regulators, payers, providers and patients.
BL: Computer vision is probably the area with the most activity, measured in terms of patents, number of start-ups and use cases. Deep learning has proven to be adept at several perception tasks involving images and video. The applications span across security, logistics, manufacturing, healthcare and medicine, media and advertising, agriculture and many more.
Share this article