Elevating healthcare

AI is increasingly being used in all facets of healthcare: “It’s everywhere – from drug discovery, the science behind developing healthcare products and running the clinical trials, through to predictive analytics and clinical decision making, and of course within consumer facing products such as wearables and health related smartphone applications,” says Dr Sarah Kavandi.

To ensure AI is used most effectively, our panellists suggest it’s important to be very clear exactly how it will provide better care and improve patient outcomes, whether through creating efficiencies in the system, improving diagnostic accuracy, or allowing healthcare professionals to better allocate resources.

Imaging is one area in which AI is already having a big impact, providing results faster than a practitioner could. However, Professor Henry Potts cautions that there are some areas where AI doesn’t perform more efficiently than an experienced clinician: “AI won’t necessarily be better – in many cases it is quicker, but we need to ensure it does what it’s designed to do. That’s why evaluation and regulation is crucial.”

While there are considerable benefits associated with harnessing the power of AI within healthcare, there is of course also the inevitable risk of doing harm. This raises some important ethical questions about the way AI is used.

“It’s encouraging to see increasing amounts of literature and research on ethics and AI, particularly as it draws on the basic pillars of biomedical ethics: beneficence, non-maleficence, justice and autonomy" explains Sarah.

She adds: “The core principle when you practice medicine is to ‘first do no harm’, and that is what we need to remain focused on as we innovate and develop products using AI.”

It’s also important to consider how much harm could be being caused by current health systems when weighing up the options. Giles Tully of PinPoint Data Science suggests this is part of the case for change: “We need to ask ourselves, ‘what is the ethical problem with doing nothing?’ It’s a balancing act between doing something, or doing nothing, and this is a key ethical consideration in terms of the use of AI in healthcare.”

Another ethical question surrounds the use of personal data for AI applications. Lucy Lehane of BIVDA says communication with patients is crucial: “We need to explain how AI will be used, and how it will add value. It’s about taking people on a journey so they understand exactly what’s been collected and why.”

Misunderstandings about what AI is and how it works can also lead to apprehension in terms of healthcare professionals’ roles. Rather than replacing them, Lucy suggests AI can support clinicians by taking on more time consuming tasks and can even increase access to healthcare in areas that don’t have enough trained professionals. For example, AI can help interpret HIV and TB lateral flow tests in under-resourced communities, providing faster access to results and therefore quicker treatment.

“I like to think of AI as the healthcare professional’s supportive friend, helping to streamline processes, promote efficiencies, reduce waiting times, free up resources and ultimately improve outcomes,” Lucy says. “It’s not about role replacement, it’s about making sure healthcare professionals are where they need to be, and can do their jobs more easily, with more attention and more time.”

While there is broad agreement around the need to regulate AI, the regulatory environment is complex and evolving. “Because technology innovation is so rapid, regulation is struggling to catch up,” Henry explains.

Different jurisdictions also have different rules, and while there is some cooperation between different regulatory bodies, Henry suggests it’s not clear yet whether products developed in one regulatory jurisdiction will be accepted in another. He says: “For example, in the UK, AI rules for medical devices sit with the Medicines and Healthcare Products Regulatory Agency (MHRA), while data privacy sits within the General Data Protection Regulation (GDPR).”

Henry says: “If you get it wrong in healthcare, it can cause harm. It’s not about stifling innovation, but this is a high-reliability field where you have to get things right. Although that may make product development harder, the right checks and mechanisms ensure that innovative solutions can be accessed by the NHS and the wider healthcare system in a way that really adds value for people.”