Artificial Intelligence, or AI, is on the front pages every day, but what does it really mean? How can AI best be harnessed to improve medical research and transform patient care in the NHS, and what does it mean for the HRA?
What does AI actually mean?
Computers can build on an algorithm (a set of instructions), based on the patterns in the information that the computer has used to learn, and therefore execute tasks: this is artificial intelligence.
An algorithm or series of algorithms can integrate vast amounts of structured information, far more than any human, but will be limited to the information on which it has been trained. In future, AI will increasingly be able to understand context and learn more widely from related experiences.
For the purposes of healthcare, AI is best thought of as augmented intelligence. There are few settings in which AI can yet replace the wide set of experience and perception of a healthcare professional, and thus AI tools can be used to support clinical decision-making and not to replace healthcare professionals or the personal care they can provide. The UK is a world-leader in this field, as explained in the Life Sciences Industrial Strategy.
How AI is currently used in healthcare?
AI is currently used throughout healthcare: from ScriptSwitch driving prescribing decisions, to diagnostic software selecting patients for recall. In health research, AI can be used to accelerate design of drugs, to select trial participants, to monitor participant outcomes and to synthesise trials outputs.
So many aspects of our everyday lives are already enhanced by AI, and healthcare and health research will be no exception. In his recent review, ‘Preparing the healthcare workforce to deliver the digital future’, Eric Topol outlined how AI can be used to create “the gift of time” for healthcare professionals, releasing clinicians from administrative tasks to spend time facing patients, but also the extensive need for training across the NHS.
Early indications are that AI in healthcare will be most effectively focused on productivity and workflow, including healthcare resource prioritisation and process-based decisions, releasing healthcare professionals to care. In research meanwhile, the CEO of Novartis has said that AI is already delivering in clinical trials operations and in back office areas such as finance, with areas such as diagnostic imaging holding promise.
What could the future hold and what does this mean for the HRA?
As Novartis has begun to demonstrate, AI will drive the evolution of clinical trials as new types of digital products are tested, products being tested are iterated more quickly, fewer pre-profiled patients are required, protocols can become more flexible, and results are immediately available.
New processes will be possible, and the workforce will need time and space to embrace change. Regulators are looking to flex their approaches in discussion with researchers, but this will take time and careful discussion.
At the HRA, we are working to embrace the opportunities which technology creates for us as an organisation, as well as to consider how we may need to operate in the future so that digital products can be efficiently, ethically assessed and trialled in our NHS. Listening to and working with the people in our Research Ethics Committees across the NHS to refine and iterate our approach will be a critical element.
Dr Nicole Mather, April 2019