Artificial intelligence (AI) is a term that’s being increasingly used in relation to the health service, but its meaning is increasingly becoming misunderstood, says Ian Jackson, medical director and clinical safety officer at Refero.
It tickles up a feeling of unease in many households as something to be feared when applied to medicine and it’s turning the public opinion of everything it promises into a dead end of mistrust. It doesn’t need to be this way – in this article Ian explains how AI – or, more accurately, machine learning – can transform healthcare
The Department of Health and Social Care has refreshed its code of conduct for the use of artificial intelligence (AI) within the NHS. Standards and regulations are good, but I’m one of the growing number of healthcare technologists concerned by the frequency and misuse of the term.
I think it’s all in the nomenclature. We should discount the word ‘artificial’ when we talk about how ‘artificial intelligence can save the lives of sick people’ – let’s use the more useful and accurate term, ‘machine learning’. The impact that the word ‘artificial’ is having on machine learning – and on the public’s perception of its potential to bring positive changes to healthcare – is too great.
Latent intelligence lurks
Let’s set aside NHS progress on paper records and comms for a moment, as machine learning can’t harness handwritten knowledge. Your local hospital probably has about 150 distinct software applications running right now. Each one contains important data that could save or change a life, so let’s call it the hospital’s total ‘intelligence’, which largely remains dormant unless called on. The potential in that latency for saving present and future lives is huge – enter machine learning.
There’s no doctor, nurse or midwife in that hospital who can access all 150 applications, all at once, and use what the hospital really ‘knows’ about an illness, a crisis, or a patient. In addition, almost none of this intelligence is readily available to the general practices which service the same patients. Machine learning could connect this knowledge and present the statistical information that no clinician will ever be able to compute themselves. Now the hospital’s intelligence is no longer simply ‘artificial’, it becomes real. Crucially, it also becomes widely available too, to all clinicians, and that’s where the ‘learning’ part of ‘machine learning’ is so important.
Though practices tend to have a single software supplier, it can still be difficult for staff to access and collate information held about a patient, especially when it’s held by another source in the health and social care network. Often, because of time or staffing pressures, patient intelligence is not accessed and, therefore, not acted on. Linking what is available on the patient record with notifications from hospital of visits to A&E, or identifying patients ‘at risk’ due to age or complex drug therapy, all fit within the art of the possible that machine learning brings. When truly linked to information available from a hospital system and social services, then it can become an extremely powerful aid to preventative care.
Machine learning to build care pathways
This real, not artificial, intelligence – drawn together by machine learning technology – can now be treated as another string to the clinician’s bow, rather than a technology that is destined to supersede them.
The unified intelligence from machine learning can be used to build bridges between healthcare and other life-changing public services such as social care, policing and mental health services. Potentially, it could solve the delayed transfer of care with its ability to connect systems and master the patterns of crisis periods.
Long-term conditions, such as cystic fibrosis, could be treated at home and recurring illness could be triaged there too. Clinicians can be connected with patients at their university halls of residence, their retirement community or their hospice. Mental illnesses can be more accurately integrated and available to a patient’s treatment pathway across all sources of care, from the maternity unit to student welfare officers.
Unlocking value, both societal and financial
Matt Hancock’s support for AI/machine learning in healthcare is valuable but it’s unlocking the value that the technology can bring which will make the difference. GPs, clinicians, nurses and social workers themselves have no ability to unlock the value of information without technology giving them access to it. We must secure their support, their enthusiasm, and pass them the key to commercialising their specialities, in the same way that the private sector has been able to. Machine learning has changed banking, customer service and retail, making commercial products out of business processes. This can also be done within healthcare.
Machine learning could bridge the commissioner/provider split, accelerating sustainability and transformation partnerships to success. It could also transform the services commercialised by global digital exemplars within the NHS, at the highest level.
The immediate, societal value of machine learning is clear. Let the technology in, to analyse data and risk, and link health and social care intelligence so that bed management becomes simpler and less emotionally charged. Allow the technology to analyse data associated with long-term conditions and patient appointments, and it can determine when appointments are necessary, and where, and how they should take place – no clinician has time to map each case they are responsible for.
The technology is the way forward, just not the term for it
The Department of Health and Social Care, unified and renamed to build better care pathways, has recognised the need for clarity around AI, or machine learning. The code of conduct, particularly around development by technology firms, is essential.
Life and death, and everything in between, becomes easier if medical data becomes intelligent – and there’s nothing artificial about that.