Louise Flintoft associate director at Onyx Health considers the role of artificial intelligence in healthcare going forward, following a report from Health Education England that found 56 new technologies are scheduled for large-scale deployment in the NHS next year.
Artificial intelligence (AI) has experienced a post-pandemic boom. But what does this mean for its future use in healthcare? Will we be talking to robot doctors and nurses in the years to come? How will AI fit into the future of healthcare?
The use of AI in healthcare is increasingly important. While this isn't a new phenomenon, it is one of many healthcare trends accelerated by the pandemic.
In response to this acceleration, Health Education England (HEE) has published the first UK roadmap on using AI in the National Health Service (NHS). It found that there are 56 AI technologies scheduled for large-scale deployment in the next year, 77% of which will be used in secondary care.
Digital transformation is also a key area of focus identified in the NHS Long Terms Plan, whereby AI will act as a catalyst for change. This will involve greater use of technology like chatbots and virtual assistants to change the way services are delivered.
AI chatbots and digital dialogue
Chatbots can fulfil various traditional healthcare functions, including booking appointments, ordering prescriptions, checking symptoms, and providing basic medical information and advice.
Babylon Health’s use of an AI chatbot based consultation service based on the patient’s medical history and standardised medical knowledge, is a striking example of this. Patients can initially report symptoms on an app, which is checked against an official medical database, using speech recognition technology. They will then receive a recommended course of action based on the information provided. The success of their work led them to partner with the NHS to develop a digital triage system for referrals and dispensing health advice during the COVID-19 crisis.
There is evidence to suggest that AI chatbots could be used to remove the stigma with sensitive health conditions like sexually transmitted infections (STIs). Research from the University of Westminster suggests that patients would rather speak to an AI chatbot than a GP about their sexual health. In addition to service delivery transformation, there is also evidence to suggest that AI is having an increasing use in NHS diagnostics.
Using AI to deliver diagnosis
The British Heart Foundation recently reported AI being used as a tool to improve patient care by detecting heart disease more quickly and accurately. The imaging technology works by analysing heart MRI scans while the patient is still in the scanner. It takes just 20 seconds for the scan results to be analysed, compared to 13 minutes for a doctor to manually assess the images. The new tool can also identify heart structure and function issues with 40% more accuracy than the human eye. AI technology like this can have a particularly important role supporting time poor clinicians as healthcare services work to recover from the critical care backlog caused by the pandemic.
AI imaging technology is also a central part of the National Optimal Stroke Imaging Pathway (NOSIP). It is used to accelerate the clinical decision-making process by providing a real-time interpretation of scan images more quickly and accurately. Diagnosis and treatment interventions are very time-sensitive for stroke patients, and technology like this could help save lives.
Tackling healthcare inequalities
AI-led innovation is also being used to address various structural system issues associated with access to healthcare. NHSX AI Lab (now part of the NHS Transformation Directorate) and the Health Foundation are currently undertaking joint project work using AI to address the racial and ethical healthcare inequalities in the NHS.
There are several technological innovations in the pipeline, such as I-SIRch, which uses AI technology to make identifying the clinical factors that lead to adverse maternity incidents in mothers from ethnic minority backgrounds easier. Black women are five times more likely to die in the UK due to pregnancy complications than white women. I-SIRch will assess how different factors combine to cause these maternity issues and enable better forms of intervention to be designed.
AI diagnostics for ethnic minorities are also under development to improve the accuracy of diabetic retinopathy screening. Recent studies have shown that people of Indian, Pakistani, Bangladeshi, and Caribbean ethnic groups are at increased risk of developing diabetic retinopathy compared to white people and receive poorer diagnosis due to their different retinal composition. As well as revolutionising diagnostic methods, AI technology is also being used in the health service to address some of the underlying healthcare inequalities exposed by the pandemic.
A greater role for AI in a post pandemic health system
Whatever the future may hold for AI in the NHS, it will clearly have an enhanced role in the post-pandemic healthcare system. The latest AI innovations have the potential to transform the way healthcare and diagnostic services are delivered and address some of the underlying healthcare inequalities exposed by the pandemic.
