Artificial intelligence (AI) in health and care: mitigating the risks and capitalising on the opportunities

Artificial intelligence (AI) in health and care: mitigating the risks and capitalising on the opportunities

Artificial intelligence (AI) has the power to transform healthcare.

It has the ability to improve patient care, enhance the quality of people’s lives, and make the jobs of healthcare professionals more rewarding.

A report based on a survey by NHS England and the Academic Health Science Network (AHSN) AI Initiative, demonstrates the potential for AI to contribute to improved patient care.

The 2018 report, Accelerating artificial intelligence in health and care, showed that 94 per cent of the UK’s AI thought-leaders cite AI as being extremely important or very important for diagnostics; 89 per cent support this view for operational and administrative goals; and 79 per cent have this opinion in regards to the benefits for health promotion and preventative health.

“The NHS is prepared to use AI to improve its efficiency, deliver better outcomes and prevent ill health,” the report states.

However, there are challenges.

The public needs to have the confidence that AI, and the supporting health data that fuels the progression of new algorithms, is being used safely, legally and ethically.

This naturally points to the need for changes to data infrastructure, organisational structures, commercial arrangements and models of consent.

Some important steps forward have already been made, such as the introduction of a national data opt-out and a proposed Bill to put the National Data Guardian on a statutory footing.

Industry members, academics, innovators and commissioners are also working together to ensure that the NHS receives the maximum benefits of AI, within the necessary regulatory framework while assessing the future needs of healthcare in the UK.

Defining AI

While there is no single, universal definition of AI, the report said, in a healthcare context, it is agreed as “problem-solving and an intelligent system as one which takes the best possible action in a given situation”.

AI makes it possible for machines to learn from new experiences, adjust outputs and perform human-like tasks.

Examples include tools that support radiologists in reviewing a large number of scans, or systems that allow for a better understanding of patients’ current and potential healthcare needs.

AI can focus on a narrow task within in a restricted set of parameters, such as reading radiology scans or optimising hospital workflows, or it can be used in a more general way where it learns to perform several tasks with consciousness.

Addressing the risks of AI

One of the biggest risks of adopting AI in healthcare is truth. If AI garners new information to decide on a new method of treatment, for instance, who now validates whether that information is a source of truth?

Speaking at the NHS Health and Care Innovation Expo 2018 in Manchester, managing director of AHSN for Kent and Surrey Sussex, Guy Boersma, said it will be important to have the right framework in place so that the easy aspects of AI can progress without obstacles while the more challenging aspects can receive the attention and regulation it needs.

“We acknowledge the challenges of building public trust and confidence in use of data… and it’ll take a long time to shift public attitudes,” Boersma said.

“We need the active participation of patients and other interested parties at all stages of the development process (of AI).

“We need to help unblock the real-life work that is currently going on. Then we need to reward and publicise it, so that people notice what AI is doing. And we need to raise awareness, generally, to break down fear, suspicion and lack of knowledge about AI.”

Looking to the future

Some of the best cases of AI-enabled outcomes always solve a problem.

Overall, respondents of the survey said game-changing areas of AI in the immediate future would be related to diagnostics, non-clinical aspects (e.g. administration and operational efficiencies), and health promotion and preventative health.

Within diagnostics, radiology is already seeing results in assisted reporting and screening, and it is predicted that 80 per cent of all dermatology diagnoses will be conducted using AI within three years because it will be better than dermatologists at diagnosing.

AI is viewed as extremely important for non-clinical aspects, such as saving time with administration. It is expected there will be a reduction in administrative staff overheads and machine learning will be increasingly used to process images and text.

Health promotion and preventative health was also cited by survey respondents as an extremely important aspect for the future of AI in healthcare, with the majority expecting AI to be used in a more predictive way.

This will allow for a shift from reactive patient care to a preventative care model in which patients will feel more empowered to take care of their own health.


TELL US: How do you think AI will impact healthcare in the future? Share your views in the comments section below.

Our site uses cookies to distinguish you from other users of our site. This helps us to provide you with a good experience when you browse our site and also allows us to improve our site. By continuing to browse the site, you are agreeing to our use of cookies. For detailed information on the cookies and how you can opt out of the cookies we use and the purposes for which we use them, see our Cookies Policy.

I agree