Scroll for more
blog Nov 5, 2016

In five years' time… Key trends in AI healthcare

Author: Jackie Hunter

10 years ago I attended a symposium looking at the future of pharmacology and healthcare.

A senior American academic gave a presentation on what healthcare would look like in 2050.  Nearly everything he talked about - remote consultations, 24-hour monitoring through wearable devices, whole genome sequencing at an individual level is happening 34 years earlier than the academic predicted.  It is happening now.

This set me thinking about the impact that AI and machine learning is currently having on healthcare, the speed the technology is being applied at and what realistically could happen in just the next five years – forget about 10 or 20!  The more I considered the implications, the more transformational the impact seemed.  

Here are the current trends I see dominating AI healthcare over the next five years: 

Intangibles will be equally valuable

As the CEO of a company using AI to change the efficiency of the drug discovery process the early impact has been clear - not only in the tangible outputs of patents and compounds but importantly the intangible outputs such as the way AI can change how a scientist approaches a problem.  When you can access millions of data points and relationships, you can begin to ask much bigger and deeper questions.

AI in a hospital setting will develop strongly

Large organisations such as IBM-Watson and Google-DeepMind are focusing on healthcare in a hospital setting.  IBM has teamed up with Memorial Sloan Kettering Hospital to offer Watson oncology for improving cancer diagnostics whilst DeepMind Heath has partnered with Moorfields Eye Hospital to use AI to look at eye scans to better detect macular degeneration and diabetic retinopathy. The automatic analysis of pathology samples has been an area where machine learning has already demonstrated an impact and this is also translating to other imaging technologies such as CT scans and MRI. Of course, how the physicians adopt and use this technology will be critical.

Mobile health and AI will see exponential growth  

Healthcare delivery in developing countries has already been revolutionised by mobile phone technology – for example Vodafone is partnering with healthcare workers in South Africa to reduce time spent in administrative activities, and in Mozambique to improve delivery of vaccines to remote areas.

The combination of AI and mobile technology could not only have amazing implications for health and wellness in the general population in the developing world but could also be transformational for developed countries. For example, Butterfly Network’s stated aim is to reduce all the components of an ultrasound machine onto a small chip which could be incorporated into a mobile phone.  Then deep learning algorithms would be used to enable the images to be analysed without the need for an experienced ultrasound physician.  I expect to see many more, new AI enabled technologies linked to mobile devices.

Understanding the human to machine interaction will be critical

Whilst medical applications are the largest class of mobile device app developed, their use by patients to shape behaviours and habits to better enable human health has not met with overwhelming success.  Human factors and motivations are important barriers or enablers – studies of human-robot interactions are beginning to unravel some of the key attributes that machine interfaces need to have to maximise their interactions and impact with humans. These will be important in determining the acceptability of remote consultations with a nurse or GP ‘chat-bot’ or with robotic assistants to enable the elderly to live more independent lives.  Of course, companies like Babylon Health are already allowing more rapid access to GP and specialist services via video or audio or use the AI powered app to ask for advice about symptoms.

AI will play a major role in the management of personalised medicine

Personalised medicine will be enabled by AI – soon everyone will know their own genetic code.  Using this, and other information, to allow individuals to receive the best medicine for them, will mean that AI based systems will have to be employed at all levels of the health system in order to best integrate the huge amounts of data available.

Given the positive impact of AI is already being felt, AI is likely to have a tremendous effect moving forward -  new and better medicines, better health monitoring and healthcare delivery which in the end will lead to better health outcomes for you and I.

It is going to be an interesting five years!