July 13, 2018
AI in healthcare: Because we don’t know what we don’t know
By: Niamh McKenna

Our world is changing …. we’re used to people retiring at 60-65 and (sadly) dying at seventy-something – but that isn’t the case anymore. However, as Professor Dame Nancy Rothwell has said “a third of people born in the UK this year can expect to live to 100.” The society we will live in will, very soon, be utterly different to the one we know. The question is, how do you manage the health of a population that includes millions of people over 100? (The Queen will have to industrialise her greeting card process, for starters!)

In healthcare, we often draw on the experience of the past to manage the present and address the needs of sick patients. This experience based model has served us well in the past, but I believe there are some fundamental shifts in society which will change our focus. Especially if our public healthcare is to remain viable under ever increasing strain on the purse strings.

To manage this ever-aging population, we’ll be forced to adapt on the fly. We really don’t know what society will be like, because we simply don’t have the data – there is little or no longitudinal data for the 100+ age-group. Two key factors may save our bacon. First – the plethora of sensors and wearable devices that can provide real time data on patient behaviours and key health indicators and, second, the advent of artificial intelligence that has the ability to interpret all this data on the fly, and provide both administrative/logistical and clinical diagnosis/decision support in real time.

And we seem to be ready for it. The results of the Accenture 2018 Consumer Survey on Digital Health indicate that tech has become more important to British consumers in the last two years, including mobile devices (up from 37 to 48%), social media (up from 20 to 28%) and wearables (up from 22 to 31%).

This increase in the number of devices means a wide range of data will have to be absorbed, analysed and converted into diagnosis and treatment support data. Pathology reports will need to be combined with data from wearable sensors, home-based pathology tests, diet and exercise data, and heart rate and blood pressure information. Based on this data, patients could be warned of necessary lifestyle changes to avoid the onset of chronic conditions, or prevent existing conditions getting worse.

Non-clinical data is also a crucial part of the picture. Socio-economic data can be woven into the picture, so that traits that might indicate, for example, that a patient can be more prone to a particular type of disease, or is unlikely to adhere to medical prescriptions, could receive additional prompts from carers, or the AI itself.

The challenges to society are often lifestyle-related. At least to some extent, obesity, asthma, diabetes and other chronic conditions can be prevented or better managed by means of lifestyle changes that could pre-empt health events. Prevention will be more effective than cure to deal with the ever-increasing pressures on the UK national health system which can, by leveraging AI, manage this more effectively.

For me, there’s no alternative – the volume of data to analyse, the changing nature of society means we need to serve the needs of the public better, faster and cheaper. If you have any comments or thoughts, please get in touch.

Popular Tags

    More blogs on this topic