FST JOURNAL
Technologies in Healthcare
The debate
Many people have already embraced the use of apps and technology. Yet there are always more opportunities to harness data. In Scotland there is also a focus on mental health and using data-driven technology to support the individual.
Public Health Scotland is using a series of datasets to help them advise and convince policymakers, supporting their arguments with evidence. It is very important that this data is independently audited. That helps to build trust.
One of the aims is to aggregate different datasets about individual patients in a safe way so that they can still be linked. Precision medicine refers to using essentially all the available data about an individual patient and then developing a personalised care or intervention plan for them based on their characteristics. To do that involves bringing all the relevant data together from different sources. That is not just the healthcare data, but also social care data, the socio-economic grouping, even the postcode – all can have an effect on health.
Some applications of AI will be relatively easy to fit into everyday life in the kinds of systems already in use. But in others, developers have to work together with professionals on new areas and that will mean training people specifically to use these methods. Regulation is important and not just in fields like medical diagnostics. It will be just as necessary for AI.
Part of the power of conversational AI is that it provides a universal interface between the systems and the users.
The implementation of many of these new systems will not be provided by their academic developers but by commercial companies. Academics will be working with companies to get training packages together and to meet the medical legislation which requires that users are trained to use devices or technology.
The NHS is not monolithic: it is, in fact, a very large conglomeration of different organisations, each with its own culture, each of them want their own data. Government has failed to roll out new devices or technologies because there are no requirements on all these players to adopt them. The savings are there to be made but there is no coordination or ‘push’ from the top. At the moment, people can simply choose not to adopt something because it does not suit their existing routine.
The scientific community is in the process of learning what AI should – and should not – be able to do, yet there are already amazing results. A general model for image recognition, trained on pictures of cats and dogs on the internet, is actually very good at recognising anatomical structures in radiology imaging. It just needs a small amount of additive data that is specific to radiology, added to the vast data of cats and dogs on the internet. The key is to understand the capabilities in order to use it in a safe and appropriate way.
FURTHER INFORMATION
CivTech – www.civtech.scot
Digital Front Door - www.digihealthcare.scot/our-work/digital-front-door
Digital Health and Care Innovation Centre – www.dhi-scotland.com
Digital Lifelines Scotland – https://digitallifelines.scot
Doccla – www.doccla.com
Health and social care: data strategy - www.gov.scot/publications/data-strategy-health-social-care-2
NHS Scotland: Near Me - www.nearme.scot
RESQ+ – https://www.resqplus.eu
Spring – https://spring-h2020.eu
The Data Lab - https://thedatalab.com
The National Robotarium – https://thenationalrobotarium.com
FST PODCASTS
Professor Dame Anna Dominiczak - Innovation in healthcare and in the NHS