Health is a large and complex area. Unfortunately, we were not able to schedule interviews with representatives from health organisations during the discovery period in which we conducted our primary research. We have relied on secondary research in this space as a result.
Applications in healthcare – in Wales and elsewhere – were outlined as follows:
AI has been used to predict gene expression, in the development of predictive risk management tools to support proactive care, in virtual assistants for patient care, such as RITA at Velindre Cancer Centre, and in the AI Pathologist at Betsi Cadwaladr University Health Board.
We also spoke with the Digital and Transformation Health and Social Services Group from Welsh Government.
They were not aware of any centrally funded efforts to make use of RPA in health, but that there may be pockets of activity in various health boards. They felt it was interesting that there are more operational instances of AI that they could point at in health than there are of RPA.
They also explained some of the existing groups set up around AI and how they are already working towards sets of standards and guidance for health.
Both these conversations identified further AI applications in Wales – either operational or in trials – including:
A stroke triage tool being trialled in Cardiff and Vale. It helps identify those potential stroke patients that should bypass the emergency department and proceed directly to CT scanning, getting them on the correct treatment path more quickly.
A wound care scanning app in use in Swansea Bay University Health Board, which scans and measures wounds so staff can monitor them virtually.
The Life Sciences Hub Wales has trialled a pain assessment app in Gwent care homes. It scans facial movements to assess someone’s pain level, allowing caregivers to carry out evidence-based pain assessments for non-verbal residents.
Overall, it is felt that health is further advanced than other sectors. Factors contributing to this may include:
- They have had their own community of practice running for roughly a year.
- Potential applications of AI to health are more heavily researched, with multiple commercial organisations in the market.
- They now have more operational deployments of AI, and feel more confident in identifying other areas the technology can be used in.
Both Public Health Wales and Welsh Government wanted to see better sharing and collaboration between health and other sectors.
Support requested
We asked respondents what support they would value from CDPS and the Chief Digital Officer for Local Government.
Answers were varied but congregated around a common theme of bringing organisations together to foster better collaboration and knowledge sharing.
For example, answers included:
- Understanding what’s possible, what works well, what to avoid and what others are doing with these technologies. Most organisations are seeking lessons learned, relevant case studies and examples of good practice.
- Directly sharing use cases, approaches, technologies, and processes between organisations trying to do the same tasks. Multiple local authorities expressed frustration at Wales solving problems 22 times.
“Encourage us to work together, collaborate, share ideas, and help us to go out and experience technologies that work well outside of our sector.”
“What could work really well is, if a particular programme was picked up around AI and we got all 22 to agree and collaborate on something new, something that we've all never done before... This is such new technology; I think there is an opportunity to do it once and do it for everyone.”
Multiple organisations also mentioned they would value support with:
- building the business case for AI and RPA
- procurement – some mentioned the possibility of negotiating from a stronger position by coming together
"We find it very hard selling concepts internally. I'd find it easier selling solutions that are tried and tested and proven. To do that we need figures to back it up. So, this costs 20 grand. It will save you X, Y and Z in these ways."
“Even getting to the point we have with Microsoft has been challenging. If others are struggling too, we can help each other. Influencing the likes of Microsoft together as a group of councils. We can't do it individually - CDPS could help facilitate that.”
Several organisations pointed to formats already used by CDPS services – such as webinars and training – and suggested they could be usefully offered on the topics of automation and AI.