Q Exchange
Using AI to manage high-intensity users for the ambulance service
- Proposal
- 2024
Meet the team
Also:
- Jack Watson
- Naomi Birchall
- Angela McNally
- Matt Dugdale
What is the challenge your project is going to address and how does it connect to the theme of 'How can we improve across system boundaries?
The North West Ambulance Service (NWAS) face challenges in dealing with over 1.7 million 999 calls each year; amidst the ever-growing, wider pressures on health services in the UK.
One specific challenge is identifying and managing high-intensity users (HIU). HIU are defined as an individual who calls 999 five or more times in a calendar month, related to single episodes of care. We have hundreds of patients meeting this criterion, yet only a small number of practitioners focused on reducing this demand.
These patients face similar challenges across the healthcare system, however, we continue to work in silos and take a reactive approach.
We aim to enhance collaboration across system borders, extending beyond the healthcare sector to include other agency partners. We will do this by bringing data together, using AI to identify trends and make recommendations then being able to share high-level overviews with system partner.
What does your project aim to achieve?
Our ambitious use of AI would allow us to proactively manage HIU’s at an early stage to reduce long-term demand on 999 services and work with partners for the benefit of our patients. This project has the following objectives:
- Early identification of high-intensity users – AI can make suggestions using ‘Directory of Services’ e.g. Social Prescribing.
- Bringing relevant data together– social factors, health conditions, demographics, time/date etc. AI will allow us to spot trends which would take hours to manually locate, saving time and creating efficiencies.
- Working collaboratively with system partners – overcome silo working and be ‘on the same page’ with a digital platform.
- Reducing health inequalities – using demographic data to identify trends and navigate the complex mix of poor health outcomes, repeat 999 use, and social factors.
Through enhanced identification we can proactively manage HIU’s earlier, preventing long-term demand on 999 services and the wider healthcare system. A case study example is below:
How will the project be delivered?
The project will be delivered with a cross-working group, within our organisation; involving Digital and Innovation Team alongside our specialist HIU team.
We intend on delivering a proof-of-concept solution focused on two organisations within one ICB, with significant user testing and comparison to the current and historic ways of working.
We recognise that service users who are determined to be ‘high intensity users’ are often facing significant challenges which has a demand upon primary care, social care, police, mental health and other agencies. We plan to engage with these partners from the start, engaging with them to provide an AI solution that offers seamless working across system boundaries – not only improving productivity, but also patient experience.
Evaluation would be a key component of the project delivery, with ongoing evaluation and deciding key metrics; this would allow us to measure the success, both from a combined productivity and patient benefit perspective.
How is your project going to share learning?
We believe this project could be shared through regional and national ambulance QI and innovation forums, and the Frequent Caller National Network (FreCaNN) ultimately for other ambulance services to adopt and develop.
Whilst the geography of each ambulance service will differ, we all face the same challenges in regard to high-intensity users and need to consider new ways of working. Planning, testing and delivering an AI solution for this challenge would build a new evidence base and encourage rapid acceleration on a national scale if the test of concept was successful.
We will share learning with wider Q network as well as specific national groups in the emergency services sector such as the Association of Ambulance Chief Executives (AACE) and the Northern Ambulance Alliance (NAA).
How you can contribute
- Providing experience of how high-intensity users in other parts of the health system are managed
- Sharing any examples of using AI to proactively monitor patient trends and then deliver valuable reports
Plan timeline
| 1 Aug 2024 | Engagement events with system partners |
|---|---|
| 1 Oct 2024 | Finalise project scope and user requirements |
| 1 Dec 2024 | UX wireframing |
| 1 Feb 2025 | User testing and revisions |
| 1 May 2025 | Platform development and AI model planning |
| 1 Jun 2025 | User acceptance and quality testing |
| 1 Jul 2025 | Pilot launch and interim evaluation |
| 1 Jul 2026 | 12-month project evaluation and sharing learning |
Comments
Graham Dodd 20 Mar 2024
Hi,
I think the use of AI would be of great benefit and process the data we have efficiently. It could save the NHS a number of man hours and allow those hours to be spent providing care to the users aimed within this project.
Carl Rees 20 Mar 2024
Excellent idea that would make the system much more efficient and enable the ambulance service to focus on those with greatest clinical need. Best of luck with the bid.
Jonathan David Hammond-Williams 19 Mar 2024
According to Lehrer, Eseryel and Rieder. et al. (2021), people with medical conditions are more motivated to use tech to treat problems through changing behaviours.
This article talks about wearable tech and what motivates people to change behaviours.
Artificial Intelligence could help to recognise behaviours/patterns and then adapt how technology supports a person, depending on the self-leadership strategies of the person.
Lehrer, C., Eseryel, U.Y., Rieder, A. et al. (2021). Behaviour change through wearables: the interplay between self-leadership and IT-based leadership. Electron Markets, 31, 747–764. https://doi.org/10.1007/s12525-021-00474-3Forum
Jonathan David Hammond-Williams 19 Mar 2024
Absolutely brilliant, team.
I am in the process of designing an infographic and writing a critical appraisal about this very thing for my MSc!
There is plenty of evidence about the use of AI to better personalise care (anything from 3D printing of drugs that are better at interacting with exiting prescriptions and ensuring a better dose, rather than splitting tablets, to integrating systems [e.g. SCR + presentations given to chatbot). The use of Deep Learning (DL) and AI algorithms to accurately and safely predict behaviours to better identify call behaviours is a real opportunity for this area and, with appropriate clinical review, we could really personalise care and empower the individuals to self-manage the majority of their conditions. However, it is important to ensure there is clinical/medical oversight, as AI cannot and should not replace human interaction (Altamimi, Altamimi, and Alhumimidi, 2023). With the development of the Care Act and the recognition of complex presentations and social needs, HIU work is often falling under the safeguarding/welfare banner and to truly make safeguarding personal, we need to involve the person in their care and not only leave it to AI.
Currently, in SWAS, we use an algorithm to auto-identify any person that meets the definition of a person that calls ambulance services frequently. Once that is done, we look to hold MDTs/MAMs to agree on a multi-agency plan to ensure our approaches are consistent and aligned with one another. However, the difficulty here is that it does not include input from the person who the plan is about - meaning it is not as personalised as we'd like it to be.
I think AI presents a wonderful opportunity to join what is reported via the call/chatbot/other with the existing data and then provide advice about what steps the person can take next to manage their own condition and also note when its not a 'normal' presentation.
The challenge will be to ensure AI can get the right data (Qian, Dong, Yuan et al., 2021) through appropriate integration of systems (e.g. NCRS/SCR + CAD).
Eileen Dudley 18 Mar 2024
Interesting idea on a challenging and sensitive subject. best of luck with developing the project
Adele Markland 17 Mar 2024
This is a really innovative idea . I think there is an opportunity to work in partnership with other health and care organisations and patients to understand why they are accessing 999/111 services. Quite often it is because services are very fragmented and not accessible with the changing needs of our patients. When I worked at place level in a Greater Manchester locality a lot of work was focused on the public health agenda to support patients e.g. if a patients home is cold because they can't heat it, this can cause respiratory issues and healthcare services are accessed for the health issue but the root cause of a cold home is not addressed. I think this learning is transferable to other services and could also be shared at the AmbulanceQ Network. Good luck team.
Stacey Hart 13 Mar 2024
A fantastic idea for a big issue to intelligently use the data . It'll be great to hear how it's going and see how many patients can be helped to get the right care in a preventative/proactive way and reduce some pressure on the system. As Michelle also mentioned and you have commented, it would be great to understand their 'why' - is it access to services? Is it something that can be worked on with system partners? We can use this to educate the public too.
Jack Watson 17 Mar 2024
Thanks Stacey, agreed!
It's about looking at the problem with both an 'individual stories' and 'system wide' view, side by side... With the support of technology, understanding the 'why' for each individual service user will then drive huge improvements and increased efficiency in dealing with the large-scale system problem .
Not only will this project support local/North West based system partners but could be scaled across the UK - acknowledging that all UK ambulance services have the same challenge in supporting this cohort of patients who ring 999 repeatedly.
Jane Clayton 12 Mar 2024
This is a great idea to use digital technology to work intelligently and collaboratively with a number of different organisations across health and social care. Proactive support and management of high intensity users can only be of benefit to the service users themselves, as well as to those working in this sphere.
Jack Watson 17 Mar 2024
Thanks Jane, the proactive aspect is so important... by identifying and supporting high-intensity users as early as possible, we can then work with system partners to find suitable resolutions.
Lucy Birtwell 5 Mar 2024
This is a really interesting idea, and I can definitely see the benefits to the trust! With other private companies and other sectors of healthcare beginning to utilise the full capability of AI, I think this is definitely something we should be looking into too.
Jack Watson 6 Mar 2024
Thanks Lucy,
No doubt in the future, AI will have a significant role in clinical triage/contact centre environments but there is considerable clinical risks associated with such...
On the other hand, we hope our project idea shows a use case for AI which, whilst still ambitious, has a specific focus. It will help support our clinical teams/system partners to gain situational awareness - ultimately, improving our care of individual high-intensity users and understanding their 'why?' for frequent 999 use.
Michelle Beattie 5 Mar 2024
Brilliant idea for a challenging problem. Great to proactively identify those frequent users. I wonder if it would be possible to also capture the stories of those frequent users to explore their why? Not sure if this is something that can be done using AI. Im guessing some reasons will be challenges in relation to accessing other services but there would likely be some things that could be addressed.
Jack Watson 5 Mar 2024
Thanks Michelle, definitely agree that understanding the ‘why’ is crucial.
Of course this can be done at an individual level by clinicians face-to-face, however our ambitious project plans to discover this at a system level, using AI to our advantage and analysis of vast amounts of data dynamically and proactively.
Whether it be a medical need, social need or usually combination of both, we see that AI can deliver crucial insights into the ‘why’ for each individual service user whom are high-intensity users… for example, detecting trends such as 999 calls always being made between 22:00 -00:00 on specific days which could ultimately indicate an increased care need for that individual.
That’s just using basic call data so by bringing in other data sources which we don’t currently utilise, we see the possibilities are endless and ultimately aim to achieve system-wide benefits!
Naomi Birchall 4 Mar 2024
I believe this innovative project has the potential to transform how ambulance services and the wider health and social care system work together to recognise, support and manage high intensity users, and to enable services to be much more proactive than reactive in identifying patients early on who require intervention to avoid complexity and reduce the demand and pressure on services.
Jack Watson 27 Feb 2024
We want to use AI in a trailblazing way, to benefit all involved across the wider healthcare system. Keen to hear your thoughts on our idea...
Comments are now closed for this post.