Skip to content

Q Exchange

Utilising AI to reduce Ambulance Conveyance

Take data from the Ambulance Data Set and use AI to identify incidents where an Ambulance has conveyed to A&E where a different outcome may have been possible.

Read comments 2
  • Proposal
  • 2024

Meet the team

Also:

  • Graham Norton - Northern Ambulance Alliance Managing Director
  • Andy Moody - Northern Ambulance Alliance Digital Lead

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?​

This project will contribute to addressing the risks to patients when ambulances are queued at EDs.  Delays in handover result in the lack of availability of ambulances and therefore longer response times. Recent changes to the ambulance data set (ADS) will allow the ambulance service to receive data that includes the outcome for a patient that they have previously handed over. The patient journey for 999 call via despatch & assessment on scene and is shown on the attached diagram. This will allow   ambulance service is able to analyse the outcomes asses if their decision to convey was appropriate. This pilot project will use AI to analyse thousands of patients journeys in order to identify any trends that suggest different despatch convey decisions can be made and therefore reduce the pressures on EDs

What does your project aim to achieve?

Aim is to use AI data analysis to help reduce unwarranted conveyance to A&E  and   to free up ambulance capacity, reduce waiting times  and improve patient outcomes  by using AI to analyse  data inputs throughout a patient journey.

How will the project be delivered?

Aim would be to extract patient information from individual patient data from Computer Aided Despatch, Electronic Patient Register and Ambulance Data Set (inclusive of Hospital Patient Record Data) and then use machine learning to identify where a&e resulted in no further action or onward referral which would indicate a conveyance may not have been necessary and determine any data initially captured by the Ambulance Service that could indicate a potential change to triage procedures resulting in less use of vital resources.
This would include involvement from Operational, Clinical and Data leads across the NAA’s boundaries.

How is your project going to share learning?

The Northern Ambulance Alliance (NAA) aims to create a collaborative approach of working together to improve health outcomes for patients and deliver greater benefits for the populations it collectively serves.
One of the main principles of the NAA is to improve quality and service delivery for all patients across the North of England by maximising the standardisation and shared learning outcomes across the North of England and the wider NHS.

How you can contribute

  • Support from the QI network to support access to Ambulance Data Set information for an AI machine learning capability.

Plan timeline

3 Jul 2024 Award Announced
1 Aug 2024 Establish project board
1 Sep 2024 DPIA endorsement
1 Sep 2024 Recruiting activity
1 Oct 2024 Appoint BA lead for the work
1 Dec 2024 Initial report for consideration
1 Jan 2025 Final report completion
1 Apr 2025 Commence roll out to NAA Trusts

Comments

  1. Great idea, and I'll be keen to learn from what the journey of discovery. I know the evidence base in this area is starting to grown - worth considering how this work can be published and build on evidence.

  2. This sounds like an interesting project, there is certainly something to be said for shared learning and knowing patient outcomes for those on the frontline.

    I'm curious, could you give examples for patients whom go to A&E but resulted in "no further action or onward referral" and how you identify who didn't require A&E? Would this include those who had investigations which they did require but results didn't indicate any necessary treatment?

    Have you considered liaising with other services with regards to alternative destinations for ambulances to convey patients to, such as mental health patients directly to mental health units for assessment?

Comments are now closed for this post.