Using AI to support triage and navigation in the NHS

What is the issue?

Getting access to the right healthcare, at the right place and the right time is not only beneficial to patients but also helps the NHS run more efficiently, which is particularly important in the context of funding cuts and workforce challenges. When a patient first contacts an NHS service for a medical concern, the information that they provide helps clinicians prioritise patients based on the urgency of their need and direct patients to the most appropriate point of care. This process is called triage.

Increasingly, digital tools are helping manage the triage process with patients completing online forms at their general practice or using NHS 111 online. The introduction of Artificial Intelligence (AI) into this process has already begun in certain areas of the NHS, but little is known about how this is being done, what works well or not so well, for whom and in what context. The technology is new and rapidly changing, so there is an urgent need for the generation of evidence to support evidence-based practice for national programmes to increase the likelihood of achieving positive outcomes for patients and NHS services, whilst minimising risk and inequality.

How are we helping?

DECIDE (Digitally Enabled Care in Diverse Environments) is a centre for rapid evaluation of technology-enabled remote monitoring in health and care, funded by the National Institute for Health and Care Research Health and Social Care Delivery Research (NIHR HSDR). DECIDE is a partnership between the University of Oxford and RAND Europe.

The DECIDE centre is conducting a rapid evaluation that focuses on AI-supported triage and navigation at the point of first contact with the NHS, particularly within primary care pathways and in an evolving technological and policy context. More specifically, the evaluation aims to:

1. Examine the current and potential role of AI in supporting triage and navigation in the NHS in the context of optimising AI and digitally-enabled triage pathways.
2. Provide practical and formative recommendations for policy and practice looking to design, embed, scale and sustain, evaluate and monitor AI-enabled triage pathways.

Our focus is two-fold: on the one hand considering the use of AI within current triage systems at first point of contact with the NHS (largely focusing on general practice and primary care), and on the other considering future design and implementation in light of rapidly evolving AI-enabled triage systems, the intersections with wider NHS data and platforms (e.g. relating to NHS 111, Single Patient Record) and triage via national portals like the NHS App.

Our key research questions are:

1. What is AI-enabled triage (i.e. what does it constitute), how do different stakeholders understand and envision AI-enabled triage, and why the growing interest in its use in the NHS?
2. What is the evolving landscape of technologies and suppliers like at present in the UK and what is it likely to look like in the near-term future?
3. How is AI-enabled triage being used in the NHS at present, how well is this working (including in relation to issues of equity and access), how is it being evaluated and monitored, and what influences effective implementation, use and impacts?
4. What would a vision for ‘good AI-enabled triage’ in primary care look and feel like across stakeholder groups (e.g. patients, providers/staff, suppliers, system leads/policy), and how can it be achieved in ways that meet the common and unique needs of different stakeholders, helps address and does not exacerbate inequalities (considering issues of candidacy and equity)?
5. How can AI-enabled triage in primary care be integrated with AI triage into the NHS App, in a way that mitigates and manages risks of inequity, safety concerns, or unintended consequences?
6. What type of an evaluation and monitoring approach could help inform decisions about adoption, implementation, continuation or discontinuation of AI-enabled triage and navigation services?