This site has been optimized to work with modern browsers and does not fully support your version of Internet Explorer.

Augmenting clinical decision making with Chronosig

One of the latest innovations to emerge from Oxford’s ideas factory, Chronosig is a digital triage for secondary mental healthcare that uses machine learning to deliver clinical decision-making tools that work alongside the clinician.

The system is designed to improve the speed and efficiency of triaging patient referrals by assisting with reviews of patient notes and making recommendations for treatment pathways.

Overburdened waiting lists and the repeated cycle of referral

Across England, there are around 400,000 referrals each month to secondary mental healthcare services, with community teams providing 96 per cent of assessments and treatment. People are often referred from a number of routes including their GP, local general hospitals and social care services.

Almost all contact between patients, NHS mental health services and referring professionals are captured as written narratives, with the patient interview used by clinicians as the primary means to understand their patient’s situation. These detailed clinical notes and correspondence are recorded in the patient’s electronic health record, resulting in voluminous, ‘free-text’ documents that can be time-consuming to review.

Patient referrals converge on multidisciplinary community mental health teams, who come together to review, or triage, a person’s referral documentation and clinical notes before deciding on appropriate treatments, how urgently they need to be seen, and which specialist teams need to be involved in their care.

The way NHS services are organised means that individual cases can be reviewed by multiple community-based teams before a treatment pathway is decided, locking patients into a repeated cycle of referral bouncing that creates a “hidden waiting list” for mental health services.

This lengthy process of triaging referrals places a significant burden on mental health teams, and in 2019, referrals and assessments cost the NHS £326 million.

An intelligent automation tool to alleviate workforce pressure

A streamlined, rapid review of a person’s electronic health record could provide a long-term solution to overburdened waiting lists, and recent advances in artificial intelligence and natural language processing can help.

Currently in development, Chronosig is an early-stage digital triage tool that sits alongside the clinician. It applies novel, natural language processing technology to distil the electronic health record into a fingerprint of a person’s difficulties, symptoms and needs over time.

Chronosig uses this fingerprint to suggest a treatment pathway, so community mental health teams only need to review the tool’s recommendation and accompanying justification, rather than the full electronic health record.

This operational tool is being developed in Oxford by a multidisciplinary team, which includes patient representatives, and was brought together by Dr Andrey Kormilitzin, Dr Dan Joyce and Professor Andrea Cipriani from Oxford Health NHS Foundation Trust, the Oxford Health NIHR Biomedical Research Centre and the University of Oxford’s Department of Psychiatry.

Partnerships for data access and scale-up

Chronosig’s reliability hinges on the viability of the data fed into it. Its underlying algorithms are being trained and tested using vast quantities of anonymized, historical clinical data from across a range of NHS Trusts, acting as a sample of the UK population and different mental health conditions.

The project is made possible through an NIHR Artificial Intelligence and Health and Care Award, which has enabled new clinical research partnerships with four NHS Trusts, including Oxford Health NHS Foundation Trust and other NHS Trusts in the UK Clinical Record Interactive Search Network.

Scaling Chronosig from a laboratory model into a validated, working digital triage platform requires the backing of commercial partners along with an economic analysis of its cost-effectiveness.

As a first step towards realising their goal of making the outputs of the Chronosig project available to the NHS in the future, Dr Joyce and his colleagues approached Oxford Academic Health Partners to open up new opportunities for commercial development and to help navigate Oxford’s network of support for new innovations.

“Oxford creates a fertile ground to deliver this type of project, with streamlined access to patient data, expertise and support for scale-up so innovations can be adopted and spread across the wider NHS. The NIHR require a well-conceived commercial strategy for new innovations, so we were able to reach out to the OAHP for support, who brought the Oxford Academic Health Science Network’s/Health Innovation Oxford and Thames Valley’s commercialisation team onboard. We are now working together on the strategy that will help us to transform our academic project into a widely-adopted, cost-effective technology for the NHS.”

Dr Dan Joyce

The Chronosig team are further evaluating the quality of Chronsig through a series of simulations with triage teams and patient and public representatives from the NIHR Oxford Health Biomedical Research Centre.

How this project will benefit the NHS

Across UK secondary mental health services, which currently face referral bottlenecks due to significant shortfalls in workforce and resources, improving the efficiency, equity, transparency and fairness of triage with AI-powered tools like Chronosig could be transformative.

Streamlined access to academic and clinical partners during these phases of Chronosig’s development is a pivotal step to gaining the backing and funding required to scale-up and spread this automated digital triage tool across multiple NHS settings.