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

Evaluating AI-augmented decision support tool to assist referrals into secondary mental healthcare

Overall summary

Across England, more than 1.3 million referrals are made each month to secondary mental health services, a process that cost the NHS £326 million in 2019. Each referral must be assessed and triaged by a clinician, starting with a review of the patient’s referral documents and, often, their historical medical records. Reviewing often extensive electronic health records (EHR), collating all contacts with services and progress notes before making a referral can be time-consuming for hard-pressed clinicians and patients alike. Chronosig is an innovative AI-augmented support tool being developed by clinicians and scientists at the University of Oxford. Using natural language processing, relevant medical information can be extracted from patient EHR and clinical notes. The Chronosig support tool summarises the patient’s clinical and treatment history and makes recommendations to assist clinicians’ decision-making. The tool also provides summary information alerting clinicians to patients with more complex or difficult to treat disorders. The Oxford AHSN team conducted a barrier to adoption study, speaking to key stakeholders in relevant roles along the care pathway to determine the potential of the technology to improve triaging of patients in the secondary mental healthcare pathway, ensuring that patients quickly get onto the right treatment pathway.

What is the challenge?

In the UK, mental healthcare is stratified into primary, secondary and tertiary services. In 2019/20, more than 2.8 million people in England were in contact with secondary mental health services.

Delays and inefficiencies in referral, triage and assessment lead to more patients in crisis resorting to attending emergency departments (EDs), shifting the care provision from secondary mental healthcare to acute general hospitals. In 2013/14, 6.2 million ED attendances across England were for mental health problems, with 50% discharged home. 1  Timely and streamlined referral and triage could prevent delays and benefit patients, as well as reducing the burden on acute hospitals that are often used as a ‘last resort’ by patients unable to access mental healthcare.

What did we do?

The Oxford AHSN performed a barrier to adoption study to evaluate the need, perceived usefulness and level of acceptance of the AI-augmented support tool Chronosig for triaging patients referred to secondary mental health services, and to gain a better understanding of its capability and limitations in routine care settings. Secondary research was conducted to understand the current approaches used for triaging patients referred to secondary mental health services. Stakeholders were also interviewed to confirm the outputs of the secondary research and to gain their views on its potential benefits and barriers to adoption in their clinical practice.

What has been achieved?

Stakeholders agreed that Chronosig has the potential to help standardise the referral process so that all patients can access the same offer of care. This barrier to adoption study showed that Chronosig was seen as a potentially useful addition to the triage process, particularly for services such as single point of access. Providing a summary of the patient’s past psychiatric history was one of the main perceived benefits of the technology in a context where reviewing patients’ notes is time-consuming and secondary mental health services are under high pressure. The treatment history feature of Chronosig was also found to be useful, especially for informing future clinical decisions for difficult-to-treat patients. However, more evidence of its technical capability and clinical relevance is needed to demonstrate its usefulness in triaging patients to the right treatment team.

What people said

“Our project included a substantial piece of work on NHS scalability and how the proposed technology would integrate with clinical services and the standard operating processes involved in triaging people referred to secondary care mental health services. After speaking with the Oxford AHSN, they generously agreed to conduct a deep dive into the project’s work packages and devised a meticulous programme of research to identify the barriers to adoption. The resulting package of work was so in-depth and driven by the AHSN’s expertise that we added the AHSN as a formal partner on the project. The resulting report delivered a thorough evaluation of the project, its mesh with NHS practices and contained invaluable feedback from clinical stakeholders. Consequently, we were able to rapidly revise and pivot the project to deliver what clinicians want, rather than what we had expected.”

Dan Joyce, Honorary Senior Research Fellow, Chronosig

What next?

A real world health economic evaluation is required to determine the cost-effectiveness of the proposed introduction of Chronosig into routine clinical practice and to inform a business case for adoption and spread of the technology.

Contact

Florence Serres, Project Manager florence.serres@healthinnovationoxford.org