Ongoing CRIS projects: general psychiatry

Discriminating N-Methyl-D-Aspartate Receptor-antibody encephalitis (NMDAR-Ab-E) from primary psychosis

NMDAR-antibody encephalitis (NMDAR-Ab-E) is a condition in which the body mistakenly attacks a person’s NMDA receptors in the brain. These receptors are important in normal brain function including learning, memory, and normal conscious awareness. The antibodies reduce available receptors and so affect these functions. As the disease worsens breathing and heart function are affected meaning that some patients become critically unwell.

People with NMDAR-Ab-E typically present with psychiatric symptoms and the symptoms can overlap with common severe mental illnesses. This makes diagnosis challenging and NMDAR-Ab-E is commonly misdiagnosed as a primary psychosis, delaying life-saving neurological treatments. Establishing which clinical features differentiate NMDAR-Ab-E from primary psychoses will improve diagnosis and care   for both groups.

The mental state examination (MSE) is a standardised clinical interview which forms part of the assessment of psychiatric and neurological disorders. We have looked at MSE data from a cohort of NMDAR-Ab-E positive patients and coded their symptoms using a standardised coding scheme. We now wish to collect the same data for patients presenting with a first episode of primary psychosis. To do this, we will extract mental state data such as symptoms and specific behaviours from case records of a psychosis early intervention service, using expert annotators to derive the information from free-text and code it using the same coding tool . The tool is a simple list of the features being present or absent and the level of confidence for that impression. We will then compare the two groups to determine which clinical features best discriminate between these diagnoses.

CHRONOSIG P1

Aims: This research will build clinical-decision tools to support clinicians in making decisions and triaging referrals in NHS clinical practice. The tools will identify and improve access to care for people who have traditionally been excluded from, underrepresented, or poorly served by mental health services in the UK, ensuring they receive equitable care.

Background: People needing treatment for mental health problems are often first seen by their general practitioners, in hospital Emergency Departments or even in educational settings and staff then refer them to mental health services. This network of referral routes converges on community mental health teams who make decisions about suitability for treatment, how urgently to see people and which specialist teams need to be involved. In routine care, all referrals are assessed or ‘triaged’ by a multi-disciplinary team and often require the person is clinically reviewed over multiple contacts to eventually decide on treatment pathways. In 2019 this referral process cost the NHS £326 million. Electronic health records contain data from clinical encounters between patients and professionals, over multiple visits and this creates a longitudinal ‘fingerprint’ of the patient’s mental health over time. These clinical records are a rich source of information, which is particularly pertinent to mental health because mental illness is largely expressed as language and behaviour and the primary tool for the clinician is the interview, recorded as clinical notes and correspondence. However, this information is voluminous and time consuming to process (even by expert clinicians) when triaging referrals. Recent advances in artificial intelligence and natural language processing will help unlock this information to make triage more efficient and access to care more robust.

Work plan: Using a decade’s worth of observational data from the largest collection of the secondary care mental health electronic health records in the UK alongside state-of-the art neural network algorithms for natural language processing, we will develop, train and validate models capable of characterising a patient’s trajectory directly from clinical notes. The model’s effectiveness, safety and efficiency will be evaluated in simulated multidisciplinary triage teams at four NHS Foundation Trusts for pre-clinical use.

Patient and public involvement: Through the Patient and Public Involvement group of the Oxford Health Biomedical Research Centre, patients, carers and clinicians will be involved throughout with regular meetings for co-producing the project scope, milestones and evaluating the clinical efficacy, safety and ethical boundaries of the clinical decision support tools.  Deliverables: An intelligent operational automation tool, based on a longitudinal patients’ trajectory derived from historical observations, will help streamline the process of being referred to an NHS mental health team. It will also help identify people with the highest care needs and suitable for services and clinical trials.

Page last reviewed: 2 November, 2021