Early Intervention and Prevention

The UK is a global leader in the delivery of national, evidence-based early intervention in psychosis. Our goal is to extend the breadth of this impact to common mental health disorders, such as depression, bipolar disorder and anxiety disorders. By intervening earlier in the life-course we are taking a truly preventative approach with focus on potential causes including bullying, deprivation, substance misuse and childhood trauma.

The development of early intervention services for psychosis across the UK has been a uniquely effective NHS innovation, led by pioneering service developments in Birmingham and Australia in the 1990’s; aiming to improve clinical outcomes, reduce the societal burden of psychosis and the healthcare costs. The UK has a universal healthcare system, and Birmingham the largest 0-25 mental health service in the UK. Our comprehensive clinical infrastructure has contact with thousands of young people with early psychosis and developing mental health disorders every year and therefore provides a unique opportunity for such impact.

Please contact the Early Intervention and Prevention Research Theme Leads for further information:
IMH Research Theme Lead - Prof Rachel Upthegrove
MH Research Theme Co-Lead - Prof Matthew Broome

Our current research:

Psychosis Immune Mechanism Stratified Medicine Study

Key people: Professor Rachel Upthegrove

In collaboration with the University of Cambridge, we have recently been awarded a major grant from the Medical Research Council to investigate the link between increased brain inflammation and psychosis. Evidence suggests that inflammation may be present before and during the early stages in some, but not all young people with psychosis. In this multicentre Psychosis Immune Mechanism Stratified Medicine Study, Professor Upthegrove and Dr Khandaker will lead a team of investigators, including University of Birmingham MDS Professors Nicholas Barnes, and George Gkoutos, to examine how immune dysfunction could cause psychosis and use advanced AI techniques to identify who might benefit most from novel immune targeted treatments.

Learn more

Preventing depression following a diagnosis of first episode psychosis 

Key people: Professor Rachel Upthegrove

Antidepressant for the prevention of DEPression following first episode Psychosis trial (ADEPP)

Existing treatments for psychosis are only partially effective, with only 30% of young people recovering to previous levels of role and work functioning. First episode of psychosis (FEP), which includes illnesses such as schizophrenia,
typically begins in the late teens or early 20s and is diagnosed by the presence of hallucinations (such as hearing voices) and delusions (unusual beliefs held with conviction). Medication and cognitive behavioural therapy help to treat
these symptoms, but young people struggle to return to previous social and work roles, have suicidal thoughts and are at high risk of relapse. 

We want to know whether antidepressant medication can reduce the risk of depression happening at all after FEP, and whether by preventing depression we can improve recovery and reduce relapse.    

Learn more about the study

Midlands Engine: Mental Health Pilot

Key people: Professor Steven Marwaha

Studies show that at any one time, a sixth of the population in England aged 16 to 64 has a mental health problem, which costs employers between £33 billion and £42 billion a year in lost productivity. Recognising the huge impact mental health issues have on employees’ wellbeing and employers’ productivity. Our Professor Steven Marwaha is part of a consortium of researchers working as part of the Midlands Engine Mental Health and Productivity Pilot to break down the barriers facing people with mental health problems, and facilitate their return to work.

Learn more

Using a machine learning approach to develop and validate a prediction model for the onset of hypomania

Key people: Professor Steven Marwaha and Dr Pavan Mallikarjun

Bipolar disorder (BD) is a debilitating mental health condition, characterised by severe shifts in mood, that can range from disabling highs (i.e., mania/hypomania) to extreme lows (i.e., depression). Approximately 1-2% of the population are affected by bipolar, with most people experiencing the onset of mood symptoms prior to their 20s. Despite this, little is known about the predictors to bipolar disorder and hypomania symptoms, particularly among young people. Intervening early in the development of bipolar is a top clinical priority, and one that may have the potential to limit its functional and symptomatic impact on those affected. Thus, predicting the onset of bipolar/hypomania prior to its onset, may help clinicians/researchers to develop novel, tailored preventative strategies and interventions for young people.

We aim to develop and validate a clinically useable prediction model to predict the risk of onset of hypomania in young people aged 21/22. We will use machine learning approaches to develop prediction models to predict the onset of hypomania in young people aged 21/22.

Anxiety and mood disorders in young people: a multivariate approach using the ALSPAC cohort

Key People: Professor Steven Marwaha

In this project, we will investigate the early life psychological, environmental and biological factors that may increase the risk of developing mood disorders (broadly defined) in late adolescence and early adulthood and the associated poor outcomes. Multi-factorial analyses will examine prospective impacts of diet, sleep, parental mental health, cognitive function, inflammatory markers, metabolomics and genetic susceptibility on the earliest symptoms of mood disorder.The aims of this project are:1. To investigate the longitudinal relationship between early explanatory factors and the development of affective symptoms, mood and anxiety disorders2. Develop and test predictive models of mood and anxiety disordersDescriptive statistics will be used to summarise data. A combination of univariate and multivariate modelling and advanced AI techniques will be used; for example, risk relationships will be investigated using general linear modelling (GLM), path analysis and supervised machine learning to predict depression, anxiety and recovery outcomes accounting for important confounding variables.

Adult Psychiatric Morbidity Survey (APMS) 2014: Mood Disorders and Personality Disorders

Key People: Professor Steven Marwaha

The purpose of the study is to complete research on mood disorders and personality disorders, in order to ultimatelyimprove the health care that this group receive. The aims of the research are to analyse the APMS data to betterunderstand the causes of mental disorders, health services access, inequalities and links to other disorders.


Key People: Professor Steven Marwaha

Background: Attention-Deficit/Hyperactivity Disorder (ADHD) is a common neurodevelopmental disorder involving inattention, hyperactivity and impulsivity that starts in childhood and frequently persists into adulthood. Stimulant and non-stimulant medication are the mainstay of treatment in adults. ADHD in adults is commonly comorbid with bipolar disorder (bipolar) and psychosis. There is substantial uncertainty over the effectiveness of stimulant and non-stimulant medication in adults with ADHD and comorbid psychosis or bipolar. There is also concern that they could trigger psychotic or manic symptoms. A randomised controlled trial (RCT) is needed to address this evidence gap.

Aims and Objectives: The aim of the study is to evaluate the clinical and cost-effectiveness of stimulant compared with non-stimulant medication for adults with ADHD and a history of either psychosis or bipolar. Primary objective: To evaluate separately in adults with 1) ADHD and a history of psychosis and 2) ADHD and a history of bipolar, whether stimulant vs. non-stimulant medication reduces ADHD symptom severity at 12 months. Secondary objectives: To evaluate separately in adults with 1) ADHD and a history of psychosis and 2) ADHD and a history of bipolar the impact of stimulant vs. non-stimulant medication: a] on ADHD symptom severity at 6 months; b] on the emergence of symptoms of bipolar or psychosis over 12 months; c] on health-related quality of life, occupational, functional, substance misuse and other outcomes at 6 and 12 months; d] on cost-effectiveness

Methods: We will complete a pragmatic, observer-blind, multi-centre, stratified, 2-arm, parallel group, RCT comparing Lisdexamfetamine (stimulant) with Atomoxetine (non-stimulant). Strata are defined according to a history of: (1) bipolar or (2) psychosis. 648 participants will be recruited in total, meaning there will be 324 participants with ADHD and bipolar, and 324 with ADHD and psychosis. We include an internal pilot with clear progression criteria that will be independently assessed for each stratum. Participants will be recruited from mental health Trusts / Health Boards in the Midlands, London and South East, the North East and Scotland. A PPI co-applicant and lived experience advisory panel will aid the study throughout. The analysis will be on an intention to treat basis and use ADHD symptoms measured at 12 months post-randomisation as the primary outcome. The primary analysis of all outcomes will be based on the separate treatment effects for each stratum. A health economic trial-based analysis of the cost-effectiveness of Lisdexamfetamine vs. Atomoxetine medication will be undertaken for each stratum. The trial will take 58 months to complete.

Impact: We will provide definitive evidence as to the clinical and cost-effectiveness of Lisdexamfetamine vs. Atomoxetine in the treatment of adults with ADHD and a history of psychosis or bipolar, as well as the risks of adverse events associated with each medication. The results will provide clear guidance to NICE, clinicians and service users. Given that untreated ADHD is associated with poor clinical outcomes, unemployment and criminal justice system involvement, clear effectiveness evidence in this area is likely to improve recovery for individuals with ADHD and a history of psychosis or bipolar, reduce costs for the individual, the NHS and society

PARTNERS2: collaborative care for people with severe mental illness

Key people: Professor Steven Marwaha

PARTNERS2 is a five year NIHR programme grant awarded to Birmingham and Solihull Mental Health Foundation Trust in conjunction with the Dept of Primary Care at the University of Birmingham. The study aims to help primary care and community based mental health services work more closely together in order to develop an evidenced based model of care to support individuals with severe mental illness in primary care with secondary care support.

Discover the PARTNERS2 study

Recent projects:

Testing a model of altered memory processes in psychosis and autism-spectrum disorder

Key people: Dr Kareen Heinze

This study looks into the mechanisms and shared neurobiology of memory impairments in young people with a diagnosis or autism and/or psychosis.

Long-term outcomes of children with borderline personality disorder traits at 11-12 years

Key people: Professor Steven Marwaha

The aim of this study was to explore to what extent bipolar disorder (BPD) symptoms at 11-12 years of age could predict negative mental health outcomes in the future and investigate which factors might increase this risk.

Learn more about the study

Our PhD research:


Neuroimaging & Metabolic Biomarker Predictors of Recovery from Depression and Psychosis

Depression is the leading cause of disability world-wide with 300 million people being affected. Adolescents and young people are particularly vulnerable to developing this illness. In the UK, the one-year prevalence of depression in adolescents is 5%. While around 60% of young people with depression fully recover, a large proportion will have ongoing difficulties. Depression is also the most common co-morbidity seen with other mental disorders such as psychosis. Complex psychopathology presents clinicians with the challenge of correctly identifying co-morbidities, avoiding misdiagnoses, and tailoring therapeutic options for the individual. 

Currently, clinicians treat depression with co-morbid psychosis the same way they treat major depressive disorder on the presumption that they have the same neural basis. Data from the PRONIA study, an EUFP7 funded 8 centre study recruiting recent onset depression and recent onset psychosis participants will be used to identify common and distinct neuroimaging, clinical, and metabolic features of depression across diagnostic groups. Features that provide significant information will be used to build predictive machine learning models that predict recovery from depression with co-morbid psychosis as well as diagnostic machine learning models.

PhD student: Paris Lalousis

Supervisors: Professor Rachel UpthegroveProfessor Stephen WoodProfessor Nikolaos KoutsoulerisDr. Lianne SchmaalDr. Renate Reniers

Developing and validating outcome prediction models for people with First Episode Psychosis using Supervised Machine Learning (Precision Medicine)

Psychosis is an illness affecting mainly young people with an incidence rate of 31 per 100,000 people. Evidence from meta-analysis suggests that more than half of young people with First Episode Psychosis have poor outcomes. Despite the heterogeneity in remission and recovery outcomes, there are currently no validated tools available for early identification of these patients. Identifying individual patients with poor outcomes at initial clinical contact would help in stratifying interventions, which may lead to improved outcomes and utilisation of clinical resources.

Previous studies have identified predictors of outcome at group level including socio-demographic variables, clinical and treatment response variables, comorbidity, functional and cognitive deficits with inconsistent reliability. There is a lack of clarity on how or which of these group level factors should be combined for prediction at the individual level, or as to the performance of prediction models in clinical settings that may be very different to the primary research study site. In order to maximise clinical utility, an approach is needed that can be applied to stratify the risk of poor outcome in an individual patient at the initial clinical contact. One solution to this is the use of machine learning, where algorithms can sift through a large array of predictor variables, looking for combinations that can reliably predict individual patient outcomes.

This research will involve empirically testing clinical prediction models in clinical settings to validate their utility in stratifying interventions based on risk prediction of outcomes and whether, by such stratification, outcomes can be improved.

PhD student: Rebecca Lee

Supervisors: Dr Pavan MallikarjunProfessor Georgios GkoutosDr Sarah-Jane FentonProfessor Stephen Wood