“When I started working in psychiatry, the last few large UK asylums were still operational,” says Rachel Upthegrove, Professor of Psychiatry and Youth Mental Health at the University of Birmingham. “These were places where people with severe and chronic mental illnesses were generally not treated with hope or recovery in mind.”

Fortunately, attitudes towards psychosis and mental health more generally are changing. After years of underfunded services and lack of investment in research, the outlook  crucially for  patients is also improving.

It could not come at a more important time. Every year, around 8,000 young people in the UK experience a first episode of psychosis. Early interventions services, which are designed with the aim of the best evidences treatments given early, instil hope of recovery and good prognosis, are at the heart of a move towards more supportive treatment pathways for those diagnosed with a psychotic disorder.

The human costs of psychotic disorders are immense, but so too are the financial costs. For example, the cost of schizophrenia alone to English society is £11.8 billion per year. (Effective Interventions in Schizophrenia: The Economic Case. LSE 2012). Affective and non-affective psychoses account for 6.3% of the global burden of disease and cost €207 billion per year in Europe alone, making them the most expensive brain-related disorders and even more expensive than cardiovascular diseases.

The right treatment for the right patient

Diagnosis itself has long been a challenge for mental health practitioners.

“The NHS is overwhelmed by people who come to mental health services, in desperate need for help,” explains Professor Upthegrove. “But diagnosis itself is very difficult, understanding when someone crosses the threshold from health to ill health, from order to disorder lacks clear thresholds, particularly in early stages of illness when interventions may be potentially more effective. The labels that we use; bipolar disorder, schizophrenia, general anxiety disorder and others, may be reliable and useful because people understand what we mean by them. But they aren’t related to specific biological markers or causal factors, and they may not be valid. There are no biomarkers or genomics tests that we can use in the clinic, and we may not have them in the future either due to the nature of these disorders. For all that we know about the changes that happen in the brain during early episodes, we can’t really say whether it is different between schizophrenia and bipolar disorder, for example, because each person is different. Getting the right treatment to the right patient is therefore very complex.”

Get Quest updates directly into your inbox

Subscribe to Quest to receive research emails

The idea of staging within psychosis has helped. Led by Professor Alison Yung at the University of Manchester, the stratifying of patients with psychosis into four key groups; those at risk, first episode psychosis, recurrent psychosis and chronic schizophrenia, has proven invaluable in understanding patient pathways and outcomes. However, where the psychiatric community falls behind the staged models seen in the treatment of cardiovascular disease or cancer, is that the interventions that follow are not aligned to these stages. The treatments, both pharmacological and psychological, are fundamentally the same at each stage.

“We’re at a point where we have to embrace the idea of personalised medicine to help patients, but perhaps this is an area where the challenges we face in psychiatry become a virtue,” adds Professor Upthegrove. “Because we cannot easily rely on validated biological tests, we have to go that bit further to understand the individual patient and their needs.”

Investigating the role of inflammation

For Professor Upthegrove, early intervention is key to bridging the unmet need of thousands, and Birmingham represented an obvious choice due to its history of strength in the field.

“One of my first clinical placements was in the very first UK Early Intervention Service, led by Professor Max Birchwood and Dr Dermot McGovern,” she says. “Birmingham was at the forefront of that movement, it was a very different approach that centred around hope and recovery. These are vital, particularly with young people experiencing their first episode.”

“What I started to become interested in is the relationship between psychosis and other areas of health and mental health. Particularly the relationship between mood and psychosis, and the relationship these symptoms may have with novel biological pathways such as inflammation. The positive symptoms that drive people into seeking care – things like hearing voices and hallucinations – often overshadow the mood-related symptoms like social isolation and anhedonia.”

She points to the Cambridge led study into the comorbidity of depression and inflammation, where they identified inflammation as a driver of depression in patients, as a pivotal step in the process.

“This sparked what is now an established body of work on links between inflammation and mood. Combined with our previous work on the importance of depression in early psychosis, it made us wonder if inflammation could explain some of the poor prognoses that can happen in psychosis. Perhaps we could improve outcomes for patients who present with inflammation during their first psychotic episode by treating them with anti-inflammatories alongside their normal medication.”

The interface between psychosis and inflammation is a new frontier for research but, according to Professor Upthegrove, the idea of staging to aid the diagnostic challenge was also seen as boundary-pushing only a few years ago. Anti-inflammatory treatments might not be suitable for everyone within the broad heterogenous group of people with a psychotic disorder but the impact of identifying clinical profiles of patients who might benefit from such interventions would be significant.

“When you think about treatment for psychosis, it is not just about prevention but also reducing the harm of an episode. There is evidence to support the idea that each episode has a cumulative impact on a person, with each recovery period potentially harder and less likely to achieve the same recovery in function. Most people recover quite well from the first episode, but where multiple episodes are experienced, there may well be irreversible changes caused in the shadow of the event. If inflammation is a contributing factor in this poor recovery and pharmacological treatment could help that, it would be key in preventing the worst outcomes.”

Towards a data science driven approach

The BeneMin trial, fronted by Professor Bill Deakin at the University of Manchester but with national collaboration from teams including Birmingham, sought to discover if minocycline, an antibiotic with neuroprotective and anti-inflammatory actions that have attracted attention as potential treatments for several neurodegenerative disorders, could be beneficial in reducing the negative symptoms of schizophrenia in patients with first episode psychosis. While the outcome of the trial was negative, Professor Upthegrove saw the knowledge it generated as a success.

“It’s really the high quality of the trial that determines its value, and BeneMin gave us a clear answer but also really strong national network and infrastructure that we could build on. The questions we can now ask is if the outcome was negative because minocycline doesn’t work? Or was it because we didn’t sufficiently target it to specific patient groups? That’s the next step for us, and our research.”

Understanding a young person’s experience is key to the Birmingham-led PIMS (Psychosis Immune Mechanism Stratified Medicine Study) trial. PIMS will investigate the association, and explore causal mechanisms, between increased innate inflammation and psychosis. If low-grade inflammation is causally linked to psychosis it will offer innovative treatment targets.

Importantly, it will use machine learning and advanced AI techniques to identify who might benefit most from novel immune targeted treatments.

If there is evidence that elevated circulating IL-6 (a glycoprotein produced by leucocytes for regulating immune responses and indicates elevated immune function) is causally linked with psychosis, the team will then use data from large clinical and epidemiological samples to identify illness stage and symptom dimension most closely linked with inflammation. Also, they will identify the relevance of peripheral blood levels of markers of inflammation to brain structure. Effectively, it would identify the biotype of immune-related psychosis.

Interleukin 6 (IL-6) molecule, chemical structure

If successful, it would lead to the team being able to carry out experiments on immune cells collected from patients, using a dose of Tocilizumab, an anti-inflammatory medication used in Rheumatoid arthritis that blocks IL-6 signalling. If it proves beneficial to the functional ability of patients, it would suggest a new treatment target and potentially move towards better outcomes for some psychosis patients.

“Bringing data science to help identify the right patients for treatment is really interesting and cutting edge,” says Professor Upthegrove. “It’s also really refreshing to have so many different disciplines involved in PIMS. We have representatives from the Institute of Mental Health, imaging experts, pharmacologists, mathematicians, psychiatrists, psychologists, and young people. It requires that much input because it is not just about identifying the possible treatments but understanding how you might practically deliver them. How can we take the idea of patient stratification within the stage model and use data science to implement related treatments within something as complex as the NHS?”

That is quite the challenge. Researchers know the difficulty of conducting research within the NHS, where very hard pressed clinical services are already working overtime to provide clinical care, so time and resource to support innovative research is quite hard. Fortunately, the Early Intervention with Psychosis Network – delivered by one trust across the city of Birmingham – represents a crucial ally for the research team.

For individuals with psychosis, the upshot of a data science driven approach to personalised care might be as simple as a blood test that is reviewed alongside their clinical symptoms, and then them being prescribed an anti-inflammatory medication alongside their anti-psychotic medication, psychological and supportive interventions. For the health service as a whole, the benefits would be very far-reaching indeed.

“Yes, it would be a great boon to the NHS and be so helpful in improving a young person’s recovery trajectory,” says Professor Upthegrove. “But I wouldn’t lose sight of the huge implications it would have in our efforts to engage young people with psychosis in new modes of thinking. With early intervention, psychosis is something you can recover from, something you can live well with, and with the right treatment and support you can still achieve your goals.”


The research mentioned is this article is funded by: MRC DPFS, NIHR-HTA, EUFP7.

It is made possible thanks to the work of Golam Khandaker, Nicholas Barnes, George Gkoutos, Andreas Karwarth, Stephen Wood, Nikos Koutsouleris, Bill Deakin, Peter Jones, Steven Burgess, Ed Bullmore, John Suckling, Joanne Neill, Anthony David, Carmine Pariante, Gary Donohoe and the PRONIA research project



Discover more stories about our work and insights from our leading researchers.