PsyLife research

Welcome to the PsyLife group research page. On this page you will find out more about the research we undertake. We conduct epidemiological research to investigate whether social and environmental factors are causally related to experiencing mental health problems. The main focus of our research is on identifying, understanding and preventing the social and environmental determinants of psychosis.

In particular, we specialise in two main research problems. First, we are seeking to determine and remove the causes of inequalities in psychosis risk by ethnicity and migrant status, one of the grossest injustices in public mental health. Second, we are seeking to understand whether being born, brought up and living in certain social environments (for example more urban, or more deprived) is a causal risk factor for psychosis.

To investigate these issues, we apply both traditional and causal inference methods in epidemiology to large, population-based observational datasets of mental health problems.

Epidemiology is a quantitative, population-based science which involves comparing differences in the risk of experiencing a given outcome between people who are, and who are not, exposed to a given risk factor. It involves assessing whether the association could have arisen by chance, may be attributable to other factors, or may have arisen as a result of biases in the way in which the data were collected and analysed. Understanding whether an association between a risk factor (or exposure, or determinant) and an outcome (such as a mental health disorder).

Wherever possible, we ensure our research is translated into evidence to help policymakers, mental health service providers and clinicians to make better decisions about care for people with mental health problems. You can read more about one example, the PsyMaptic prediction tool, on these pages, which is used by NICE and NHS England to plan Early Intervention in Psychosis [EIP] services in England according to our epidemiological models of need for care in different local populations.

Click on a project below to find out more about our work

PsyLife - Understanding the social determinants of psychosis & other mental health problems over the life course

Duration: 2014-present

Setting: Various, including Swedish national register data, ALSPAC birth cohort (UK), and the National Longitudinal Study of Children & Youth (NLSCY, Canada)

Funders: Wellcome Trust & Royal Society, MHRUK

Conducted at: Division of Psychiatry, UCL

PI: Prof James B. Kirkbride

Partners: Prof Ian Colman (University of Ottawa); Prof Christina Dalman (Karolinska Institutet)

Aim: To delineate the burden and social and environmental determinants of psychosis & other mental health problems

Method: Various methodologies, including causal inference methods applied to longitudinal datasets of mental health problems. Methods include causal mediation analysis, genetically-informed modelling, multilevel survival analysis, and other multilevel models.

Main findings: In ALSPAC, we demonstrated that the association between greater neighbourhood deprivation at birth and psychotic symptoms in adolescence was not explained by polygenic risk for schizophrenia [13]. We have also shown that children who live in neighbourhoods with higher levels of perceived neighbourhood stress and discord were more likely to report psychotic experiences during adolescence [12].

In nationwide Swedish data, we used causal mediation analysis to show that up to a quarter of the association between urban birth and later risk of psychotic disorder could be attributed to effects on cognition, if causal [10].

In other Swedish register studies, we have shown increased risk of mental health problems amongst migrants and their children appear specific to psychotic rather than affective domains of neuropsychopathology [3], and are influenced by family [2] and neighbourhood [1] social support during and after migration. Moving house more often and over longer distances in childhood and adolescence (but not adulthood) also predicts increased future risk of psychotic disorder [11].

We were the first group to conclusively demonstrate that refugees were more likely to experience psychosis than Swedish-born individuals and non-refugee migrants from the same regions [7]. We have shown elevated risks in refugee and other migrant groups do not exist for suicide [6] or substance use disorders [4], but rates of these mental health problems converge to Swedish rates over time. Migrants also experience more adverse pathways to psychiatric care [5,14].

Using Canadian data, we have also shown that children growing up in more adverse neighbourhoods are more likely to report mental health difficulties in adolescence [8]. However, we have also shown that growing up in positive environments, with high levels of social capital, can protect children against mental health problems in adolescence following exposure to stressful life events [9].

Impact: Results have furthered our understanding of the social aetiology of psychotic disorders, and could aid population-level prevention strategies for improving mental health in different communities.

Selected publications

PsyMaptic-A - Psychiatric Mapping Translated into Innovations for Care

Duration: 2017-2020

Setting: England

Official website: www.psymaptic.org

Funders: Wellcome Trust & Royal Society

Conducted at: UCL, University of Cambridge, Hull-York Medical School, Manchester Metropolitan University

PI: Prof James B. Kirkbride

Co-Is: Profs Peter B. Jones, Gianluca Biao, Paul French, David Osborn, Jeremy Coid; Dr Pia Wohland

Aim: To develop a population-level prediction tool to predict need for Early Intervention in Psychosis [EIP] care in England up to 2025, using epidemiological data on first episode psychosis [FEP].

Method: Seed data from almost 1,500 people with FEP in England were used to estimate robust risks for disorder in the English population. These were combined with prior empirical information and projections of the future population in England in Bayesian Poisson regression models to forecast need in local populations (electoral ward level) up to 2025.

Main findings: We developed and validated a population-level prediction tool to forecast the number of people in England requiring care for FEP each year up to 2025. We additionally forecast extra demand placed on EIP services resulting from additional referrals, assessments and treated cases who would not fulfil criteria for FEP.

Impact: Since 2012 our free, open access predictions of likely need for EIP care in England have been used by NHS England to inform service planning. Our prediction tool, PsyMaptic, has informed and is recommend in NICE guidelines for commissioning and providing EIP services in England.

Publications

SEPEA - The Social Epidemiology of Psychoses in East Anglia

Duration: 2009-2013

Setting: East Anglia, UK

Funder: Wellcome Trust

Conducted at: Department of Psychiatry, University of Cambridge

PI: Prof James B. Kirkbride; Co-I: Prof Peter B. Jones

Aim: To precisely estimate the treated incidence of psychotic disorders in a rural setting & investigate variation by social determinants for the first time.

Method: Naturalistic cohort study of all people aged 16-35 years old presenting to Early Intervention Psychosis [EIP] services in the region for the first time. Denominator obtained from 2011 Census. Basic sociodemographic & clinical data collected at baseline and discharge. Analyses included multilevel Poisson regression.

Main findings: Incidence of psychotic disorders higher than previously expected in rural region (Kirkbride et al, 2017, Am J Psych), but varied by age (highest risk in late adolescence/early 20s), sex (higher in men) and ethnicity (higher in several ethnic minority groups). Immigration during adolescence associated with higher risk (Kirkbride et al, 2017, Schiz Bull). More deprived, socially isolated rural areas had higher rates (Richardson et al, 2018, JAMA Psychiatry).

Impact: Results have furthered our understanding of the epidemiological landscape of psychotic disorders, showing substantial need for care in diverse, rural populations. Data have been used to inform future planning of EIP services in England as part of the PsyMaptic study

Publications

EU-GEI - The European Union Gene-Environment Interaction study

Duration: 2010-2014

Setting: 6 countries, 17-centres (WP2) including Cambridgeshire (UK)

Official website: www.eu-gei.eu

Funder: EU Framework Programme 7

Conducted at: Department of Psychiatry, University of Cambridge

Co-PI: Prof James B. Kirkbride; Co-PI: Prof Peter B. Jones

Aim: To understand the roles of genetic and environmental factors in contributing to the risk of experiencing psychotic disorders.

Method: Incidence and case-control study of ~2400 people in their first episode of psychosis, ~1200 of their unaffected siblings and ~2400 controls. We collected detailed biological, clinical and social data on our participants, and analyse the data in a variety of ways including multilevel Poisson and logistic regression.

Main findings: The first direct international comparison of incidence of psychotic disorders for over 30 years revealed considerable variation in rates between settings and countries [2]. We have shown that this is highly correlated with social determinants of health including economic disadvantage [2] and cannabis use [3]. Incidence rates are also raised in many ethnic minority groups. Using the case-control data from the study [4], we have also shown for the first time that exposure to structural racism (indexed via ethnically patterned social disadvantage, sociocultural exclusion and discrimination) can account for the excess risk of psychosis in several migrant and minority ethnic groups which have been consistently observed for over 140 years.

Impact: Findings from this study point towards new understanding of social and racial inequalities in psychosis risk, and provide a potential roadmap for public health practitioners to devise prevention strategies capable of removing these social injustices.

Publications

OLA - Building resilience and resources to overcome depression and anxiety in young people from urban neighbourhoods in Latin America

Duration: 2018-2023

Setting: Peru, Colombia, Argentina

Twitter: OLA study

Funder: MRC (UK)

Conducted at: Queen Mary University London, UCL, Peruvian University Cayetano Heredia, University of Buenos Aires, Pontifical University Javeriana, Kings College London

PI: Prof Stefan Priebe (QMUL); Co-I: Prof James B. Kirkbride

Aim: To understand the factors that contribute to resilience against depression in Latin American youth.

Method: Mixed methods study including cross-sectional and cohort survey of adolescent youth living in deprived parts of Lima (Peru), Bogota (Colombia), Buenos Aires (Argentina).

Main findings: The study is currently underway and results are expected to be published over the next few years. Check our Twitter feed for latest news.

Publications

EPICare - Early Psychosis Informatics into Care

Duration: 2021-present

Setting: UK

Twitter: EPICare study

Funder: NIHR (UK)

Conducted at: UCL, University of Birmingham, University of Cambridge, University of Manchester, University of Bristol, Norfolk and Suffolk NHS Foundation Trust

Co-PI: Prof Rachel Upthegrove (University of Birmingham); Prof James B. Kirkbride

Co-I: Prof Sonia Johnson (UCL), Dr Graham Murray (University of Cambridge); Prof Justin Waring (University of Birmingham); Dr Richard Drake (University of Manchester); Mr Matthew Machin (University of Manchester); Dr Sarah Sullivan (University of Bristol); Prof John Ainsworth (University of Manchester); Dr Sophie Allan (Norfolk and Suffolk NHSFT)

Aim: To improve access and delivery of timely, evidenced based interventions for people with first episode psychosis in Early Intervention in Psychosis (EIP) services.

Method: To co-design and test a national patient-centred digital psychosis registry (EPICare) and clinical decision support system; to pilot and implement EPICare in several NHS Trusts; and to evaluate the effectiveness and cost effectiveness of EPICare.

Main findings: The study is currently underway and results are expected to be published over the next few years. Check our Twitter feed for latest news.

Latest on Twitter

EPITOME - Evaluating Policy Implementations TO Predict MEntal health: A Bayesian hierarchial framework for quasi-experimental designs in longitudinal settings

Duration: 2021-2025

Setting: UK

Funder: Wellcome Trust (UK)

Conducted at: UCL, Imperial College London, London School of Economics and Political Science

PI: Prof Gianluca Baio (UCL); Co-I: Prof James B. Kirkbride; Prof Marta Blangiardo (ICL); Dr Sara Geneletti (LSE)

Aim: To provide robust evidence-based evaluation of government policies on mental ill-health with specific focus on ethnic minorities.

Method: Developing a hierarchical statistical framework and applying this to several UK-based longidutinal datasets

Main findings: The study is currently underway and results are expected to be published over the next few years.

Systematic reviews

Duration: 2008-present

Setting & funders: Various

PI: Prof James B. Kirkbride

Aims: We conduct systematic reviews relevant to the incidence, prevalence and clinical, social and economic determinants of psychosis and other mental health conditions. We also collaborated on a review of the age-at-onset of all mental disorders.

Method: We have an established systematic review methodology, and provide training to our MSc and PhD students with the skills and knowledge they need to conduct high quality reviews. Our reviews are pre-registered on PROSPERO and conform to Prisma guidelines. Where suitable, we also perform meta-analysis and meta-regression.

Main findings: In 2012 we published a highly-cited review of the incidence of psychotic disorders in England between 1950-2009 [2], showing robust variance in rates by person and place. This has been followed by an updated and influential review of the international incidence of psychotic disorders up to 2019 [6], and a separate review of the incidence of very-late onset schizophrenia-like psychosis [5]. We have also been involved in a review of the role of air pollution in mental health, finding robust associations between PM2.5 and depression and – potentially – anxiety [7]. We have also reviewed the literature between psychosis risk and place [1].

Impact: Our systematic review findings contribute to the consolidation of knowledge within the field. They have also been used to help guide empirical development of our population-level prediction tool for psychosis, PsyMaptic.

Publications

DEPICt - Delineating the Epidemiology of first episode Psychosis In Chile: a nationwide register study of 30,000 incidenT cases between 2005 and 2018

Duration: 2019-20

Setting: Chile

Funder: UCL Grand Challenges of Global Health

Conducted at: UCL

PI: Prof James B. Kirkbride; Co-I: Dr Christina Ding (UCL)

Partners: Prof Lucia Valmaggia (IoPPN), Dr Alfonso González-Valderrama (Universidad Finis Terrae), Dr Hannah Jongsma (UCL), Rayanne Jean-Baptiste Bastien

Aims: (1) To conduct a systematic review of the incidence of psychotic disorders in Global South countries. (2) To estimate the incidence of non-affective disorders in Chile using nationwide register data, and investigate variation by available social determinants of health. (3) To use this data to develop a version of the PsyMaptic prediction tool to estimate future need for psychosis care in Chile.

Method: Systematic review of the incidence of psychotic disorders in the Global South following PRISMA guidelines (Prospero ref: ). Incidence rates in Chile modelled from Chilean First-Episode of Schizophrenia Programme register, which records all people who presented to primary, secondary or tertiary care with a non-affective disorder between 2005-2018. Incidence modelled by multilevel Poisson regression with fractional polynomial regression terms fitted to inspect nonlinear associations with deprivation, time period and latitude. A version of the PsyMaptic tool followed a similar methodology to that described above.

Main findings: Incidence of non-affective disorders in Chile followed a typical age-sex distribution as observed in other studies [1]. Rates were higher in more deprived regions, but no association with population density was observed [1]. Incidence rates declined non-linearly, before increasing again over the time period, a trend which was more pronounced in women. No linear association with latitude was observed [1].

Impact: Our findings are informing psychosis care planning in Chile. The lack of association between population density and incidence rates of non-affective psychosis differ from many findings in Global North contexts, but should be examined more precisely in future studies.

Publications

PsyMap-Zn - PSYchosis MAPping in kwaZulu-Natal

Duration: 2019-2022

Setting: KwaZulu-Natal, South Africa

Funder: MRC (UK), SA-MRC (South Africa)

Conducted at: University of KwaZulu-Natal, University of Exeter, UCL, King’s College London, Stellenbosch University

PI: Profs Bonga Chiliza (KZN), Jonathan Burns (Exeter); Co-I: Prof James B. Kirkbride (UCL)

Aim: To estimate the incidence of psychotic disorders in a defined catchment area in KwaZulu-Natal, South Africa, and investigate the social determinants of any variation in risk.

Method: Incidence and case-control study working with traditional and formal healthcare providers to identify cases over a defined period of time. Multilevel Poisson regression analyses will be used to formally investigate variance in incidence.

Main findings: The study is currently underway and results are expected to be published over the next few years

Meet the team

Other - various research

Duration: 2005-present

Setting & funders: Various

Summary: We are involved in various other studies investigating the epidemiology and public health impact of mental disorders on different populations.

Our early research, conducted by Dr James Kirkbride as part of his PhD and postdoctoral career, used epidemiological data from the AESOP and East London First Episode Psychosis studies to precisely characterise variation in incidence of schizophrenia and other psychotic disorders. This work has shown that people from Black Caribbean, Black African and Pakistani and Bangladeshi ethnicities in particular experience rates of psychotic disorder that are 2-4 times higher than in the White British population, even after accounting for differences in age, sex and socioeconomic position. This research has also shown higher rates occur in more urban communities with higher levels of inequality and deprivation.

Recent research we collaborated on from Australia has also found inequalities in psychosis risk by migrant status and ethnicity, for the first time in this context. Migrants from North and Sub-Saharan Africa were 3-5 times more likely to experiencing psychosis than the Australian-born population, but rates for Asian migrants were 2-4 times lower than for the Australian-born group.

In cross-sectional survey from China, we have also investigated the prevalence of psychotic symptoms, showing that people reporting several psychotic symptoms were more likely to have been born in urban areas.

We’ve also worked on research studies which have used epidemiological methods to investigate important questions, which at first glance might appear to be light-hearted. For example, we investigated a long-held hypothesis that cat ownership could increase risk of psychosis, putatively because of the association between toxoplasmosis (which can be harboured in cat poo) and psychosis. We used longitudinal data from ALSPAC to refute this hypothesis, finding no association between cat ownership in childhood and psychotic experiences in adolescence. In other research, published in a BMJ Christmas Special, we designed a cohort study to show that top-grossing movies for children had higher levels of death, including murder, than a corresponding sample of movies for adults!