Research

My research examines how mental health develops across lives, places, and systems, why trajectories and outcomes differ, and which processes may be modifiable. I work across psychiatric epidemiology, health data science, causal inference, and health services research.

Much of this work is grounded in linked administrative and electronic health-record data, longitudinal studies, and routinely collected service data. Within MINDSET, I extend that foundation by combining it, where useful, with intensive longitudinal, environmental, behavioural, and biological evidence. The substantive question determines the mix: multimodality is a means to stronger inference rather than an end in itself.

How the programme connects

Each project enters the programme at a different point, but the work connects three layers, with substantive mental-health questions leading.

The hierarchy is deliberate but iterative: substantive questions determine which evidence and methods are needed; inferential work establishes what the measurements can support; and better measurement, governance, and data systems create new possibilities for research. This helps distinguish processes underlying mental-health trajectories and inequalities from the ways they are measured, recorded, and observed.

What shapes mental health and inequalities over time?

I study how conditions and experiences within people, between people, and across the environments and institutions they encounter combine to shape mental health, service use, and wider outcomes over the life course.

This work spans severe mental illness, neurodivergence, addiction, and mental-health services. Recurring questions include how disadvantage accumulates, how service contact and exclusion shape later outcomes, and why trajectories differ between people, populations, and places.

Current work includes WISDOM, examining social, emotional, environmental, and institutional influences on severe and enduring mental illness, and a process evaluation of a co-produced anti-stigma intervention for health and care staff working with people who use substances. Recent work through HOPE-SEN used linked health and education data to investigate inequalities in the experiences and outcomes of neurodivergent children and young people.

Areas of interest include:

• longitudinal and life-course psychiatric epidemiology
• social and environmental determinants of mental health
• inequalities in access to and outcomes of care
• severe mental illness and co-occurring health and social needs
• neurodivergence, education, and health
• addiction, treatment pathways, and service use
• spatial and environmental epidemiology
• ecological momentary assessment and intensive longitudinal data
• multimodal measurement and integration across social, behavioural, and biological scales

How can heterogeneous evidence support credible inference?

No source directly captures a person’s mental health or the processes that shape it. Evidence may arise from:

• experience: surveys, interviews, ecological momentary assessment, and narrative or clinical text
• behaviour and physiology: repeated digital measures, sensors, and wearables
• biology: biomarkers, genomic and other molecular data, and neuroimaging
• place and context: social, relational, environmental, spatial, and network information
• institutions and care: linked administrative records, electronic health records, education data, and treatment or service systems

Useful combinations depend on the question. Every source has distinct timescales, selection processes, measurement error, missingness, and interpretive limits. Administrative and clinical records, for example, reflect access, recognition, documentation, coding, linkage, and institutional practice as well as underlying phenomena.

My methodological work therefore includes:

• longitudinal, repeated-measures, multilevel, and fixed-effects designs
• causal inference and observational epidemiology
• data linkage and natural language processing, including large language models in governed research environments
• spatial and network analysis
• principled integration and triangulation across heterogeneous sources
• explicit analysis of selection, measurement, missingness, and data quality
• transparent and reproducible research in secure environments

The aim is to strengthen inference and make explicit what each source, and each combination of sources, can and cannot establish.

Research infrastructure as an enabling foundation

Research using sensitive human data depends on how evidence is recorded, governed, linked, accessed, documented, and interpreted. I therefore treat research infrastructure as part of the scientific work, not simply a technical prerequisite.

As Deputy Lead of the North London NHS Foundation Trust Research Database, I support research using routinely collected mental-health records, including research development, governance, methodological support, and collaboration across clinical, informatics, academic, and public partners.

I also work with colleagues at South London and Maudsley NHS Foundation Trust on harmonisation across mental-health research infrastructures and approaches that support cross-site replication and comparative research.

Within UNITED (Using a National, Interdisciplinary Team to Enhance Drug and alcohol treatment data), I lead work examining how England’s National Drug Treatment Monitoring System (NDTMS) is used for research and how treatment data can better support epidemiology, service improvement, policy, and clinical practice.

See also people and partnerships, publications, research funding, open research and resources, and leadership and recognition.