Adverse built, social and physical environments are associated with worse mental health outcomes which first emerge in adolescence, offering potentially modifiable population-level targets for prevention. While early detection has become the cornerstone of UK and Australian youth mental health provision, which has led to improved downstream clinical and social outcomes for young people, primary prevention remains an elusive goal, resulting in lifelong physical and mental health disparities which affect whole communities. Hitherto, most studies of the environment and mental health have considered selected indicators from a single domain (built, social or physical), making unobserved and residual confounding major obstacles to causal inference and primary prevention. Methods to characterise the way in which the exposome affects mental health and well-being are now required.

We will address this in two phases.

First, by linking large, geocoded epidemiological and clinical data from ALSPAC with a comprehensive set of built, social and physical environmental exposures via the ALSPAC-PEARL/ALGAE linkage, we will identify the pathways through which these factors affect various mental health outcomes. We will also consider how physical health and activity in childhood may mediate the relationship with later mental health, and vice versa. Environmental data will be linked from multimodal sources available in PEARL/ALGAE and via integration of other available environmental data to characterise the built (building quality, indoor air quality, density, land use, overcrowding, transportation links), social (population density, social isolation and cohesion, inequality, deprivation, ethnic diversity, homelessness, crime) and physical (air, light & noise pollution; accessibility to and quality of green or blue spaces, walkability) environment.

Second, we will develop a simulation platform (SAPPHIRE) to evaluate putative intervention strategies in the built environment to prevent selected adverse mental health outcomes and physical health comorbidities. We will develop a simulation approach based on the ALSPAC sample and the wider population of the Bristol region. We will synthesise theoretical and empirical evidence to build this platform via a hybrid Dynamics/Agent-Based Modelling approach, consistent with capturing the interplay between geospatial, household and individual factors which affect mental health. SAPPHIRE will be open-source, so that decision makers can readily adapt, deploy and test prevention strategies in different contexts based on parameter values representing their local circumstances. This has the potential to unlock primary or secondary prevention strategies in the built environment, which would otherwise be prohibitively expensive, time-consuming or impossible to deploy in the wild.