Using a Complex Systems Approach to Understanding Population Mental Health.

There is a growing consensus that it is not just individual characteristics and behaviours that influence health and wellbeing, but the types of neighbourhoods we live in, and the opportunities they afford us. This consensus is reflected by published evidence of inequalities in people's life chances between different places, the Marmot review "Fair Society Healthy Lives" demonstrates by association that neighbourhoods that are more deprived also tend to have populations with lower life expectancies.

The overall evidence for neighbourhood effects is somewhat limited, only recently has research begun to demonstrate an unambiguous association between neighbourhood and health. Health outcomes and behaviours studied have particularly focused on diet and physical activity, which have been linked to a range of neighbourhood contextual factors including access to green space, places to buy fresh fruit and vegetables (or conversely fast-food), and the 'walkability' of the built environment. Mental health and wellbeing has received markedly less focus, although some high quality studies have looked at the relationship between neighbourhood green space and wellbeing. Emerging epidemiological evidence has benefitted from researchers working across disciplines, particularly with geographers, in order to produce robust study designs. Nonetheless, evidence of a causal link between our changing neighbourhoods, and our changing health is rather less forthcoming, owing to the complexity of the social-ecological processes hypothesised to be taking place.

This research project aims to make a significant contribution to empirical evidence for neighbourhood effects by applying 'systems thinking.' This means treating neighbourhoods as complex, dynamic multi-factorial processes, where existing research treats them as simple, stationary snapshots with limited depth. Specific training in 'Urban Informatics' will greatly augment my existing expertise in Geographic Information Science and Social Epidemiology, and give me the tools required to operationalise dynamic, rather than static, representations of neighbourhoods. Capturing how neighbourhoods change over time alongside how individuals change over time is fundamental to unpicking causality.

In particular, the research will focus on the effects on mental health and wellbeing of people moving, expecting to move, or preferring to move, at both the neighbourhood scale, and the individual scale. We know that neighbourhoods are not equal, and yet in terms of population movement - people moving into or out of a neighbourhood - we tend to treat them equally. Studies have suggested that failing to account for this dynamic neighbourhood process - the residential movement of people - can have implications on our expectations for people's health through what are termed 'selection effects'. These neighbourhood compositional factors will be studied alongside contextual aspects of neighbourhoods, notably neighbourhood crime and disorder and neighbourhood economic insecurity, in order to explain how mental health and wellbeing is affected by neighbourhood social and environmental changes.

This type of research is only just becoming possible, due to a range of complementary factors. Firstly, world-class training at New York University's Centre for Urban Science and Progress (NYU CUSP). Secondly, the availability of the UK's rich longitudinal "Understanding Society" survey data. Thirdly, the proliferation of open, and free government administrative data, and finally, the growing support from research councils in promoting 'Big Data' and informatics approaches. This fellowship takes advantage of all of these factors, with the hope of producing world-leading research on neighbourhoods and health.

AIM: The research project will aim to describe, model and explain the implications for general health and mental health of dynamic neighbourhood contextual and compositional factors. This will be achieved using longitudinal and spatial statistical approaches augmented with methods from Urban Informatics.

OBJECTIVES: (i) To explain the pattern of flows of individuals between neighbourhoods according to psychological health and its social and biological antecedents, over time. (ii) To analyse the health and socio-economic trajectories of neighbourhoods over time, according to individual exposure to neighbourhood change. (iii) To model the key individual and environmental factors complicit in explaining population change, particularly focusing on neighbourhood systems related to crime and disorder and economic insecurity, and to assess whether these factors exacerbate inequalities in mental health and wellbeing. (iv) To synthesise findings from objectives i-iii in specifying a mathematical model that operationalises the neighbourhood as a dynamic complex system using spatial flows describing population mobility, disaggregated according to neighbourhood health status and trajectory. Population movement will be explored subject to a range of neighbourhood change scenarios that will be developed.

METHODOLOGY: The outcomes of interest are: general health (SF-12), psychiatric morbidity (GHQ-12); and positive mental wellbeing (WEMWBS). Exposures of interest relate to changes in the individual and neighbourhood social and biological antecedents of health and wellbeing.
The project will make use of the longitudinal survey "Understanding Society", the UK Census of Population for 2001 and 2011, Census Flow data for 2001 and 2011, a range of open and free administrative data resources, authoritative Ordnance Survey Spatial Reference Data, and the newly funded Administrative Data Research Centre for England (ADRC-E) and Consumer Data Research Centre (CDRC).

As per the communications plan, primary beneficiaries of this research will be academic, however this academic community will be highly interdisciplinary. Academic disciplines within which impact could be expected include: social epidemiology, public health, and, health and medical geography. Interdisciplinary contexts include those relevant to both the use of population-level survey data, such as Understanding Society, and the application of spatial analysis and Urban Informatics. These might include: Urban Studies, Geocomputation, Planning and Built Environment Studies, and analysts of mathematical complexity and system approaches.

The findings of the proposed research will benefit a number of public sector organisations. In particular those invested in understanding communities, allocating resources appropriately, making interventions in health and wellbeing, and evaluating or anticipating the effects of those interventions. I would expect health outcome specific findings to be of interest to the Department of Health, and the NHS, with evidence of divergent mental health outcomes by neighbourhood dynamics likely to be important in directing resources and strategy. Mapping neighbourhood change would be of substantive interest to the Department for Communities and Local Government (DCLG), and could help advance indices aimed at capturing deprivation and social change. The DCLG has previously commissioned exploratory work in this area, albeit under a previous government. Additionally, the social care functions of Local Authorities would likely be interested, as would primary care providers. During my PhD I worked with Southwark Primary Care Trust on a range of topics related to accessing health care. I know that the research outputs would have found a home in that context, but with the transition to Clinical Commissioning Groups (CCGs) I'm unsure of how connections would be made, this is something that I will seek to follow up during the fellowship.

Other national bodies, such as Public Health England (PHE), and the Office for National Statistics (ONS) would be relevant to the communication of the research. PHE sets the agenda for public health priorities, and is an important body for positioning and commissioning evidence. Persuasive evidence on neighbourhood dynamics and mental health would be hugely beneficial to a range of social policy interventions, from regeneration and renewal, to promoting or limiting neighbourhood resources like green spaces, and managing crime and disorder. The ONS, have an important role to play in the creation and dissemination of public administrative data, such as the Index of Multiple Deprivation, and the Census itself; being able to feedback approaches to using data that go beyond simple, established cross-sectional methods via the ONS would be beneficial to a great many users.

Significant findings on the topic of mental health and wellbeing are likely to be of interest to public sector institutions, research bodies and charities specifically concerned with mental health, such as Mind and CentreForum's mental health commission. I will engage with such institutions as required, and in line with recommendations from the LSHTM public engagement team. I am keen that significant research findings be presented in an accessible light and will approach Sense about Science, a charity concerned with the public understanding of science, in order to achieve this.

Finally, there is likely to be a commercial interest in the approaches of the research. Many companies, from insurance to supermarkets, base customer projections on an understanding of people's neighbourhoods and communities. New approaches to thinking about neighbourhoods that are more effective than those commonly used could be good for business. I have previously engaged with businesses on the topic of geodemographics, notably with CACI and Experian, and so there may be scope to promote my research to these kind of users.