Neighbourhood effects, local crime and mental health: longitudinal analyses over the life course

Mental health conditions are major contributors to global disability and suffering, with a substantial social and economic burden. Besides well-established individuallevel biopsychosocial determinants, emerging literature emphasises that social and physical features of the residential environment are associated with mental health. However, there is a limited understanding of how, where and for whom neighbourhood matters for mental health, partly because of the methodological shortcomings of existing literature. This thesis takes a longitudinal approach to examine the links between place-based factors, in particular neighbourhood crime, and anxiety, depression and psychosis. A systematic review and meta-analysis on the association between local crime and mental disorders sets the context for the thesis. Based on 50 studies meeting the inclusion criteria, random-effects meta-analyses indicated higher risk of depression and psychological distress in high crime areas; for anxiety and psychosis there was only limited evidence. Associations varied by study design (longitudinal versus cross-sectional), type of crime measurement (perceived versus objective) and between different age groups. Importantly, the review identified research gaps, which were the focus of the following chapters. The thesis was structured around four longitudinal investigations, two utilizing perceived and two objective neighbourhood measurements. First, available evidence in the field is limited to a few countries and there is no information on country-level heterogeneity. Data on perceived neighbourhood conditions and depression from 16 countries were utilised across three ageing cohorts (English Longitudinal Study of Ageing; Health and Retirement Study; Survey of Health, Ageing and Retirement in Europe [SHARE]) capturing adults aged 50 and over (n=32,531). Findings indicated elevated risk of depression amongst participants living in an area with perceived neighbourhood disorder (including crime) or with lack of social cohesion. Further analyses uncovered cross-level interactions by income inequality, population density and air pollution for social cohesion and by forest coverage for neighbourhood disorder. Second, neighbourhood effects might be determined by vulnerability build up over the life course; however, no information is available on the long-term impact of childhood stressors, a sensitive period in human development. Prospective and retrospective data on adults aged 50 and over (n=10,328), were analysed from the SHARE survey. In addition to a higher risk of depression when living in areas with perceived neighbourhood nuisances (including crime), and lower in areas with good access to neighbourhood services, childhood socioeconomic conditions modified neighbourhood effects. Older adults who grew up in better childhood circumstances benefited more from neighbourhood resources, but they were at higher risk of depression when exposed to neighbourhood problems. Third, there is a lack of understanding of different neighbourhood crime and mental health associations across psychiatric conditions. A large data-linkage study (Scottish Longitudinal Study [SLS]), on small area-level crime rates and prescribed psychotropic medications was carried out (n=129,945). Findings indicated higher risk of antidepressants and antipsychotics, but not anxiolytics medications in high crime areas. Moreover, there was higher risk of antidepressants prescriptions among adults aged 24-53 in 2009, antipsychotics among men aged 44-53 in 2009, and among those in the middle of the social ladder, when living in high crime areas. Fourth, although changing levels of neighbourhood exposure may help in understanding the causal relationship between context and health, very few studies have utilised repeated measurements of small area-level crime. Analysis based on the previous study (SLS) with additional linkage for three consecutive area crime measurements were carried out, to explore the association of changing crime rates with self-reported mental illness and prescribed medications among residential stayers and movers (n=112,251). Recent increase in crime exposure was associated with mental health problems among stayers aged 16-30 (self-reported mental illness, antidepressants), and among movers aged 31-45 (self-reported mental illness, antipsychotic medication). After excluding individuals with pre-existing mental health conditions, findings suggested causation for the former, and health selective migration for the latter group. Neighbourhood crime and other contextual factors in the residential area are significant determinants of mental health, but associations differ by childhood and adult socioeconomic conditions, across sex and age groups, and between anxiety, depressive and psychotic disorders. Place-based interventions aimed at reducing crime, supporting social cohesion and allocating targeted mental health preventions and services in the vicinity of high crime areas, may have long-term benefits for residents’ mental health, especially for those more vulnerable. Future research should investigate the relationship between area stressors and mental health by exploring direct and indirect pathways, studying crime effects at different geographical levels, and applying the life course framework