Testing geographic regression models for analysing impacts of urban structure on mental health

Mental health is one of most common disorders associated to socioeconomic individual and neighborhood characteristics. The aim of this study is to test the association at area level between specific characteristics of the... [ view full abstract ]

Mental health is one of most common disorders associated to socioeconomic individual and neighborhood characteristics. The aim of this study is to test the association at area level between specific characteristics of the urban structure (urban density, functional mix, green and pedestrian areas, cultural and leisure services, accessibility by public transport) with antidepressant consumption, in Turin (Italy).
Data was extracted from the Turin Longitudinal Study (TLS)—a census-based cohort study following up prospectively the mortality and morbidity of the population. Estimates were adjusted by individual socio-demographic variables (education, housing tenure, employment) and contextual social environment (SE) variables (social and physical disorder, crime rates).
A previous study, run with traditional statistic regression models, highlighted accessibility and urban density as the most influential features. Using geographically weighted regression models we compare results obtained, highlighting differences and suitability of the two methods. Geographic regression models show an innovative insight on health impact analysis, and offer the opportunity to customize and refine analysis at area level.
In agreement with previous literature, our results suggest that BE has a stronger effect on MH for people who spend more time in it. We conclude with some considerations on the equity impacts related to distribution of services and urban features in the city of Turin, as a result of urban renewal policies developed in the last 20 years.