Der Einfluß kontextueller Merkmale auf die Entstehung gesundheitlicher Ungleichheit: Theoriebildung und Methodenentwicklung

点击翻译按钮获取中文摘要

The relationship between social inequalities and health is well documented. Until now, social factors at the individual level have predominantly been considered. In our previous DFG project we could demonstrate that factors of the small-scale environment or context of persons play a role in the production of health inequalities, and include them in a theoretical model. Due to data availability, administratively defined spatial levels had to be used; the question what geographical levels are the most relevant for understanding the association between contextual factors and health still remains to be answered. A definition of such small areas (neighbourhoods) from the point of view of an individual might be more appropriate to understand how contextual factors influence individual health and contribute to the development of health inequalities. In the present project, such a notion of neighbourhood is investigated. We will study two aspects: the neighbourhood s mediating role between place and health of an individual (subject matter aspect) and its quantitative characterisation (methodological aspect).The aim is to assess, within five distinct German and European urban areas, how the neighbourhoods are characterised and which among a range of individual and contextual variables affect those characteristics. Five different cohorts providing measures of geo-location and of individuals health are available. The covariates of interest will include demographic data, individual and contextual measures of socio-economic status and subjective measures of perception of the environment. Also a comparison of the results obtained will be performed between the German and European urban areas in order to assess which parameters could further refine our general theoretical model explaining the mediation between place and health. In terms of methodology, such a neighbourhood will be defined, on average over all individuals in a single urban area, by a disk containing neighbours whose health is correlated to the health of the individual at its centre. The neighbourhood is characterised by its range and the strength of the correlation. We will model the correlation structure using semi-variograms. The effect of covariates will be assessed by modelling the correlation of residuals. Issues around sample size and distribution of the data over the area (cluster or uniform) will be investigated through simulations in order to advise on the optimal size and sample form to obtain estimates of strength and range. This project will deepen the understanding of the mechanisms which mediate the relationship between place and health. A strong correlation between health outcomes in the neighbourhood indicates context-dependent health inequities. Understanding the factors at work will inform Public Health strategies to tackle those inequities.