Mirela BARROS SERAFIM, Department of Geography, University of São Paulo, Brazil
Ligia VIZEU BARROZO, Department of Geography, University of São Paulo, Brazil
Neighborhood effects caused by shared socio-environmental factors are more intense in slum areas. The city of São Paulo is home to approximately 12 million inhabitants, and 11% of its population were living in precarious settlements in 2010. While some socioeconomic indices used by national studies to predict health outcomes do not include essential dimensions, others, such as the Municipal Human Development Index (MHDI) (PNUD, 2013), contain demographic or health data that create mathematical redundancy into regression models. In this context, Barrozo et al. (2020) developed the GeoSES index with applicability in three scales of the Brazilian territory. Despite its innovation, health differences between the precarious settlements are hardly captured by the GeoSES, and the index does not include the environmental context. In this study, we present a socio-environmental index capable of contrasting the health conditions between and even within the slums of the São Paulo municipality. The index was calculated for the 1.593 Units of Human Development (UHD) from the Atlas Brazil project (PNUD, 2013). The initial dataset was composed of 41 variables collected from the 2010 Brazilian Census and different sources of environmental data. Using the same methodology of the GeoSES, by successively applying the principal component analysis (PCA) technique, we kept six socioeconomic and six environmental variables. The variables that captured the maximum variance were schooling and sewage collection. The index was internally consistent (Cronbach’s Alpha of approximately 0.7 in both dimensions) and presented a correlation of p = 0.88 with the MHDI. The high correlation (p > 0.85) with the health data from the Atlas Brazil project (infant mortality, life expectancy, and others) points to the significant explanatory potential of the index for future research in slum health. Finally, the proposed index can be easily applied to all metropolitan regions of Brazil.
Mots clés : Multidimensional poverty|Neighborhood effects|Socioeconomic deprivation|Slum health|Small area
A104983MS