A Multilevel Analysis to Disentangle the Effect of Neighborhood on Health and Mortality Risks

Michel Poulain, Université Catholique de Louvain and Tallinn University

Health and mortality inequalities are commonly observed in large cities as for example in Montreal (Ross et al. 2004), New York (Karpati et al. 2006) and more recently in Cincinnati (Maseru et al. 2011). Already in 1830 Villermé demonstrated that mortality levels were different between the 12 “quartiers” of Paris and investigated the causes of such differences by considering several ecological variables. This contribution is based on individual and aggregated data on the population of Namur (Belgium), a 100,000 inhabitants’ city divided in 46 neighborhoods. The outcome variable is the 4 years survival after 2001 census. That survival is observed through the Population Registration System. A statistical analysis is applied involving explanatory variables at three levels: individual, household/housing and neighborhood. The first results confirm a limited impact of the neighborhood environment on the mortality risk but such results could have important policy implications.