Back to Annual Meeting Page		American Public Health Association 133rd Annual Meeting & Exposition December 10-14, 2005 Philadelphia, PA

Abstract #103536

Basile Chaix, PhD¹, Juan Merlo, MD PhD Asso Prof², Lennart Råstam, Prof MD PhD², and Pierre Chauvin, MD, PhD, DSc¹. (1) Research team on social determinants of health and utilization of care, National Institute of Health and Medical Research, INSERM U707, Faculté Saint-Antoine, 27 rue Chaligny, Paris Cedex 12, 75571, France, 33144738443, chaix@u707.jussieu.fr, (2) Department of Community Medicine, Malmö University Hospital, Faculty of Medicine, Lund University, S-205 02, Malmö, Sweden

The multilevel approach used in most studies of neighborhood effects on health may often fail to provide optimal epidemiological information, since it does not incorporate any notion of space. We compared the advantages and drawbacks of cluster-recognition approaches, multilevel models, and hierarchical geostatistical models to gain insight into the spatial distribution of outcomes. We also endeavored to measure contextual factors in more local areas than administrative neighborhoods. Data on all 89,285 individuals ages 40–69 in Malmö, Sweden, 2001, geolocated at their exact residence, were used to investigate (i) mental disorders due to psychoactive substances, (ii) affective disorders, and (iii) neurotic disorders. The spatial scan statistic indicated the existence of a large cluster of disease in the north of Malmö for the three disorders. However, the method of Bayesian Detection of Clusters and Discontinuities showed that affective disorders were more clustered in space than the other disorders. We then used multilevel models to compare the magnitude of neighborhood variations of the three mental health outcomes. Hierarchical geostatistical models, however, provided information on not only the magnitude but the scale of neighborhood variations, indicating that substance-related and affective disorders varied in space on a much larger scale than that of administrative neighborhoods. The prevalence of disorders increased with neighborhood deprivation, but far stronger associations were observed when using indicators measured in circular areas of smaller size than administrative neighborhoods. In many neighborhood studies, viewing space in a continuous way may yield more complete information on the spatial distribution of health outcomes.

Learning Objectives: