A multilevel analysis of the socio-spatial pattern of assault injuries in greater Vancouver, British Columbia.

OBJECTIVES The purposes of this study are to a) determine the extent to which individual and neighbourhood-level socio-economic indicators broadly reflect the social conditions associated with assault injuries within an urban Canadian city, b) examine the significance of this relationship and c) determine if this relationship is best explained at the individual or neighbourhood scale. METHODS Assault-related hospitalization data (2001-2006) were obtained from the British Columbia Trauma Registry (BCTR). Data from the 2001 Census were used as proxy measures of individual and neighbourhood socio-economic status (SES). A generalized hierarchical nonlinear model was used to differentiate between individual and neighbourhood effects. RESULTS A social gradient according to individual and neighbourhood SES and frequency of assault injuries was observed for adults of all ages. After controlling for age and individual SES, probability of greater risk of assault injury among individuals living in progressively less privileged neighbourhoods remained 1.5-3 times higher than individuals living in the least deprived neighbourhoods. For adults under the age of 35, neighbourhood SES was a more statistically significant indicator of increased odds of assault injury than individual income. DISCUSSION Assessing compositional and contextual variations in health outcomes provides health researchers engaged in injury surveillance a way of showing how, and for which type of people, neighbourhood environments influence the likelihood that an individual will be hospitalized due to an intentional injury. This analysis suggests that prevention efforts exclusively focused on the individual may have a limited effect in reducing the occurrence of assault-related injuries, especially among young adults.