Andrea Di Rocco

This research examines the spatial distribution of fatal injuries in Ontario, with a focus on the Toronto Census Metropolitan Area (CMA). The Ontario Coroner's database from 2000 to 2004 was the source for injury data. Emphasis was placed upon the death rates from falls, homicides, motor vehicle collisions, and suicides.

The dataset underwent a detailed descriptive analysis by mechanism of injury mortality. Multiple regression models to predict death rate using socio-economic and demographic variables from the 2001 Canadian census were created for the four types of injuries yielding fairly weak models. The best model was the prediction of death rate due to falls using 11 variables and explaining about 30% of the variation in the death rate.

The kernel density function was used to determine areas of high density for each of the five death rates in the Toronto CMA. Global Moran's I revealed a very weak positive spatial autocorrelation. Local indicators of spatial association (LISA) generated specific areas of clustering amongst the death rate data - both on their own and when paired with socio-economic and demographic variables. Although the overall spatial pattern of fatal injuries within the Toronto CMA did not reveal strong spatial association, further investigation into its spatial clusters is needed to confirm the link between where people live and their risk to certain fatal injuries.