An improved approach for geocoding Canadian postal code-based data in health-related studies

摘要

Due to Canadian privacy and confidentiality laws, high precision addresses must be geocoded to coarser geographies such as postal codes, or randomized to different locations. This study introduces an enhanced postal code geocoding approach that improves upon the traditional approach used by Statistics Canada by considering the land use and spatial distribution of populations within postal code boundaries. The proposed and traditional postal code geocoding approaches were compared based on their distance proximity to residential locations using two study areas: the province of Ontario and the city of Kingston. The Wilcoxon signed-rank test for paired samples was performed to compare the distance measures between the two approaches for the urban and rural areas. Results showed that the proposed geocoding approach has a relatively higher positional accuracy than the traditional approach. On average, the postal code locations of the proposed approach were in closer proximity to residential areas by about 25 m in urban areas and 300 m in rural areas for the province of Ontario, and by 15 m at the urban and 70 m in rural areas for the city of Kingston. The improved method enabling an increased level of geocoding precision can be used to facilitate spatial analysis at a larger scale without sacrificing individual confidentiality in population-based health studies and other applications.