Building obesity in Canada: understanding the individual- and neighbourhood-level determinants using a multi-level approach

Submitted: 9 December 2014
Accepted: 9 December 2014
Published: 1 November 2014
Abstract Views: 2477
PDF: 1244
Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.


The objective of this paper was to identify heterogeneities associated with the relationships between the body mass index (BMI) and individual as well as socio-environmental correlates at the individual- and area-levels. The data sources used were: (i) the 2003 Canadian Community Health Survey; (ii) the 2001 Canadian Census; and (iii) the Enhanced Points of Interest (EPOI) database from the Desktop Mapping Technologies Inc. Participants were adults (‰¥20 years; n = 12,836; based on a survey weight scheme Nweighted = 5,418,218) from Toronto and Vancouver census metropolitan areas with no missing BMI records. In addition to conventional 1 km-buffers, we constructed activity-space-buffers to better assess the walkability and potentially increased BMI of individuals. Multi-level analysis was then applied to estimate the relative effects of both individual- and area-level risk-factors for increased BMI. The findings demonstrate a negative association between BMI and energy expenditure, mixed land uses, residential density and average value of dwellings, while a positive association was found with low educational attainment. Relationships were independent of individual characteristics such as age and ethnic- ity. Although the majority of the variation in these outcomes was found to be due to individual-level differences, this study did show significant differences at the area-level as well. The activity-space-buffers presented a vast improvement compared to the conventional 1 km-buffers. The results presented support the rationale that targeting high-risk individuals will only address a portion of the increasing BMI problem; it is essential to also address the characteristics of places that compel indi- viduals to make unhealthy choices.



PlumX Metrics


Download data is not yet available.


How to Cite

Pouliou, T., Elliott, S. J., Paez, A., & Newbold, K. B. (2014). Building obesity in Canada: understanding the individual- and neighbourhood-level determinants using a multi-level approach. Geospatial Health, 9(1), 45–55.

List of Cited By :

Crossref logo

Similar Articles

You may also start an advanced similarity search for this article.