Introduction
Youth initiation may drive differences in smoking prevalence across Canadian provinces. Provincial differences in initiation relate to tobacco control strategies and public health funding, but have also been attributed to population characteristics. We test this hypothesis by examining the extent to which seven characteristics—immigration, language, family structure, education, income, home ownership and at-school status—explain differences in initiation across provinces.
Methods
We used data from 16 897 youth aged 12 to 17 years in the Canadian Community Health Survey collected from 2015 to 2018. To examine the proportion of provincial differences explained by population characteristics, we compared average marginal effects (AMEs) from partially and fully adjusted models regressing “having ever initiated” on province and other characteristics. We also tested interactions to examine differences in the association between population characteristics and initiation across provinces.
Results
Initiation varied from 4% in British Columbia to 10% in Quebec. Being born in Canada, speaking French, not living in a two-parent household, being in the lowest household income quintile, having parents without postsecondary education, living in rented accommodation and not being in school were each associated with initiation. Taking these results into consideration, the AME of residing in another province compared with Quebec was attenuated by between 3% and 9%. Family structure and household income were more strongly associated with initiation in the Atlantic region and Manitoba, but not in Quebec.
Conclusion
Differences in initiation between Quebec and other provinces are unlikely to be substantially explained by their demographic or socioeconomic composition. Reprioritizing tobacco control and public health funding are likely key in attaining the “tobacco endgame” across provinces.