OBJECTIVES: This study aimed to examine whether variations among regions in Quebec existed after we controlled for individual characteristics in the prevalence of 1) alcohol, cannabis, and gambling behaviours and 2) substance-related disorders and pathological gambling. METHODS: Using data derived from the Canadian Community Health Survey: Mental Health and Well-Being (CCHS 1.2), we nested 5332 respondents from the province of Quebec within 374 regions equivalent to census subdivisions (CSDs). Outcome variables included 1) drinking status (past 12 months), alcohol consumption (last week), and 12-month diagnosis of alcohol dependence; 2) cannabis use (past 12 months and lifetime) and diagnosis of illicit drug dependence; and 3) gambling status, severity of gambling problems, and number of reported gambling activities (past 12 months). Multilevel regression models with individuals (Level 1) nested in regions (CSDs, Level 2) assessed the variations among regions in the prevalence of various outcomes and disorders when individual characteristics were controlled for. RESULTS: Variance component models revealed that all alcohol-related variables, the prevalence of cannabis use (12 months), and problem gambling did not vary among areas. Gambling rates and the average number of reported gambling activities varied among areas, even when individual-level variables were accounted for in the models, whereas for lifetime cannabis use, variations among areas became nonsignificant. CONCLUSION: Intervention programs may need to address the environment as a relevant determinant of health-related behaviours and lifestyles.
OBJECTIVES: This study aimed to examine whether variations among regions in Quebec existed after we controlled for individual characteristics in the prevalence of 1) alcohol, cannabis, and gambling behaviours and 2) substance-related disorders and pathological gambling. METHODS: Using data derived from the Canadian Community Health Survey: Mental Health and Well-Being (CCHS 1.2), we nested 5332 respondents from the province of Quebec within 374 regions equivalent to census subdivisions (CSDs). Outcome variables included 1) drinking status (past 12 months), alcohol consumption (last week), and 12-month diagnosis of alcohol dependence; 2) cannabis use (past 12 months and lifetime) and diagnosis of illicit drug dependence; and 3) gambling status, severity of gambling problems, and number of reported gambling activities (past 12 months). Multilevel regression models with individuals (Level 1) nested in regions (CSDs, Level 2) assessed the variations among regions in the prevalence of various outcomes and disorders when individual characteristics were controlled for. RESULTS: Variance component models revealed that all alcohol-related variables, the prevalence of cannabis use (12 months), and problem gambling did not vary among areas. Gambling rates and the average number of reported gambling activities varied among areas, even when individual-level variables were accounted for in the models, whereas for lifetime cannabis use, variations among areas became nonsignificant. CONCLUSION: Intervention programs may need to address the environment as a relevant determinant of health-related behaviours and lifestyles.
Authors: Arijit Nandi; Thomas A Glass; Stephen R Cole; Haitao Chu; Sandro Galea; David D Celentano; Gregory D Kirk; David Vlahov; William W Latimer; Shruti H Mehta Journal: Am J Epidemiol Date: 2010-01-21 Impact factor: 4.897