Key points are not available for this paper at this time.
AIM: To investigate the effects of demographic factors and drinking location on contextual drinking in a comparison of US and Canadian adults. DESIGN: Multi-level techniques were used to model the two-level hierarchical structure of drinking contexts (level 1) nested within individuals (level 2). PARTICIPANTS: Two random samples of current drinkers aged 18 years or older were drawn from Canada's Alcohol and Other Drugs Survey (CADS, 1994) and the 1995 National Alcohol Survey (NAS 9). The US sample included 2304 respondents (level 2) who reported a total of 5956 drinking contexts (level 1); in Canada, 5394 respondents reported 13 235 drinking contexts. MEASUREMENTS: Participants reported usual alcohol intake in up to four drinking locations. Demographic data included age, gender, education level, income and marital status. FINDINGS: Significant variation in usual alcohol intake was observed between drinking locations in both the US and Canada. The full multi-level models explained 25% of the variance at the contextual level and 25% and 22% at the individual level in the US and Canada. Contextual drinking was determined by a complex relationship between demographic characteristics and drinking locations. Some interactions between locations and demographic variables were observed for both the US and Canada, whereas others were observed only in the US sample. CONCLUSIONS: There is a need, from a population health perspective, for a multi-level approach in epidemiology and prevention that considers drinking setting as a relevant level for analysis and intervention.
Kairouz et al. (Wed,) studied this question.