Retail alcohol availablity and youth drinking -- a critical assessment

Background: Alcohol outlets dictate many aspects of the sale of alcohol and may affect drinking behaviors. Limits on alcohol outlets could curtail consumption by limiting physical access to alcohol, increasing the cost of alcohol, limiting exposure to alcohol marketing, and affecting social aggregation in and around alcohol outlets. A recent systematic review found widely discrepant measures of alcohol outlet density, making it difficult to summarize effect sizes and rule out modifiable areal unit problem. Moreover, few of the existing studies are national in scope, making regional differences another likely contributor to the observed heterogeneity. Methods: We developed a national dataset from geocoded addresses of alcohol outlets from a commercial business information service. We calculated 21 different measures of alcohol outlet density using circular buffers, ZIP code areas, Census polygons and kernel density estimation techniques as well as Euclidian and network proximity around the homes of a national sample of adolescents (N = 3618) for whom drinking behaviors and covariates were assessed. Results: Correlation between the density and proximity varied widely. After controlling for individual level factors (friend use of alcohol, sensation seeking, age, race/ethnicity, movie exposure to alcohol depictions, parent monitoring and average household size), none of the measures of alcohol availability were significantly associated with binge drinking. We plan to conduct a sensitivity analysis using 2010 NAICS data in which we validate our business list data on alcohol outlets. Conclusion: Preliminary results suggest that alcohol outlet density and proximity are not associated with adolescent binge drinking.