Abstract
OBJECTIVE: To use geographic information systems data and analyses to describe locations and characteristics of tobacco billboards in a large metropolitan area, and to assess the extent to which tobacco companies are locating billboards in close proximity to minority neighbourhoods and schools. DESIGN: Observational study of billboards in a large metropolitan region. SETTING: City and county of St Louis, Missouri. PARTICIPANTS: All stationary billboards in the city and county of St Louis were eligible to be observed, with the exception of bus stop and street side retail advertising signs (for example, cigarette advertising at gas stations). A total of 1239 non-blank billboards were observed. All data were collected in early 1998. MAIN OUTCOME MEASURES: Tobacco and non-tobacco billboard geographic distribution; billboard type and product brand frequencies; and billboard neighbourhood characteristics. RESULTS: Almost 20% of the billboards contained tobacco advertising. Four of the top five and nine of the top 22 brands displayed on billboards were tobacco products. Billboards were located in all areas of St Louis except for the communities with the highest average incomes. Tobacco billboards were more likely to be found in low income areas and areas with a higher percentage of African Americans. Images of African American figures on tobacco billboards were concentrated in the most heavily African American populated regions of the city. Approximately 74% of all billboards in the city of St Louis were within 2000 feet (700 metres) of public school property. CONCLUSIONS: Tobacco products were the single most heavily advertised type of product on billboards in St Louis. The geographic distribution of tobacco billboards, as well as the types of images found on these billboards, is consistent with the hypothesis that tobacco companies are targeting poor and minority communities with their advertising. Methods employing geographic information systems are a powerful technique for examining outdoor tobacco advertising.