Have the Associations between State Per Capita Alcohol Sales and Self-Reported Alcohol Measures in the United States Changed Over Time?

There is well-recognized value for tracking alcohol consumption and other alcohol behaviors on a population-wide basis for surveillance, epidemiologic, and other purposes. Population-level measures of alcohol use are usually based on per capita consumption estimates (alcohol sales data) or self-reported measures from surveys. Evidence from many countries has shown that alcohol sales data are associated with certain adverse health effects on a population basis (e.g., cirrhosis rates), but they have certain disadvantages, as they cannot distinguish differences in alcohol use among population groups. Self-reported survey data can identify drinking patterns and populations at higher risk, but individuals underreport and underestimate alcohol use. Limited research, particularly from recent years in the U.S., has been conducted comparing state data from alcohol sales with survey data. We correlated state alcohol sales data from the Alcohol Epidemiologic Data System with adult survey data on alcohol measures (consumption; prevalence of current use, heavy use, and binge drinking) from the Behavioral Risk Factor Surveillance System for the years 1993-2006. As expected, alcohol consumption estimates from surveys were substantially lower than those from sales data (range for medians: 22.3% to 31.6% of sales). Nevertheless, state self-reported alcohol measures and sales data were consistently moderate-to-strongly correlated in all years (range: 0.55-0.71 for consumption; 0.57-0.65 for current use; 0.33-0.66 for heavy use; and 0.45-0.61 for binge drinking). These findings demonstrate that despite underreporting, self-reported state alcohol measures are valid for monitoring alcohol use trends over time. Potential explanations and practical implications of these findings will be discussed.