Alcohol-Related Harm in the U.S. General Population: A Risk Function Analysis from the 2000 and 2005 National Alcohol Surveys

To assess the relationship of alcohol use and alcohol-related harms during the last 12-months, risk curves were developed using CHAID (Chi-Squared Automatic Interaction Detection) analysis for current drinkers by gender in two national datasets. Data for analysis involve 8,877 respondents reporting at least one drink in the last 12 months, in a pooled sample from the 2000 and 2005 National Alcohol Surveys. Alcohol use was measured by average daily volume of consumption (adjusting for drink size); heavy episodic drinking was measured by the frequency of consuming >5 and >8 drinks in a day. Reporting any of 6 alcohol-related harms (friendship/social life, outlook on life, home life/marriage, financial position, work/employment opportunities, and health) was used as the dependent variable. For men, the >8 drinks measure differentiates those who report alcohol-related harms at the lowest consumption level (17% versus 2% at <1 drink/week) and at the highest drinking level (44% vs 29% at >4 drinks/day). For several mid-level volumes, frequency of >5 drinks is significantly associated with reports of any alcohol-related harm (but with various frequencies—yearly, monthly, weekly depending on volume). For women, the >5 drinks measure significantly differentiates reports of alcohol-related problems at all consumption levels except the highest (>4 drinks/day). At this level, 40% of women respondents report an alcohol-related harm. These findings underscore the importance of patterns of alcohol use when examining risks of alcohol-related harm and point to the importance of including multiple ranges of heavier episodic drinking given the differences in patterns between women and men.