The findings in this report are subject to several limitations. First, data on alcohol consumption used to calculate indirect estimates of AAF are based on self-reports and may underestimate the true prevalence of excessive alcohol consumption because of underreporting by survey respondents and sampling noncoverage (17). A recent study that used BRFSS data found that self-reports identify only 22% to 32% of presumed alcohol consumption in states on the basis of alcohol sales (18). Second, risk estimates used in ARDI were calculated by using average daily alcohol consumption levels that begin at levels greater than those typically used to define excessive drinking in the United States. Third, deaths among former drinkers, who might have discontinued their drinking because of alcohol-related health problems, are not included in the calculation of AAF, even though some of these deaths might have been alcohol-attributable. Fourth, ARDI does not include estimates of AAD for several causes (eg, tuberculosis, pneumonia, hepatitis C) for which alcohol is believed to be an important risk factor, but for which suitable pooled risk estimates were not available. Fifth, ARDI exclusively
