Assessment methods for alcohol consumption, prevalence of high risk drinking and harm: a sensitivity analysis

Background There are no standardized ways to assess alcohol consumption in epidemiological studies. The main objective of the present study was to compare three widely used methods for assessing alcohol consumption with respect to resulting prevalence estimates for high risk drinking and harm as defined by morbidity and mortality indicators. Methods A within-subjects design was used to compare a quantity frequency, a graduated frequency, and a weekly drinking recall measure. Data consisted of a representative sample of 3961 adult residents of the province of Ontario, Canada, who participated in a multi-wave cross-sectional survey between 1990-1994. Cross-tabulation, Spearman correlation, and standard methodologies for prevalence-based cost-of-illness studies were used. Results The graduated frequency measure consistently yielded higher estimates of the prevalences of high risk drinking and harm. Differences were marked on all indicators, but were most pronounced for harmful drinking as defined by consuming an average of >60 g pure alcohol per day for males, and >40 g per day for females. Prevalence estimates of harmful drinking were almost five times higher for graduated frequency versus weekly drinking measures, and almost three times higher for graduated frequency versus quantity frequency measures. Conclusions The characteristics of different measures of alcohol consumption should be considered in future research in epidemiology.