Abstract
Background and Aims
The alcohol purchase task (APT), which presents a scenario and asks participants how many drinks they would purchase and consume at different prices, generates indices of alcohol reward value that have shown robust associations with alcohol-related outcomes in numerous studies. The aim was to test its prospective validity at 4-year follow-up.
Design
Prospective cohort study.
Setting
General population sample of young Swiss men.
Participants
4594 Swiss young men completed baseline questionnaires; among those, 4214 (91.7%) were successfully followed-up 4 years later (mean age [standard deviation] = 21.2 [1.2] and 25.4 [1.3]).
Measurements
Alcohol reward value parameters (i.e., Intensity, the planned consumption when drinks are free; Breakpoint, the price at which consumption would be suppressed; Omax, the maximum alcohol expenditure; Pmax, the price associated with Omax; and Elasticity, the relative change in alcohol consumption as a function of the relative change in price) were derived from the APT at baseline and used to predict self-reported weekly drinking amount, monthly binge drinking, alcohol-related consequences, and DSM-5 alcohol use disorder criteria.
Findings
Regression analyses, adjusting for the baseline alcohol measure, age, linguistic region, and socio-economic indicators showed that Intensity, Breakpoint, Omax, and Elasticity significantly predicted all tested outcomes in the expected direction (e.g., standardized incidence rate ratio [95% confidence interval (CI)] = 1.11 [1.07-1.15], 1.07 [1.03-1.10], 1.08 [1.04-1.11], and 0.92 [0.89-0.95], respectively, for weekly drinking amount, all P < 0.001). Pmax did not significantly predict any outcomes. Non-adjusted correlations, baseline adjusted regression, and ancillary analyses using 1) latent alcohol variables, 2) multiple imputation for missing data, and 3) replications in training and testing subsamples to evaluate predictive accuracy provided consistent findings.
Conclusions
The alcohol purchase task demand curve measures of alcohol reward value are useful in characterizing alcohol-related risk in young men and have long-term predictive utility.