Utilizing the commodity purchase task to evaluate behavioral economic demand for illicit substances: a review and meta-analysis

Background andAims: The commodity purchase task is a simulated demand procedure that is easy and quick to complete (< 5 minutes) as well as adaptable for remote delivery and use with varied study populations. The purpose of this meta-analysis was to synthesize research using the commodity purchase task with illicit substances to evaluate the magnitude of omnibus effects sizes and moderators of the correlation of demand indices with quantity–frequency (QF) and severity measures. Design: Random-effects meta-analyses and meta-regressions involving studies with cross-sectional correlational designs. Setting and Participants: Eleven studies, 10 outcomes and 2146 participants from two countries (USA and UK) published up to 1 October 2018. Measurements: Omnibus effect sizes (correlation coefficients) of five demand indices from the commodity purchase task [intensity (unconstrained consumption), elasticity (price sensitivity), O_max (maximum expenditure), P_max (price at maximum expenditure) and breakpoint (first price of zero consumption)] with QF and severity measures. Meta-regression models tested moderators of effect sizes (i.e. sample age and sex composition, commodity type and number of prices used in the commodity purchase task). Findings: Significant omnibus effect sizes were observed with QF and severity measures for intensity (r = 0.32/0.28, QF/severity, respectively), elasticity (r = −0.14/−0.18), O_max (r = 0.30/0.29) and breakpoint (r = 0.17/0.22) values. P_max was only significantly associated with severity measures (r = 0.15). The percentage of female participants and number of prices used in the purchase task significantly moderated P_max and breakpoint effect-size estimates in that stronger associations were observed in samples with a greater percentage of women and in studies using tasks with more price points. Commodity type (cannabis versus cocaine) did not significantly moderate associations involving any demand index. Conclusions: Behavioral economic demand as measured by the commodity purchase task is consistently correlated with measures of illicit substance use quantity–frequency and severity.