Exponentiated model of drug demand is preferred over exponential models in people with daily/near daily cannabis use.

Abstract

Cannabis use is a growing health concern emphasizing the need to better understand the complexities of drug choice in people with daily/near daily cannabis use. Hypothetical purchasing tasks provide a means to collect data on drug consumption behavior without requiring drug administration and have been used to isolate behavioral economic factors of choice, including facets of drug demand in substance using populations. Various models are used for analyzing hypothetical purchasing task data, but challenges exist in modeling data sets with consumption values of zero. Additionally, a single model or approach may not be best for all commodities and drug classes. This study compared two common demand models (exponential vs. exponentiated) applied to identical hypothetical purchasing task data from 21 (n = 21) individuals with daily/near daily cannabis use. The exponential model was fit using three common levels of replacement values for zero consumption (.1, .01, .001) and compared to the exponentiated model without replacement values. We found that the exponentiated model produced significantly better model fits for individual data, compared to all exponential models. Additionally, significant differences for model derived values of demand elasticity and intensity were found between the exponentiated model and different levels of the exponential model. We conclude that the exponentiated model is preferred over the exponential model for performing demand analysis on hypothetical purchasing task data from individuals with daily/near daily cannabis use.