Cannabis substitution for prescription drugs: understanding longitudinal trends in the era of legal cannabis

Abstract

The effects of changes in cannabis laws continue to be of great interest to researchers, medical professionals, and policymakers as individual US states implement legal cannabis markets and cannabinoids are better studied. Cannabis may serve as a substitute for a number of substances including several classes of prescription drugs. In order to evaluate cannabis-for-prescription drug substitution in the period following legal dispensary opening in Washington State (July 2014-December 2017), we evaluated the effect of increased cannabis demand on opioid dispensing outcomes in individual counties. We instrumented county-quarter level cannabis sales using state-regulated dispensary density, estimating within-county changes in the number of individuals dispensed opioids per quarter. Effects of increasing retail cannabis sales on short- and long-term dispensing were small in magnitude and not statistically significant. Small effects for long-term high dose dispensing (approximately 3% of dispensing) may reflect substitution among the individuals at the highest risk for opioid use disorder; however, we did not find consistent evidence for substitution in the early period during recreational cannabis rollout in Washington. We also evaluated post-cannabis policy change substitution effects for antianxiety and antidepressant medication use in their respective indications among individuals with employer-sponsored insurance within US states (2010-2017). Medical cannabis policy did not lead to measurable changes in the level of use of antianxiety medications or antidepressants within their respective conditions; however, there may be noteworthy state and time-related heterogeneity as evidenced in event studies. In synthetic control models for Colorado and Washington, which legalized recreational cannabis in 2012, the use of antianxiety medications in anxiety and antidepressants in depression was not impacted by recreational cannabis policy change. Together, these findings may reflect little substitution for these classes in the first few years after cannabis policy enactment.