Recreational And Medical Cannabis Legalization And Opioid Prescriptions And Mortality

Background: There is mixed evidence when examining the association between cannabis legalization and opioid-related outcomes in the US.

Objectives: To quantify changes in opioid prescriptions and opioid overdose deaths associated with recreational and medical cannabis legalization in the US.

Methods: This is a quasi-experimental, generalized difference-in-differences analysis, using data from the US Centers for Disease Control and Prevention (CDC) and IQVIA Xponent database. Our exposure is recreational and medical cannabis law implementation between January 2006 and December 2020 across US states. For each treated state with cannabis law, we identified control states as those that had not yet implemented the cannabis law and had the same confounding opioid laws as the treated state during the year before cannabis law implementation. The outcome of interest were the number of opioid prescriptions per 100 persons and opioid overdose deaths per 100 000 population. Linear regression and traditional (difference in difference) DD analyses were used to model all outcomes, adjusting for state-level economic indicators, time-fixed state characteristics, and secular changes in outcomes. We did sensitivity analyses by excluding time-varying state-level economic indicators, changing the coding of exposure, accounting for additional opioid laws, narrowing study period, and using event study to examine the pre-policy trend.

Results: Between 2006 and 2020, 13 states legalized recreational cannabis and 23 states legalized medical cannabis. There was no statistically significant association of recreational or medical cannabis laws with opioid prescriptions or overall opioid overdose mortality across the 15-year study period, although the results also suggested a potential reduction in synthetic opioid deaths associated with recreational cannabis laws (4.9 fewer deaths per 100 000 population; 95% CI, −9.49 to −0.30; P = .04). Sensitivity analyses excluding state economic indicators, accounting for additional opioid laws and using alternative ways to code treatment dates yielded substantively similar results, suggesting the absence of statistically significant associations between cannabis laws and the outcomes of interest during the full study period.

Conclusions: The results of this study suggest that, after accounting for biases due to possible heterogeneous effects and simultaneous assessment of recreational and medical cannabis legalization, the implementation of recreational or medical cannabis laws was not associated with opioid prescriptions or opioid mortality, except for a possible reduction in synthetic opioid deaths associated with recreational cannabis law implementation.