Colorado county-level cannabis access showed mixed effects on opioid prescribing

Using county-level variation in recreational dispensary allowances across Colorado, recreational cannabis exposure was associated with significantly fewer 30-day opioid fills (coefficient: -117.6, P<0.01) and inpatient visits (coefficient: -0.8, P=0.03), but not with reductions in total morphine milligram equivalents or ED visits. Counties without prior medical cannabis access experienced greater reductions.

Observational quasi-experimental study using county-level Colorado data (2011-2018) with a differences-in-differences framework. Combined Prescription Drug Monitoring Program data, Colorado Department of Revenue licensing data, and Colorado Hospital Association records across 2,048 county-quarter observations.

By exploiting natural variation in dispensary access at the county level, this study provides more granular evidence than typical state-level analyses, revealing that the cannabis-opioid substitution effect may have limits.

The mixed results suggest that cannabis access may reduce some opioid metrics (number of prescriptions, hospitalizations) without necessarily reducing overall opioid consumption, complicating the narrative that cannabis simply substitutes for opioids.

Colorado-specific findings may not generalize to other states. County-level analysis may miss within-county variation. Cannot determine whether individuals actually used cannabis instead of opioids. Time period may not capture longer-term effects.