Study Found No Evidence That Oregon and Washington Drug-Possession Policy Changes Increased Fatal Overdoses
A 2023 study published in JAMA Psychiatry found no evidence that changes reducing criminal penalties for drug possession in Oregon and Washington were associated with higher fatal overdose rates during the first year after implementation.
Researchers analyzed monthly overdose mortality data from January 2018 through March 2022. They used a synthetic-control method, constructing comparison states that matched Oregon and Washington’s pre-policy overdose trends and then estimating how those states might have fared without the legal changes.
Oregon’s Measure 110 took effect on February 1, 2021, changing possession of small amounts of previously criminalized drugs from a criminal offense to a Class E violation punishable by a maximum $100 fine. The fine could be waived if the person completed a health assessment. Measure 110 also directed cannabis-tax revenue toward treatment, harm-reduction and recovery services.
Washington’s policy shift began with a February 2021 state Supreme Court decision, State v. Blake, which invalidated the state’s drug-possession statute. The Legislature later replaced the felony penalty with a misdemeanor and required law-enforcement officers to offer referrals to assessment or treatment before making many low-level possession arrests.
The estimated difference in average monthly fatal-overdose rates between Oregon and its synthetic control was 0.268 deaths per 100,000 people. In Washington, the difference was 0.112 deaths per 100,000. Neither result was statistically significant, meaning the study did not detect a reliable increase or decrease attributable to the policy changes.
That conclusion is narrower than saying decriminalization improved public health. The research was observational, covered approximately one year after implementation and could not establish what would have happened under every possible policy alternative. The authors also noted that Oregon’s treatment and recovery funding was distributed slowly, with most of the approximately $287 million in additional awards reaching providers after the study period. The analysis did not evaluate nonfatal overdoses or outcomes by race, ethnicity or socioeconomic status.
The findings were consistent with a separate study of arrest patterns in Oregon and Washington, which reported substantial declines in drug-possession arrests and found no evidence of a corresponding increase in arrests for violent crimes. Together, the studies suggest that reducing criminal penalties did not produce the predicted short-term rise in overdose deaths or violent-crime arrests, but they do not resolve broader questions about treatment access, public drug use, community safety or long-term health outcomes.
Fentanyl further complicates comparisons between the states. Overdose deaths were rising across the country as fentanyl spread through unregulated drug markets, and the timing of that spread differed by region. A later JAMA Network Open analysis of Oregon that accounted for fentanyl’s rapid expansion also found no association between Measure 110 and fatal overdose rates.
The policy landscape has since changed. Oregon’s Measure 110 program was substantially revised by House Bill 4002, which restored criminal penalties for possession of small amounts of controlled substances effective September 1, 2024, while creating deflection and treatment pathways. A 2025 state audit later cited problems with stability, coordination and clear results in the treatment system. Those developments fall outside the original study period and cannot be used to alter its findings.
The study’s central takeaway remains limited but important: in Oregon and Washington, the first year after removing or reducing criminal penalties for low-level drug possession was not associated with a statistically significant change in fatal overdose rates. Longer-term research is needed to assess how these policies affect treatment engagement, nonfatal overdoses, racial disparities, public safety and mortality after the programs and their surrounding systems have had more time to develop.
