Large U.S. Claims Study Found No Overall Rise in Psychosis Diagnoses After State Cannabis Legalization
A large analysis of U.S. health insurance claims found no statistically significant overall increase in psychosis-related diagnoses or antipsychotic prescriptions after states adopted medical or recreational cannabis policies.
The study, published in JAMA Network Open in 2023, examined records for 63,680,589 beneficiaries aged 16 and older from 2003 through 2017. The database included commercial and Medicare Advantage insurance claims from all 50 states and the District of Columbia, covering more than 2 billion person-months of follow-up.
Researchers compared monthly rates of psychosis-related diagnoses and filled prescriptions for antipsychotic medications as states moved from having no cannabis policy to adopting medical or recreational legalization. Their models accounted for differences between states, changes over time, seasonal patterns and several time-varying state characteristics. The analysis also distinguished between policies that permitted retail cannabis sales and those that did not.
Overall, the researchers found no statistically significant increase in either outcome associated with medical or recreational cannabis policies. Compared with states without a legalization policy, states with recreational policies and retail outlets had an estimated 39% higher rate of psychosis-related diagnoses, but the result did not meet the study’s threshold for statistical significance. The corresponding estimate for antipsychotic prescriptions was a 14% increase, also not statistically significant.
Exploratory subgroup analyses produced different results. Psychosis-related diagnoses were significantly higher among men, adults ages 55 to 64 and Asian beneficiaries in states with recreational cannabis policies. These analyses were not the study’s primary outcome, and the authors said the findings require further investigation. No statistically significant association was found between cannabis policy and antipsychotic prescriptions in the subgroup analyses.
The findings do not show that cannabis use is harmless or that cannabis cannot contribute to psychosis in some individuals. The Centers for Disease Control and Prevention notes that cannabis use is associated with psychosis and schizophrenia, particularly among people who begin using cannabis earlier or use it more frequently. The JAMA Network Open study addressed a narrower question: whether changes in state cannabis policy were followed by measurable changes in psychosis-related health care claims.
The study also has important limitations. Its data came from insured beneficiaries and may not represent people without insurance or those whose symptoms never led to medical care. Claims data cannot reliably distinguish newly developed psychotic disorders from conditions diagnosed before a person entered the database. In addition, the study period ended in 2017, when recreational cannabis markets were less established and retail sales were available in relatively few state-months.
Because the research was observational, it cannot establish that legalization causes—or prevents—psychosis. The authors also noted that unmeasured differences between states, including patterns of substance use, health care access and population movement, could have influenced the results. Their conclusion was therefore measured: the analysis did not support an overall association between state cannabis legalization and psychosis-related health outcomes, but continued research—especially using more recent data and better measures of cannabis exposure—is warranted.
