Table 3.
Studies investigating the relationship of recreational cannabis legalization and health-related outcomes.
| Author | Year | Location | Date of legalization | Study design | Sample | Brief findings |
|---|---|---|---|---|---|---|
| Emergency service utilization | ||||||
| Baraniecki et al | 2021 | Ontario | 2018 | Retrospective chart review | N = 173 | There was no difference in rate of cannabis intoxication related visits pre to post RCL. RCL was associated with an increase in patients 18 to 29. Post-RCL, the patients needing only observation increased, and the number of patients ordered for bloodwork or imaging decreased. |
| Calcaterra et al | 2019 | Colorado | 2012 | Archival administrative data | N = 38 406 | Rates of cannabis related emergency visits significantly increased from 2009 to 2015. Alcohol related visits also increased, but to less of an extent than cannabis. Cannabis related emergency visits did show an abrupt increase following RCL. |
| Callaghan et al | 2022 | Alberta & Ontario | 2018 | Archival administrative data | N = 230 206 | The rate of emergency department visits with cannabis-induced psychosis did not change pre- to post-RCL. Further, there was no change in admissions with amphetamine or alcohol induces psychosis. |
| Delling et al | 2019 | Colorado, New York, & Oklahoma | 2012 | Archival administrative data | Colorado: N = 2 088 909, New York: N = 11 726 283, Oklahoma: N = 2 334 988 | The rate of change for cannabis diagnoses was greater in Colorado than New York and Oklahoma post-RCL. There were decreased admissions for cannabis abuse in Colorado compared to Oklahoma post-RCL. Healthcare costs and length of patient stay showed no significant difference across state. Colorado also had increased motor vehicle accidents, alcohol abuse, injection overdose injuries, and decreased chronic pain admissions post-RCL compared to both states. |
| Grigorian et al | 2019 | California | 2016 | Archival administrative data | N = 21 173 | Post-RCL also had significantly higher adult trauma activation. Both adults and pediatrics had increased mortality rates post-RCL. |
| Kim et al | 2022 | Ontario | 2018 | Interrupted time series | N = 14 900 820 | Cannabis-related emergency department visits increased for individuals under 65 post-RCL. RCL was associated with immediate visits for men 45 to 64, women 25 to 44, and women 45 to 65. However, RCL was not associated with trend level increases in emergency visits. |
| Masonbrink et al | 2021 | U.S. | Cohort | N = 1 898 432 | RCL was associated with increased adolescent cannabis-related admissions from 2008 to 2019. While there was an increasing trend pre-RCL, the rate of increase in admissions accelerated post-RCL. | |
| Mennis & Stahler | 2020 | Colorado & Washington | 2012 | Archival administrative data | N = 653 232 | Adolescent cannabis treatment admissions rates decreased in both states over time, with steep declines post-RCL. The decrease in admissions for both states was greater than non-legal states but not significantly. |
| Myran et al | 2022 | Ontario | 2018 | Repeated cross-sectional | N = 13 853 396 | Cannabis-related emergency visits in youth and young adults were increasing pre-RCL, but RCL was associated with an immediate spike followed by a monthly attenuation in rate of visits. |
| Myran et al | 2022 | Ontario | 2018 | Repeated cross-sectional | N = 14 375 697 | Rates of cannabis hyperemesis related emergency visits were increasing pre-RCL. Post-RCL there was no significant change in rates of emergency visits, but the increasing trend continued. |
| Myran et al | 2022 | Canada | 2018 | Archival administrative data | N-581 | Children hospital admissions for cannabis poisonings increased 2.6x post-RCL for all provinces examined (British Columbia, Alberta, Ontario, Quebec). |
| Pusateri et al | 2022 | Colorado & Washington | 2012 | Archival administrative data | N = 18 545 | Rates of steroid use and need for total parenteral nutrition in irritable bowel disease patients decreased post-RCL. Total hospital costs in patients also dropped post-RCL. In cannabis users specifically, there was less patients needing total parenteral nutrition and lower hospital costs post-RCL. |
| Roth et al | 2022 | California | 2016 | Archival administrative data | N = 12 108 | Post-RCL monthly cannabis-exposure poisons control calls significant increased. By age, exposures in youth under 13 significant increase post-RCL, but there was no change for those 13+. |
| Sokoya et al | 2018 | Colorado | 2012 | Archival administrative data | N = 2164 | There was no change in number of facial fractures pre to post RCL. Maxillary and skull base fractures were the only type to significantly increase post-RCL. |
| Thomas et al | 2019 | Washington | 2012 | Archival administrative data | N = 161 | The number of unintentional pediatric cannabis exposures per month increased post-RCL. |
| Wang et al | 2018 | Colorado | 2012 | Archival administrative data | N = 4202 | Overall, 67% of adolescent patients had THC positive urine drug screens. The rate of annual cannabis-related visits to emergency care significantly increased over time. Behavioral health evaluations from visits also increased over time. |
| Wang et al | 2022 | Colorado | 2012 | Archival administrative data | N = 262 699 | Cannabis-related pregnancy admissions significantly increased from 2011 to 2018, with spikes in 2012 and 2014. |
| Wang et al | 2017 | Colorado | 2012 | Archival administrative data | N = 7 440 392 | Cannabis related hospitalizations increased over time, with the greatest increases in 2009 and 2014. Visits associated with mental illness were more common in cannabis related visits. Poison control calls remained stable but there were significant increases in 2010. There were increases in calls for those under 17 and over 25 after 2014. Unintentional cannabis exposure increased for those 0 to 8 from 2008 to 2014 and for 9+ year old’s from 2013 to 2015. |
| Wang et al | 2016 | Colorado | 2012 | Archival administrative data | N = 244 | Unintentional cannabis exposure in children increased 2 years post-RCL compared to 2 years pre-RCL. There was also a significant increase in poison control cases over time. This increase was significantly greater compared to the rest of the U.S. |
| Yeung et al | 2021 | Alberta | 2018 | Archival administrative data | N = 1920 | Overall pediatric cannabis-related emergency department visits did not change pre- to post-RCL. For specific age groups rate and proportion of visits for children under 12 increased post-RCL. Emergency visit rates for cannabis and other substances decreased in adolescents 15 to 17. For cannabis co-diagnoses, the proportion of cannabis hyperemesis presentations increased post-RCL in adolescents 15 to 17. Unintentional cannabis ingestion rates did increase post-RCL for children and older adolescents, but not for younger adolescents. |
| Yeung et al | 2020 | Alberta | 2018 | Archival administrative data | N = 14 732 | The volume of cannabis-related emergency department visits and poison control calls increased post-RCL. Cannabis and other substance admissions and co-diagnoses decreased post-RCL. |
| Opioid use | ||||||
| Dranitsaris et al | 2021 | Canada | 2018 | Archival administrative data | Public and private prescription claims | There was a steady decline in volume of opioids prescribed for public and private drug plans. Post-RCL there was a significant spike in the rate of declines (5.4x greater than pre-RCL). |
| Geoffrion et al | 2021 | British Columbia | 2018 | Archival administrative data | N = 3705 | Post-RCL women were less likely to consume opioids and other narcotics. |
| Livingston et al | 2017 | Colorado | 2012 | Interrupted time series | CDC and Prevention WONDER from 2000 to 2015 | There was a significant decrease in opioid-related deaths post-RCL. Even after controlling for trends in comparison states there was still a significant reduction. |
| Lopez et al | 2021 | U.S. | Archival administrative data | N = 144 000 | There was no significant association between RCL and opioid prescriptions by an orthopedic surgeon. RCL states had non-significant increases in daily doses of opioid and hydrocodone prescriptions respectively. | |
| Shi et al | 2019 | U.S. | Archival administrative data | Medicaid State Drug Utilization Data | RCL states had slightly greater, not significantly, Schedule II and III opioid prescriptions compared to medical only states. States with RCL in 2015 to 2017 had reduced Schedule III prescriptions while states with RCL to 2012 had increases. RCL was not associated with number of prescriptions, total doses, or spending of Schedule II opioids. However, RCL in 2015 was associated with the former two and Schedule III spending. | |
| Adverse birth outcomes | ||||||
| Siega-Riz et al | 2020 | Colorado & Washington | 2012 | Archival administrative data | N = 1 347 916 | The rate of small for gestational age births did not change pre to post RCL in both Washington and Colorado. Pre-term births did increase post-RCL but only in Colorado. Congenital anomalies significantly increased for both states pre to post-RCL. |
| Straub et al | 2021 | Washington | 2012 | Archival administrative data | N = 5343 | The prevalence of positive THC screens in women giving birth did not change over time. The prevalence of low-birth-weight births did increase from pre to post-RCL. However, RCL was not associated with small for gestational age births. |
| Mental health outcomes | ||||||
| Callaghan et al | 2022 | Alberta & Ontario | 2018 | Archival administrative data | N = 230 206 | Emergency visits with schizophrenia and related conditions codes did not change pre- to post-RCL. |
| Geoffrion et al | 2021 | British Columbia | 2018 | Archival administrative data | N = 3705 | Post-RCL, women had higher anxiety scores than pre-RCL. |
| Hawke & Henderson | 2021 | Ontario | 2018 | Cohort | N = 269 | There were no significant differences for the pre and post-RCL cohorts for internalizing or externalizing disorders or crime/violence screenings. |
| Rusby et al | 2019 | Oregon | 2014 | Ecological momentary assessment | N = 466 | Cannabis users had higher mood lability scores compared to non-users. RCL had no impact on the association of anxious mood and cannabis use. |
| Vignault et al | 2021 | Quebec | 2018 | Archival administrative data | N = 2615 | Prevalence of psychotic disorders did not differ pre- to post-RCL, but personality disorders and other psychiatric disorders were more prevalent post-RCL. |
| Yeung et al | 2021 | Alberta | 2018 | Archival administrative data | N = 1920 | Personality and mood related co-diagnosis decreased post-RCL for adolescents 15 to 17. |
| Miscellaneous health outcomes | ||||||
| Fedorova et al | 2022 | California | 2016 | Longitudinal | N = 668 | Approximately half of medical cannabis patients remained so from pre- to post-RCL. The most common transition group pre- to post-RCL was out of medical cannabis patient status, followed by never been issued a medical cannabis recommendation, with into medical cannabis patient at the smallest transition group. RCL was the most common reason reported for transitioning out if medical cannabis patient status. |
| Geoffrion et al | 2021 | British Columbia | 2018 | Archival administrative data | N = 3705 | Post-RCL, women had higher pain catastrophizing scores than pre-RCL. Post-RCL women were less likely to consume anti-inflammatories, and nerve medications to treat pelvic pain, but more likely to use herbal pain medication. |
| Jordan et al | 2022 | New Brunswick | 2018 | Retrospective chart review | N = 3060 | The proportion of post-mortem positive cannabis screens did increase from pre- to post-RCL but was not significant following Benjamini-Hochberg correction. The only age group with a significant increase in proportion of positive screens post-RCL was 25-44-year-olds. Those who died post-RCL did have higher odds of cannabis present post-mortem. Tests for cannabinoid detection, did find an increase in positive detection over time, with the steepest increases occurring pre-RCL. There was no change in detection of other drugs. |
Author, Author of article; Year, Publication year of article; Location, Jurisdiction article data was collected in; Date of Legalization, Year legalization was enacted in jurisdiction; Sample, Total N of article sample; CDC, Center for Disease Prevention; WONDER, Wide-Ranging Online Data for Epidemiologic Research; RCL, Recreational Cannabis Legalization.