Research suggests that cannabis use is associated with potential therapeutic benefits but also individual health harms, particularly for adolescent populations.
Evidence is inconsistent and, in some cases, inconclusive regarding how recreational cannabis legalization affects outcomes significant to public health, including cannabis use, cannabis use disorder, cannabis-related hospitalizations and poisonings, driving safety, and other substance use.
Researchers can improve on existing evidence by targeting understudied areas, including features of legal cannabis markets, heterogeneous policy effects across populations, and characteristics of cannabis supply such as price, potency, and product type.
In the face of uncertainty regarding the health consequences of legalization, policy makers should adopt regulatory designs informed by public health priorities and work with researchers to evaluate policy effects.
States are adopting policies that expand cannabis access to larger proportions of the US population than ever before, as described in an accompanying Health Affairs Health Policy Brief. Past-year cannabis use increased from 10.4 percent of US adults in 2002 to 15.3 percent in 2017, and the proportion of past-year users reporting near daily use doubled between 2006 and 2016. During that same period, perceived great risk from smoking cannabis declined among those ages twelve and older, going from 38.3 percent in 2002 to 26.1 percent in 2017. It is important to consider whether perceptions about risk are accurate and what may be the larger population health benefits and harms associated with expanded cannabis access.
In this brief, we summarize research on how cannabis legalization relates to use of the substance and key population health outcomes. We focus on recreational laws because of their broad application and ability to affect public health, although we include some studies on medical cannabis laws with generalizable or relevant findings. A detailed summary of representative studies in these domains is provided in supplemental appendix tables 1–7. We conclude by outlining areas ripe for future research and policy consideration related to safely legalizing cannabis.
Cannabis Use And Health
Research suggests that cannabis use is associated with both positive and negative health effects. The term cannabis refers to parts of or products derived from the plant Cannabis sativa that contain substantial amounts of tetrahydrocannabinol (THC)—the substance primarily responsible for cannabis’s high-inducing effect. Although it is claimed that cannabis and its derivative substances improve outcomes for many disorders, there is only strong scientific evidence suggesting that it is an effective treatment of three conditions: in treating chronic neuropathic pain in adults, as an antiemetic after chemotherapy treatment, and for improving patient-reported multiple sclerosis symptoms.
Demonstrated adverse effects of short-term cannabis use include impaired short-term memory, altered judgement that increases engagement in risky behaviors, and impaired driving. Heavy and long-term cannabis use in adolescents carries substantial risks, including altered brain development and cannabis dependence, which is correlated with elevated risk of using other illegal drugs. Emerging literature suggests that higher-potency cannabis—with a greater THC concentration—may intensify cognitive impairment, severity of dependence, and adverse psychological outcomes. Although cannabis use has been associated with poor educational outcomes and mental illness, it is challenging to attribute causality to these complex, multifactorial outcomes.
The mechanism of consumption likely affects cannabis’ health effects. Although large population studies have not identified an association between cannabis smoking and lung cancer, the link cannot yet be ruled out, as cannabis smoke contains carcinogens.
Effects of Cannabis Legalization
Evidence is emerging on the public health impacts of cannabis legalization. Here we summarize current research regarding the relationship between recreational and, to a lesser degree, medical cannabis legalization and various outcomes that is key to understanding the public health policy implications. We also highlight three areas ripe for future research: additional measures of cannabis use, use disorder, and product type; heterogenous policy design; and differential effects by population subgroups.
Cannabis use is best operationalized through measures of both prevalence (past-month or past-year use) and intensity (for example, number of days used, total grams consumed, potency per dose). Findings on the relationship between recreational legalization and cannabis use among adults are inconclusive, and effects may differ by age group. Using a large, nationally representative survey across several years, two studies observed increased prevalence of past-month cannabis use and frequent use for adults ages twenty-six and older in states with recreational legalization but did not report similar changes among young adults ages 18–25. Other research has identified increases in cannabis use prevalence and intensity among college students who fall within the 18–25 age range. Evidence on the impact of legalization on youth cannabis use remains inconclusive, with research identifying increases, decreases, and no change in use prevalence and intensity measures.
As Rosanna Smart and Rosalie Pacula note, medical cannabis laws, which apply to more limited sectors of the population than do recreational laws, have not been associated with increases in the prevalence of youth (ages younger than eighteen) cannabis use. Although these laws appear to correlate with increased use among adults, subgroup analyses suggest that the evidence remains mixed for young adults ages 18–25. Some studies find that medical legalization has no effect on past-month use or use intensity for young adults, although Christine Mauro and colleagues observed increases in use intensity for males in this age range.
Notably, product characteristics of the legal supply affect the relationship between legalization and cannabis use and use intensity. Early research describes the evolution of product type, potency, and price after legalization in Washington State. Similar research is needed on the legal medical and recreational markets in other jurisdictions to capture how cannabis market characteristics affect consumption patterns.
CANNABIS USE DISORDER
Research on the effects of cannabis legalization on cannabis use disorder is relatively nascent, offering inconclusive findings and suggesting that effects may differ by age group. As Smart and Pacula summarize, medical legalization increases, decreases, or has no effect on self-reported prevalence of or treatment admissions for cannabis use disorder. The few studies that consider heterogenous policy effects of medical cannabis laws suggest that the presence of commercial dispensaries increases both overall and youth cannabis use disorder treatment admissions.
This literature is inconclusive partly because the common outcomes used to measure cannabis use disorder—self-reported symptoms or treatment admissions—are likely influenced by legalization without actually changing prevalence of cannabis use disorder. For instance, by changing social norms around problematic cannabis use, legalization may reduce the likelihood that an individual will self-report symptoms of cannabis use disorder. Similarly, given that treatment admissions for cannabis use disorder often occur through the criminal justice system, legalization may affect written and de facto policies governing law enforcement treatment referrals for cannabis use disorder with or without affecting cannabis use disorder prevalence.
CANNABIS-RELATED HOSPITALIZATIONS AND POISONINGS
Cannabis legalization may result in increases in hospitalization and emergency department visits related to cannabis abuse and dependence and injuries occurring under the influence of the substance. A 2020 narrative review reported that cannabis-related hospitalizations in Colorado increased after recreational legalization, above and beyond earlier additions associated with medical legalization. A rigorously designed study found that the presence of recreational cannabis dispensaries, but not enactment of a recreational cannabis law, is statistically positively associated with poisonings involving cannabis dry plant products overall and in those younger than twenty-one. However, because of data limitations, this study fails to consider exposures to other cannabis product forms that are of particular concern for youth, such as edibles.
Potential increases in car accidents involving cannabis use are a chief concern among those disfavoring legalization. Simulation studies suggest that cannabis intoxication impairs driver reaction time, spatial perceptions, and decision making. Detection of cannabis in drivers has tripled from 4.2 percent of fatally injured drivers in 1999 to 12.2 percent in 2010, although it remains unclear how much of this increase can be attributed to cannabis policy liberalization.
Research investigating the relationship between cannabis legalization and driving safety typically leverages fatal crash data. However, less than 0.5 percent of crashes are fatal, so research using data sets that fail to capture nonfatal injuries underreport traffic accidents associated with driving under the influence of cannabis.
Even among studies that use fatality data and consider policy heterogeneity, findings vary from significant positive to significant negative to insignificant relationships between medical cannabis laws and traffic fatalities. Research on recreational cannabis laws is similarly mixed, likely as a result of methodological differences and confounding variables.
USE OF OTHER SUBSTANCES
The extent to which cannabis interacts with other substances heavily influences the public health implications of legalization. In particular, whether cannabis is a complement or substitute to alcohol, tobacco, or opioids is a critical consideration for legalization policy. The effect of legalization on substance use likely differs by substance and user age.
Overall, the literature on the effects of cannabis laws on alcohol use remains mixed for both adolescents and adults, suggesting both substitution and complementary relationships. Studies of recreational cannabis laws have identified both significant declines and insignificant effects on alcohol use with cannabis and without cannabis, as well as intensity of alcohol use. The emerging literature on tobacco is similarly inconclusive but has begun to assess differential effects of legalization across tobacco products. For example, Rebekah Coley and colleagues find that recreational legalization increases e-cigarette use among adolescents but has no effect on cigarette use. Early evidence on medical cannabis laws suggests that the complementary or substitutive nature of the relationship between alcohol or tobacco and cannabis depends on policy restrictiveness and the age of the consuming population.
Research has also focused on the potential for cannabis legalization to address opioid-related harms. A growing minority of states recognize opioid dependence as a qualifying condition for medical cannabis access. Although older studies identified a negative association between medical cannabis laws and opioid mortality, subsequent replications using additional data years have suggested that omitted variable bias may have driven earlier findings. More recent research highlights the importance of considering policy dimensions in evaluations of legalization policies. For example, the existence of dispensaries in medical or recreational legal markets, not legalization alone, may contribute to any observed effect on opioid mortality. Although researchers find (with important exceptions) that opioid prescribing is negatively associated with medical legalization or certain features of medical cannabis markets, this literature may inadequately control for changes in public and private policies and programs targeting opioid misuse.
For more detailed information about legalization and noncannabis substance use, we refer readers to reviews from Gabrielle Campbell and colleagues, Smart and Pacula, and Meenakshi Subbaraman.
Future Research Directions
Although more research is needed regarding the public health consequences of a legal cannabis supply for adults, we have elucidated what is known about four key health outcomes: cannabis use and use disorder, cannabis-related hospitalizations and poisonings, driving safety, and other substance use. Expansion of cannabis research into understudied areas may help address some of the existing inconsistent evidence.
For instance, future research should evaluate specific provisions that govern the legal market. Features of medical and recreational cannabis laws, such as whether a jurisdiction allows dispensaries, likely affect health outcomes and explain discrepancies between studies that do not stratify states by relevant provisions. Although some studies assess the effect of recreational dispensaries on cannabis use, unintentional cannabis exposure, motor vehicle fatalities, and opioid mortality, differentiating states by provisions is not as common in recreational legalization research compared with the more established literature on medical cannabis laws. Further, many studies only evaluate provisions related to legal supply methods (for example, dispensaries, home cultivation), ignoring other elements of the regulatory framework that may have significant implications for public health outcomes. These include provisions such as taxation and advertising restrictions. Evaluations of provisions that govern legal markets could benefit from a categorization system that differentiates states on the basis of multiple characteristics. In addition, analyses should distinguish between policy implementation and effective dates.
Researchers should also consider the public health effects of different characteristics of the legal cannabis supply, including product type, price, potency, and sourcing. Cannabis legalization, particularly commercialization, has the potential to transform the cannabis market. For instance, there has been a proliferation of high-potency products in legal cannabis markets, both in the US and abroad. Novel extracts also make up a rapidly growing market segment. Given public perceptions that alternatives to smoked cannabis products are healthier and more efficient to consume, it is critical to assess their public health effects. Use will also be shaped by the post-tax retail price, and as anticipated, cannabis prices have fallen steeply in states that have legalized recreational use. Thus, researchers should prioritize evaluating the effects of different taxation approaches (for example, ad valorem or based on THC content) on use.
Further, the 2020 vaping crisis associated with diluted THC-containing products raises concerns around unregulated cannabis markets in both legalized and nonlegalized states, heightening the need to study the public health effects of legally versus illegally produced cannabis products. Notably, cannabis markets vary at the local level, and the cannabis industry will continue to evolve in response to federal and state policy changes. Thus, establishing data systems capable of capturing local characteristics of cannabis markets over time will be essential.
It is also critical that researchers expand the study of heterogeneous policy effects across population groups. Our review highlights several studies that evaluate differential effects by age. Researchers should also build on the notable research studying effects by race and ethnicity to investigate why legalization may exert differential effects on cannabis and other substance use for different populations. Cannabis legalization is promoted as a tool to advance social equity, yet legalization has the potential to create or exacerbate socioeconomic and health inequities. Having a criminal record has implications both for health and for economic well-being. Although declines in adult arrests follow cannabis legalization, racial disparities in arrest rates persist, and more research is needed on how the financial and remaining criminal penalties for cannabis violations are distributed across demographic groups. Whether health and other harms associated with cannabis-related criminal justice contact continue to be concentrated among racial and ethnic minorities after legalization deserves careful study.
Beyond criminal justice contact, the design of cannabis markets has health equity implications. Low-income and racial/ethnic minority communities, as well as youth, are especially vulnerable to the commercialization of cannabis. For instance there is some evidence that cannabis dispensaries are concentrating in minority communities. Researchers should prioritize an examination of the differential impact of attributes of legalized markets—including taxation mechanisms, location of cannabis outlets, marketing strategies, product design, and potency—
by social class, race, ethnicity, and age. Further, recent controversies over the effectiveness of social equity provisions embedded within cannabis policy reform call for researchers to evaluate whether these initiatives are achieving their intended effects or producing new inequities.
Even if researchers fill these gaps in the literature, data and methodological challenges limit our ability to draw conclusions about other impacts of legalization. For instance, research could be strengthened by adoption of a clear, consistent, and expansive definition of cannabis “use” across data sources. As Beau Kilmer and Rosalie Pacula highlight, research must move beyond measures of prevalence to include measures of use intensity, such as days used, total grams consumed, and potency per dose. Although researchers should work to develop standardized definitions for and methods to collect dosage information that is comparable across products, it is important to acknowledge that these measurement improvements may add significant costs to study designs and must be weighed against feasibility.
Another important area for methodological and data improvements is in understanding how cannabis legalization relates to driving safety. There exist challenges to reliably measuring THC impairment in drivers (for example, in blood), and uncertainty remains over what level of THC in the blood leads to impairment. In addition, future research should clarify how alcohol and cannabis co-use affects driving impairment. Improved data sources that include information on crashes and fatalities by substance type are essential to fully assessing the impact of legalization on driving safety.
As policy makers design or revise legal markets, it is important to note that research on early recreational legalization adopters may not be generalizable to all states and localities. The states subject to the most research so far (Colorado, Oregon, and Washington) all had expansive medical programs established before recreational legalization and mostly adopted commercial approaches toward the recreational market. It is unclear whether states with less commercialized medical programs or that adopt different supply chain architectures for recreational supply will experience similar effects on use and public health outcomes.
Nevertheless, there are opportunities for policy makers to incorporate a public health perspective in the design of legal cannabis markets. These include suggestions based on lessons learned from tobacco and alcohol policy and recommendations that focus on youth cannabis use. Other important policy goals may include minimizing drugged driving, unwanted contaminants, and co-substance use. Policy makers may also consider adopting alternative legalization models beyond commercial markets and should thoughtfully incorporate social equity considerations into legalization design and oversight through provisions that address socioeconomic and health disparities. Finally, policy makers should work with researchers to evaluate the impacts of their own legalization schemes as they unfold.
(Editor’s Note: This article was conceived and drafted when Haffajee was employed at the RAND Corporation, and the findings and views in this article do not necessarily reflect the official views or policy of her current employer, the US Department of Health and Human Services, nor the US government.)