Recreational marijuana policies have mixed effects on opioid use, may affect other substance use patterns
Most states have adopted marijuana policies that allow for medicinal and/or recreational use. Some prior research suggests such policies may increase marijuana use as a substitute for other more immediately harmful substances such as alcohol, opioids, and stimulants. In this study, researchers examined the relationship between recreational marijuana laws and subsequent substance use patterns, with a focus on prescription opioid misuse.
Permissive marijuana policies are becoming more common across the United States. As of November 2022, 21 states have adopted recreational marijuana laws, where the possession of small amounts of marijuana by adults aged 21 and older is legal, and most states have adopted medicinal marijuana laws. Most states that have legalized recreational marijuana use also have retail outlet dispensaries where the substance can be purchased. Previous research has suggested that the adoption of recreational and medicinal marijuana laws is associated with changes in substance use. For example, recreational marijuana laws are associated with substantial increases in marijuana use, including among adolescents, whereas medical marijuana laws are associated with small increases in marijuana use, with no effect among adolescents.
Recent research has also highlighted the potential for more permissive marijuana policies to help address the opioid crisis, although findings are mixed. The theory here is that marijuana can help substitute for more immediately harmful substances, including opioids and alcohol. Early adoption of medical marijuana laws (1999-2010) was associated with reduced opioid-related mortality although extending the study period to 2017 found that, over time, the trend reverses and these laws are associated with increased opioid-related mortality. The research on recreational marijuana laws is also mixed, with some studies suggesting a protective effect on opioid-related mortality and other studies suggesting a detrimental effect. More permissive marijuana laws have been associated with lower opioid prescribing, highlighting a potential mechanism for lower opioid-related harm.In addition, since most opioid-related overdose deaths involve non-opioid substances, if more permissive marijuana laws decrease the co-use of a contributing substance, such as alcohol, then they may indirectly decrease opioid-related mortality. Research can help characterize the potential short-term and long-term impacts – whether helpful or harmful – of marijuana policy on substance use in the US.
In this study, researchers used advanced statistical techniques to address many of the methodological limitations in prior studies to examine the relationship between both recreational and medical marijuana laws and subsequent substance use patterns, with a focus on prescription opioid misuse. Sound scientific inquiry into the impact of permissive marijuana policies is paramount to inform policymakers on potential solutions to the opioid crisis as well as prepare prevention and treatment professionals for changing substance use patterns.
HOW WAS THIS STUDY CONDUCTED?
This study examined the relationship between both medical and recreational marijuana laws and subsequent substance use patterns, including prescription opioid misuse, among adults aged 19 and older using several advanced statistical techniques. Substance use trend data from 2004 to 2020 were obtained at the state level from the National Survey on Drug Use and Health. This is an annual survey conducted by the Substance Abuse and Mental Health Services Administration to provide nationally representative data on prevalence and trends for a wide range of substance use, including alcohol, marijuana, and heroin, as well as misuse of prescription opioids and stimulants. Restricted data on these substance use and misuse trends are also available at the state level. The researchers used the variation of state adoption of medical and recreational marijuana laws to examine how substance use and misuse trends changed in states that implemented these laws compared with states that did not implement these laws.
The researchers were interested in the relationship between both medical and recreational marijuana laws and five types of substance use: marijuana use, alcohol use, heroin use, stimulant use, and prescription opioid misuse. In addition to the prevalence and trends in the use of these substances, the degree of substance use was measured (i.e., intensity), which included use in the past year, use in the past month, frequency of use in the past year, and if the individual had a substance use disorder.
Researchers defined a “frequent user” as reporting using the substance at least once per week, on average (i.e., 52 days or more per year). Due to a change in the way that the National Survey on Drug Use and Health measured prescription opioid misuse beginning in 2015, the researchers had to use two mutually exclusive indicators to measure prescription opioid misuse before and after this survey change.
Researchers used two alternative indicators for recreational marijuana laws. The first indicator measured whether the state had adopted a recreational marijuana law by the end of 2020. Twelve states and the District of Columbia had adopted these laws at that time. The second indicator measured whether the state allowed recreational sales of marijuana. Ten states and the District of Columbia were in that category. The researchers also used an indicator for the presence of medical marijuana laws in their statistical models. By the end of 2020, 37 states and the District of Columbia had adopted these laws.
The researchers used advanced statistical modeling techniques and several different models which accounted for various differences between states at both the individual level and the population level. These included a standard difference-in-difference model, a model with an indicator for whether states allow recreational sales of marijuana rather than just having a recreational marijuana law, a model with statistical corrections to account for dynamic effects due to policies being adopted at different points in time, and an event study to capture change over time. Individual-level characteristics controlled for in the model included age, gender, educational attainment, race, ethnicity, marital status, income, and insurance coverage status. Population-level characteristics included a state’s unemployment rate, whether that state had a naloxone (the overdose reversal drug, commonly known as “Narcan”) access law, and whether the state had implemented a prescription drug monitoring program. The researchers also included a statistical technique to capture differences between states and over time that were not characteristics mentioned above (called fixed effects).
WHAT DID THIS STUDY FIND?
Recreational marijuana laws, but not medical marijuana laws, were associated with increased marijuana use.
Recreational marijuana laws were associated with an increase in the probability of past-year marijuana use, an increase in the probability of past-month marijuana use, and an increase in the frequency of past-year use. However, these laws were not associated with an increase in marijuana use disorder. Medical marijuana laws were not associated with any of the measures of marijuana use.
Medical marijuana laws were associated with decreased alcohol use while evidence for recreational marijuana laws was mixed.
Medical marijuana laws were associated with decreases in the probabilities of past-year alcohol use, past-month alcohol use, and frequent (i.e., weekly or greater) alcohol use. In contrast, recreational marijuana laws were associated with a decreased probability of alcohol use disorder and an increased frequency of alcohol use in the past year.
Both medical and recreational laws were associated with increased stimulant use.
Medical marijuana laws were associated with increases in the probabilities of past-year stimulant use, past-month stimulant use, and frequent stimulant users along with increased frequency of stimulant use in the past month. Recreational marijuana laws were associated with an increased probability of frequent stimulant use and an increased frequency of stimulant use in the past year.
Recreational marijuana laws were associated with some measures of decreased prescription opioid misuse and some measures of increased heroin use in the standard model.
Recreational marijuana laws were associated with both decreased probabilities of frequent prescription opioid use and having a prescription opioid use disorder during the 2004 to 2014 study period, but these measures were not statistically significant in the subsequent study period. Recreational marijuana laws were associated with a decreased probability of heroin use disorder and an increased frequency of heroin use in the past year.
Figure 1. Effects of marijuana laws on non-medical opioid use and heroin use outcomes. Overall effect of recreational marijuana laws (Column 1), overall effect of recreational marijuana sales, meaning when actual commercial sales began, not law passage (Column 2), effect of recreational marijuana laws accounting for prior patterns of outcomes (Column 3), and effect of recreational marijuana laws over time, meaning how patterns changed over the 4 years since policy implementation (Column 4).
The inverse relationship between recreational marijuana laws and prescription opioid misuse is found to be inconsistent when additional measures are added to the analyses.
The researchers performed checks on the relationship between more permissive marijuana laws and substance use patterns to see if the results change with different approaches and assumptions. When the indicator for recreational marijuana laws was replaced with an indicator for recreational marijuana sales, in general, the associations for alcohol use, marijuana use, and stimulant use stayed the same while the associations for prescription opioid misuse and heroin use were no longer significant. Accounting for staggered policy adoption and effects over time provides further evidence for the association between recreational marijuana laws and increased marijuana use but calls into question the associations for prescription opioid misuse and heroin use. In calculating event study estimates (a way to see how substance use patterns changed over the study period), marijuana use measures were increasing before recreational marijuana laws were implemented (a violation of the parallel-trends assumption and calling into question the relationship between these laws and increased marijuana use). As such, these estimates provided little evidence that legalizing recreational marijuana permanently reduces prescription opioid misuse or heroin use.
WHAT ARE THE IMPLICATIONS OF THE STUDY FINDINGS?
This study found that recreational marijuana laws have a weak association with measures for prescription opioid misuse and heroin use, though these laws potentially may have led to increases in marijuana use and stimulant use. Medical marijuana laws may have led to increases in stimulant use and potentially led to decreases in alcohol use.
The most important finding, and the primary research interest of this study, is that there are inconsistent associations between more permissive marijuana laws and measures of prescription opioid misuse and heroin use. These results, along with the current body of literature, do not support more permissive marijuana laws as a solution to the opioid crisis. In fact, other research shows that expanded access to marijuana could exacerbate the opioid crisis.
In this still-emerging field of research, policymakers should interpret recommendations through the lens of the entire body of literature rather than individual studies, which has not established the claim that making marijuana more readily available can reduce opioid-related harm or that medical marijuana is an appropriate treatment for opioid use disorder.
Worth noting is that the impact of marijuana laws extends well beyond substance use to issues of criminal justice, race, and health equity. Studies like these are important to uncover the impacts of cannabis policy on substance use and its harms to inform strategies that can counter some of its negative consequences.
Associations between more permissive marijuana laws and measures of other types of substance use are of great public health significance. This study found that recreational marijuana laws were associated with increased marijuana use, although use of this substance was already trending upward before these laws were implemented – suggesting that changing attitudes toward marijuana broadly may be accounting for both – drawing into question the actual policy change impact. In addition, both medical and recreational marijuana laws were associated with increased stimulant use, suggesting that some individuals may be complementing marijuana use with stimulant use, and medical marijuana laws were associated with decreased alcohol use, suggesting that individuals may be substituting alcohol use for marijuana use. More scientifically rigorous research is needed, however, to test these possibilities to help determine if they are actually causal. Future research could also explore the potential mechanisms and impacts of these rather complex substance substitution and complementary effects. For example, if marijuana is made more readily available through more permissive marijuana policies, young people may be more inclined to substitute marijuana for alcohol as a socially acceptable alternative. In turn, decreased alcohol use among young people may decrease the litany of harms caused by alcohol intoxication, toxicity, and addiction, including but not limited to its impact on opioid-related overdose deaths.
This study is subject to theecological fallacy, where individual data is aggregated to the state level and averages are compared across the study period.Control variables in this paper including both individual and “ecologic” population-level measures, somewhat reducing the potential for the ecological fallacy to be an explanation for the findings. While these types of studies can be very informative in terms of overall trends in a population, caution should be made when extrapolating from these group-level data to individuals.
The cross-sectional nature of the National Survey on Drug Use and Health warrants caution in drawing conclusions on the causality between more permissive marijuana laws and substance use outcomes.
The National Survey on Drug Use and Health may underestimate substance use, especially illegal substances, due to stigma, social desirability bias, and non-response bias. This impact may have changed for substances that became more acceptable over time, such as marijuana, and may at least partially explain the association between recreational marijuana laws and increased marijuana use.
The National Survey on Drug Use and Health measured prescription opioid misuse differently beginning in 2015. Before this date, prescription opioid misuse was measured as taking the medication nonmedically to get high. Beginning in 2015, this misuse was broadened to measure any prescription opioid use not prescribed by a healthcare provider.
BOTTOM LINE
This study examined the relationship between both medical and recreational marijuana laws and subsequent substance use patterns, including prescription opioid misuse, among adults in the United States. Results suggest that recreational marijuana laws have a weak association with prescription opioid misuse and heroin use, though these laws may have led to increases in marijuana use and stimulant use. Medical marijuana laws may have increased stimulant use and decreased alcohol use.
For individuals and families seeking recovery: According to this study, and the body of literature as a whole, there is not clear evidence that expanding access to marijuana is a solution to the opioid crisis nor is medical marijuana an evidence-based treatment for opioid use disorder. Because these topics are often highly charged with many different financial and other interests at stake, this study and others using similar types of study designs that cannot directly speak to truly causal effects resulting from different policies, serve as a reminder to be wary of reports that may overstate the implications of study results.
For treatment professionals and treatment systems: According to this study, and the body of literature as a whole, there is not clear evidence that expanding access to marijuana is a solution to the opioid crisis nor is medical marijuana an evidence-based treatment for opioid use disorder. More permissive marijuana policies may have an impact on marijuana use as well as other types of non-opioid substance use such as stimulants and alcohol, although more research is needed to confirm these findings. These changes in substance use patterns may disproportionately affect some demographics, like adolescents and young adults. Treatment professionals and treatment systems may benefit from increased awareness of substance use patterns in a changing drug policy landscape.
For scientists:This study used a standard difference-in-difference approach but also used a two-stage procedure to account for staggered policy adoption and dynamic effects. The researchers also calculated event study estimates to show trends over time. The association between more permissive marijuana laws and marijuana use violated the parallel-trends assumption, so this direct relationship should be interpreted with caution. This is a still-emerging field of research that has important policy implications. Scientists could contribute to this field by using more rigorous research methods, such as prospective studies on individual behaviors in response to these policies, and more years of data after policy implementation. A question also, from a causal perspective, is how exactly increases in marijuana use could plausibly lead to increases in stimulant use and decreases in alcohol use.
For policy makers: The findings of this study, along with the current body of studies in this area, do not support more permissive marijuana laws as a solution to the opioid crisis. In fact, other research shows that expanded access to marijuana could exacerbate the opioid crisis. In this still-emerging field of research, public health policy based on the entire body of literature rather than individual studies, is not consistent with the claim that making marijuana more readily available can reduce opioid-related harm or that medical marijuana is an appropriate treatment for opioid use disorder. Associations between more permissive marijuana laws and non-opioid substance use are of great public health significance, and future research should be funded to more rigorously evaluate the potential causal impact of these policies.
Permissive marijuana policies are becoming more common across the United States. As of November 2022, 21 states have adopted recreational marijuana laws, where the possession of small amounts of marijuana by adults aged 21 and older is legal, and most states have adopted medicinal marijuana laws. Most states that have legalized recreational marijuana use also have retail outlet dispensaries where the substance can be purchased. Previous research has suggested that the adoption of recreational and medicinal marijuana laws is associated with changes in substance use. For example, recreational marijuana laws are associated with substantial increases in marijuana use, including among adolescents, whereas medical marijuana laws are associated with small increases in marijuana use, with no effect among adolescents.
Recent research has also highlighted the potential for more permissive marijuana policies to help address the opioid crisis, although findings are mixed. The theory here is that marijuana can help substitute for more immediately harmful substances, including opioids and alcohol. Early adoption of medical marijuana laws (1999-2010) was associated with reduced opioid-related mortality although extending the study period to 2017 found that, over time, the trend reverses and these laws are associated with increased opioid-related mortality. The research on recreational marijuana laws is also mixed, with some studies suggesting a protective effect on opioid-related mortality and other studies suggesting a detrimental effect. More permissive marijuana laws have been associated with lower opioid prescribing, highlighting a potential mechanism for lower opioid-related harm.In addition, since most opioid-related overdose deaths involve non-opioid substances, if more permissive marijuana laws decrease the co-use of a contributing substance, such as alcohol, then they may indirectly decrease opioid-related mortality. Research can help characterize the potential short-term and long-term impacts – whether helpful or harmful – of marijuana policy on substance use in the US.
In this study, researchers used advanced statistical techniques to address many of the methodological limitations in prior studies to examine the relationship between both recreational and medical marijuana laws and subsequent substance use patterns, with a focus on prescription opioid misuse. Sound scientific inquiry into the impact of permissive marijuana policies is paramount to inform policymakers on potential solutions to the opioid crisis as well as prepare prevention and treatment professionals for changing substance use patterns.
HOW WAS THIS STUDY CONDUCTED?
This study examined the relationship between both medical and recreational marijuana laws and subsequent substance use patterns, including prescription opioid misuse, among adults aged 19 and older using several advanced statistical techniques. Substance use trend data from 2004 to 2020 were obtained at the state level from the National Survey on Drug Use and Health. This is an annual survey conducted by the Substance Abuse and Mental Health Services Administration to provide nationally representative data on prevalence and trends for a wide range of substance use, including alcohol, marijuana, and heroin, as well as misuse of prescription opioids and stimulants. Restricted data on these substance use and misuse trends are also available at the state level. The researchers used the variation of state adoption of medical and recreational marijuana laws to examine how substance use and misuse trends changed in states that implemented these laws compared with states that did not implement these laws.
The researchers were interested in the relationship between both medical and recreational marijuana laws and five types of substance use: marijuana use, alcohol use, heroin use, stimulant use, and prescription opioid misuse. In addition to the prevalence and trends in the use of these substances, the degree of substance use was measured (i.e., intensity), which included use in the past year, use in the past month, frequency of use in the past year, and if the individual had a substance use disorder.
Researchers defined a “frequent user” as reporting using the substance at least once per week, on average (i.e., 52 days or more per year). Due to a change in the way that the National Survey on Drug Use and Health measured prescription opioid misuse beginning in 2015, the researchers had to use two mutually exclusive indicators to measure prescription opioid misuse before and after this survey change.
Researchers used two alternative indicators for recreational marijuana laws. The first indicator measured whether the state had adopted a recreational marijuana law by the end of 2020. Twelve states and the District of Columbia had adopted these laws at that time. The second indicator measured whether the state allowed recreational sales of marijuana. Ten states and the District of Columbia were in that category. The researchers also used an indicator for the presence of medical marijuana laws in their statistical models. By the end of 2020, 37 states and the District of Columbia had adopted these laws.
The researchers used advanced statistical modeling techniques and several different models which accounted for various differences between states at both the individual level and the population level. These included a standard difference-in-difference model, a model with an indicator for whether states allow recreational sales of marijuana rather than just having a recreational marijuana law, a model with statistical corrections to account for dynamic effects due to policies being adopted at different points in time, and an event study to capture change over time. Individual-level characteristics controlled for in the model included age, gender, educational attainment, race, ethnicity, marital status, income, and insurance coverage status. Population-level characteristics included a state’s unemployment rate, whether that state had a naloxone (the overdose reversal drug, commonly known as “Narcan”) access law, and whether the state had implemented a prescription drug monitoring program. The researchers also included a statistical technique to capture differences between states and over time that were not characteristics mentioned above (called fixed effects).
WHAT DID THIS STUDY FIND?
Recreational marijuana laws, but not medical marijuana laws, were associated with increased marijuana use.
Recreational marijuana laws were associated with an increase in the probability of past-year marijuana use, an increase in the probability of past-month marijuana use, and an increase in the frequency of past-year use. However, these laws were not associated with an increase in marijuana use disorder. Medical marijuana laws were not associated with any of the measures of marijuana use.
Medical marijuana laws were associated with decreased alcohol use while evidence for recreational marijuana laws was mixed.
Medical marijuana laws were associated with decreases in the probabilities of past-year alcohol use, past-month alcohol use, and frequent (i.e., weekly or greater) alcohol use. In contrast, recreational marijuana laws were associated with a decreased probability of alcohol use disorder and an increased frequency of alcohol use in the past year.
Both medical and recreational laws were associated with increased stimulant use.
Medical marijuana laws were associated with increases in the probabilities of past-year stimulant use, past-month stimulant use, and frequent stimulant users along with increased frequency of stimulant use in the past month. Recreational marijuana laws were associated with an increased probability of frequent stimulant use and an increased frequency of stimulant use in the past year.
Recreational marijuana laws were associated with some measures of decreased prescription opioid misuse and some measures of increased heroin use in the standard model.
Recreational marijuana laws were associated with both decreased probabilities of frequent prescription opioid use and having a prescription opioid use disorder during the 2004 to 2014 study period, but these measures were not statistically significant in the subsequent study period. Recreational marijuana laws were associated with a decreased probability of heroin use disorder and an increased frequency of heroin use in the past year.
Figure 1. Effects of marijuana laws on non-medical opioid use and heroin use outcomes. Overall effect of recreational marijuana laws (Column 1), overall effect of recreational marijuana sales, meaning when actual commercial sales began, not law passage (Column 2), effect of recreational marijuana laws accounting for prior patterns of outcomes (Column 3), and effect of recreational marijuana laws over time, meaning how patterns changed over the 4 years since policy implementation (Column 4).
The inverse relationship between recreational marijuana laws and prescription opioid misuse is found to be inconsistent when additional measures are added to the analyses.
The researchers performed checks on the relationship between more permissive marijuana laws and substance use patterns to see if the results change with different approaches and assumptions. When the indicator for recreational marijuana laws was replaced with an indicator for recreational marijuana sales, in general, the associations for alcohol use, marijuana use, and stimulant use stayed the same while the associations for prescription opioid misuse and heroin use were no longer significant. Accounting for staggered policy adoption and effects over time provides further evidence for the association between recreational marijuana laws and increased marijuana use but calls into question the associations for prescription opioid misuse and heroin use. In calculating event study estimates (a way to see how substance use patterns changed over the study period), marijuana use measures were increasing before recreational marijuana laws were implemented (a violation of the parallel-trends assumption and calling into question the relationship between these laws and increased marijuana use). As such, these estimates provided little evidence that legalizing recreational marijuana permanently reduces prescription opioid misuse or heroin use.
WHAT ARE THE IMPLICATIONS OF THE STUDY FINDINGS?
This study found that recreational marijuana laws have a weak association with measures for prescription opioid misuse and heroin use, though these laws potentially may have led to increases in marijuana use and stimulant use. Medical marijuana laws may have led to increases in stimulant use and potentially led to decreases in alcohol use.
The most important finding, and the primary research interest of this study, is that there are inconsistent associations between more permissive marijuana laws and measures of prescription opioid misuse and heroin use. These results, along with the current body of literature, do not support more permissive marijuana laws as a solution to the opioid crisis. In fact, other research shows that expanded access to marijuana could exacerbate the opioid crisis.
In this still-emerging field of research, policymakers should interpret recommendations through the lens of the entire body of literature rather than individual studies, which has not established the claim that making marijuana more readily available can reduce opioid-related harm or that medical marijuana is an appropriate treatment for opioid use disorder.
Worth noting is that the impact of marijuana laws extends well beyond substance use to issues of criminal justice, race, and health equity. Studies like these are important to uncover the impacts of cannabis policy on substance use and its harms to inform strategies that can counter some of its negative consequences.
Associations between more permissive marijuana laws and measures of other types of substance use are of great public health significance. This study found that recreational marijuana laws were associated with increased marijuana use, although use of this substance was already trending upward before these laws were implemented – suggesting that changing attitudes toward marijuana broadly may be accounting for both – drawing into question the actual policy change impact. In addition, both medical and recreational marijuana laws were associated with increased stimulant use, suggesting that some individuals may be complementing marijuana use with stimulant use, and medical marijuana laws were associated with decreased alcohol use, suggesting that individuals may be substituting alcohol use for marijuana use. More scientifically rigorous research is needed, however, to test these possibilities to help determine if they are actually causal. Future research could also explore the potential mechanisms and impacts of these rather complex substance substitution and complementary effects. For example, if marijuana is made more readily available through more permissive marijuana policies, young people may be more inclined to substitute marijuana for alcohol as a socially acceptable alternative. In turn, decreased alcohol use among young people may decrease the litany of harms caused by alcohol intoxication, toxicity, and addiction, including but not limited to its impact on opioid-related overdose deaths.
This study is subject to theecological fallacy, where individual data is aggregated to the state level and averages are compared across the study period.Control variables in this paper including both individual and “ecologic” population-level measures, somewhat reducing the potential for the ecological fallacy to be an explanation for the findings. While these types of studies can be very informative in terms of overall trends in a population, caution should be made when extrapolating from these group-level data to individuals.
The cross-sectional nature of the National Survey on Drug Use and Health warrants caution in drawing conclusions on the causality between more permissive marijuana laws and substance use outcomes.
The National Survey on Drug Use and Health may underestimate substance use, especially illegal substances, due to stigma, social desirability bias, and non-response bias. This impact may have changed for substances that became more acceptable over time, such as marijuana, and may at least partially explain the association between recreational marijuana laws and increased marijuana use.
The National Survey on Drug Use and Health measured prescription opioid misuse differently beginning in 2015. Before this date, prescription opioid misuse was measured as taking the medication nonmedically to get high. Beginning in 2015, this misuse was broadened to measure any prescription opioid use not prescribed by a healthcare provider.
BOTTOM LINE
This study examined the relationship between both medical and recreational marijuana laws and subsequent substance use patterns, including prescription opioid misuse, among adults in the United States. Results suggest that recreational marijuana laws have a weak association with prescription opioid misuse and heroin use, though these laws may have led to increases in marijuana use and stimulant use. Medical marijuana laws may have increased stimulant use and decreased alcohol use.
For individuals and families seeking recovery: According to this study, and the body of literature as a whole, there is not clear evidence that expanding access to marijuana is a solution to the opioid crisis nor is medical marijuana an evidence-based treatment for opioid use disorder. Because these topics are often highly charged with many different financial and other interests at stake, this study and others using similar types of study designs that cannot directly speak to truly causal effects resulting from different policies, serve as a reminder to be wary of reports that may overstate the implications of study results.
For treatment professionals and treatment systems: According to this study, and the body of literature as a whole, there is not clear evidence that expanding access to marijuana is a solution to the opioid crisis nor is medical marijuana an evidence-based treatment for opioid use disorder. More permissive marijuana policies may have an impact on marijuana use as well as other types of non-opioid substance use such as stimulants and alcohol, although more research is needed to confirm these findings. These changes in substance use patterns may disproportionately affect some demographics, like adolescents and young adults. Treatment professionals and treatment systems may benefit from increased awareness of substance use patterns in a changing drug policy landscape.
For scientists:This study used a standard difference-in-difference approach but also used a two-stage procedure to account for staggered policy adoption and dynamic effects. The researchers also calculated event study estimates to show trends over time. The association between more permissive marijuana laws and marijuana use violated the parallel-trends assumption, so this direct relationship should be interpreted with caution. This is a still-emerging field of research that has important policy implications. Scientists could contribute to this field by using more rigorous research methods, such as prospective studies on individual behaviors in response to these policies, and more years of data after policy implementation. A question also, from a causal perspective, is how exactly increases in marijuana use could plausibly lead to increases in stimulant use and decreases in alcohol use.
For policy makers: The findings of this study, along with the current body of studies in this area, do not support more permissive marijuana laws as a solution to the opioid crisis. In fact, other research shows that expanded access to marijuana could exacerbate the opioid crisis. In this still-emerging field of research, public health policy based on the entire body of literature rather than individual studies, is not consistent with the claim that making marijuana more readily available can reduce opioid-related harm or that medical marijuana is an appropriate treatment for opioid use disorder. Associations between more permissive marijuana laws and non-opioid substance use are of great public health significance, and future research should be funded to more rigorously evaluate the potential causal impact of these policies.
Permissive marijuana policies are becoming more common across the United States. As of November 2022, 21 states have adopted recreational marijuana laws, where the possession of small amounts of marijuana by adults aged 21 and older is legal, and most states have adopted medicinal marijuana laws. Most states that have legalized recreational marijuana use also have retail outlet dispensaries where the substance can be purchased. Previous research has suggested that the adoption of recreational and medicinal marijuana laws is associated with changes in substance use. For example, recreational marijuana laws are associated with substantial increases in marijuana use, including among adolescents, whereas medical marijuana laws are associated with small increases in marijuana use, with no effect among adolescents.
Recent research has also highlighted the potential for more permissive marijuana policies to help address the opioid crisis, although findings are mixed. The theory here is that marijuana can help substitute for more immediately harmful substances, including opioids and alcohol. Early adoption of medical marijuana laws (1999-2010) was associated with reduced opioid-related mortality although extending the study period to 2017 found that, over time, the trend reverses and these laws are associated with increased opioid-related mortality. The research on recreational marijuana laws is also mixed, with some studies suggesting a protective effect on opioid-related mortality and other studies suggesting a detrimental effect. More permissive marijuana laws have been associated with lower opioid prescribing, highlighting a potential mechanism for lower opioid-related harm.In addition, since most opioid-related overdose deaths involve non-opioid substances, if more permissive marijuana laws decrease the co-use of a contributing substance, such as alcohol, then they may indirectly decrease opioid-related mortality. Research can help characterize the potential short-term and long-term impacts – whether helpful or harmful – of marijuana policy on substance use in the US.
In this study, researchers used advanced statistical techniques to address many of the methodological limitations in prior studies to examine the relationship between both recreational and medical marijuana laws and subsequent substance use patterns, with a focus on prescription opioid misuse. Sound scientific inquiry into the impact of permissive marijuana policies is paramount to inform policymakers on potential solutions to the opioid crisis as well as prepare prevention and treatment professionals for changing substance use patterns.
HOW WAS THIS STUDY CONDUCTED?
This study examined the relationship between both medical and recreational marijuana laws and subsequent substance use patterns, including prescription opioid misuse, among adults aged 19 and older using several advanced statistical techniques. Substance use trend data from 2004 to 2020 were obtained at the state level from the National Survey on Drug Use and Health. This is an annual survey conducted by the Substance Abuse and Mental Health Services Administration to provide nationally representative data on prevalence and trends for a wide range of substance use, including alcohol, marijuana, and heroin, as well as misuse of prescription opioids and stimulants. Restricted data on these substance use and misuse trends are also available at the state level. The researchers used the variation of state adoption of medical and recreational marijuana laws to examine how substance use and misuse trends changed in states that implemented these laws compared with states that did not implement these laws.
The researchers were interested in the relationship between both medical and recreational marijuana laws and five types of substance use: marijuana use, alcohol use, heroin use, stimulant use, and prescription opioid misuse. In addition to the prevalence and trends in the use of these substances, the degree of substance use was measured (i.e., intensity), which included use in the past year, use in the past month, frequency of use in the past year, and if the individual had a substance use disorder.
Researchers defined a “frequent user” as reporting using the substance at least once per week, on average (i.e., 52 days or more per year). Due to a change in the way that the National Survey on Drug Use and Health measured prescription opioid misuse beginning in 2015, the researchers had to use two mutually exclusive indicators to measure prescription opioid misuse before and after this survey change.
Researchers used two alternative indicators for recreational marijuana laws. The first indicator measured whether the state had adopted a recreational marijuana law by the end of 2020. Twelve states and the District of Columbia had adopted these laws at that time. The second indicator measured whether the state allowed recreational sales of marijuana. Ten states and the District of Columbia were in that category. The researchers also used an indicator for the presence of medical marijuana laws in their statistical models. By the end of 2020, 37 states and the District of Columbia had adopted these laws.
The researchers used advanced statistical modeling techniques and several different models which accounted for various differences between states at both the individual level and the population level. These included a standard difference-in-difference model, a model with an indicator for whether states allow recreational sales of marijuana rather than just having a recreational marijuana law, a model with statistical corrections to account for dynamic effects due to policies being adopted at different points in time, and an event study to capture change over time. Individual-level characteristics controlled for in the model included age, gender, educational attainment, race, ethnicity, marital status, income, and insurance coverage status. Population-level characteristics included a state’s unemployment rate, whether that state had a naloxone (the overdose reversal drug, commonly known as “Narcan”) access law, and whether the state had implemented a prescription drug monitoring program. The researchers also included a statistical technique to capture differences between states and over time that were not characteristics mentioned above (called fixed effects).
WHAT DID THIS STUDY FIND?
Recreational marijuana laws, but not medical marijuana laws, were associated with increased marijuana use.
Recreational marijuana laws were associated with an increase in the probability of past-year marijuana use, an increase in the probability of past-month marijuana use, and an increase in the frequency of past-year use. However, these laws were not associated with an increase in marijuana use disorder. Medical marijuana laws were not associated with any of the measures of marijuana use.
Medical marijuana laws were associated with decreased alcohol use while evidence for recreational marijuana laws was mixed.
Medical marijuana laws were associated with decreases in the probabilities of past-year alcohol use, past-month alcohol use, and frequent (i.e., weekly or greater) alcohol use. In contrast, recreational marijuana laws were associated with a decreased probability of alcohol use disorder and an increased frequency of alcohol use in the past year.
Both medical and recreational laws were associated with increased stimulant use.
Medical marijuana laws were associated with increases in the probabilities of past-year stimulant use, past-month stimulant use, and frequent stimulant users along with increased frequency of stimulant use in the past month. Recreational marijuana laws were associated with an increased probability of frequent stimulant use and an increased frequency of stimulant use in the past year.
Recreational marijuana laws were associated with some measures of decreased prescription opioid misuse and some measures of increased heroin use in the standard model.
Recreational marijuana laws were associated with both decreased probabilities of frequent prescription opioid use and having a prescription opioid use disorder during the 2004 to 2014 study period, but these measures were not statistically significant in the subsequent study period. Recreational marijuana laws were associated with a decreased probability of heroin use disorder and an increased frequency of heroin use in the past year.
Figure 1. Effects of marijuana laws on non-medical opioid use and heroin use outcomes. Overall effect of recreational marijuana laws (Column 1), overall effect of recreational marijuana sales, meaning when actual commercial sales began, not law passage (Column 2), effect of recreational marijuana laws accounting for prior patterns of outcomes (Column 3), and effect of recreational marijuana laws over time, meaning how patterns changed over the 4 years since policy implementation (Column 4).
The inverse relationship between recreational marijuana laws and prescription opioid misuse is found to be inconsistent when additional measures are added to the analyses.
The researchers performed checks on the relationship between more permissive marijuana laws and substance use patterns to see if the results change with different approaches and assumptions. When the indicator for recreational marijuana laws was replaced with an indicator for recreational marijuana sales, in general, the associations for alcohol use, marijuana use, and stimulant use stayed the same while the associations for prescription opioid misuse and heroin use were no longer significant. Accounting for staggered policy adoption and effects over time provides further evidence for the association between recreational marijuana laws and increased marijuana use but calls into question the associations for prescription opioid misuse and heroin use. In calculating event study estimates (a way to see how substance use patterns changed over the study period), marijuana use measures were increasing before recreational marijuana laws were implemented (a violation of the parallel-trends assumption and calling into question the relationship between these laws and increased marijuana use). As such, these estimates provided little evidence that legalizing recreational marijuana permanently reduces prescription opioid misuse or heroin use.
WHAT ARE THE IMPLICATIONS OF THE STUDY FINDINGS?
This study found that recreational marijuana laws have a weak association with measures for prescription opioid misuse and heroin use, though these laws potentially may have led to increases in marijuana use and stimulant use. Medical marijuana laws may have led to increases in stimulant use and potentially led to decreases in alcohol use.
The most important finding, and the primary research interest of this study, is that there are inconsistent associations between more permissive marijuana laws and measures of prescription opioid misuse and heroin use. These results, along with the current body of literature, do not support more permissive marijuana laws as a solution to the opioid crisis. In fact, other research shows that expanded access to marijuana could exacerbate the opioid crisis.
In this still-emerging field of research, policymakers should interpret recommendations through the lens of the entire body of literature rather than individual studies, which has not established the claim that making marijuana more readily available can reduce opioid-related harm or that medical marijuana is an appropriate treatment for opioid use disorder.
Worth noting is that the impact of marijuana laws extends well beyond substance use to issues of criminal justice, race, and health equity. Studies like these are important to uncover the impacts of cannabis policy on substance use and its harms to inform strategies that can counter some of its negative consequences.
Associations between more permissive marijuana laws and measures of other types of substance use are of great public health significance. This study found that recreational marijuana laws were associated with increased marijuana use, although use of this substance was already trending upward before these laws were implemented – suggesting that changing attitudes toward marijuana broadly may be accounting for both – drawing into question the actual policy change impact. In addition, both medical and recreational marijuana laws were associated with increased stimulant use, suggesting that some individuals may be complementing marijuana use with stimulant use, and medical marijuana laws were associated with decreased alcohol use, suggesting that individuals may be substituting alcohol use for marijuana use. More scientifically rigorous research is needed, however, to test these possibilities to help determine if they are actually causal. Future research could also explore the potential mechanisms and impacts of these rather complex substance substitution and complementary effects. For example, if marijuana is made more readily available through more permissive marijuana policies, young people may be more inclined to substitute marijuana for alcohol as a socially acceptable alternative. In turn, decreased alcohol use among young people may decrease the litany of harms caused by alcohol intoxication, toxicity, and addiction, including but not limited to its impact on opioid-related overdose deaths.
This study is subject to theecological fallacy, where individual data is aggregated to the state level and averages are compared across the study period.Control variables in this paper including both individual and “ecologic” population-level measures, somewhat reducing the potential for the ecological fallacy to be an explanation for the findings. While these types of studies can be very informative in terms of overall trends in a population, caution should be made when extrapolating from these group-level data to individuals.
The cross-sectional nature of the National Survey on Drug Use and Health warrants caution in drawing conclusions on the causality between more permissive marijuana laws and substance use outcomes.
The National Survey on Drug Use and Health may underestimate substance use, especially illegal substances, due to stigma, social desirability bias, and non-response bias. This impact may have changed for substances that became more acceptable over time, such as marijuana, and may at least partially explain the association between recreational marijuana laws and increased marijuana use.
The National Survey on Drug Use and Health measured prescription opioid misuse differently beginning in 2015. Before this date, prescription opioid misuse was measured as taking the medication nonmedically to get high. Beginning in 2015, this misuse was broadened to measure any prescription opioid use not prescribed by a healthcare provider.
BOTTOM LINE
This study examined the relationship between both medical and recreational marijuana laws and subsequent substance use patterns, including prescription opioid misuse, among adults in the United States. Results suggest that recreational marijuana laws have a weak association with prescription opioid misuse and heroin use, though these laws may have led to increases in marijuana use and stimulant use. Medical marijuana laws may have increased stimulant use and decreased alcohol use.
For individuals and families seeking recovery: According to this study, and the body of literature as a whole, there is not clear evidence that expanding access to marijuana is a solution to the opioid crisis nor is medical marijuana an evidence-based treatment for opioid use disorder. Because these topics are often highly charged with many different financial and other interests at stake, this study and others using similar types of study designs that cannot directly speak to truly causal effects resulting from different policies, serve as a reminder to be wary of reports that may overstate the implications of study results.
For treatment professionals and treatment systems: According to this study, and the body of literature as a whole, there is not clear evidence that expanding access to marijuana is a solution to the opioid crisis nor is medical marijuana an evidence-based treatment for opioid use disorder. More permissive marijuana policies may have an impact on marijuana use as well as other types of non-opioid substance use such as stimulants and alcohol, although more research is needed to confirm these findings. These changes in substance use patterns may disproportionately affect some demographics, like adolescents and young adults. Treatment professionals and treatment systems may benefit from increased awareness of substance use patterns in a changing drug policy landscape.
For scientists:This study used a standard difference-in-difference approach but also used a two-stage procedure to account for staggered policy adoption and dynamic effects. The researchers also calculated event study estimates to show trends over time. The association between more permissive marijuana laws and marijuana use violated the parallel-trends assumption, so this direct relationship should be interpreted with caution. This is a still-emerging field of research that has important policy implications. Scientists could contribute to this field by using more rigorous research methods, such as prospective studies on individual behaviors in response to these policies, and more years of data after policy implementation. A question also, from a causal perspective, is how exactly increases in marijuana use could plausibly lead to increases in stimulant use and decreases in alcohol use.
For policy makers: The findings of this study, along with the current body of studies in this area, do not support more permissive marijuana laws as a solution to the opioid crisis. In fact, other research shows that expanded access to marijuana could exacerbate the opioid crisis. In this still-emerging field of research, public health policy based on the entire body of literature rather than individual studies, is not consistent with the claim that making marijuana more readily available can reduce opioid-related harm or that medical marijuana is an appropriate treatment for opioid use disorder. Associations between more permissive marijuana laws and non-opioid substance use are of great public health significance, and future research should be funded to more rigorously evaluate the potential causal impact of these policies.
