“Marijuana Legalization Is Not Harmless” – Does Legalization Impact Opioid Mortality?

Today, we’re going to talk about the harms of cannabis. Again. 

Time to get on with it.


A new working paper from the National Bureau of Economic Research contains some inconvenient news for the rosy worldview of those who claim that marijuana is a completely harmless drug.

The paper reviews data on opioid and marijuana use and makes two key findings — first, that “medical marijuana, particularly when available through retail dispensaries, is associated with higher opioid mortality.” The second finding is that data “for recreational marijuana, while less reliable, also suggest that retail sales through dispensaries are associated with greater death rates relative to the counterfactual of no legal cannabis.”

The increase in opioid deaths associated with marijuana use is greater for men, nonwhites, and young people.


I’m curious about the source material (the study) since a link was not provided within this editorial. Let’s go hunting.

Our analyses show that RMLs increase adult marijuana use and reduce drug-related arrests over an average post-legalization window of three to four years. There is little evidence to suggest that RML-induced increases in marijuana consumption encourage the use of harder substances or violent criminal activity, and some evidence that RMLs may aid in reducing opioid-related mortality.


*Raises eyebrow*

It’s a bit hard to determine what paper the source article is referencing (here is the PDF of the full paper sourced above), but hardly a skim indicates a totally different set of results than those reported in the article. At least if this is the source, which it may not be. 

I also came across THIS paper, which also seems to indicate results contrary to the narrative that the examiner is aiming for. As goes for THIS one.

Altering my query a bit (and using scholar mode), I finally seem to be getting closer.


Recent studies have concluded that state laws legalizing medical marijuana can reduce deaths from opioid overdoses. Using data from the National Survey on Drug Use and Health, a survey uniquely suited to assessing drug misuse, we examine the relationship between recreational marijuana laws (RMLs) and the use of opioids. Standard difference-in-differences (DD) regression estimates indicate that RMLs do not affect the likelihood of misusing prescription pain relievers such as OxyContin, Percocet, and Vicodin. Although DD regression estimates provide evidence that state laws legalizing recreational marijuana can reduce the frequency of misusing prescription pain relievers, event-study estimates are noisy and suggest that any effect on the frequency of misuse is likely transitory.


This one?


For other public health outcomes such as mortality involving prescription opioids, the effect of legalizing medical marijuana has proven more difficult to gauge and, as a consequence, we are less comfortable drawing firm conclusions.



Nope. Next!


. . . in most cases, the inclusion of more comprehensive controls, longer analysis periods and more correctly defined dependent variables results in less favorable estimates, often including predicted increases in opioid deaths.



Though it might not be apparent, I committed some Washington Examiner-level journalistic manipulation in the last quote.

Well, it is pretty obvious. Nonetheless, here is what I (and they) failed to tell you. Keep in mind that the following is from the abstract (I have not even touched the actual PDF!):


Over the last two decades there has been considerable movement at the state-level to legalize marijuana, initially for medical purposes and more recently for recreational consumption. Despite prior research, it is unclear how, if at all, these policies are related to rates of opioid-involved overdose deaths, which have trended rapidly upwards over time and represent a major public health problem. We provide two types of new information on this question. First, we replicate and extend upon previous investigations and show that the empirical results of those studies are frequently fragile and that, in most cases, the inclusion of more comprehensive controls, longer analysis periods and more correctly defined dependent variables results in less favorable estimates, often including predicted increases in opioid deaths. Second, we present new estimates from generalized differences-in-differences and event study models that incorporate more recent data and improvements developed in our replication and extension of early research. These results indicate that legal medical marijuana, particularly when available through retail dispensaries, is associated with higher opioid mortality. The results for recreational marijuana, while less reliable, also suggest that retail sales through dispensaries are associated with greater death rates relative to the counterfactual of no legal cannabis.


My first thought upon reading that is “What does this even mean?!”.

For the Washington Examiner editorial team, it’s a handy reference that no one (aside from some Canadian running a blog no one reads) will realize is falsely cited. But for me, it strikes me as citing a question of correlation vs causation.

Since cannabis dispensaries and the majority of drug users are likely to be clumped in higher population areas (such as cities), could the correlation be related to nothing more than geographical dynamics?

This is highly doubtful. Let’s cite the paper itself to see if we can gain some insights.


From the paper:

More than 930,000 Americans died of drug overdoses from 1999-2020 (Hedegaard et
al. 2021). A large majority of these involved opioids, and both all drug mortality and deaths
implicating opioids accelerated markedly during the first year of the COVID-19 pandemic.

In response to these alarming trends, there have been multiple federal, state and local efforts
to reduce opioid deaths and related problems including: better tracking of prescribing
through drug monitoring programs; improved access to non-opioid pain care, naloxone, and
medications treating opioid use disorder; assistance to high-risk persons following release
from incarceration; physician and prescriber education programs; improved data
surveillance; Good Samaritan laws that reduce barriers to calling for help during opioid
emergencies; and multiple federal grant programs that provide states and local governments
with assistance in funding these and other endeavors (Purington 2019; Harris and
Mukkamala 2020; Katcher and Ruhm 2021).

I can already start to see an answer to my question in this paragraph. The pandemic drove everyone down to new levels of misery, which would account for drug-seeking behaviour in regards to all substances. While not mentioned, I wouldn’t be surprised to see that alcohol misuse also shot up.

It’s a coping mechanism, after all.

Nonetheless, I’m not going to put the answer I want to hear into the paper. After all, if I did that, I would be no better than the Washington Examiner editorial team.

At the same time, policies not directly related to opioid use or deaths may affect
these outcomes. An important potential example are state laws that legalize the
consumption and retail sale of medical or recreational marijuana. 1 Prior to 1999, the first
year analyzed below, three states (California, Oregon, and Washington) had legalized
medical marijuana, but none permitted retail sales through dispensaries. By the end of 2019,
the last year studied, 33 states had legalized medical cannabis, 29 with medical dispensaries
in place, 11 states permitted recreational marijuana, and eight of these states had operating
retail dispensaries.

A rapidly growing body of scholarship examines the relationship between marijuana
legalization and various aspects of public health. 

                                                                                             * * *

There has been more limited study of its effects on opioid-related outcomes such as prescribing behavior
(Bradford and Bradford 2016; Bradford et al. 2018; Wen and Hockenberry 2018; McMichael,
Van Horn, and Viscusi 2020) and admissions to substance abuse treatment programs,
emergency departments, or hospitals (Chu 2015; Powell, Pacula, and Jacobson 2018;
Conyers and Ayres 2020; Jayawardhana and Fernandez 2021).

Finally, researchers have examined how marijuana legalization is related to opioid
deaths. These studies, some of which are summarized in the next section, while not
voluminous, have been influential. Particularly prominent is Bachhuber et al.’s (2014)
conclusion that “medical cannabis laws are associated with significantly lower state-level
opioid overdose mortality rates” (p. 1668). This study has been widely cited (over 760
Google Scholar citations as of February 2022) and has played an important role in
arguments that led some states to approve medical marijuana as a treatment for opioid use
disorder (Shover et al. 2020).  However, as discussed below, these findings turn out not to
be robust to changes in the analysis period, with subsequent research yielding ambiguous


And there it is.

I didn’t answer the question that I was looking to answer earlier, but I found the purpose of the study. The goal of which seems less about questioning the harmlessness of cannabis legalization, and more about cautioning that more study is required in terms of utilizing medicinal cannabis therapy for the treatment of opioid use disorder. Or maybe, utilizing cannabis as a method of treating opioid misuse disorder?

I’m not a doctor, so take this for what you will.

Since we’re this deep in the paper, we may as well explore some of their findings in this area.


The current study provides more definitive information on the relationship between
marijuana legalization and opioid deaths. We first show that prior empirical results are frequently fragile and that, in most cases, the inclusion of more comprehensive controls,
longer analysis periods and more correctly defined treatment variables results in less
favorable estimates or deleterious predicted effects of legal cannabis. We then present new
estimates, from generalized differences-in-differences (DiD) and event study (ES) models,
that incorporate more recent data and the improvements developed in our replication and
extension of previous research.

These results indicate that legal medical marijuana, particularly when available
through retail dispensaries, is associated with higher opioid death rates. The estimates for
recreational marijuana while less reliable – probably because most such policies have been
only recently enacted and in a lower number of states than for medical marijuana – also
suggest that retail sales through dispensaries are associated with greater opioid mortality,
relative to the counterfactual of no legal cannabis.

There is also suggestive evidence of
heterogeneity across demographic groups, with stronger deleterious recreational marijuana
effects for males, nonwhites, and relatively young adults than for their counterparts. Retail
cannabis sales also likely increase deaths involving non-opioid drugs such as stimulants and
sedatives. Finally, we indicate that more favorable findings previously observed when
analyzing deaths from 1999-2010 may reflect idiosyncratic and unreliable findings when
considering short time periods rather than, as suggested by some researchers, changes over
time in the stringency of the regulatory approaches.


Given this observation, we now know for sure that this is the paper that the Examiner editorial team was referencing. And the observation makes me think of a whole new hypothesis/guess for the correlation. People that are inclined to utilize or necessitate the consumption of medicinal or recreational marijuana are also likely to be open to seeking other substances. Though no line is drawn to chronic pain, mental health or any other variable, these variables tend to be a common throughline to drug-seeking behaviour, no matter what the substance.

I recall the Washington Examiner article I opened with bringing up the cannabis as a gateway drug argument when considering all of this data. What they fail to note is that many people who got slash get addicted to opioids didn’t even start out as typical drug users. Some may not have even touched cannabis once in their lives (or if they did, not for a VERY long time).

Many people in the past decade or so suffered some sort of injury and were prescribed some form of an opioid to help with pain relief. Unknown to these patients (who put their trust in the doctors and pharmaceutical companies within the US medical system), profitability was often pushed at all costs when it came to selling opioid medications. While far from an exhaustive list, Insys Therapeutics and Perdue Pharma are 2 very egregious examples of this malpractice in action.

While not mentioned by the Washington Examiner article or in the paper they cited (at least not in the pages I referenced), all of this plays into the outcomes we are looking at.

Though simple minds like easy-to-digest conclusions, humans are very messy as far as all things medical, physical and mental are concerned. One could say that it is one of the biggest drawbacks of being human. We’re all complex, but our understandings are often extremely simplistic. Not a good recipe for a complex society that is evolving on a daily (if not hourly!) basis.

But THAT is a whole other topic.

Bachhuber et al. (2014), mentioned above, used public-use National Vital Statistics
System (NVSS) data from 1999-2010 to examine the relationship between medical marijuana
legalization (MML) and opioid deaths.  Their estimates suggest that MML reduced age-

adjusted opioid analgesic mortality by almost 25% and a broader measure of opioid deaths
by 23%, in models with state and year fixed effects, although with some attenuation when
state time trends were also controlled for. However, this result is sensitive to the analysis
period. Shover et al. (2019) replicated Bachhuber’s analysis and obtain a similar 21%
reduction over the 1999-2010 timespan, but they also demonstrate that the relationship
reverses when extending the investigation through 2017, with medical cannabis legalization
predicting a 23% increase in prescription opioid deaths over this longer period.

Powell, Pacula, and Jacobson’s (2018) innovation is to distinguish between the
legalization of medical marijuana and the availability of retail sales to qualified patients
through authorized medical marijuana dispensaries (MMD). Using non-public NVSS data,
they confirm Bachhuber’s (2014) negative relationship between legalization of medical
marijuana and opioid deaths from 1999-2010 but, consistent with Shover et al. (2019), show
that the effects weaken and become statistically insignificant when extending the period
through 2013. However, their key finding is that the availability of medical marijuana sales
through retail dispensaries is associated with a 28% reduction in deaths involving
prescription opioids or heroin, relative to states without legal cannabis.


The analysis of the data from 1999 thru 2010 and 2013 makes sense given that the opioid epidemic really started to take off in the early 2010s. As does the noted drop in opioid fatalities in localities with a dispensary since people with the option are more likely to try alternatives (such as CBD, or normal THC) in their pursuit of pain relief. Law-abiding people outside of areas with dispensaries are more likely to follow the law and thus take a prescription from their medical doctor. A prescription which is likely to involve an opioid medication.

Keep in mind that not all prescriptions for opioids are unnecessary and that not all doctors prescribing them are malicious. Just as the public was, many doctors were lied to in terms of the potential harms of the medications they were prescribing, often learning the hard way that their attempt to help patients only lead to their ultimate detriment. The opioid crisis has many victims, and the good ethical people of the medical establishment are one of them.

Using similar methods and data for 1999-2017, Chan, Burkhardt and Flyer (2020),
add controls for the legalization of recreational marijuana (RML), as well as corresponding
dispensaries (RMD). In their preferred specification, which limits analysis to 28 states, the
coefficient on recreational marijuana dispensaries is -0.23 and significant at the 10 percent
level, which they interpret to imply a 21% decrease in opioid death rates. 5 However, this
conclusion depends critically on the counterfactual comparison. Specifically, the
corresponding RML coefficient is 0.19, implying that while RMD reduces predicted opioid

mortality rates by 21% compared to an otherwise equivalent state that legalized recreational
marijuana but without retail sales, the decrease is just 4% relative to one not allowing any
type of recreational cannabis.
In recent work, Sabia et al. (2021) uses data from 2000-2019 to examine how the
legalization of recreational marijuana relates to a variety of outcomes, including mortality
rates. They provide suggestive evidence of beneficial effects, but the estimates attenuate
and frequently become statistically insignificant or detrimental with the inclusion of more
comprehensive controls or if recreational marijuana sales, rather than legalization, is used
as the treatment variable. They also do not control for the legalization of medical marijuana
in any of their models, so that the counterfactual combines states without legal marijuana
and those allowing medical cannabis.



I’m going to end my analysis of the paper here, finding little need to go further. Though I ended up going down many tangential rabbit holes, I feel like I’ve made it clear that the goal of the paper is very different from the goal of the Washington Examiners’ interpretation of it. It’s no wonder they didn’t link directly to the document.

Speaking of the OP article . . .


That may be somewhat disturbing, but the details appear even more devilish. The study importantly addresses earlier results, based on data from 1999 to 2010, that had seemed to suggest a more beneficial effect. It turns out, though, that the results abruptly changed. If you include data from 2010 to 2017, the period when medical and/or recreational marijuana legalization began in earnest in the states, the results swing from a 21% reduction in opioid deaths to a 23% increase.

The results are complex, but the study undercuts a key claim of marijuana legalization proponents who argue that marijuana is a harmless substance that causes a cheap, temporary high and nothing more.

Here, the relationship between cannabis and opioid deaths is interesting in that it reinforces a much-mocked description of marijuana as a “gateway drug.” Different drug habits might well be related in ways we do not yet understand.


1.) Notice that nothing is mentioned of the massive increase in opioid prescriptions in the 1999-2017 timeframe. One would seem that would be an important factor to consider.

Unless you don’t care about context.

2.) Only idiotic marijuana proponents claim it to merely be a harmless, cheap high. The associations between mental health complications and lung health are well known.

3.) As explored earlier, the gateway drug argument, in this case, is stupid. The opioid epidemic didn’t become the massive epidemic it is now just because of stoners moving up the pharmacological spectrum to better shit.

Again, context matters. Well, unless you are a right-wing rag with an agenda.


And of course, this is not the only pitfall associated with marijuana use that marijuana campaigners work hard to minimize. For example, habitual marijuana use as late as one’s mid-20s can cause permanent brain damage. That’s because it prevents proper development of the frontal cortex, which the American Psychological Association describes as one of the last regions of the brain to develop fully. This brain structure is “critical to planning, judgment, decision-making, and personality.”


For once, we have a link. And it seems to bare out what is being communicated.

Indeed, a number of studies have found evidence of brain changes in teens and young adults who smoke marijuana. In 2013, Rocío Martín-Santos, MD, PhD, at the University of Barcelona, and colleagues reviewed 43 studies of chronic cannabis use and the brain. They found consistent evidence of both structural brain abnormalities and altered neural activity in marijuana users. Only eight of those studies focused on adolescents, but the findings from those studies suggested that both structural and functional brain changes emerge soon after adolescents start using the drug. Those changes may still be evident after a month of abstaining from the drug, the researchers reported (PLOS ONE, 2013).

Some of those brain abnormalities have been linked to cognitive differences. Gruber found that regular, heavy marijuana users — those who reported smoking five of the last seven days, and more than 2,500 times in their lives — had damage to their brains’ white matter, which helps enable communication among neurons. Those white matter changes were correlated with higher impulsivity, she found, particularly in people who began smoking before age 16 (Psychopharmacology, 2013).

Much of Gruber’s work compares heavy, regular marijuana users who began before and after age 16. Her results suggest there’s greater risk in starting young. Compared with users who began after 16, early-onset smokers made twice as many mistakes on tests of executive function, which included planning, flexibility, abstract thinking and inhibition of inappropriate responses. As adults, those who started using before 16 reported smoking nearly 25 times per week, while those who started later smoked half as often, about 12 times per week. The early-onset smokers also reported smoking an average of nearly 15 grams each week, versus about 6 grams for their late-onset counterparts (Psychology of Addictive Behaviors, 2012).

Gruber’s participants had reported using marijuana at least five times in the past week. But other labs have found structural differences in the brains of less frequent users. Jodi Gilman, PhD, at Massachusetts General Hospital/Harvard Center for Addiction Medicine, and colleagues used MRI to look for brain changes in 18- to 25-year-olds who smoked marijuana at least once per week, but were not dependent on the drug.

Compared with nonusers, the smokers had changes in the shape, volume and gray matter density of two brain regions associated with addiction: the nucleus accumbens (which plays a role in motivation, pleasure and reward processing) and the amygdala (a region involved in memory, emotion and decision-making). Participants who smoked more often had more significant differences (Journal of Neuroscience, 2014).



However . . .


But the case against marijuana isn’t closed. Other studies have failed to turn up evidence that marijuana use results in brain abnormalities. In one recent example, Barbara Weiland, PhD, at the University of Colorado at Boulder, and colleagues attempted to replicate Gilman’s study in adolescents and adults who smoked marijuana daily. But Weiland’s team argued that previous studies, including Gilman’s, failed to adequately control for alcohol use by the participants. After carefully matching for alcohol intake in the control and experimental subjects, the researchers failed to find physical differences in the nucleus acumbens or the amygdala of daily marijuana smokers (Journal of Neuroscience, 2015).

On the other hand, says Lisdahl, Weiland’s subjects were primarily male — and some research suggests females might be more sensitive to marijuana’s effects during adolescence.

In other cases, too, the evidence against marijuana is frustratingly mixed. While some studies have found increased risk for mood disorders and psychotic symptoms among marijuana users, for instance, a new study by Jordan Bechtold, PhD, at the University of Pittsburgh Medical Center, and colleagues found that chronic use among teenage boys did not raise the risk of later depression, lung cancer, asthma or psychotic symptoms (Psychology of Addictive Behaviors, 2015).


Context. It matters. But on the bright side, the researchers are doing what they do best and attempting to close the information gap.

In hopes of painting a clearer picture of marijuana’s potential risks to youth, NIDA plans to launch the Adolescent Brain and Cognitive Development (ABCD) study later this year. The prospective longitudinal study will follow 10,000 individuals across the United States over a decade, starting when they’re 9 or 10. “The idea is to look at what these kids are like before they start using substances, and then follow over time what happens to their brains,” Weiss says.


With that out of the way, let’s see what else the Examiner editorial has for us to sus out.


This means that use among teenagers, which is a lot more common than people would like to admit, and even use among young adults has deleterious and permanent health effects.


The only people that don’t want to admit that teenagers use (or are prone to use, at very least) cannabis are those with their heads up their ass. I suspect that this group overlaps with the Washington Examiners’ audience base, but of course, this is just a hypothesis.

One legislative action that tends to decrease teen cannabis use, however, is legalization. Selling cannabis through age-restricted dispensaries.

Whoda thought. Forcing sales through underground sources that don’t require an ID for the purchase of cannabis results in many more teenagers accessing the drug.


This is something to remember when these campaigners come to your state and try to sell you on the idea of cannabis as something completely harmless. One need not exaggerate the dangers of marijuana to acknowledge that they at least exist.


And YET, here we are.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.