Amid the search for solutions to the opioid epidemic, which kills an estimated 130 Americans every day, some argue that increased access to cannabis could reduce this devastating toll. Part of their reasoning? A 2014 paper published in JAMA Internal Medicine reported lower opioid overdose death rates in states with medical marijuana laws.

A study published Monday in PNAS contradicts that widely cited paper, raising new questions about whether and how medical marijuana can affect the opioid crisis.

The 2014 study found that between 1999 and 2010, states with medical cannabis laws had a nearly 25% lower average rate of opioid overdose deaths than states without such laws. Much has changed since 2010 — 34 states have now legalized medical marijuana and the number of opioid overdose deaths was six times higher in 2017 than it was in 1999 — so Stanford University researchers decided to replicate the original study and expand its analysis to include seven more years of data.


When they limited their analyses to the same time period as the original study, they saw the same trend. But when they expanded the time frame through 2017, the association between medical marijuana laws and opioid overdose deaths reversed: States with medical marijuana laws had average rates of opioid overdose deaths that were nearly 23% higher than those without these laws.

Lead author Chelsea Shover, a postdoctoral research fellow in psychiatry at Stanford, said this apparent reversal is cause for concern. The 2014 study has been cited in more than 350 scientific articles and drew significant public and media attention. Advocates, industry representatives, and policymakers point to the 2014 findings to argue that marijuana liberalization could help address the opioid epidemic.

“It’s become such a pervasive idea,” said Shover. “It would be amazing if it was this simple, but the evidence is telling us now that it’s not.”

It’s unlikely that the change in findings means that medical marijuana laws were once helpful and are now harmful, Shover said. Instead, as she and her colleagues wrote, the findings suggest that the relationship between medical marijuana laws and opioid overdose deaths “is spurious.”

“This isn’t to say that cannabis was saving lives 10 years ago and it’s killing people today,” said Shover. “We’re saying these two things are probably not causally related.”

Since the Stanford study used the same methods as the original, Shover said it should garner just as much attention. Advocates and policymakers that pointed to data from the 2014 should be interested to know that a longer time frame led to a very different result.

Wayne Hall, a professor at the University of Queensland Centre for Youth Substance Abuse Research in Australia, said the 2014 study made an impact outside the United States. “It’s been cited in my own country as compelling evidence that medical cannabis reduces opioid overdose deaths,” he said. “It will be interesting to see what the response is to the new findings.”

Dr. Sharon Levy, director of the adolescent substance use and addition program at Boston Children’s Hospital and associate professor of pediatrics at Harvard Medical School, has previously expressed concerns about marijuana liberalization as a method to address the opioid epidemic. The relationship between marijuana use and opioid use is extremely complex, Levy said, and poorly understood.

“My life’s work is devoted to helping people with addiction,” said Levy. “But this is the wrong approach.”

Like opioids, medical marijuana is often sought out to manage pain. In fact, according to a 2018 report by New York’s medical marijuana program, the majority of people registering for the program listed severe or chronic pain as their primary condition. If medical marijuana is available as an option for dealing with pain, some people might use it instead of opioids — at least that’s the hypothesis. Several states have even added opioid dependency or “all conditions for which opioids could be prescribed” into their list of qualifying conditions for medical marijuana.

It’s a plausible idea, Shover noted, but these studies should not be considered evidence that it would work. Both the 2014 study and her research were ecological studies. These examine exposures and outcomes at the population level, but cannot capture what is happening at the individual level. In this case, these studies can’t demonstrate that individuals decided to replace opioids with cannabis but only show the number of overdose deaths after states passed medical marijuana laws. Any number of other changes in the state could be factors.

There’s another important limitation. “These large population studies don’t show causality,” said Yasmin Hurd, a neuroscientist and director of the Addiction Institute at Mount Sinai in New York. “There is really no evidence that the changes in medical cannabis laws correlate directly to opioid overdose mortality.”

To get a better idea about how these two factors relate, Hurd said large-scale clinical trials could better reveal how individuals use marijuana and respond to it, and whether cannabis availability could actually reduce the use of opioids. One of Hurd’s recent studies, for example, found that cannabidiol (CBD) could decrease craving and anxiety. But, she said, more studies and larger clinical trials are needed to confirm this finding.

“In a time of an epidemic, we have to think differently,” said Hurd. “We have to be more bold in pushing forward clinical trials on a much faster timeline than we have in the past.”

“This isn’t to say that cannabis was saving lives 10 years ago and it’s killing people today. We’re saying these two things are probably not causally related.”

Stanford's Chelsea Shover, the paper's lead author

The issue of one study contradicting another is not unique to marijuana liberalization and the opioid crisis. Reproducibility is an important and often overlooked aspect of scientific research. Preliminary studies with new and exciting findings understandably garner attention, but it often takes some time before they can be replicated.

The hype surrounding new studies can be especially problematic when they are used to shape public opinion and influence policy decisions. In this case, if the 2014 study is being used to make decisions about how to address the opioid crisis, it’s important to verify its findings.

Hurd agreed. “Is cannabis less of a mortality risk than opioids? Absolutely. Hands down,” she said. “But there’s really no research that says cannabis use per se decreases opioid overdose. You can’t make your medical cannabis laws based on that [hypothesis].”

Medical marijuana does have other therapeutic uses, Shover said, and there are plenty of other, well-supported reasons to consider medical and recreational marijuana laws. “It just doesn’t seem like reducing opioid overdose mortality is one of those benefits,” she said.

There are several strategies for preventing opioid overdose that have more evidence behind them, including increased access to known treatments like methadone and overdose reversals like naloxone. If medical marijuana laws don’t actually affect the number of opioid overdose deaths, then passing them could give policymakers and the public a false sense of addressing the issue.

“The risk is distracting attention away from much more effective interventions that should be implemented to reduce overdose deaths,” said Hall.

Leave a Comment

Please enter your name.
Please enter a comment.

  • Regarding Diane Armstrong’s comment of BlATANT LIE:

    Who knows? The opioid use may have declined due to its effectiveness on pain, and/or maybe some people just transitioned to a new and safer way to stay HIGH!

  • Did you take into consideration that during the 2010 to 2017 period doctors were stopping prescribing opiods, because of the CDC guidelines and local legislation?? This may have skewed your data.

  • Also consider that, while sev several states have passed laws , few have implemented such laws. The heavy restrictions will do nothing to change people who are already end-stage addicts or chronic users. More “yellow journalism”. Is this FOX News?

  • Are you ignoring the AMA studies that indicate that less opiods are needed when supplemented by cannabis? Reducing the necessary amount of opiods to achieve the same level of relief seems to clearly suggest a decrease in addiction.

  • This article is a BLATANT LIE. And has already been proven wrong. There is already scientific proof that it absolutely DOES reduce opioid use. YOU really need to verify facts before spreading more false propaganda.

  • It does not help that opioids are handed out like candy from doctors. I know that is has become more restrictive when receiving an opioid prescription yes but I also feel that if someone is completely addicted to opioids and if they are in rehab, cannabis is not even an option in most circumstances. A state run rehab facility will absolutely not counter an opioid addiction with cannabis because of it still being federally illegal. There are some rehab facilities in Colorado and other legal states that do use cannabis for addiction recovery but they are so few and far between that the ratio of addicts to cannabis usage is extremely low due to the fact if inaccessibility to cannabis for the same reason and that shows in the numbers. End of the day, legalize cannabis nationwide=more access to this miracle medicine for everyone.

  • Neither negative nor positive effects of long-term or regular MJ use have been studied, but just because it can play with the mind the use of it became criminalized. Speaking of bias ……. It is time for more research, this is one of them, but let’s keep an open mind and not just simply denounce the millions of users (seniors a large group of them) that swear by the merits of CBD. The debilitating stigma needs to be removed !!

    • Before you can simply dismiss inconvenient data as “biased”, you must explain how the study is inaccurate. This is a well-designed study by a leading researcher in the field. It is completely normal for leading marijuana researchers to join organizations like Smart Approaches to Marijuana.

    • @Dex Neither study was designed to determine a causal relationship. The problem with SAM is they use correlations as evidence and even use thoroughly debunked myths as arguments to keep cannabis illegal. The reason they chose to replicate this study is so they can use it to further push their prohibition agenda. Even though they admit correlation is not causation I can guarantee they’ll still keep using the same old arguments which are all based on correlation. If they cared about scientific integrity that would be one thing but it’s clear to me based on their history that this is not the case.

  • “The hype surrounding new studies can be especially problematic when they are used to shape public opinion and influence policy decisions. In this case, if the 2014 study is being used to make decisions about how to address the opioid crisis, it’s important to verify its findings.”

    Funny how this doesn’t apply to the harms of cannabis. There is absolutely no causation when it comes to any of the harms yet we’ve based all of our public policy on correlations that may not exist.

  • Marijuana might be fine for people that just want to stay HIGH all day, and I don’t mean for those that just use it recreationally on occasion, but it has no place on the Chronic Pain venue! Pot may not kill people initially, but I wonder how many people look back in their rear view mirror, or even think about about its impact on their family?

Sign up for our Daily Recap newsletter

A roundup of STAT’s top stories of the day in science and medicine

Privacy Policy