Marijuana and the Opioid Crisis
[Updated 2/19 to add further discussion of Powell et al. study.]
The nation’s opioid crisis has sparked a lively debate about the relationship between marijuana and the use of harder drugs like opioid painkillers. Richard Friedman is the latest to weigh in on this debate. In a New York Times op-ed posted February 8, Marijuana Can Save Lives, Friedman lauds marijuana’s potential to reduce the nation’s use of opioids (call this the beneficial gateway hypothesis).
In relevant part, Friedman writes:
This week, Attorney General Jeff Sessions paused a discussion of the opioid epidemic to, once again, go after marijuana. He suggested that addictive pain medication wasn’t the only problem and that many heroin addicts start out “with marijuana and other drugs.”
There is a relationship between cannabis and opioids, but Mr. Sessions has it backward. Marijuana isn’t a gateway drug to opioid addiction; it’s a safer alternative to pain medicines. Mr. Sessions’s vow to crack down on marijuana will only make the opioid epidemic worse.
We know that 40 percent of all opiate overdose deaths involve a prescription opiate. So having legal access to cannabis as another option for pain relief may actually reduce consumption of opiates.
I know it sounds counterintuitive, but consider the evidence. To start, a large study [by Marcus Bachhuber, et al. (here)] assessed the effect of medical-marijuana laws on opiate-related deaths between 1999 and 2010 in all 50 states and reported a 25 percent decrease in opiate overdose mortality in states where medical marijuana was legal, compared with those where it wasn’t. . . .
Other epidemiologic studies found similar results.
The op-ed is well written and worth a full read. I agree in part and disagree in part with Friedman’s argument.
On the one hand, Friedman is right to criticize the Attorney General (or anyone else) for claiming that using marijuana leads people to use other, harder drugs like heroin. As Friedman notes in his piece, there is very little scientific evidence to support this traditional version of the gateway hypothesis. Indeed, as I point out in my book (page 218), even the Drug Enforcement Administration is unconvinced by the traditional gateway theory. In its most recent response to a scheduling decision, for example, the agency acknowledged that “research does not support a direct causal relationship between regular marijuana use and other illicit drug use” (page 218).
On the other hand, I think it somewhat premature to claim that legalizing medical marijuana will instead reduce use of those harder drugs. To be sure, there are studies supporting this beneficial version of the gateway hypothesis (including the Bachhuber study Friedman mentions). But there are reasons to take these studies with a grain of salt.
Perhaps most importantly, the studies linking medical marijuana laws to lower opioid deaths generally fail to control for significant differences in medical marijuana laws both across states and across time (see Chapters 4 and 8 for discussion of these differences). The failure to control for these differences means that comparisons across states might overstate or even understate the beneficial effects of medical marijuana.
Fortunately, one important new study—not mentioned by Friedman—begins to control for some of these differences across states and time. The study, by David Powell et al., can be found in the most recent issue of the Journal of Health Economics here.
Powell et al. examine whether the opening of legal medical marijuana dispensaries in a state (and not just the passage of a medical marijuana law) has any impact on opioid harms like overdose deaths. The hypothesis is that dispensaries should make it easier to acquire marijuana and should thereby promote any beneficial effects marijuana use might have on opioid harms. The variation in when or even whether dispensaries became operational in different states during the study period (1999-2013) gave the authors a great opportunity to test this hypothesis.
Powell et al. provide some nuanced support for the beneficial gateway hypothesis, but even this study leaves many questions unanswered.
To begin, consistent with other studies, the authors found that the passage of a state medical marijuana law is correlated with a reduction in opioid deaths, at least for the period from 1999-2010 (the same time period covered by other studies).
Unlike earlier studies, however, Powell et al. find that the beneficial result is largely if not entirely dependent on having operational medical marijuana dispensaries to supply the drug. In other words, simply legalizing possession of the drug is not enough—states need to provide for supply of the drug as well to reduce opioid harms:
“[T]he estimates imply that dispensaries reduce opioid mortality rates by about 20% while the main effect of having a law is relatively small in magnitude, implying declines of about 5%, and not statistically distinguishable from zero.” (p. 34)
So why, exactly, does marijuana use reduce opioid harms? On this key question, Powell et al. raise more questions than they provide answers. Surprisingly, the authors found no relationship between the introduction of medical marijuana laws or operating dispensaries and the supply of legal opioid painkillers:
“[O]ur findings for overdose deaths and treatment admissions are not driven by a decrease in the legal supply of opioids. This implies that access to medical marijuana through dispensaries must decrease the harm from opioids (overdoses and addiction), but not necessarily the medical use of opioids.” (p. 36)
The failure to find any reduction in the legal supply of opioids in medical marijuana states is surprising because many proponents of the beneficial gateway hypothesis (including Friedman) think that marijuana is beneficial precisely because it reduces demand for more dangerous painkillers. For example, Friedman explains in his op-ed that:
Why might cannabis work so well as an alternative to opioids? It does offer some mild pain relief. But more significant, both opiates and cannabis — like all recreational drugs — cause the release of dopamine in the brain’s reward pathway. That signal conveys a powerful sense of pleasure and craving. Thus, cannabis might preempt some of the rewarding effects of opiates, decreasing the general desire to use them.
To explain the null results on supply of opioid painkillers, Powell et al. suggest that “pharmacies may stockpile or be generally slow to adjust their inventory [following adoption of a medical marijuana law] or . . . there may be less illegal diversion when individuals can access marijuana through dispensaries.” I think this explanation is plausible, but not convincing. As for the latter explanation, the authors are essentially suggesting that many opioid users are willing to buy opioids–but NOT marijuana–on the black market. In other words, illicit opioid users are only willing to try medical marijuana if they can buy it from a dispensary the state (but not federal government) declares legal. This dynamic seems far-fetched, or at least, needs some further explanation. As for the former explanation, the authors are essentially suggesting that Walgreens et al. failed to anticipate a decline in demand for lawful opioids that would follow from the legalization of medical marijuana. This explanation seems a bit more plausible and will be amenable to testing soon (i.e., one would expect sophisticated companies like Walgreens to adjust their inventories in short order). Testing the other hypothesis — i.e., whether medical marijuana laws curb illicit use of opioids — will be more difficult, given data limitations.
I think the second finding in Powell et al. just highlights the fact that the causal mechanism behind the gateway theory is complicated and not yet fully understood.
The study is interesting and worth a read.
The bottom line: I think there is promising evidence of a beneficial gateway effect from well-designed medical marijuana laws, but the jury is still out. Long ago, the DEA prematurely bought into the harmful gateway hypothesis. (As noted above, it has since disavowed reliance on the gateway theory in its scheduling decisions.) Proponents of marijuana reform should not repeat the mistakes of the past, even when those mistakes now favor their position. As always, care should be taken in describing the potential benefits (and harms) of marijuana use.