A STUDY PUBLISHED IN JAMA INTERNAL MEDICINE IN MAY 2018 seemed to offer a clear way forward in the fight against opioid addiction. When a state opens a legal dispensary for medical cannabis, the researchers found, Medicare Part D prescriptions for opioids in that state drop by an average of 3.7 million daily doses per year—an average decline of nearly 15%.

The reason opioid prescriptions drop, the authors suggest, is that people switch to cannabis to manage pain. That’s not a current indication for a cannabis prescription, although the JAMA paper includes links to studies exploring its potential as an analgesic.

To the JAMA paper’s authors, the call to action seems clear. “If you take an evidence-based approach, you would do well to liberalize cannabis policies at state and federal levels,” says the study’s lead author, W. David Bradford, a professor in the Department of Public Administration and Policy at the University of Georgia.

Another paper, however, published in April in the Journal of Addiction Medicine, casts doubt on the notion that increased access to cannabis may reduce use of opioids. The authors—Keith Humphreys, a professor of psychiatry and behavioral sciences at Stanford University, and Theodore Caputi, George J. Mitchell Scholar at University College Cork in Ireland—combed through the data of a 2015 national survey about drug-use habits among Americans age 12 and older. They found that medical marijuana users are about twice as likely to misuse prescription drugs—including opioids—as those who don’t use medical marijuana. “And these associations persist after adjusting for population differences in age, race, sex, self-reported health status, family income and living in a state allowing medical marijuana,” says Caputi.

So which is correct—does legal cannabis decrease opioid misuse or lead to more of it? Lipi Roy, medical director of a New York City addiction-treatment center and clinical assistant professor at NYU Langone Health, is among those trying to make sense of the two findings. Bradford’s paper appears more rigorous, she says, because it’s based on long-term data rather than the self-reported results of a single survey.

“Several other studies also now point to a correlation between medical marijuana programs and lower rates of opioid use and overdose,” says Roy. “It’s not just one study versus another. And the ‘gateway theory’ of drug use—that marijuana is a stepping stone to harder drugs—has been refuted repeatedly.”

But Caputi argues that the approach of those other studies—statewide research that uses population data to look at drug-use habits—has pitfalls, with many confounding factors that may skew the numbers. You might theorize that a state with more liberal marijuana laws might also be more likely to address the opioid problem with education or treatment programs, which could in turn cause an overall dip in addiction numbers that is unrelated to cannabis dispensaries.

Caputi feels that looking at individual user habits, as his study does, is a step in the right direction. “But until better, individual longitudinal-level data is available, the question will remain unanswered,” he adds.

Roy agrees that more research is needed. While many medical and public health officials have warmed to the idea of access to medical cannabis as a response to opioid overuse, studies supporting or disputing the effectiveness of this approach remain relatively inconclusive, largely because conducting that kind of research is held back by federal laws, says Roy.

Though cannabis is the most widely used illicit drug in America and is available for medical use in 30 states, it is federally classified as a Schedule I drug, which means that the U.S. government deems it to have no accepted medical use and, like heroin or LSD, a high potential for abuse. This makes it difficult for researchers to secure federal grants to study medical applications of cannabis, and it limits opportunities to examine whether it does treat pain, or what a cannabis regimen for opioid recovery might look like.

“The longer marijuana remains a Schedule I drug, the longer we will have to wait for these answers, and the longer it will be until we can understand its potential as a therapeutic drug,” says Sheila Vakharia of the Drug Policy Alliance.