Despite the growing popularity of medical cannabis for mental health and addiction, a new comprehensive review shows little evidence that these products actually help treat most of these conditions. A comprehensive analysis revealed that while cannabis-based medicines may provide mild relief for some specific problems, they do not improve symptoms such as depression and anxiety and carry a higher risk of side effects. These findings were recently published in the journal lancet psychiatry.
Cannabinoids are active compounds found in the cannabis plant. The best known of these are tetrahydrocannabinol, which causes the high associated with marijuana, and cannabidiol, a non-intoxicating compound often sold as a wellness product. In recent years, more and more people are turning to these substances to manage their mental health.
In the United States and Canada, approximately 27 percent of people between the ages of 16 and 65 report using cannabis for medical reasons. About half of those individuals use it specifically to treat mental health issues. Australia is seeing a surge in prescription approvals for cannabinoid medicines for mental illness and addiction.
This surge in use far exceeds medical evidence. Researchers wanted to know whether these plant-based medicines and drug prescriptions were really justified based on science. Jack Wilson, a researcher at the University of Sydney in Australia, led a team investigating whether these treatments actually work and are safe.
The research team sought to provide clarity as clinical use expands. They found that there was a huge gap between how often these products were prescribed and the scientific evidence to support it. To fill this gap, Wilson and his colleagues set out to collect and evaluate all of the best available data from the past 40 years.
To see if cannabinoids are effective, researchers conducted a systematic review and meta-analysis. A systematic review involves searching major scientific databases to collect all studies that meet rigorous criteria. Meta-analysis then combines numerical data from all these individual studies into one large statistical model.
Combining data in this way allows researchers to see the big picture. This gives us a clearer idea of the true effects of a treatment than looking at a single study alone. For this project, the team searched for randomized controlled trials published between January 1980 and May 2025.
Randomized controlled trials are considered the gold standard of scientific research for testing treatments. In these studies, participants are randomly assigned to either receive the treatment being studied or a placebo. Placebos are inert substances, like sugar pills, that look like real drugs but have no physical effect.
The researchers ultimately collected 54 of these trials, with a total of 2,477 participants. They specifically looked for trials where a mental health condition or substance use disorder was the primary reason for treatment. They evaluated how well cannabinoids reduced symptoms and tracked adverse events, which are unwanted side effects such as dizziness and nausea.
Results for most mental health conditions were not statistically significant. The data showed no real benefit for people suffering from anxiety, psychotic disorders, or post-traumatic stress disorder. There were also no randomized trials testing cannabis as a primary treatment for depression.
This lack of evidence is especially noteworthy because anxiety, depression, and post-traumatic stress disorder are among the most common reasons people seek medical cannabis. Researchers found similar effects for obsessive-compulsive disorder and bipolar disorder. In trials testing treatments for attention deficit hyperactivity disorder, the improvements were not statistically significant.
Similarly, no supporting results were found for eating disorders. Two studies looked at people with anorexia nervosa, an eating disorder characterized by an intense fear of gaining weight. Researchers found that there was no real difference in weight gain or physical activity levels between those who took cannabinoids and those who took a placebo.
The researchers also looked at substance use disorders, which occur when people are unable to stop using drugs or medications despite causing health or social problems. They found it was ineffective in treating opioid addiction and tobacco dependence. In fact, results showed that for people with cocaine use disorder, consuming cannabinoids actually increases their craving for cocaine.
Wilson pointed to this specific risk in a recent statement about the study. He cautioned against applying one drug to all addictions. “However, when medical cannabis was used to treat people with cocaine use disorder, their cravings increased. This means medical cannabis should not be considered for this purpose and may actually worsen cocaine dependence,” he said.
This analysis found several areas where cannabinoids offer potential benefits, although the quality of the evidence was generally considered to be low. One such area was the treatment of cannabis use disorder itself. People with this condition have a hard time controlling their marijuana use.
Data showed that using a pharmaceutical-grade combination of tetrahydrocannabinol and cannabidiol can help reduce withdrawal symptoms and the total amount of cannabis consumed by a person. Wilson compared it to other addiction treatments. “Medical cannabis can be part of an effective treatment for people with cannabis use disorder, similar to how methadone is used to treat opioid use disorder. Oral formulations of cannabis have been shown to reduce cannabis smoking when administered alongside psychotherapy,” he said.
Another area that showed modest improvement was the treatment of tic disorders and Tourette syndrome. These symptoms cause sudden, uncontrollable movements and sounds. Participants who received the combination of tetrahydrocannabinol and cannabidiol had reduced tic severity compared to those who took a placebo.
Researchers also looked at autism spectrum disorder, a developmental disorder that affects how people communicate and interact with the world. Across two studies, cannabinoids were found to be associated with a reduction in certain traits associated with autism. However, the researchers cautioned that these particular studies had a high risk of bias, meaning the results could have been skewed by the way the trials were designed or reported.
Sleep issues were another condition where cannabinoids showed promise. For people with insomnia, taking any type of cannabinoid led to an increase in total sleep time. This was measured both by an electronic sleep tracker and by participants writing sleep diaries.
Despite these several positive indications, the safety data raised some concerns. Participants who took cannabinoids were more likely to experience common adverse events than those in the placebo group. For every seven people treated with these drugs, one additional person experienced side effects such as dry mouth, nausea, and diarrhea.
The researchers noted that the drug did not increase the chance of serious adverse events. A serious adverse event is a serious medical problem that may require hospitalization or pose a serious health threat. People taking cannabinoids were also no more likely to drop out of the study than people taking a placebo.
Even though the side effect profile is relatively mild, Wilson cautioned that these findings have far-reaching implications. He pointed out that unproven treatments carry hidden risks. “Although our paper does not specifically address this, routine use of medical cannabis may do more harm than good by worsening mental health outcomes, for example by increasing the risk of developing psychotic symptoms and cannabis use disorder, and by delaying the use of more effective treatments,” he said.
There are several caveats to consider when interpreting these results. Studies included in reviews are often very small, which makes it difficult to draw firm conclusions. The researchers also noted that many trials are at high risk of bias, as the companies that make cannabis products are often involved in funding and running the studies.
Furthermore, this review only considered results from the longest follow-up period of each study. This means that short-term benefits or harms that occurred early in the treatment process may have been missed. The analysis was also limited by a lack of data on whether these treatments affect men and women differently.
The researchers emphasized the need for better designed future studies. Future trials should include larger groups of participants to provide clearer and more accurate results. Scientists also need to conduct research that is free from industry influence to ensure that their findings are completely independent.
Researchers are urging medical professionals to use extreme caution until better evidence arrives. They hope their work will lead to safer prescribing practices. “Our research provides a comprehensive and independent assessment of the benefits and risks of cannabis medicines, helping clinicians make evidence-based decisions and ensuring patients receive effective treatment while minimizing harm from ineffective or unsafe cannabis products,” Wilson said.
The study, “Efficacy and Safety of Cannabinoids in the Treatment of Mental and Substance Use Disorders: A Systematic Review and Meta-Analysis,” was authored by Jack Wilson, Olivia Dobson, Andrew Langcake, Parkesh Mishra, Zachary Bryant, Janie Leung, Daniel Dawson, Myfanwy Graham, Marie Teeson, Tom P. Freeman, Wayne Hall, Gary CK Chan, and Emily. stockings.
