Cannabis use linked to violence in a review spanning more than 265,000 people
A large systematic review and meta-analysis has tied cannabis use to a higher likelihood of both committing violent acts and becoming a victim of violence. The analysis pooled 63 studies and more than 265,000 people, separating findings from the general population and from patients receiving psychiatric care. Across those groups, the association ran in both directions: cannabis use tracked with more violence toward others and more violence experienced by the user.
The review does not establish causation, and it does not claim that cannabis alone explains the behavior. Even so, the size of the dataset gives the findings real clinical weight. In mental health settings, where substance use, trauma exposure, psychosis, mood disorders, and social instability often intersect, a signal that broadens the violence conversation beyond intoxication alone deserves careful attention.
One result stands out sharply. In the general population, cannabis users were around twice as likely to have committed a violent act as people who did not use cannabis.
Psychiatric care showed a stronger association
Among people receiving psychiatric care, the link was even more pronounced. Cannabis users in psychiatric settings were around two and a half times as likely to be violent as psychiatric patients who did not use cannabis. For violence that ended in criminal conviction, cannabis use carried a three- to fourfold higher likelihood of conviction. The review also found that people who used cannabis were about one and a half times as likely to become victims of violence, with the victimization signal stronger among women than among men.
Those differences matter clinically because psychiatric populations are not a simple extension of the general population. Baseline rates of psychosis, trauma, impulsivity, substance co-use, and housing insecurity are often higher, and each of those factors can complicate both violent behavior and vulnerability to assault. A substance association inside this setting can therefore reflect layered risk rather than a single direct biological pathway. The strength of the association in psychiatric care suggests cannabis use should be treated as part of a broader risk assessment, not as a standalone explanation.
Why the association may not mean direct causation
The review used the World Health Organization’s broad definition of violence, covering physical, sexual, and psychological violence directed at another person, both inside and outside intimate relationships. It also distinguished self-reported violence from criminal convictions, and when studies measured more than one type of violence, the review focused on the most serious outcome reported.
That approach captures meaningful clinical behavior, but it also introduces heterogeneity. Different studies measured different forms of violence, and the review had limited data on how much cannabis participants used or how often they used it. Most studies were also conducted before high-potency cannabis products became common, which limits how confidently these findings can be translated to today’s daily users of stronger products.
A plausible clinical explanation is not hard to imagine. Cannabis use often travels with alcohol use, other drug use, and difficult social circumstances, each of which can elevate the risk of aggression or victimization. In psychiatric settings, cannabis may also worsen paranoia, disinhibition, or mood instability in susceptible people, and those effects can compound preexisting clinical risk. None of that proves causality, but it does explain why clinicians cannot treat the association as incidental.
Bedside implications for psychiatry and public health
The practical consequence is a broader screening conversation. The investigators argue that mental health services should routinely ask about cannabis use when assessing risk of violence. That approach fits everyday psychiatric care, where risk formulation already depends on substance history, prior trauma, psychotic symptoms, access to weapons, and recent changes in behavior.
Public health messaging may also need to widen. Cannabis policy discussions often focus on intoxication, dependence, cognitive effects, or respiratory harms. Violence risk, including the possibility of becoming a victim, rarely receives equal attention. The review suggests that omission leaves out a clinically relevant part of the picture, especially as cannabis becomes more commercialized and legalized in more jurisdictions.
For women, the stronger victimization association deserves particular care. Victimization risk can reflect intimate partner violence, exploitation, or broader social exposure, and cannabis use may function as a marker for environments in which safety is already compromised. In that context, asking about cannabis use is not merely a substance-use checkbox; it can open a conversation about immediate safety, coercion, and support needs.
What clinicians should take from the data
The findings support vigilance, not stigma. Heavy cannabis use may be one factor among many that changes violence risk, and the clinical response should reflect that complexity. In psychiatric practice, a positive cannabis history should prompt attention to co-use of alcohol and other drugs, recent escalation in use, psychotic symptoms, and social instability that could increase harm to the patient or others.
The review’s limits are important. Only a small number of studies measured dose or frequency of use, so the evidence cannot yet define a clear threshold at which risk rises. The conviction data came from a relatively limited set of studies, and the older evidence base may not fully represent current high-potency products. Even so, the pattern across a large body of observational research is difficult to ignore: cannabis use tracked with more violence, more victimization, and stronger associations in psychiatric care than in the general population.
The next question is not whether cannabis should be discussed in violence risk assessment; it already should be. The harder issue is how risk changes with dose, potency, co-use, and psychiatric diagnosis, and whether today’s products intensify the signal seen in older studies.