Thirteen years ago, Uruguay became the first country on the planet to officially legalise cannabis for recreational (adult-use) purposes. In December 2013, state-controlled cannabis retail sales, home cultivation, and social clubs were fully legalised by President José Mujica.
Unlike the growing number of countries across the world that have legalised one or more of these elements in the years since, this was not the result of a zeitgeist shifting towards more liberal drug policies. It was a reluctant security measure, emerging from a region that, in Mujica’s own words, was ‘losing the battle against drugs and criminality’.
At the time, the move was criticised both internationally and domestically. Raymond Yans, then president of the International Narcotics Control Board (INCB), accused Montevideo of ‘pirate attitudes’ and said the government had ‘knowingly decided to break universally agreed and internationally endorsed legal provisions’ of the 1961 Single Convention on Narcotic Drugs.
More than a decade later, data from Uruguay and the growing number of countries that have since reformed their cannabis laws allow us to examine the impact of cannabis liberalisation on public health, and whether the critics have been proven correct.
A landmark review of global cannabis policy
Last week, a new landmark ‘narrative review’ published in The Lancet Psychiatry (vol 13: 615–26) synthesised 25 years of existing research from the Americas, Europe, Africa, Oceania, and Asia, drawing on natural experiments created by policy changes to assess what actually happened in each case.
The study, led by Professor Tom Freeman and Dr Rachel Lees Thorne at the University of Bath, who collaborated with 11 research institutions, sought to measure three outcomes: changes in cannabis use prevalence, rates of cannabis use disorder, and hospital admissions for psychiatric conditions.
It provides the most comprehensive international assessment ever conducted on what tends to accompany different types of cannabis reform, illuminating consistent patterns, but never claiming to establish causation.
Ultimately, it found that public health outcomes are less associated with whether cannabis is legal, and far more with how the market is structured and regulated.
Commercial legalisation
The most robust evidence comes from North America, where the data infrastructure is strongest. In Canada, past-12-month cannabis use among adults aged 16 and over rose from 22% before legalisation in 2018 to 27% by 2020, stabilising at 26% by 2023–24.
In Ontario between 2014 and 2021, emergency department visits for cannabis-induced psychosis rose 220.7%, a trend that had begun before legalisation but accelerated during the commercialisation phase, as retail stores expanded and higher-potency products entered the market.
In Colorado, hospitalisations for patients with comorbid psychosis and cannabis use disorder rose from 3.4 to 8.5 per 100,000 following non-medical legalisation, with retail sales beginning in 2014.
Throughout the US, the study notes that the number of people using cannabis on a daily or near-daily basis overtook those drinking alcohol at the same frequency ‘since the mid-1990s and especially after 2015’, a shift the authors regard as historically significant.
On the psychiatric evidence, the authors stated: “There is no consistent evidence for associations with the prevalence or incidence of psychotic disorders, however, commercialised legalisation was associated with an increase in hospital visits related to psychosis, and psychotic disorders comorbid with cannabis use disorder.”
Controlled legalisation models
Conversely, the data from countries with more tightly controlled markets showed little to no evidence of increased use.
In Uruguay, two studies found either no significant increase or reductions in past-year use, past-month use, and high-risk use among young people aged 12–21 compared with a matched Chilean control group.
Germany’s interim evaluation, published in September 2025, found no association between its April 2024 legalisation and changes in cannabis consumption across three population surveys.
One Swiss randomised controlled trial, which assigned 378 participants which randomly assigned 378 participants either to an illegal market control group, or a group that had legal access to cannabis for non-medical purposes, including regulated cannabis products, safer use information, and voluntary counselling, found that cannabis use disorder scores at six-month follow-up were not significantly different between two groups.
In the UK, reduced enforcement between 2004 and 2009 produced no substantial change in use among adolescents or adults, while hospital admissions for cannabis-induced psychosis declined during the same period.
However, the data from European states is thinner. Germany’s evaluation covers only months of a new framework, the Netherlands pilot has yet to report findings. Switzerland’s randomised controlled trial represents the most methodologically robust European data point so far, but covers only six months of follow-up.
As such, the authors remain cautious about drawing firm conclusions from these limited data sets, suggesting the data points to an absence of harm, not a guaranteed absence of harm over time.
On medical cannabis, now by far the dominant model of access to cannabis throughout the world, the authors raise concerns about poorly regulated markets.
As an example, Australia has generated over 700,000 prescriber approvals across more than 250 medical indications, many with limited efficacy or safety data.
They conclude that legal medical access can provide genuine benefit when supported by adequate evidence, but that wider access without that foundation may offer little benefit while increasing harm.
Thailand
The starkest case study in the review is Thailand, which in 2022 became the first country in Asia to legalise non-medical cannabis, but has since reversed course.
Following legalisation, a rapid commercialised proliferation of dispensaries drove a five-fold increase in patients diagnosed with cannabis-related disorders.
Inpatient cases of cannabis-induced psychotic disorder rose from 477 in 2017 to 2,713 in 2023, cannabis poisoning cases from 39 to 637, and acute cannabis intoxications from 76 to 1,039 over the same period.
In June 2025, Thailand’s Ministry of Public Health mandated a medical certificate for all cannabis purchases from licensed shops. The authors describe this as ‘an unprecedented test of how a commercialised legal market for non-medical use might be reversed.’
Conclusions
The review’s central finding is that the model of legalisation matters as much as legalisation itself. Commercial legalisation, characterised by retail store proliferation and the sale of higher-potency products, is consistently associated with increased cannabis use and cannabis use disorder among adults.
Tightly controlled, non-commercial legalisation, as in Uruguay, was associated with reductions in use among young people and higher-risk user groups. Decriminalisation, across Europe, Africa, Oceania, and Asia, showed little to no association with changes in use or psychiatric outcomes.
On psychosis, the picture is more nuanced. The paper finds no consistent evidence that cannabis policy changes the prevalence or incidence of psychotic disorders in the general population.
It does, however, find that commercial legalisation is associated with an increase in hospital visits related to psychosis, and with psychotic disorders comorbid with cannabis use disorder.
Although these patterns appear consistently, the study is a narrative review of natural experiments. As such the findings are synthesised through best available evidence from real-world policy changes, rather than from controlled clinical trials.
As the authors acknowledge, causation cannot be established. What the review demonstrates is association between certain policy structures and certain outcomes. Whether those outcomes are driven by legal access, by the commercial environment, by product characteristics, or by underlying social and demographic factors that also shaped the policy decision cannot be conclusively separated.
The evidence base also has a geographic imbalance. The majority of high-quality research originates from the United States and Canada, limiting the generalisability of these findings to the many countries now weighing cannabis reform.
“In a rapidly changing global cannabis policy landscape it is increasingly important to ask how policy will change, rather than if it will change at all,” said lead author Daniel Freeman.
“The type of policy change is critical. Alternative policies, such as decriminalisation or strictly regulated legalisation, can remove the harms of criminalising people who use cannabis, while limiting changes in use.”
