Cannabis in your 30s carries a bigger psychosis risk than in your teens

What just happened

A meta-analysis of 149 international studies involving more than 71,000 participants has produced a finding that challenges a persistent assumption about cannabis and psychiatric risk: the people whose psychosis onset is most accelerated by cannabis are not teenagers — they are people in their 30s.

The overall picture from the analysis is that cannabis users develop psychosis on average 2.5 years earlier than non-users. But the age-specific breakdown is where the data gets uncomfortable. Among people in their early 20s, cannabis use advances psychosis onset by about 1.6 years. Among people under 20, this analysis found no statistically significant evidence of earlier onset. Among people in their 30s: six or more years earlier.

That gradient runs the wrong way from what most people — including most clinicians — would expect.

Both-and

The assumption that adult use is safer

The conventional narrative around cannabis and mental health has been heavily weighted toward adolescent risk. Adolescent brains are still developing, particularly the prefrontal cortex — the region involved in executive function, impulse control, and social cognition. The theoretical argument for why teenagers might be especially vulnerable to cannabis-related psychiatric harm is coherent and well-publicised.

But lead researcher Carly Stevens, writing in The Conversation, suggests the new findings point to a cumulative model rather than a developmental-window model. The harm does not appear to depend primarily on how mature the brain was when cannabis use began. It appears to accumulate with years of use — gradually lowering the threshold at which a susceptible person tips into a first-episode psychosis.

If that model holds, then the period between a person’s 20s and 30s — often the years of most sustained, regular cannabis use, frequently for sleep, anxiety, or pain management — is not the safe harbour many people assume. It may be where the accumulation happens.

The Australian potency concern

Jack Wilson from the University of Sydney highlighted a specific and timely local concern: cannabis products in Australia have become substantially more potent in recent years. Delta-9-tetrahydrocannabinol (THC) concentrations in Australian flower and concentrate products are higher than at any prior point. The same pattern applies in the UK, Europe, and North America.

This matters because the meta-analysis was not able to control for cannabis potency — a limitation the authors acknowledge. If the psychosis-timing effect is dose-related, as theory suggests, then higher-potency products could produce the same outcome in a compressed timeframe. The people using increasingly potent cannabis for sleep or anxiety in their 30s — and not mentioning it to their doctor — are the ones this research speaks to most directly.

What this research does not tell us

Association is not causation in a simple linear sense. Cannabis does not cause psychosis in the majority of people who use it. The proportion at greatest risk appears to be those with genetic vulnerability — a personal or family history of psychosis, schizophrenia, or bipolar disorder with psychotic features.

What the research does tell us is that “adult use is lower-risk” is not well-supported by the data. And in a landscape where cannabis has entered mainstream wellness culture — routinely recommended online for anxiety, sleep, and chronic pain, often without a GP conversation — that framing matters.

Lead author Dr Matthew Large noted this is “a finding that has not been demonstrated in any other study”, which is both a claim worth holding carefully and a reason to watch whether replication follows.

My two cents

Cannabis has separated itself from serious psychiatric risk in mainstream wellness culture. People in their 30s using it for sleep or anxiety tend not to frame it as a risk behaviour. The narrative that adult use is safer than teenage use has given it a kind of implicit safety certification it does not warrant.

This research does not mean cannabis use in adulthood will cause psychosis. What it means is that the conversation about risk needs to be honest about where the evidence actually sits. Six years of psychosis onset brought forward is a meaningful clinical event — for the person experiencing it, for their relationships, for their work, for everything that follows.

If you use cannabis and have a personal or family history of psychosis, schizophrenia, or significant mood disorders — or if you have noticed changes in your thinking, perception, or mental state since using cannabis — this is a conversation worth having with your GP.

Verdict: maybe — the risk is real and the potency concern in Australia is legitimate, but individual susceptibility varies considerably.

---

Sources cited

1. Cannabis use and psychosis are linked, but study finds age matters. ABC Health, 19 August 2026. https://www.abc.net.au/news/health/2026-08-19/cannabis-linked-to-earlier-onset-of-psychosis/107048934
2. Cannabis isn’t ‘safer’ in your 20s and 30s — it’s linked to an earlier onset of psychosis. The Conversation, 19 August 2026. https://theconversation.com/cannabis-isnt-safer-in-your-20s-and-30s-its-linked-to-an-earlier-onset-of-psychosis-289960