Childhood adversity and daily cannabis carry the largest attributable shares of psychosis

Design

Umbrella review of 69 systematic reviews with meta-analyses of observational studies

Population

970 studies examining 221 factors for non-organic psychotic disorders

Primary outcome

Grade of evidence for association with psychotic disorder

Effect

6 class I factors, e.g. birthweight <2,000 g OR 1.77 (95% CI 1.48–2.13); PAF childhood adversity 40.7%

An umbrella review pooled 69 systematic reviews with meta-analyses, covering 970 studies and 221 putative risk or protective factors for non-organic psychotic disorders published between 2017 and March 2025. Each association was graded from convincing (class I) to weak (class IV), with sensitivity analyses restricted to prospective studies and using stricter uncertainty methods. It appeared in World Psychiatry on 1 October.

Six factors reached class I: Black-African ethnicity in England (OR 4.89), South Asian ethnicity in England (OR 2.15), paternal age under 20 (OR 1.34), birthweight under 2,000 g (OR 1.77) or under 2,500 g (OR 1.50), and a mother with three or more previous pregnancies (OR 1.32). Eighteen more were highly suggestive, including maternal psychosis, first-generation migration, cannabis use 5–7 days a week, ADHD and Toxoplasma gondii IgG positivity.

The population attributable fractions for potentially modifiable factors were largest for childhood adversity (about 41%), Toxoplasma seropositivity (about 20%), near-daily cannabis (about 13%) and the clinical high-risk state (about 13%). These are estimates from observational data and assume causation, so they mark targets for prevention research rather than proven savings. Clinically, they support asking about trauma and cannabis frequency in every young person with emerging psychotic symptoms.

• Ask young people with unusual experiences how many days a week they use cannabis, not just whether they use it.
• Take a history of childhood adversity in first-episode psychosis; it was the largest attributable factor.
• Record family history of psychosis and other maternal mental illness, both highly suggestive risk factors.
• Refer people meeting clinical high-risk criteria to early intervention services where they exist.
• In migrant and minority patients, treat raised risk as a reason for early review, not a diagnostic shortcut.

Why it matters

It separates the few psychosis risk factors with convincing evidence from the many that are weak or null.

Don't overread it

These are associations — the attributable fractions assume causation the data cannot prove.

The statistics, in plain English

An odds ratio of 1.34 means modestly higher odds, while 4.89 is a strong association. A population attributable fraction of 41% is the share of cases that would disappear if the factor were removed and if it truly caused psychosis — a large assumption for observational data.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.