Is Cannabis Really Linked to Psychosis—Or Is It the Other Substances?

Cannabis use was initially associated with psychotic-like experiences in 62,000 Americans, but the association disappeared when researchers accounted for methamphetamine, cigarette, and opioid use.

Johnson, Emma C et al.·medRxiv : the preprint server for health sciences·2025·Moderate EvidenceCross-Sectional·1 min read
RTHC-06765Cross SectionalModerate Evidence2025RETHINKTHC RESEARCH DATABASErethinkthc.com/research

Quick Facts

Study Type
Cross-Sectional
Evidence
Moderate Evidence
Sample
N=62,153
Participants
N=62,153 adults, diverse population from the All of Us Research Program in the US.

What This Study Found

Drawing from the All of Us Research Program—a massive, diverse U.S. biobank—this study examined whether cannabis use is associated with psychotic-like experiences (PLEs) such as hearing things, seeing things, referential ideation, and persecutory thoughts.

The initial results seemed to confirm the cannabis-psychosis link: ever using cannabis was associated with higher odds of all four types of PLEs (odds ratios 1.21–1.44). More frequent cannabis use showed a dose-response relationship—the more someone used, the more likely they were to report PLEs. These associations held even after adjusting for personal and family history of schizophrenia and genetic risk (polygenic scores).

Then came the critical analysis. When the researchers adjusted for lifetime use of other substances—methamphetamine, cigarettes, and opioids—the cannabis association vanished entirely. Methamphetamine use, cigarette use, and opioid use remained independently associated with PLEs (odds ratios 1.22–1.65), but cannabis did not.

Another telling finding: the schizophrenia polygenic score predicted whether someone was prescribed medication for their PLEs, but cannabis use frequency did not. This suggests that while cannabis users may report more psychotic-like experiences, these experiences may be less clinically severe or less likely to be true psychotic symptoms.

The study can't prove that cannabis doesn't cause psychosis—cross-sectional data can't establish causation in either direction. But it does suggest that much of the observed cannabis-psychosis link may be confounded by co-occurring use of other substances.

Key Numbers

N = 62,153. Cannabis ever-use initially associated with PLEs (ORs 1.21–1.44). After adjusting for other substance use: cannabis association no longer significant. Methamphetamine (OR 1.22–1.65), cigarettes, and opioids remained significant. Schizophrenia polygenic score predicted PLE medication, cannabis use frequency did not.

How They Did This

Cross-sectional analysis using the All of Us Research Program (release 8), a population-based U.S. biobank. Maximum analytic N = 62,153. Self-reported cannabis use and four types of self-reported psychotic-like experiences (auditory perceptual distortions, visual perceptual distortions, referential ideation, persecutory ideation). Adjusted for personal and family history of schizophrenia, polygenic risk scores, and lifetime use of other substances.

Why This Research Matters

The cannabis-psychosis association is one of the most debated topics in cannabis research and has driven significant policy decisions. If much of that association is actually explained by co-occurring methamphetamine, cigarette, and opioid use, it would fundamentally change how we understand the risk. This doesn't mean cannabis is safe for psychosis-prone individuals, but it suggests the risk may be overstated when other substance use isn't properly accounted for.

The Bigger Picture

This challenges the straightforward interpretation of the cannabis-psychosis literature that has accumulated over decades. If the association is largely confounded by polysubstance use, then studies that didn't adjust for other substances (which is most of them) may have overestimated cannabis-specific risk. This aligns with RTHC-00162's point that understanding cannabis risks requires looking at the full picture of substance use patterns, not cannabis in isolation.

What This Study Doesn't Tell Us

Cross-sectional design—can't determine whether substance use preceded or followed PLEs. Self-reported substance use and psychotic experiences (potential underreporting). The All of Us cohort, while large and diverse, is not perfectly representative of the U.S. population. Adjusting for other substance use may over-control if cannabis is a gateway to those substances. Lifetime ever-use is a crude measure that doesn't capture timing, dose, or potency.

Questions This Raises

  • ?Would prospective studies that follow people over time show the same pattern of confounding by other substances?
  • ?Does the cannabis-psychosis link reappear with high-potency products that weren't captured in this lifetime-use measure?
  • ?If cannabis users who also use meth and opioids are driving the association, what does that mean for cannabis-only users?

Trust & Context

Key Stat:
Evidence Grade:
Large cross-sectional analysis from a major U.S. biobank with genetic data, but limited by self-report and inability to establish temporal relationships.
Study Age:
Published in 2025 as a preprint (medRxiv), meaning it has not yet undergone peer review.
Original Title:
Cannabis use and psychotic-like experiences in the All of Us Research Program.
Published In:
medRxiv : the preprint server for health sciences (2025)medRxiv is a preprint server for health sciences, known for sharing early research findings.
Database ID:
RTHC-06765

Evidence Hierarchy

Meta-Analysis / Systematic Review
Randomized Controlled Trial
Cohort / Case-Control
Cross-Sectional / ObservationalSnapshot without intervening
This study
Case Report / Animal Study

A snapshot of a population at one point in time.

What do these levels mean? →

Read More on RethinkTHC

Cite This Study

RTHC-06765·https://rethinkthc.com/research/RTHC-06765

APA

Johnson, Emma C; Luo, Zhen; Romero Villela, Pamela N; Agrawal, Arpana; Hatoum, Alexander S; Karcher, Nicole R. (2025). Cannabis use and psychotic-like experiences in the All of Us Research Program.. medRxiv : the preprint server for health sciences. https://doi.org/10.64898/2025.12.01.25341322

MLA

Johnson, Emma C, et al. "Cannabis use and psychotic-like experiences in the All of Us Research Program.." medRxiv : the preprint server for health sciences, 2025. https://doi.org/10.64898/2025.12.01.25341322

RethinkTHC

RethinkTHC Research Database. "Cannabis use and psychotic-like experiences in the All of Us..." RTHC-06765. Retrieved from https://rethinkthc.com/research/johnson-2025-cannabis-use-and-psychoticlike

Access the Original Study

Study data sourced from PubMed, a service of the U.S. National Library of Medicine, National Institutes of Health.

This study breakdown was produced by the RethinkTHC research team. We analyze and report published research findings without making health recommendations. All interpretations are based solely on the published abstract and study data.