Is adolescent THC use linked to mental illness? Let’s see what the science has to say:
Adolescent Cannabis Use and Psychiatric Risk: What the Latest Research Shows
A landmark study published in JAMA Health Forum in February 2026 tracked nearly half a million adolescents over nearly a decade to answer an increasingly urgent question: does cannabis use during the teenage years increase the risk of developing psychiatric disorders? The answer appears to be yes, and the effects are stronger than many expected.
The Study: Scale and Scope
Researchers at Kaiser Permanente Northern California followed 463,396 adolescents aged 13 to 17 years from 2016 through 2023, documenting their cannabis use through routine screening during well-child visits and tracking their psychiatric diagnoses through electronic health records. About 5.7% of adolescents reported past-year cannabis use at baseline, and researchers followed participants through age 26 to examine whether cannabis use predicted the development of clinically diagnosed psychiatric disorders.
Key Findings
The associations were substantial. Adolescents who reported cannabis use had:
2.19 times greater risk of developing psychotic disorders
2.01 times greater risk of bipolar disorder
1.34 times greater risk of depressive disorder
1.24 times greater risk of anxiety disorder
Notably, these risks remained significant even after the researchers adjusted for other substance use (alcohol and other drugs), socioeconomic factors, and neighborhood characteristics. The time from first reported cannabis use to psychiatric diagnosis averaged 1.7 to 2.3 years, suggesting that cannabis use preceded illness onset rather than resulting from it.
An important age-specific pattern emerged: the associations between cannabis and depression or anxiety disorders weakened substantially as adolescents aged. By the early twenties (ages 21-25), cannabis use was no longer significantly associated with depression or anxiety—though associations with psychotic and bipolar disorders remained robust across ages.
Strengths of This Research
This study has several methodological advantages over prior work. It examined a large, diverse, population-based sample rather than small clinical cohorts. Universal screening during routine care reduces the selection bias and recall problems that plague retrospective studies. The researchers measured any cannabis use not just cannabis use disorder diagnoses at scale, which is more practically relevant. They also adjusted for multiple confounders and other substance use, examined age-specific effects, and found that results held up across sensitivity analyses that excluded participants with pre-existing psychiatric conditions.
Critical Limitations Worth Understanding
However, several limitations warrant candid discussion. Self-reported cannabis use may underestimate actual consumption, and the study captured only past-year use, not frequency, mode of delivery, or THC potency, factors likely to matter substantially. The study was conducted in California, where cannabis is legal and widely available with high average THC content (>20%), potentially limiting generalizability to other contexts.
Psychiatric diagnoses relied on EHR documentation; some youth with symptoms may not have received formal evaluation. While cannabis use preceded diagnoses temporally, causation cannot be definitively established. Unmeasured confounders—like genetic vulnerability, adverse childhood experiences, or parental mental health—could contribute to both cannabis use and psychiatric outcomes. The researchers acknowledge this honestly: an unmeasured confounder would need substantial strength (hazard ratios of 1.79 to 3.79) to fully explain the associations, but such confounders could exist.
The "reverse causation" question also matters: some adolescents may use cannabis to self-medicate early psychiatric symptoms before diagnosis, making it impossible to know whether cannabis causes illness or emerges as a symptom management attempt.
Clinical and Patient-Facing Implications
This study strengthens the evidence that adolescence represents a vulnerable developmental window for cannabis exposure, particularly for severe outcomes like psychosis and bipolar disorder. The findings align with neuroscientific understanding that THC disrupts the endocannabinoid system during critical brain development periods.
For clinical practice, these data support early screening for cannabis use during adolescent visits and warrant age-appropriate conversations about risks, particularly for youth with family history of psychotic or bipolar disorders, who may carry genetic vulnerability. The data don't prove cannabis causes these conditions in any given individual, but they do suggest the risk is real and measurable at the population level.
