| Audience | Patients, caregivers, primary-care clinicians, mental-health clinicians, cannabis-medicine professionals, and public-health readers |
| Primary Topic | National trends in self-reported poor mental health among U.S. adults reporting past-month cannabis use |
| Source | Read the full study |
Table of Contents
- Cannabis Use and Poor Mental Health: What the New U.S. Trends Show
- Why a Large Association Still Cannot Tell Us Which Came First
- The Same Study Can Mean Different Things Depending on the Question Being Asked
- Use the Finding as a Prompt to Track Your Own Pattern
- Screen Without Assuming the Direction of Cause
- The Outcome Is Broad and the Exposure Is Blunt
- Important Confounders and Time Order Remain Unresolved
- The Results Fit a Larger, Complicated Literature
- Product Details Matter More Than the Survey Could Capture
- Longitudinal and Product-Specific Research Is Needed
- Public Health Messaging Should Inform Without Stigmatizing
- Frequently Asked Questions
Cannabis Use and Poor Mental Health: What the New U.S. Trends Show
A nationally representative analysis found that poor-mental-health days became more common among adults reporting cannabis use between 2016 and 2023, especially among younger adults and women. The study identifies a screening signal, not proof that cannabis caused the trend.
| Study Type | Repeated cross-sectional analysis of a nationally representative telephone survey |
| Data Source | Behavioral Risk Factor Surveillance System, 2016-2023 |
| Population | U.S. adults aged 18 years and older |
| Sample Size | 865,178 unweighted respondents, representing 448,137,324 weighted respondents |
| Cannabis Measure | Any self-reported cannabis use during the previous 30 days; frequency and route were also categorized |
| Mental Health Measure | At least one day in the previous 30 days when mental health was reported as not good |
| Analysis | Survey-weighted prevalence comparisons and multivariable logistic regression |
| Main Adjustment Variables | Year, age, sex, race and ethnicity, education, marital status, income, and employment |
| Journal | Social Psychiatry and Psychiatric Epidemiology |
| Published | June 11, 2026 |
| PMID | 42274737 |
| DOI | 10.1007/s00127-026-03145-w |
| Funding / Conflicts | The authors reported no supporting funding and no competing interests |
The investigators analyzed BRFSS responses collected from 2016 through 2023. Respondents were classified as using cannabis if they reported at least one day of use during the previous 30 days. Frequent use meant more than six days, and daily use meant all 30 days.
The mental-health outcome was broad. A respondent counted as having past-month poor mental health after reporting at least one day when mental health, including stress, depression, or emotional problems, was not good. This is useful for population surveillance, but it is not equivalent to a diagnosis of depression, anxiety, psychosis, or cannabis use disorder.
Among people reporting cannabis use, the weighted prevalence of at least one poor-mental-health day rose from 54.54% in 2016 to 67.88% in 2023. Across the complete study period, cannabis users reported the outcome more often than non-users: 61.34% compared with 35.14%.
The increase appeared across routes of use. In 2023, reported prevalence reached 76.96% among people who vaped cannabis and 77.18% among those who dabbed. Those categories may involve higher THC exposure, but the survey did not directly measure potency or dose.
Adults aged 18 to 24 had the highest reported prevalence in 2023 at 80.49%. Adults aged 25 to 34 showed the largest increase, rising from 51.42% in 2016 to 77.67% in 2023. These patterns support closer attention to coping motives, product strength, and functional effects in younger adults.
Women reported higher prevalence than men in 2023, 77.08% compared with 60.14%. Hispanic respondents had the highest reported prevalence among the racial and ethnic groups presented, at 69.17%. These differences may reflect many factors beyond cannabis, including baseline mental-health disparities, care access, social stressors, and reporting patterns.
After adjustment, past-month cannabis use was associated with higher odds of reporting poor mental health. Frequent use was also associated with increased odds. Daily use, however, was not significantly associated in the authors’ adjusted model.
That apparent inconsistency should discourage easy conclusions. Daily users may differ from other users in age, medical use, tolerance, product selection, reasons for use, or other unmeasured characteristics. A cross-sectional survey cannot turn these categories into a clean biological dose-response curve.
Cannabis may be used before, during, or after the emergence of distress. Some people use it to cope with anxiety, insomnia, trauma symptoms, or low mood; others experience worsening anxiety, impaired motivation, panic, or other adverse effects. Both directions can exist within the same population.
The clinically useful question is therefore not whether cannabis is categorically good or bad for mental health. It is whether a particular person’s pattern, product, dose, timing, goals, and symptoms form a helpful, neutral, or harmful relationship over time.
A finding like this should change the quality of the conversation, not end it. When cannabis use and emotional distress travel together, the responsible response is curiosity: What is being used, how often, for what purpose, and what happens to mood, sleep, anxiety, motivation, and function afterward?
The study does not give clinicians permission to blame cannabis for every symptom. It does give us a reason to screen consistently and to help patients notice patterns that broad population data cannot resolve for them.
Why a Large Association Still Cannot Tell Us Which Came First
Large samples can estimate population patterns with impressive precision. They cannot, by size alone, solve the problem of time order. In this study, cannabis use and poor mental health were measured over the same recent period.
That means several explanations remain compatible with the result: cannabis could worsen symptoms for some people, distress could lead some people to use cannabis, shared social or medical factors could influence both, or all three processes could occur in different subgroups.
Four Questions Between Association and Causation
Association -> Time Order
To infer cause, we first need to know which exposure came before which outcome. This survey cannot establish that sequence.
Cannabis Use -> Actual Exposure
A day of low-dose CBD and a day of high-potency THC dabbing are both counted as cannabis use, despite very different pharmacology.
Poor Mental Health -> Clinical Diagnosis
One self-reported difficult day is meaningful, but it is not the same as a structured diagnosis or a measure of symptom severity.
Population Pattern -> Individual Decision
A national association cannot predict whether one person’s regimen is helping, worsening, or unrelated to that person’s symptoms.
The Same Study Can Mean Different Things Depending on the Question Being Asked
Scientific papers rarely answer a single question. Patients, clinicians, researchers, and critics can read the same data differently. These evidence-based lenses show where this trial is useful, where it remains uncertain, and how easily it can be overstated.
Use the Finding as a Prompt to Track Your Own Pattern
The study does not say that cannabis is causing a mental-health problem for every person who uses it. It says that emotional distress is common in this population and has become more common over time.
A practical response is to track product type, THC and CBD content, dose, timing, reason for use, short-term effects, next-day effects, and changes in sleep or function. Patterns are more informative than labels.
Screen Without Assuming the Direction of Cause
Cannabis use should open a brief, nonjudgmental mental-health assessment, particularly for younger adults and people using frequent or high-potency inhaled products.
The assessment should include symptom timing, coping motives, substance-use history, suicidality when indicated, sleep, other medications, and whether use changes function. The survey does not justify a predetermined conclusion.
The Outcome Is Broad and the Exposure Is Blunt
At least one poor-mental-health day is a low threshold that combines stress, depression, and emotional difficulty. Cannabis use likewise combines different products, doses, motives, and frequencies.
These broad categories can reveal trends but cannot explain mechanism. The very large sample narrows statistical uncertainty while leaving measurement uncertainty intact.
Important Confounders and Time Order Remain Unresolved
The adjusted model included several demographic and socioeconomic variables, but residual confounding remains likely. Trauma, chronic pain, other substance use, psychiatric history, treatment access, and reasons for cannabis use could influence both exposure and outcome.
Repeated longitudinal measurements within the same individuals would be more informative about whether changes in cannabis use precede changes in mental health.
The Results Fit a Larger, Complicated Literature
Prior research has linked cannabis use, especially frequent or high-THC use, with several adverse mental-health outcomes. Other research documents symptom relief or coping motives in selected patients.
The new paper does not reconcile those literatures. It adds contemporary national trend data and identifies groups in whom closer assessment may be especially useful.
Product Details Matter More Than the Survey Could Capture
Route categories provide some information, but vaping and dabbing do not reveal THC concentration, inhaled amount, contaminants, or CBD content. Edibles likewise vary widely in dose and timing.
Clinically useful counseling requires details the survey lacks: actual product, dose, frequency, co-use, symptom targets, adverse effects, and whether benefits persist beyond intoxication.
Longitudinal and Product-Specific Research Is Needed
Future studies should follow individuals over time, measure baseline psychiatric symptoms, and capture THC and CBD exposure, potency, route, medical indication, and reasons for use.
Research should also test whether changes in product or frequency predict subsequent changes in symptoms and whether those relationships differ across demographic and clinical groups.
Public Health Messaging Should Inform Without Stigmatizing
The observed trends support accessible screening, potency education, honest labeling, and mental-health resources. They do not support treating all cannabis use as equivalent or all users as disordered.
Policy claims about legalization require separate analysis because the study did not link respondents to state policy conditions. Responsible messaging should state that limitation plainly.
Join the Conversation
Have a question about how this applies to your situation? Ask Dr. Caplan
Want to discuss this topic with other patients and caregivers? Join the forum discussion
Have thoughts on this? Share it:
Frequently Asked Questions
Did this study prove that cannabis causes poor mental health?
No. It was a cross-sectional survey analysis. Cannabis use and poor mental health were associated, but the study could not determine which came first or whether other factors influenced both.
How was poor mental health defined?
Respondents were counted if they reported at least one day in the previous 30 days when their mental health, including stress, depression, or emotional problems, was not good.
How large was the study?
The analysis included 865,178 unweighted respondents and used survey weights representing 448,137,324 respondents across the study years.
What changed between 2016 and 2023?
Among adults reporting cannabis use, the prevalence of at least one poor-mental-health day increased from 54.54 percent in 2016 to 67.88 percent in 2023.
Which age group had the highest prevalence?
Adults aged 18 to 24 had the highest reported prevalence in 2023 at 80.49 percent. Adults aged 25 to 34 showed the largest increase over the study period.
Did the study measure THC or CBD dose?
No. It categorized route and frequency but did not directly measure cannabinoid content, potency, dose, or product accuracy.
Were vaping and dabbing associated with higher prevalence?
Yes. In 2023, reported prevalence was 76.96 percent among people who vaped cannabis and 77.18 percent among those who dabbed, but the study could not prove those routes caused the difference.
Why was daily use not significantly associated in the adjusted model?
The study cannot fully explain that result. Daily users may differ in medical use, tolerance, product choice, reasons for use, or other unmeasured factors, so it should not be interpreted as proof that daily use is protective.
What should clinicians do with this information?
Clinicians can screen nonjudgmentally for cannabis patterns, mood, anxiety, sleep, coping motives, product potency, other substances, and functional change rather than assuming cannabis is either the cause or the solution.
Should a patient stop cannabis because of this study?
The study cannot make an individual treatment decision. Anyone concerned about worsening mood, anxiety, sleep, or function should review the pattern, product, dose, goals, and alternatives with a qualified clinician.
