| Audience | Patients, families, clinicians, public-health leaders, treatment systems, and cautious readers following cannabis risk. |
| Primary Topic | New national estimates of cannabis use disorder prevalence and severity among US adults from 2021 through 2024. |
| Source | Read the open-access JAMA Psychiatry article |
Cannabis Use Disorder Is Rising Among US Adults
National survey data show rising cannabis use disorder across US adults, but the cross-sectional trends do not identify a cause or define any individual patient’s risk.
| Publication | JAMA Psychiatry, August 19, 2026 |
| Design | Repeated cross-sectional national survey analysis |
| Dataset | 2021 to 2024 National Surveys on Drug Use and Health |
| Sample | 186,823 US adults |
| Outcome | Past-year DSM-5 cannabis use disorder and severity |
| Male CUD prevalence | 7.3% in 2021; 9.3% in 2024 |
| Female CUD prevalence | 4.5% in 2021; 5.6% in 2024 |
| Key age signal | Increases were prominent in adults aged 35 to 49 and several groups aged 50 or older |
| Evidence boundary | Population trends cannot determine cause or diagnose an individual |
A national analysis published August 19 in JAMA Psychiatry found that cannabis use disorder and moderate-to-severe disorder increased among US adults from 2021 through 2024. The study used weighted data from 186,823 adults in the National Survey on Drug Use and Health. Read the primary article.
Clinical news coverage highlighted the rise on August 26. Read the news report. The most practical implication is not that every user is at high risk, but that clinicians should ask about function, control, and severity across the adult lifespan.
Cannabis use disorder is not defined by using cannabis once, using it often, or having a medical recommendation. It involves a pattern of difficulty controlling use and continuing despite meaningful problems.
Warning signs can include unsuccessful attempts to cut down, craving, withdrawal, increasing time devoted to cannabis, or interference with work, relationships, finances, driving, or health. A confidential conversation can begin before the pattern becomes severe.
Ask about symptom goals, route, potency, frequency, dose escalation, impaired control, withdrawal, co-use, and function. Screening should be routine and nonpunitive, especially when cannabis is being used for pain, sleep, anxiety, or another persistent symptom.
A positive screen is not a complete diagnosis. Clarify severity, psychiatric comorbidity, safety-sensitive work, driving, pregnancy, cardiovascular risk, and whether the patient wants help reducing use.
Cannabis use increased during the study period, particularly in several older age groups, while alcohol trends moved differently. Those parallel changes describe the national landscape but do not prove substitution or a shared cause.
Health systems increasingly need cannabis conversations that can hold two truths at once: some patients report therapeutic benefit, and some develop clinically important loss of control or harm.
The analysis cannot identify whether legalization, potency, commercialization, medical access, pandemic-era behavior, or another factor caused the trends. It also cannot determine the risk attached to a specific product or route.
Population percentages cannot diagnose an individual. Frequency, tolerance, and withdrawal require context, particularly when cannabis is prescribed or used for persistent symptoms.
More on the topic:
SAMHSA’s 2026 Clinical Guidance for Physicians and Patients on Cannabis Use Disorder
Is Weed Addictive? Separating Dependence from Use
Insights on Cannabis Craving and Cross-Substance Use from Pooled Trial Data
Young adults still carried high prevalence, while increases among middle-aged and older adults argue against screening only younger people. Route, THC exposure, psychiatric vulnerability, sedating co-use, and driving risk remain clinically relevant.
No FDA-approved medication exists specifically for cannabis use disorder. Evidence-based care can include motivational approaches, cognitive behavioral strategies, contingency management, treatment of comorbid conditions, and sustained follow-up.
NSDUH relies on self-report and excludes unsheltered and incarcerated populations. The authors did not apply multiple-comparison correction, so subgroup findings should be interpreted with their confidence intervals rather than as equally certain signals.
The paper discusses potency and changing acceptance as possible context, but these exposures were not tested as causal drivers in the analysis.
Public-health responses should expand early identification and voluntary treatment without converting screening into punishment or using disorder prevalence to justify claims the study did not test.
Useful policy measurement would track treatment access, severity, product patterns, functional outcomes, stigma, and disparities while separating medical and nonmedical contexts where the data permit.
Cannabis use disorder is defined by clinically meaningful patterns such as impaired control, craving, role disruption, continued use despite harm, tolerance, and withdrawal. Exposure alone is not the diagnosis.
The study’s comparison with alcohol is descriptive. Different prevalence trends do not make one substance safe, nor do they show that cannabis directly replaced alcohol in any individual.
This is a reason to normalize better questions, not to stigmatize cannabis users. A patient can report benefit and still develop loss of control, just as a patient can use regularly without meeting criteria for a disorder.
The most useful response is structured follow-up: define the goal, track dose and frequency, ask what happens when use is reduced, review mood and sleep, and notice whether cannabis is beginning to displace work, relationships, finances, or safer treatment choices.
How to Read the New Cannabis Use Disorder Numbers
These are national population estimates, not predictions for a particular patient.
Four distinctions keep the findings clinically useful.
Four distinctions that matter
Use versus disorder
Cannabis exposure is not itself a diagnosis; impaired control and clinically meaningful consequences matter.
Trend versus cause
The survey shows change over time but cannot identify what produced it.
Screening versus labeling
A brief screen opens a conversation and must be followed by clinical assessment.
Medical use versus immunity
A therapeutic goal does not eliminate the possibility of problematic use or withdrawal.
Rising Cannabis Use Disorder Through Eight Practical Lenses
The same national trend creates different questions for patients, clinicians, families, health systems, and policymakers
Look for Loss of Control, Not a Moral Label
A patient faces a personal question that national prevalence cannot answer: is cannabis still serving the intended goal, or has use become harder to control? Useful signs include repeated failed efforts to cut down, escalating use, craving, withdrawal, or continued use despite clear harm.
Regular use does not automatically equal disorder, and a medical reason does not automatically prevent it. The practical step is to track function, amount, frequency, and what happens during reduction. The boundary is that self-observation can guide a conversation but cannot replace a full clinical assessment.
Make Cannabis Screening Routine and Specific
Clinicians should ask about cannabis with the same calm specificity used for alcohol, medications, sleep, and mood. Frequency alone is incomplete. Ask about route, potency, symptom target, dose escalation, craving, withdrawal, impairment, and unsuccessful attempts to reduce use.
A screen should lead to clarification rather than an automatic label. Medical use, chronic symptoms, psychiatric conditions, and social circumstances can change how criteria are interpreted. The counterweight is that avoiding the topic because it feels sensitive leaves clinically important patterns invisible until harm is more advanced.
Notice Functional Change Without Policing the Person
Families may see missed obligations, isolation, financial strain, mood change, unsafe driving, or repeated promises to cut down before the person recognizes a pattern. Those observations can be useful when offered concretely and without accusation.
Concern is not diagnosis, and surveillance can damage trust. Agree on how to discuss safety, storage, transportation, and caregiving responsibilities. The important boundary is that family support should not become coercion, especially when the person is using cannabis for a legitimate symptom and fears losing care.
Do Not Assume Problematic Use Is Only a Young Person’s Issue
The study found notable increases in several groups aged 50 years or older, even though absolute prevalence remained lower than in younger adults. Older adults may be using cannabis for pain, sleep, anxiety, or other persistent concerns and may not be asked about it routinely.
Medication interactions, falls, cognition, cardiovascular symptoms, driving, and delayed oral-product effects deserve attention. Small subgroup estimates and wide confidence intervals limit precise conclusions, particularly for older women. The practical implication is age-inclusive screening, not alarm based on a relative increase alone.
High Prevalence Calls for Earlier, Nonpunitive Support
Among adults aged 18 to 20 who used cannabis, the reported prevalence of cannabis use disorder was high, and many cases were moderate to severe. Young adults face developmental, educational, occupational, driving, and mental-health consequences that can compound quickly.
Prevention should offer accurate risk information, confidential screening, and accessible care rather than relying on fear or punishment. The study is cross-sectional and does not establish which products, policies, or personal factors caused disorder. Its strongest implication is that waiting for severe consequences misses an opportunity for earlier support.
Therapeutic Benefit and Disorder Risk Can Coexist
Patients may obtain real symptom relief and still experience tolerance, withdrawal, dose escalation, or difficulty reducing use. A medical recommendation changes the context but does not make monitoring unnecessary.
Clinicians should define the target symptom, functional goal, dose ceiling, route, review interval, and stopping conditions. The study did not separate every medical and nonmedical pattern or evaluate treatment plans, so it cannot quantify risk for a particular certified patient. The concrete implication is prospective monitoring rather than assuming either benefit or harm.
Build a Path Beyond Screening
Screening without referral capacity can identify distress while leaving patients with nowhere to go. Health systems need brief intervention, behavioral treatment, psychiatric care, pain and sleep alternatives, and follow-up that patients can access without stigma.
There is no FDA-approved medication specifically for cannabis use disorder, but absence of a medication does not mean absence of care. Systems should measure engagement and function, not only positive screens. The counterweight is workflow burden, so implementation must be concise, integrated, and supported rather than added as an unfunded checkbox.
Use the Trend to Improve Care, Not Overclaim Causation
Policymakers may be tempted to attribute the increase to legalization, potency, or commercialization. Those are plausible research questions, but this study did not compare state policies or isolate a causal driver.
A proportionate response would strengthen surveillance, product information, prevention, clinician training, and voluntary treatment access while monitoring disparities and unintended consequences. Punitive responses can deter disclosure and care. The boundary is equally important in the other direction: legal access and therapeutic potential do not erase population-level disorder risk.
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
Frequently Asked Questions
What did the new cannabis use disorder study find?
From 2021 to 2024, past-year cannabis use disorder and moderate-to-severe disorder increased among US adult males and females in national survey data.
How many adults were included?
The analysis included 186,823 adult participants in the 2021 to 2024 National Surveys on Drug Use and Health.
Does frequent cannabis use automatically mean cannabis use disorder?
No. Diagnosis depends on impaired control, craving, role disruption, continued use despite harm, and related criteria, not frequency alone.
Can medical cannabis users develop cannabis use disorder?
Yes. A medical goal does not eliminate risk, although clinical context matters when interpreting tolerance and withdrawal.
Which age groups showed notable increases?
The steepest reported increases included adults aged 35 to 49 and several groups aged 50 or older, while young adults continued to have high prevalence.
Did the study prove legalization caused the increase?
No. The cross-sectional analysis did not compare state policies or establish why prevalence changed.
What are warning signs of cannabis use disorder?
Possible signs include unsuccessful efforts to cut down, craving, withdrawal, escalating use, and continued use despite functional or health problems.
Is there an FDA-approved medication for cannabis use disorder?
No medication is currently FDA-approved specifically for cannabis use disorder, but behavioral treatments and clinical follow-up can help.
What are the study’s main limitations?
It relied on self-report, excluded unsheltered and incarcerated people, was cross-sectional, and did not correct for multiple comparisons.
What should a patient do if cannabis is becoming hard to control?
Discuss the pattern confidentially with a clinician and ask about reduction support, withdrawal planning, behavioral treatment, and care for underlying symptoms.
