A large cross-sectional study from Kaiser Permanente Washington found that daily cannabis users identified through a brief clinical screen had nearly a 45% rate of cannabis use disorder, compared to about 13% among those using less than monthly. The findings validate a simple one-question frequency screen as a reliable way to stratify clinical risk in primary care, though the cross-sectional design cannot establish that higher frequency causes the disorder.

Cannabis legalization has expanded access without a proportional expansion in clinical infrastructure for identifying problematic use. Primary care clinicians need practical tools that distinguish low-risk from higher-risk patients quickly and reliably. Until now, the brief screening instrument most commonly adopted in primary care settings had been validated for diagnostic accuracy but not mapped against actual prevalence of cannabis use disorder across the frequency categories it generates. Filling that gap determines whether a screening result is merely informative or genuinely actionable at the point of care.

As cannabis legalization proceeds across the United States, primary care practices increasingly encounter patients who use cannabis regularly but may or may not meet criteria for cannabis use disorder. The Substance Use Brief Screen with Cannabis supplement (SIS-C) was developed to provide a fast, clinically embedded way to categorize patients by their self-reported frequency of past-year cannabis use: less than monthly, monthly, weekly, or daily. Prior work validated the SIS-C against a gold-standard diagnostic interview, but the actual prevalence of cannabis use disorder within each frequency stratum remained undefined in a primary care population. Researchers at Kaiser Permanente Washington addressed this gap by linking clinical SIS-C screening data from nearly 109,000 primary care patients to a confidential diagnostic survey administered to a stratified random subsample of 1,688 respondents.

The results demonstrated a steep, statistically significant gradient. Any cannabis use disorder was present in 12.7% of less-than-monthly users and rose to 44.6% of daily users (p < 0.001). More critically, moderate-to-severe CUD, the threshold most likely to warrant clinical intervention, was essentially absent among infrequent users (0.9%) but affected roughly one in five daily users (20.3%). The clinical screen’s frequency categories also correlated with confidential reports of typical weekly and daily consumption, supporting the SIS-C’s construct validity as a proxy for real exposure. However, the 34% survey response rate and the restriction to a single integrated health system in a state with legal recreational cannabis limit how confidently these prevalence figures can be generalized to other populations or health care settings. The authors appropriately note that longitudinal data are needed to determine whether screening and subsequent intervention actually reduce the burden of cannabis use disorder over time.

What I appreciate about this study is its pragmatism. It answers the question clinicians actually face after a screening result: “My patient says they use cannabis daily. How worried should I be?” Having concrete prevalence numbers attached to each frequency category is far more useful than a binary positive or negative screening outcome. The dose-response gradient is striking and feels clinically real to anyone who sees cannabis patients regularly. That said, a 45% rate of any CUD among daily users means 55% do not meet criteria, so frequency alone is not destiny, and we need to resist the reflex to pathologize all frequent use.

In my own practice, I already use frequency of use as a core triage variable, but I pair it with questions about functional impairment, escalation patterns, and whether the patient has tried to reduce use without success. This study reinforces that approach. For patients reporting daily use, I now have evidence to frame a direct conversation: “About one in five people in your situation develop problems that meet criteria for a use disorder. Let’s talk about whether that applies to you.” That level of specificity lands better than vague warnings and opens a genuine dialogue about harm reduction or structured use.

This study sits at an important inflection point in the research arc on cannabis screening in primary care. The SIS-C has been validated, and implementation studies are underway in several integrated health systems. What was missing was a prevalence map that tells clinicians what they are likely to find when they screen. By providing that map, this work transitions the SIS-C from a research instrument to a practical clinical tool with interpretable benchmarks. It also provides indirect evidence that the four-category frequency structure captures real variation in consumption behavior, since the screen responses track closely with confidential self-reports of weekly and daily use intensity.

From a pharmacological and safety standpoint, the finding that daily users are significantly more likely to use cannabis for health reasons and to use higher-potency routes of administration (such as concentrates) raises important considerations. These patients may be self-treating chronic pain, anxiety, or sleep disturbances, and abruptly recommending cessation without addressing the underlying condition is unlikely to succeed and may cause harm. Clinicians encountering daily-use screen results should pair the CUD risk discussion with a thorough assessment of the therapeutic motivations driving use, and consider whether supervised cannabis guidance, dose reduction strategies, or alternative therapeutic approaches might be appropriate before defaulting to abstinence-based recommendations.

Have thoughts on this? Share it: