A nationally representative survey from 2018 found that nearly 1 in 5 Americans aged 50 to 64 used cannabis, with smoking as the dominant consumption method. Among those using cannabis for medical purposes, fewer than 1 in 4 had received any healthcare provider recommendation, revealing a significant gap in clinical oversight for a population at elevated risk from drug interactions and falls.

Cannabis use among adults aged 50 and older is not a marginal or emerging phenomenon. This population carries the highest burden of polypharmacy, chronic disease, and physiologic vulnerability to adverse drug effects in the United States. When millions of older Americans are using a pharmacologically active substance, predominantly by smoking, and doing so almost entirely outside the healthcare system, clinicians and policymakers face a care gap that demands immediate attention. Understanding the scope and behavioral texture of this use is the first step toward a coherent clinical response.

The rapid expansion of cannabis legalization in the United States has coincided with a growing recognition that older adults represent one of the fastest-rising segments of cannabis consumers. Yet prior national surveys, notably the National Survey on Drug Use and Health (NSDUH), have provided limited behavioral detail for this age group, typically reporting only binary use or nonuse. This study leveraged the 2018 Health and Retirement Study, a survey specifically designed to represent Americans aged 50 and older, and its experimental cannabis module to characterize not only prevalence but consumption methods, reasons for medical use, healthcare provider involvement, and attitudes. The rationale for detailed behavioral profiling rests on the clinical reality that the same dose of cannabis carries different risk depending on whether it is smoked or ingested, used once monthly or daily, and managed with or without a clinician’s awareness.

The study found that 18.5% of adults aged 50 to 64 and 5.9% of those aged 65 and older reported past-year cannabis use, translating to approximately 11.5 million and 3 million individuals, respectively. Among those aged 65 to 74, prevalence was nearly five times higher than among those 75 and older (8.84% vs. 1.78%). Smoking was the dominant consumption method across both age groups. Roughly one-quarter of lifetime users reported medical cannabis use, yet only about 22% of those medical users had received any healthcare provider recommendation. The authors acknowledge the cross-sectional design, the 2018 time point, and limited covariate adjustment as key constraints, and they call for longitudinal studies and updated surveillance to track this rapidly evolving landscape.

Cannabis Use in Adults 50+: Prevalent, Mostly Smoked, and Largely Outside Medical Care

More than 11 million Americans between 50 and 64 used cannabis in 2018. Most of them were smoking it. Most of those using it medically had never discussed it with a doctor. And yet, this is a population that takes multiple prescription medications, is at elevated risk for falls, and whose bodies metabolize drugs differently than they did at 30. The data have been sitting in a nationally representative survey for years, and they demand clinical attention. What this study from the Health and Retirement Study genuinely contributes is not just headline prevalence numbers, which are striking enough on their own, but a behavioral portrait we have not had before at this scale: how older Americans consume cannabis, why they say they use it medically, and whether anyone in the healthcare system is part of that conversation. The answer to that last question is sobering. Fewer than 1 in 4 older adults who self-identified as medical cannabis users had received a provider’s recommendation. That statistic does not tell us whether cannabis is helping or harming these individuals, but it tells us something equally important: the healthcare system has largely opted out of a conversation that millions of patients are already having with themselves.

The central methodological limitation to keep in mind is that this is a single cross-sectional snapshot from 2018, a year when the cannabis legal landscape looked meaningfully different than it does today. Asking people at one moment whether they use cannabis tells you nothing about how long they have used it, what outcomes followed, or whether their use will change as they age further. It is the equivalent of taking a single photograph of a moving river and trying to predict its course. Moreover, the study’s decision to impute all nonresponse on cannabis items as “no use” likely pushed prevalence estimates downward. If the people who declined to answer were disproportionately cannabis users, which is plausible given ongoing stigma in 2018, then the true rates were even higher than reported. The paper does not disclose the nonresponse rate or test this assumption with sensitivity analyses, which is a meaningful gap. Separately, the term “medical use” in this study is entirely self-defined by participants. There is a real difference between a patient whose physician has recommended a specific cannabis formulation for neuropathic pain and someone who picks up edibles at a dispensary because a friend suggested it for back pain. This study cannot distinguish between those scenarios, and readers should resist the temptation to treat “medical use” here as equivalent to medically supervised use.

What would I say to a patient? I would tell them that nearly 1 in 5 people their age use cannabis, and most of them never mention it to their doctor. I would rather know, so we can ensure it is not interacting with their blood pressure medication or increasing their fall risk. To a colleague, I would say these data confirm what many of us already sense in practice: cannabis use among adults over 50 is prevalent, mostly self-directed, mostly smoked, and mostly invisible to us. It belongs in our intake questions. To a policymaker, I would emphasize that this 2018 snapshot documented millions of older Americans already using cannabis for medical reasons without any clinical involvement, and that investment in clinician education, cannabis-drug interaction guidelines, and updated epidemiologic surveillance is overdue. The durable lesson from this study is worth stating plainly: prevalence data from a single cross-sectional survey can be highly valuable for public health prioritization, but it cannot substitute for longitudinal evidence on outcomes, and “high prevalence of a behavior” should not be conflated with “high evidence of harm from that behavior” or vice versa.

This study sits early in the research arc for understanding cannabis use specifically in adults over 50. While NSDUH data have documented rising prevalence for years, the behavioral granularity provided by the HRS module represents a necessary first step toward the kind of clinical characterization that can inform guidelines. It is descriptive, not evaluative: it tells us what older Americans are doing, not what happens when they do it. Longitudinal studies linking these behavioral patterns to health outcomes remain the essential next phase.

From a pharmacological standpoint, the dominance of smoking as the consumption method is concerning for a population with elevated baseline risk for cardiovascular events and respiratory disease. The near-absence of provider involvement among medical users is equally troubling given the known potential for cannabis to interact with common medications in this age group, including anticoagulants, statins, and sedative-hypnotics, through cytochrome P450 enzyme inhibition. Clinicians should consider incorporating a standardized cannabis use screening question into routine intake for patients aged 50 and older, treating it with the same clinical seriousness as questions about alcohol, tobacco, and supplement use.

This is a cross-sectional epidemiologic analysis of a nationally representative survey, positioned in the lower-middle tier of the evidence hierarchy. It uses weighted prevalence estimation and multivariable logistic regression to describe associations, not test interventions. The single most important inference constraint is that the cross-sectional design at a single time point precludes any causal or temporal claims about cannabis use and health outcomes.

The prevalence estimates from this HRS analysis are higher than contemporaneous NSDUH estimates for the same age groups, which is likely attributable to the HRS’s superior sampling precision for adults over 50 rather than an indication that one survey is correct and the other wrong. A prior brief HRS report examined cannabis use in relation to health conditions and opioid use among adults 50 and older but did not provide the behavioral detail captured here. This study extends that work by characterizing consumption methods, medical use rationale, and provider involvement, filling a gap that neither NSDUH nor the earlier HRS analysis addressed. Taken together, these data sources consistently indicate that cannabis use among older Americans is more prevalent than clinical awareness would suggest, and the current study adds critical context about how and why that use is occurring.

The most consequential analytic choice was the imputation of all nonresponse on cannabis items to “no use.” If cannabis users were disproportionately likely to decline these questions, an entirely plausible scenario given persistent stigma in 2018, then the reported prevalence of 18.5% and 5.9% may represent underestimates. A sensitivity analysis treating nonresponse as missing data, or modeling a range of plausible use rates among nonresponders, could have produced meaningfully higher prevalence figures. Additionally, the absence of state cannabis legalization status as a covariate is a notable gap; adjusting for legal environment would likely have revealed significant geographic heterogeneity in use patterns and could have altered the magnitude of sociodemographic associations reported in the multivariable models.

The most likely overinterpretation is treating these 2018 prevalence figures as current estimates. Cannabis legalization has expanded substantially since 2018, and both use rates and attitudes have almost certainly shifted. Readers should treat these data as a historical benchmark, not a present-day snapshot. A second common misreading involves the low rate of provider involvement among medical users: this finding describes a care gap, not a clinical outcome. It does not demonstrate that unsupervised cannabis use caused harm to these individuals, nor does it establish that supervised use would have been safer. The study documents behavior, not consequences.

This study provides the first nationally representative behavioral profile of cannabis use among U.S. adults aged 50 and older, documenting high prevalence, dominant smoking as the consumption route, and widespread self-directed medical use without provider guidance. It does not establish causal relationships between cannabis use and health outcomes. For practice today, it reinforces the case for routine cannabis screening in primary care for older adults and underscores the need for updated surveillance data that reflects the current legal environment.

How common is cannabis use among older adults in the United States?

According to this 2018 survey, approximately 18.5% of adults aged 50 to 64 and 5.9% of those 65 and older reported using cannabis in the past year. These figures translate to roughly 11.5 million and 3 million individuals, respectively. Current rates are likely higher given continued expansion of legalization since 2018.

Should I tell my doctor if I use cannabis?

Yes. This study found that fewer than 1 in 4 older adults using cannabis for medical purposes had received any guidance from a healthcare provider. For patients over 50 who often take multiple medications, disclosing cannabis use allows a clinician to check for drug interactions, monitor for side effects, and provide personalized recommendations about safer consumption methods.

Is smoking cannabis particularly risky for people over 50?

This study did not directly measure health outcomes from different consumption methods, but smoking was the dominant route in both age groups studied. Given the known respiratory and cardiovascular risks associated with inhaling combusted plant material, and the elevated baseline cardiovascular risk in adults over 50, clinicians generally regard nonsmoking methods as preferable for this population. A conversation with your physician can help identify the approach that carries the least risk for your individual health profile.

Does this study prove that cannabis is harmful for older adults?

No. This study describes how many older adults use cannabis and how they use it, but it was not designed to measure health outcomes or establish whether cannabis causes harm. What it does reveal is a widespread pattern of use occurring without clinical oversight in a population that is particularly vulnerable to adverse drug effects, which is a reason for greater clinical engagement rather than a conclusion about harm.

1. Livne O, Stohl M, Gilman J, Goldberg TE, Wall MM, Hasin DS. Epidemiology of Cannabis Use Among Middle-Aged and Older Adults in the U.S. Am J Prev Med. 2026;70(3):108149. doi:10.1016/j.amepre.2025.108149

2. Health and Retirement Study (HRS). 2018 Experimental Cannabis Module. University of Michigan, Survey Research Center. Supported by National Institute on Aging.

3. Substance Abuse and Mental Health Services Administration. National Survey on Drug Use and Health. Various years. Cited in Livne et al. 2026.

4. Prior HRS brief report on cannabis use, health conditions, opioid use, and service utilization in adults aged 50 years and older. Cited in Livne et al. 2026.

Have thoughts on this? Share it: