A new program report from a geriatric primary care clinic found that older adults seeking medical cannabis certification carried an average of 4.6 drug-cannabis interactions per patient, with two-thirds already taking CNS depressants. The study demonstrates that embedding cannabis oversight into geriatric practice is operationally feasible, but it measured no patient outcomes, so whether this model actually improves safety or health remains entirely unknown.

Cannabis use among adults over 65 is rising faster than the clinical infrastructure to support it safely. Older adults carry the highest burdens of polypharmacy, multimorbidity, and age-related pharmacokinetic changes, yet most medical cannabis certification pathways offer no drug interaction screening, no access to the patient’s medical record, and no coordination with existing providers. The question of whether structured, primary care-embedded cannabis oversight can actually be built within existing geriatric practice workflows is immediately relevant to every health system grappling with this growing clinical demand.

Cannabis use among older adults has expanded rapidly, yet certification pathways remain largely disconnected from primary care, leaving medically complex patients without structured oversight for drug interactions and safety monitoring. Researchers at VCU Health designed a monthly, physician-led medical cannabis certification clinic embedded within their geriatric primary care division, staffed by an interprofessional team including a pharmacist who performs systematic drug utilization reviews using two interaction databases. The model operates within the existing electronic health record, uses standard CPT billing codes, and is positioned to address the well-documented care fragmentation that arises when cannabis certification occurs outside the patient’s longitudinal medical home.

Over 30 months, the clinic completed 144 visits for 122 unique patients, with a mean age of 65 years. The population was strikingly complex: patients carried an average of 20.9 comorbidities and 14.7 medications. Pain was the primary qualifying condition in 88.9% of visits. Drug interaction screening identified a mean of 4.6 cannabis-drug interactions per patient; 66.0% of patients were co-prescribed CNS depressants, 59.0% were on pain medications, and 56.9% on psychiatric medications. All 144 visits resulted in certification or recertification. Critically, no clinical outcomes, including symptom changes, adverse events, falls, or hospitalizations, were measured. The authors acknowledge the need for prospective outcome studies and multi-site replication before effectiveness conclusions can be drawn.

4.6 Interactions Per Patient: What a Geriatric Cannabis Clinic Found When It Actually Looked

More than two-thirds of older adults seeking medical cannabis certification at this clinic were already taking CNS depressants, including sedatives, opioids, and muscle relaxants, and nearly none of them had anyone systematically checking what would happen when they added cannabis to that mix. That gap is exactly what this paper set out to close. And on the question of whether a structured, physician-led cannabis clinic can physically exist within a functioning geriatric primary care practice, the answer is a clear yes. The VCU team built a clinic that runs monthly, bills through standard CPT codes, integrates into the existing electronic health record, and deploys a pharmacist to screen every patient’s medication list against two interaction databases. The finding that patients averaged 4.6 drug-cannabis interactions is the paper’s most honest and clinically valuable contribution. It quantifies, for the first time in a geriatric-specific setting, the scope of a problem that most telehealth certification platforms do not even attempt to address. That number alone justifies the pharmacist’s seat at the table. But there is a critical difference between identifying a fire risk and proving that you have prevented fires. The clinic screened for interactions and documented them in the medical record. Whether those findings changed prescribing, altered cannabis dosing, prevented falls, reduced emergency department visits, or improved pain control is entirely unknown. The paper measures the process, not the result.

I am also struck by the 100% certification rate across all 144 visits. A security checkpoint that waves through every traveler is either proof that all travelers were safe or proof that the checkpoint was not filtering anyone, and you cannot tell which without knowing who it turned away. The paper does not report referral-to-enrollment ratios, nor does it describe any instance where the safety review led to a recommendation against certification. That absence does not necessarily indicate a problem. It may reflect careful pre-screening upstream. But it does mean we cannot evaluate the clinic’s safety function from this data. To a colleague, I would say: this is a well-designed implementation model that addresses a real care gap, and the interaction data alone makes a compelling case for pharmacist involvement in any cannabis certification workflow. To a patient, I would say: this clinic offers something genuinely valuable, a geriatrician who understands your medications and your aging body, reviewing your cannabis use within your actual medical record. That is meaningfully better than an online certification mill. But whether the clinic has improved outcomes for people like you is still an open question.

The VCU geriatric cannabis clinic is a thoughtfully constructed answer to a genuine problem: older adults are using cannabis in growing numbers, in a regulatory environment that prioritizes access over safety, without the integrated clinical oversight their medical complexity demands. What it has not yet done is demonstrate that it makes patients better or safer. That is not a critique of the clinic; it is an honest accounting of what a 30-month descriptive program evaluation can and cannot tell us. The next step is measuring what actually happens to these patients after they leave the clinic with their certification in hand. Implementation evidence and effectiveness evidence are not interchangeable. A well-designed clinic that does the right things is not the same as a proven clinic that achieves better outcomes. Both matter, but only one of them can justify broad clinical adoption.

This paper sits at the very beginning of the research arc for integrated geriatric cannabis care. It answers the question of whether such a clinic can be built and sustained, and it characterizes the population likely to walk through the door. It does not answer whether the model changes clinical trajectories. For clinicians working with older adults who use or are considering cannabis, the drug interaction data is immediately useful: the prevalence of CNS depressant co-prescription (66%), pain medication use (59%), and psychiatric medication use (57%) in this cohort should prompt any provider to conduct or request systematic interaction screening before endorsing cannabis use in a comparable patient.

From a pharmacological standpoint, both THC and CBD are metabolized through CYP enzymes, with CBD acting as a potent CYP3A4 and CYP2D6 inhibitor. In a population averaging nearly 15 medications, the potential for clinically significant interactions with warfarin, tacrolimus, certain anticonvulsants, and opioids is not theoretical but expected. The paper’s case example of a patient on tacrolimus and warfarin illustrates this well. For practices considering similar models, the single most actionable recommendation from this report is to integrate a clinical pharmacist into any cannabis certification workflow serving older adults. The interaction burden is too high, and the consequences in this population too serious, to leave drug utilization review to chance.

This is a retrospective, descriptive program evaluation published in the quality improvement and program dissemination section of a major geriatrics journal. It occupies the lower tiers of the evidence hierarchy, below observational cohort studies, controlled trials, and systematic reviews. Its design permits description of the clinic’s operational structure and the population it serves, but it does not permit causal inference, comparative effectiveness assessment, or any conclusions about whether the model improves clinical outcomes. The single most important inference constraint is the absence of any measured patient outcome.

This paper extends a small but growing body of implementation literature on structured medical cannabis programs. It is most directly comparable to the Montefiore program, which certified over 1,600 patients in five years using a six-step provider-led process, and the Canadian interdisciplinary model described by Prosk and colleagues, which processed more than 13,000 referrals. The VCU model is unique in its geriatric-specific focus and pharmacist-led drug interaction screening, but like its predecessors, it reports process and population data without clinical outcomes. Williams and colleagues previously documented that only 42% of states with marijuana laws included more than two components of traditional medical care, with 99.4% of participants enrolled in less comprehensive programs. The VCU clinic directly addresses this structural gap, but its contribution remains limited to proof of concept until prospective outcome data become available.

The most consequential analytic choice was the decision to collect no patient outcome data. Had the investigators tracked even basic patient-reported outcomes such as pain severity, sleep quality, medication changes, or adverse events at follow-up visits, the paper would have moved from pure implementation description to preliminary effectiveness evidence. The 15.3% of visits that were renewals represent a natural follow-up cohort where pre-post comparisons could have been drawn. Additionally, reporting referral-to-enrollment ratios and any instances of non-certification would have allowed readers to assess the clinic’s safety filtering function. Without these data, the paper’s evidentiary ceiling remains fixed at feasibility, regardless of how the existing descriptive statistics are presented.

The most consequential misreading of this paper would be concluding that embedding cannabis certification in geriatric primary care improves patient outcomes. The study measured no outcomes whatsoever. A related error is interpreting the 4.6 drug-cannabis interactions per patient as evidence that the clinic prevented harm. Interactions were identified and documented, but whether that identification led to dosing changes, medication adjustments, or reduced adverse events is entirely unreported. Similarly, the 100% certification rate should not be read as evidence that all patients were appropriate candidates; it may equally reflect pre-visit selection filtering or a low threshold for certification. Finally, this paper provides no comparative data to support the claim that integrated primary care models are superior to telehealth or dispensary-based certification. That claim is logically plausible but empirically unsupported by this dataset.

This paper contributes a detailed, reproducible blueprint for embedding a physician-led, pharmacist-supported medical cannabis clinic within geriatric primary care and characterizes the high-complexity population such a clinic will serve. It does not establish that this model improves symptoms, reduces adverse events, or is superior to alternative certification pathways. For clinical practice today, the most actionable takeaway is the drug interaction burden itself: older adults using cannabis alongside CNS depressants, opioids, and psychiatric medications require systematic pharmacist-level screening, and most current certification models do not provide it.

Does this study prove that a geriatric cannabis clinic is better for older patients?

No. This study demonstrates that such a clinic can be built and operated within an existing geriatric practice, and it describes who showed up. It did not measure patient outcomes, so it cannot tell us whether patients were healthier, safer, or better managed as a result of the program.

What does “4.6 drug interactions per patient” actually mean for my safety?

It means that the average patient in this clinic had nearly five medications that could interact with cannabis in clinically meaningful ways, including sedatives, pain medications, and psychiatric drugs. Not all interactions cause harm, but in older adults with multiple health conditions, each interaction represents a potential risk that should be reviewed by a knowledgeable clinician or pharmacist before cannabis is started or continued.

Should I seek out a clinic like this instead of using telehealth certification?

If you are an older adult taking multiple medications, a clinic that has access to your complete medical record and offers pharmacist-led drug interaction screening is likely to provide more thorough safety evaluation than a telehealth platform that operates without that information. However, this paper does not prove that the integrated model produces better health outcomes. The logic is sound, but the evidence for a difference in results has not yet been generated.

Why were all patients in this study certified for cannabis?

The paper reports a 100% certification rate across all 144 visits, but it does not explain why. This could mean patients were carefully pre-screened before referral, or it could raise questions about whether the safety review process led to any recommendations against certification. Without data on how many patients were referred but not enrolled, or how many received cautionary recommendations, readers cannot fully evaluate this finding.

Have thoughts on this? Share it: