Medical Cannabis Knowledge Gaps in U.S. Healthcare Providers

Abstract

Study at a Glance

Study Snapshot

Study Facts Table

What Researchers Actually Did

Between December 2021 and October 2023, investigators at the Johns Hopkins Behavioral Pharmacology Research Unit recruited U.S. healthcare professionals through social media advertising, the Johns Hopkins Center for Psychedelic and Consciousness Research website, and institutionally approved email distribution. Eligible participants must have been at least 18 years old, English-fluent, and currently working in a U.S. clinical setting. Of 2,212 total responses, 1,333 were excluded for failing quality-control criteria — including CAPTCHA failures, incorrect attention-check responses, duplicate submissions, or incomplete surveys — leaving 879 for analysis.

The survey comprised two substantive cannabis-specific sections. The first used 5-point Likert scales to capture self-rated knowledge across three domains (therapeutic indications, risks and adverse effects, mechanism of action), attitudes toward clinical use and legal access, and concern ratings for eight specific clinical issues. The second was an objective knowledge check: three targeted questions asked participants to select all correct answers regarding evidence-based therapeutic indications (muscle spasms, pain, nausea and vomiting), risks (increased blood pressure and heart rate, dizziness, nausea, psychosis), and primary mechanism of action (CB1 and CB2 receptor activity). A response was scored as correct only if all correct options were selected and no incorrect options were endorsed. Multivariable linear regression then examined predictors of openness to clinical use and predictors of objective knowledge, using profession, age, personal cannabis experience, self-rated knowledge, and mean concern score as covariates.

Key Findings: Primary Outcomes

• Objective knowledge scores were strikingly low: Only 116/879 (13%) correctly answered the risks and adverse effects check; 186/879 (21%) answered the therapeutic indications check correctly; 558/879 (64%) answered the mechanism of action check correctly.
• Almost no one passed all three domains: Only 22/879 (3%) answered all three knowledge checks correctly under strict scoring; 241/879 (27%) answered all three incorrectly.
• Self-rated knowledge did not reflect actual knowledge: Participants rated themselves highest in the risks domain (mean 4.1/5), yet objective performance in that domain was the lowest. Self-rated and objective knowledge scores were weakly but significantly correlated (r = 0.08, p < 0.05).
• Openness to clinical use was generally high but varied by profession: Overall mean openness was 4.1/5; registered nurses scored highest (4.4), while physicians and mental health professionals scored lowest (3.9 each).
• Regression model for openness (R² = 0.159): Higher self-rated cannabis knowledge was the strongest positive predictor (B = 0.35, p < 0.001). Greater concern score was a significant negative predictor (B = -0.09, p = 0.035). Compared to physicians, APPs (B = 0.24, p = 0.022) and RNs (B = 0.42, p < 0.001) were significantly more open. Openness decreased significantly with age for groups 40-49, 50-59, and 60-69 compared to the 18-29 reference group.
• Regression model for objective knowledge (R² = 0.141): Physicians outperformed all other professional groups. APPs (B = -0.31, p < 0.001), RNs (B = -0.60, p < 0.001), MHPs (B = -0.72, p < 0.001), and others (B = -0.58, p < 0.001) all demonstrated significantly lower objective knowledge than physicians. Personal cannabis experience (B = 0.18, p = 0.001) and higher concern scores (B = 0.11, p = 0.001) were both positively associated with higher objective knowledge.

Key Findings: Secondary Outcomes and Subgroup Analyses

• Support for legal access was near-universal: 95% supported legal medical cannabis use; 86% supported recreational access; 88% supported religious access. Mean legal access support was higher for RNs, MHPs, and other professionals than for physicians and APPs.
• Therapeutic promise broadly endorsed: 87% believed cannabis can be safely delivered in clinical settings; 87% endorsed its therapeutic promise; 96% supported further cannabis research.
• Concern profiles differed by profession: MHPs reported the highest overall concern scores, followed by APPs and physicians, then RNs. MHPs expressed notably higher concern about lack of trained providers (mean 3.2), administration to patients with contraindications (mean 2.7), psychosis risk (mean 2.6), and recreational misuse (mean 2.6).
• Preferred knowledge sources diverged from trusted ones: Personal experience (76%), popular media (73%), and informal conversations (71%) were actual knowledge sources; academic research centers (94%), experienced clinicians (92%), and professional organizations (80%) were most trusted. Formal clinical training was endorsed as an actual knowledge source by only 28%.
• Specialized clinics preferred for administration: 87% endorsed specialized clinics, 80% private practice, 74% outpatient clinics, and 74% supervised home use. Only 25% considered emergency departments appropriate.
• Interest in further training: 78% expressed interest in further training to use cannabis in their practice.

Results: Adverse Events and Safety Profile

This was a survey study; no adverse events in the pharmacological sense were assessed or reported. The study did document provider-reported concerns about cannabis safety, including psychosis risk (20% “very” or “extremely concerned”), cardiovascular effects, and harms to patients with contraindications (19%). These concern ratings function as a proxy for perceived, not observed, clinical risk in this dataset.

Statistical Approach and Rigor

The analysis used descriptive statistics and multivariable linear regression. Internal consistency for grouped Likert-scale domains was strong, with Cronbach’s alpha ranging from 0.80 to 0.83. The objective knowledge scoring was strict by design — credit required selecting every correct option without endorsing any incorrect option — which likely depressed scores and may not fully represent partial knowledge. The regression models explained modest variance (R² = 0.159 for openness; R² = 0.141 for objective knowledge), indicating that important predictors remain unmeasured. The correlation between self-rated and objective knowledge (r = 0.08) was statistically significant but clinically trivial. No correction for multiple comparisons was reported. Age was treated categorically, not continuously, which limited sensitivity in the regression. The sample was convenience-based, unverified, demographically homogenous, and likely biased toward favorable attitudes about cannabis — all factors the authors acknowledge.

Clinical Takeaway

Healthcare providers in this sample were broadly supportive of medical cannabis, but objective knowledge — particularly about risks and adverse effects — was far below what their self-ratings would suggest. The finding that providers rated themselves most confident in the risks domain yet performed worst on it objectively is the central clinical concern. A provider who does not know cannabis’s cardiovascular and psychiatric risks, its drug interactions via CYP inhibition (particularly with oral CBD), or its contraindications cannot counsel patients accurately, regardless of how open or supportive they are. The regression data add a nuanced layer: greater objective knowledge was associated with more concern, not less, suggesting that clinicians who actually understand the pharmacology recognize the legitimate complexity involved. Training programs that move providers from passive exposure to tested, reinforced curricula are what this dataset calls for.

Why This Matters Clinically

Cannabis is not a passive patient activity that happens outside the clinical encounter. In this survey, 89% of providers already have patients who use cannabis. That means the knowledge gap documented here is not hypothetical — it is operating in clinics today. Providers who overestimate their understanding of cannabis risks may fail to screen for contraindications (pregnancy, psychosis history, cardiac arrhythmia), fail to counsel about drug-drug interactions (notably, oral CBD inhibits CYP3A4 and CYP2C9, raising plasma levels of numerous co-medications), and fail to distinguish between regulated dispensary products and unregulated hemp-derived products that may carry heavy metal contamination. The inverse relationship between concern and openness, combined with the finding that better-informed providers expressed more concern, suggests that the most “open” clinicians may paradoxically be those least equipped to use this tool safely. This is the clinically important tension this paper surfaces.

CED Clinical Relevance

At CED Clinic, every patient interaction begins with a structured assessment that does precisely what this study shows the broader healthcare workforce has not yet systematized: verifying the type of cannabis product (THC-dominant, CBD-dominant, balanced), the source, route of administration, dose, and frequency of use. The knowledge gaps documented here — especially in the risks and drug-interaction domains — underscore why specialist cannabis consultation exists as a distinct clinical service. Patients who have been using cannabis for years, or who are considering initiating it, deserve a provider who can accurately characterize cardiovascular risk, psychosis risk in vulnerable populations, interaction potential with concurrent medications, and the distinction between evidence-supported indications (pain, nausea, spasticity) and uses that remain speculative. The finding that personal experience and popular media are the primary knowledge sources for most providers reinforces what structured cannabis medicine training programs aim to replace.

Fits What We Already Know

The findings align with the prior literature the authors themselves cite. Kruger et al. (2022, Cannabis and Cannabinoid Research) found limited provider knowledge of medical cannabis among a U.S. sample. Philpot et al. (2019, BMC Family Practice) documented gaps in primary care provider attitudes and knowledge. Weisman and Rodríguez (2021, Journal of Cannabis Research) conducted a systematic review confirming that limited knowledge alongside general openness is the modal pattern across studies. Worster et al. (2023, Cannabis and Cannabinoid Research) found variable training and belief structures across professions in an interprofessional assessment. What distinguishes the current study is the direct comparison of self-rated and objective knowledge within the same sample — a methodological advance over most prior work, which relied on self-report alone. The weak correlation (r = 0.08) between subjective and objective knowledge confirms that self-confidence is a poor surrogate for actual understanding, a finding with clear implications for how training adequacy is assessed.

What This Study Teaches Us

Healthcare providers, regardless of discipline, largely believe cannabis has legitimate medical uses — and most of them are already seeing patients who consume it. The problem is not motivation or attitude; most providers also want more training. The problem is that the training infrastructure has not kept pace with legalization. Providers are constructing their clinical cannabis knowledge from personal experience and media rather than from structured pharmacology curricula. And when tested objectively, that knowledge is weakest exactly where it matters most: understanding what cannabis can harm. The study also teaches us that opening up to cannabis clinically and knowing cannabis clinically are two different things — the data show these can move in opposite directions.

What It Does Not Show

• The study does not demonstrate that knowledge gaps cause patient harm; it demonstrates that gaps exist and that providers are unaware of them.
• It does not establish causal relationships between training, knowledge, openness, and clinical outcomes — the cross-sectional design prohibits this.
• Professional roles were self-reported and unverified, so the profession-level comparisons cannot be taken as representing those professions nationally.
• The sample is heavily White (86%), female (71%), and self-selected through channels associated with interest in psychedelic and novel medicines — generalizability to the full U.S. healthcare workforce is limited.
• The objective knowledge checks, while a methodological advance, used only three questions and were not psychometrically validated prior to administration.
• The study does not assess whether providers who report concern or lower openness deliver worse or better patient outcomes — concern, as measured here, is attitudinal, not behavioral.

Fits the Broader Conversation

This paper arrives at a moment when the regulatory and clinical landscape for cannabis is shifting faster than training infrastructure can adapt. The 2018 Farm Bill created widespread availability of hemp-derived cannabinoids; the November 2025 FY2026 Agriculture appropriations act begins to close that loophole. Over 40 states permit medical cannabis; 24 allow recreational use. Cannabis use disorder rates are rising. Product diversity — oils, edibles, inhalants, isolates, full-spectrum formulations, and an expanding list of minor cannabinoids — is expanding. Against this backdrop, a finding that only 3% of surveyed providers can pass a basic three-domain knowledge check is not a footnote. It is a structural indictment of health professional education. The call for standardized, psychometrically validated, competency-tested curricula in medical, nursing, and pharmacy programs is the natural policy conclusion this data supports.

Teaches Us: For Families

Many patients assume that if a healthcare professional seems confident discussing cannabis, that confidence reflects formal training and tested expertise. This study suggests that assumption may not always be justified. Most providers in this survey reported caring for patients who use cannabis, and most expressed support for its medical use. Yet objective knowledge scores were surprisingly low, particularly regarding risks and adverse effects.

For patients and families, this highlights the importance of asking questions. If cannabis is being considered for pain, sleep, anxiety, nausea, spasticity, epilepsy, autism-related symptoms, or any other condition, it is reasonable to ask a provider how familiar they are with cannabis pharmacology, dosing, adverse effects, drug interactions, and product selection. Expertise should not be assumed simply because cannabis is legal or commonly discussed.

The findings also reinforce the importance of transparency. Patients should tell healthcare providers about cannabis use just as they would disclose any prescription medication, supplement, or over-the-counter treatment. Accurate disclosure allows providers to identify potential interactions, monitor side effects, and incorporate cannabis use into the broader treatment plan.

Teaches Us: For Clinicians

The most important lesson for clinicians may not be that cannabis knowledge is incomplete. That fact has been recognized for years. The more important lesson is that self-assessment appears to be an unreliable measure of competence in this area.

Participants rated themselves most knowledgeable about cannabis risks and adverse effects, yet performed worst on objective testing in that domain. This pattern mirrors a broader phenomenon observed throughout medicine: confidence and competence do not always move together. In rapidly evolving areas of practice, clinicians may develop familiarity without developing mastery.

The practical implication is straightforward. Cannabis education should increasingly resemble education surrounding any other therapeutic category. Clinicians need structured training in pharmacology, dosing principles, product selection, adverse effects, contraindications, drug interactions, and documentation standards. Informal exposure through media, patients, colleagues, or personal experience cannot reliably substitute for formal clinical education.

Where This Paper Deserves Skepticism

Although the study identifies meaningful knowledge gaps, several limitations deserve attention. The sample was recruited largely through online channels and organizations likely to attract individuals already interested in cannabis, psychedelics, or emerging therapies. As a result, attitudes toward cannabis may be more favorable than those of the broader U.S. healthcare workforce.

The objective knowledge assessment was also intentionally stringent. Participants received credit only when every correct answer was selected and every incorrect answer was avoided. This scoring approach likely underestimated partial knowledge and may have exaggerated apparent knowledge deficits. A clinician who identified most correct answers but missed one detail received the same score as someone with substantially less knowledge.

Additionally, objective knowledge was measured using only three questions. While these questions addressed clinically relevant topics, they cannot fully characterize competency in a field as broad as cannabis medicine. Knowledge of dosing, formulations, cannabinoid ratios, terpene profiles, routes of administration, regulatory issues, and patient counseling was not directly assessed.

Finally, because the study was cross-sectional, it cannot determine whether low knowledge scores translate into worse patient outcomes. The study demonstrates educational gaps. It does not demonstrate patient harm resulting from those gaps.

Future Research Priorities

The next generation of research should move beyond measuring attitudes and knowledge alone. Investigators should examine whether formal cannabis education improves objective competency, changes prescribing and counseling behavior, or influences patient outcomes. Randomized educational interventions and longitudinal assessments could help determine which training models produce meaningful improvements in clinical practice.

Future studies should also develop and validate more comprehensive cannabis competency assessments. Just as medical education uses standardized approaches to evaluate proficiency in cardiology, endocrinology, infectious disease, and pharmacology, cannabis medicine may ultimately require formal competency frameworks that extend beyond self-reported confidence.

Final Perspective

This paper documents a paradox increasingly visible throughout healthcare. Cannabis has entered mainstream clinical conversations faster than formal medical education has adapted. Most providers encounter patients who use cannabis. Most believe cannabis has legitimate therapeutic potential. Most support continued research and legal medical access. Yet objective knowledge remains uneven, and confidence frequently exceeds demonstrated understanding.

The encouraging finding is that clinicians appear interested in learning more. Nearly four out of five respondents expressed interest in additional cannabis training. That suggests the problem is not resistance. The problem is infrastructure. Educational systems have not yet caught up with clinical reality.

As cannabis use becomes increasingly integrated into routine healthcare, the question is no longer whether clinicians should understand cannabis. Patients have already answered that question through their behavior. The question now is whether medical education, continuing education programs, professional organizations, and healthcare systems can provide the evidence-based training necessary to ensure that clinicians are prepared for conversations they are already having every day.

Selected References

1. Kumar L, Wang E, Mathai DS, Zamarripa CA, Spindle TR, Vandrey R, Garcia-Romeu A.
   Knowledge, attitudes, and concerns about medical cannabis among U.S. healthcare professionals.
   Journal of Cannabis Research. 2026.
   doi:10.1186/s42238-026-00450-8.
2. Kruger DJ, Mokbel MA, Clauw DJ, Boehnke KF.
   Assessing Health Care Providers’ Knowledge of Medical Cannabis.
   Cannabis and Cannabinoid Research.
   2022;7(4):501-507.
3. Philpot LM, Ebbert JO, Hurt RT.
   A survey of the attitudes, beliefs and knowledge about medical cannabis among primary care providers.
   BMC Family Practice.
   2019;20(1):17.
4. Weisman JM, Rodríguez M.
   A systematic review of medical students’ and professionals’ attitudes and knowledge regarding medical cannabis.
   Journal of Cannabis Research.
   2021;3(1):47.
5. Worster B, Valleriani J, Zawertailo L, Selby P.
   Clinician Attitudes, Training, and Beliefs About Cannabis: An Interprofessional Assessment.
   Cannabis and Cannabinoid Research.
   2023;8(3):547-556.
6. Rice J, Hildebrand A, Waslo CS, Cameron MH, Jones KD.
   Cannabis for medical purposes: a cross-sectional analysis of health care professionals’ knowledge.
   Journal of the American Association of Nurse Practitioners.
   2022;34(1):100-106.