Differences in Cannabidiol-Related Attitudes and Practice Behaviors Between U.S. Primary Care Physicians Practicing in a Single Health Care System Across States With and Without Marijuana Legalization

Abstract

Introduction:

Dramatic shifts in marijuana laws, along with federal deregulation of hemp with the 2018 Farm Bill, have resulted in increased availability and use of cannabidiol (CBD) supplements throughout the United States (US). Given the rapid increase in CBD use in the U.S. general population, in this study, we aim to characterize primary care physician (PCP) attitudes and practice behaviors and to assess whether differences in provider attitudes and behaviors vary as a function of marijuana legalization (ML) status in the state of practice.

Materials and Methods:

Data are from an online provider survey on CBD supplement-related attitudes, beliefs, and behaviors administered to 508 PCPs as part of a larger mixed methods study. Participating PCPs were recruited from the Mayo Clinic Healthcare Network and provided medical care in primary care settings across four U.S. states (Minnesota, Wisconsin, Florida, and Arizona).

Results:

The survey response rate was 45.4% (n=236/508). According to providers, CBD was frequently brought up in PCP settings, typically by patients. PCPs were generally hesitant to screen for or discuss CBD with their patients and identified multiple barriers to open patient–provider dialogue about CBD. PCPs practicing in states that had passed ML were more receptive to patients using CBD supplements, whereas PCPs practicing in states that had not passed ML were more concerned about CBD-related side effects. Regardless of state ML status, most PCPs did not feel that they should be recommending CBD supplements to their patients. Most PCPs reported believing that CBD was unhelpful for most conditions for which it is marketed, with chronic non-cancer pain and anxiety/stress being exceptions. PCP respondents generally felt that they had insufficient knowledge/training around CBD.

Conclusions:

Results from this mixed methods study show that PCPs practicing in the U.S. rarely screen for or discuss CBD use with their patients and report several barriers to engage in proactive CBD-focused practice behaviors. Furthermore, survey results show that some PCP attitudes, practice behaviors, and barriers vary as a function of state ML status. These findings may guide medical education efforts and inform primary care practice modifications aimed at enhancing screening and monitoring of patient CBD use by PCPs.

Introduction

The United States cannabinoid market is estimated in the billions of dollars, with cannabidiol (CBD)-containing products making up a significant portion of the market share. CBD is a nonpsychoactive phytocannabinoid with broad psychopharmacological effects; poorly understood mechanisms that may include antioxidant, anti-inflammatory, and central nervous system modulatory properties; and potential medical applications.¹ In the U.S., CBD has been sold as a dietary or health supplement in multiple formulations as CBD-containing tinctures, capsules, gummies, vape oil, e-juice, and disposable vape pens, and CBD-infused food and drinks. Problematically, the growing demand for CBD products and broad claims of purported benefits have greatly outpaced the body of literature substantiating its use as a therapeutic treatment for medical conditions.^2–4

More than a quarter of all American adults report having tried CBD supplements in the past 2 years.² Results from a 2019 Gallup Poll show that 14% of U.S. adults, including 20% of U.S. young adults, report using CBD products in the past year.³ CBD supplements are not risk-free.^5,6 CBD supplement use may pose significant risks to some patients, including potential supplement–drug interactions between CBD supplements and prescription medication, side effects (SE), and other adverse effects. Data from clinical and pre-clinical studies indicate that CBD use is associated with drowsiness, diarrhea, and abnormal liver functioning in some patients. Supplement–drug interactions with prescribed medications have been found to occur between CBD and several commonly prescribed medications for psychiatric and neurological conditions.⁵

Furthermore, many patients may replace conventional medication in favor of CBD supplements, despite limited evidence for safety or efficacy. For example, 42% of CBD users from a recent national survey reported replacing a prescription or over-the-counter medication for pain, anxiety, or sleep with CBD supplements.² These risks for adverse outcomes related to CBD supplement use may be further amplified by problems related to contamination, adulteration, and mislabeling reported in commonly available CBD products.^7,8

Because of these concerns, national regulatory and medical organizations such as the U.S. Food and Drug Administration, the National Institutes of Health, and the Joint Commission recommend that patients consult a health professional before starting dietary supplements (such as CBD) and physicians inquire about CBD and other dietary supplement use and document patient's use of supplements, evaluate the types/kinds/content of supplement used, discuss available safety and efficacy data with their patients, and monitor their patients who choose to take supplements for adverse events (AEs) and therapeutic responses.⁹ Despite these national recommendations, little is known about health care provider practices and attitudes toward CBD supplement use.

Most patients believe that CBD supplements are safe because they are “natural” and sold over-the-counter.¹⁰ Furthermore, the expanding CBD supplement market results in patients' desire for an alternate option, greater control over their health care treatment, and decreased utilization of evidence-based medical treatments offered by their health care providers in favor of unsupported CBD-based treatments.¹¹ As patients can obtain CBD supplements without a prescription, few disclose their CBD supplement use to their health care provider.¹² This represents a patient–provider communication gap that, if addressed, could mitigate the risk for AEs and supplement–drug interactions in patients identified as being at elevated risk for harm and maximize health benefits for patients identified as being most likely to benefit from CBD supplement treatment. One way to bridge this patient–provider gap is to increase screening and monitoring of patient CBD supplement use among primary care physicians (PCPs).¹³

In the medical marijuana space, PCPs are often the patients' first contact in health care systems for patients interested in medical marijuana and its derivative.^14,15 PCPs must certify patients based on their states' qualifying medical or psychiatric conditions for patients to receive medical marijuana.¹⁴ Related to CBD supplement: PCPs are well positioned to screen for and advise patients about indications for CBD supplement use, supplement–drug interactions, and other potential risks related to supplement use, based on their knowledge about their patient's medical history and medications and role in routine health monitoring. Understanding current PCP attitudes and practice behaviors related to CBD supplements, and perceived barriers to patient–provider communication about CBD supplements can be used to guide next-generation medical training to improve screening and monitoring by U.S. PCPs.

Dramatic changes in state marijuana laws over the past two decades have shifted societal perceptions and use patterns of cannabinoid products across the U.S. with unclear implications for U.S. health care practice and American health outcomes. At the state level, marijuana has been legalized for medical indications and recreational purposes in many U.S. states, with the number of states passing marijuana legislation growing each year.¹⁶ As of July 2022, 38 states and the District of Columbia (DC) have passed and implemented medical marijuana laws (MML), whereas 19 states and the DC have passed and implemented recreational marijuana laws (RML).

Enactment of state MML and RML has been linked with changing perceptions about the risk for harm and potential health benefits related to cannabinoid product use (including CBD products) among individuals living in those states. It goes to follow that health care providers living or practicing in states that have enacted MML or RML may also experience changes in their attitudes, risk perceptions, and health perceptions related to cannabinoid products, with these shifting perceptions changing the likelihood that they inquire about and recommend these products to their patients.

Studies of PCP practice behaviors related to medical marijuana show that PCPs perceptions, beliefs, and understanding of cannabinoid products play an essential role in whether PCPs recommend medical marijuana to patients requesting it.^17,18 For example, in a survey conducted among PCPs practicing in Colorado, which has both MML and RML, only 19% of PCPs thought that physicians should recommend medical marijuana to patients.¹⁹ In contrast, another survey querying physicians practicing in New York, another state with both MML and RML, found that most physicians believed that patients should be able to access marijuana when necessary for medical reasons.¹⁸ Given these findings, it is important to determine if health care providers' attitudes, perceptions, and practice behaviors are changing in relation to changing marijuana policies across the U.S.

Given the rapidly shifting cannabinoid health care space, it is increasingly important to understand health care provider attitudes, beliefs, and practice behaviors related to CBD and other cannabinoid-based products. In addition, it is unclear if provider attitudes, beliefs, and practice behaviors toward CBD are influenced by the state's cannabis legalization climate. Based on the dearth of evidence in this area, our group has conducted a mixed methods study to investigate attitudes and behaviors related to CBD and how PCPs and patients discuss CBD supplements. Results of the qualitative arm of our study exploring PCPs attitudes, experiences, and practices about CBD and barriers to patient–provider communication around patient CBD supplement use showed that PCPs do not regularly screen for patient CBD use, citing busy practices and lack of time in busy primary care practice.¹³ Therefore, clinical inquiry and efficient screening of CBD use by patients in a clinical space are essential.

How PCPs beliefs and perceptions about CBD affect their practice behavior is currently unknown. Building off of our qualitative work, the present article reports on the quantitative component of our mixed methods study—a survey of PCPs in a large medical institution querying attitudes, beliefs, and practice behaviors related to CBD and examining differences in these attitudes and behaviors as a function of marijuana legalization (ML) status in the state of practice. Our main objectives were to characterize PCPs attitudes and practice behaviors related to CBD supplement use by their patients and to assess whether differences in provider attitudes and behaviors existed as a function of ML status in the state of practice. We hypothesized that PCPs practicing in states where marijuana is legal (RML and MML vs. no ML) would endorse higher acceptability of patient CBD supplement use and would be more likely to openly discuss both risk and benefits associated with CBD use in their clinical practice.

Methods

Study overview

This report used data from an electronic survey developed and administered to Mayo Clinic and Mayo Clinic Health System (MCHS) PCPs as part of a larger mixed methods study examining provider and patient attitudes, beliefs, and behaviors related to CBD and medical cannabis and investigating the content and characteristics of patient–provider discussions about CBD supplements in 2021. Details of the larger mixed methods study are detailed elsewhere.¹³ Using results from the qualitative arm of the mixed methods study as a guide, the survey was designed to obtain basic sociodemographics and practice characteristics of providers and to query provider attitudes, beliefs, and practice behaviors related to CBD supplement use by their patients.

Survey instrument

The survey was developed from the themes generated from the qualitative arm¹³ of a larger mixed methods study, which aimed to gather basic sociodemographic information about PCPs and to query provider attitudes, beliefs, and practice behaviors related to CBD supplement use by their patients. Guided by the qualitative results and informed by a scoping literature review on the topic, the survey items were developed to address existing gaps in the literature related to patient and provider attitudes about CBD and patient–provider communication around CBD supplement use. All survey items were agreed upon by authors' consensus. The full survey included 24 items querying provider demographics, practice characteristics, and specific CBD-related items focused on attitudes and beliefs about CBDs effectiveness for different medical and psychiatric conditions, practice behaviors related to CBD supplements, and provider-reported barriers to open provider–patient communication about CBD (Supplementary Table S1).

Data collection/procedure

Mayo Clinic Survey Research Center (SRC) conducted all survey administration procedures. Following approval of the survey by the Mayo Clinic Institutional Review Board, a cover letter describing the survey with an attached link was distributed via an email invitation to targeted PCPs in the MCHS meeting study eligibility criteria. The email cover letter had the subject line “Factors Affecting CBD Related Communication in Outpatient Clinical Setting.” In the cover letter, participants were informed that the survey was anonymous and confidential, that their participation was optional, and that they could leave the survey or skip questions anytime. No compensation was offered for survey completion. The survey was available from January 27, 2022, to March 1, 2022. After survey closure, study investigators received de-identified data from the SRC for analysis.

Statistical analyses

Statistical analyses were conducted using Statistical Analysis Software version 9.2 (Cary, North Carolina, USA) and R Statistical Software Version 4.2.2 (www.r-project.org). The significance for all statistical tests was set at p<0.05. A multipronged approach to managing missing data was taken. First, we assessed survey data for patterns of missingness, and outlier items were identified and removed. Through this process, we removed one survey item (Q4.5.: an item querying perceived health benefits of CBD for autism spectrum disorders) because of poor response rate. Next, we stratified the sample of responders into groups of survey completers (n=199) and non-completers (n=37), using 75% item completion rate as a cutoff.

With this approach, survey completers were defined as participants who submitted the survey and answered more than 75% of the survey questions (i.e., >18 of the 24 items). In group comparisons, survey completers and non-completers did not differ in sex, Hispanic ethnicity, years in medical practice, or ML status of state of practice (all p>0.05) but did differ in racial composition, with the proportion of respondents identifying as White/Caucasian race being higher among completers compared with non-completers (76.4% vs. 17.1%, p≤0.0001). The main analysis was conducted from the completer sample (n=199) removing participants deemed non-completers. Missing values on variables of interest in the completer sample was modest (0.5–8.0%).

For analyses, missing values were included in the total sample size but excluded from the denominator when calculating the percentages for individual variables. For our a priori analyses focused on the effects of state ML status on provider attitudes, health perceptions, and practice behaviors, we had initially planned to stratify the analytic sample into three groups based upon ML status (i.e., RML, MML, no-ML) and run comparative analyses. However, during variable preparation, we found that the groups were unbalanced and too few PCP respondents from RML states were available for an appropriately powered three-way comparison. Based on this result, we decided to collapse RML and MML groups into a single ML group and shifted our primary analysis to a two-way comparison of PCPs practicing in states with any ML (n=128) and no-ML (n=67), reserving three-way RML versus MML versus no-ML group comparisons for exploratory analyses.

When the total sample was restratified along any ML status, no group differences were observed between ML and no-ML groups on race, ethnicity, sex, and length of time in practice (all p>0.05). Descriptive analyses were conducted and are presented as frequencies and simple proportions for each survey item, within and across groups. Bivariate analyses were conducted to compare responses to each survey item between PCPs practicing in ML compared with no-ML states using χ² tests and Fisher's exact tests for categorical variables and the Kruskal–Wallis one-way analysis of variance (ANOVA) for continuous variables. For outcomes that showed significant effects in bivariate analyses, exploratory post hoc analyses were conducted comparing PCPs in RML, MML, and no-ML states using the Kruskal–Wallis one-way ANOVA tests.

Given racial differences between survey completers and non-completers, two sensitivity tests were conducted to assess the influence of race on our primary outcomes and mitigate its possible confounding effect on our ML analyses. First, we ran Cochran–Mantel–Haenszel (CMH) tests in the total sample of completers (n=199), enabling us to examine ML versus no-ML group differences in CBD-related attitudes and behaviors, while controlling for the effects of race on our outcomes.

As a second sensitivity test, using a proportion matching approach, we matched survey-completing respondents from the ML and no-ML groups in a 1-to-1 ratio using self-reported White/Caucasian race (yes=1, no=0) as a grouping variable and then reran our main ML versus no-ML bivariate analyses using these racially matched subgroups (n=67 per race-matched group). An important limitation of the propensity matching approach should be noted: This approach decreases the sample size by excluding subjects during the matching process, and, in doing so, reduces the power to detect group differences as a function of ML status.

Results

Sample characteristics

The survey was distributed to PCPs who provide general primary medical care for patients in Mayo Clinic (across Minnesota [MN], Arizona [AZ], and Florida [FL]) and MCHS (across MN and Wisconsin [WI]) outpatient clinical settings across diverse urban and rural locations practicing primary care. Inclusion criteria were as follows: PCPs were invited to participate in the survey if they were a practicing medical doctor (with a valid state medical license and MD/DO/MBBS degrees) who provided medical treatment for patients in outpatient primary care clinics and specialized in Family Medicine, Internal Medicine, Pediatrics, or Adolescent Medicine. Exclusion criteria included hospitalists and those not working in an outpatient community primary practice.

The survey response rate was 45.4% (n=236/508). One hundred ninety-five (83%) of the 236 PCPs practicing primary care in Mayo Clinic and MCHS outpatient clinical settings who initiated the survey completed >75% of survey items and were included in the analytic sample for this report. All respondents from the analytic sample (100%) indicated that CBD-related topics had been brought up in their clinical practice. Respondents from the analytic sample exhibited diverse sociodemographic and practice backgrounds, with 76.4% of respondents identifying as non-Hispanic White race/ethnicity and 54.3% identifying as male sex. Length of time practicing medicine was also variable, with 43.1%, 29.2%, and 27.7% of respondents reporting >20 years, 11–20 years, and <10 years of experience practicing medicine, respectively. PCP respondents practiced medicine in one of the outpatient primary care settings in four states (WI, MN, FL, and AZ) with variable ML.

At the time of the survey, 38 states and the DC have passed and implemented MML, whereas 19 states and the DC have RML. All states that passed RML had previously passed and implemented MML. See Supplementary Table S2 for details on ML status for each state represented in the report. Of the 195 PCP respondents, 13 (6.7%) practiced in RML states, 115 (59.0%) practiced in MML states, and 67 (34.4%) practiced in states with no-ML. Survey respondents practicing in RML, MML, and no-ML states did not differ in sex, ethnicity, or years of practice (all p>0.05) but did show racial differences, with the proportion of respondents identifying as White/Caucasian race being higher among those practicing in no-ML compared with RML but not MML states (74.6% vs. 46.2% vs. 80.9%, p=0.02).

Results of descriptive and comparative analyses examining CBD-related attitudes and behaviors in PCPs in the total sample, as a function of state ML status, are shown in Supplementary Table S3 and described below.

Frequency and characteristics of patient–provider CBD discussions

All PCP respondents (100% in both groups) reported having discussions with their patients about CBD supplements during a clinical encounter. Regardless of state ML status, most CBD-related discussions were initiated by patients, with fewer than 10% of PCPs from each group reporting that they “always” or “often” initiate CBD-related discussions. In terms of frequency, respondents reported that CBD was brought up as a topic of discussion in ∼5% of outpatient clinic visits during a typical week. The frequency of CBD-related discussions did not differ between respondents who practiced in ML and no-ML states (4.9% vs. 4.9% of visits).

CBD-related beliefs about medical benefits and concerns and practice behaviors

A minority of PCPs agreed that “physicians should recommend CBD supplement use for non-FDA approved indications.” ML and no-ML groups did not differ in this belief (12.6% vs. 16.7%). Regarding their beliefs about the medical benefits of CBD, PCP respondents practicing in ML and no-ML states did not differ in their beliefs about the helpfulness or unhelpfulness of over-the-counter CBD supplements for treatment of or relief from various medical and psychiatric conditions, including depression, anxiety/stress, sleep, post-traumatic stress disorder, autism, nausea, vomiting, cancer and non-cancer chronic pain, and other neurological conditions.

While proportions varied by condition asked, most providers reported that CBD supplements are generally unhelpful for the above conditions, with a few notable exceptions in the domains of pain, anxiety, and quality of life (QOL). Across both ML and no-ML groups, the highest percentage of PCP respondents reported believing that over-the-counter CBD supplements were helpful or very helpful for relieving or treating chronic non-cancer pain (63.4% and 67.2%) and anxiety/stress (51.2% and 48.5%) and improving patient QOL (48.4% and 50.0%).

Regarding beliefs about harms/adverse effects related to CBD use, the most commonly cited SEs or AEs that PCPs in both ML and no-ML groups expressed concern about were drug-to-drug interactions (68.0% vs. 67.2%), sedation (44.5% vs. 43.3%), somnolence (43.0% vs. 44.8%), and addiction risk (33.6% vs. 26.9%). Notably, nearly a third of respondents in both groups endorsed believing that “CBD supplements can be addictive” (ML: 32.3% vs. no-ML: 33.3%). ML and no-ML groups did not differ in these concerns but did differ in concerns related to gastrointestinal SEs. A higher proportion of PCPs practicing in no-ML compared with ML states expressed concerns about CBD-related gastrointestinal SEs such as diarrhea (no-ML: 26.9% vs. ML: 11.4%, p=0.0055).

Regarding practice behaviors, PCPs in no-ML states were less likely to recommend continuing CBD for a patient taking over-the-counter CBD supplements (no-ML: 62.1% vs. ML: 78.7%, p=0.0136), and a higher proportion of PCPs from ML states reported discussing risk and benefit analysis of CBD supplement use compared with PCPs from no-ML states (no-ML: 56.1% vs. ML: 69.5%, p=0.0902).

Regarding medical training/education, PCPs in ML states, compared with no-ML states, were more likely to express the need for more training on how to discuss CBD with patients (no-ML: 82.1% vs. ML: 92.1%, p=0.0361).

Physician barriers and physician-perceived patient barriers to CBD-related discussions

Regarding barriers to initiating CBD discussions with patients, PCP respondents in both groups commonly cited provider/practice barriers of “too little time,” “not as important as other topics,” “lack of provider knowledge on CBD supplements,” and “lack of medical guidelines to appropriately advise patients” and physician-perceived patient barriers of “patient is already using CBD and does not feel the need to report it to the physician,” “patients think physicians will not recommend CBD use,” “defensiveness on patients part,” “because CBD is over-the-counter,” and “social stigma associated with marijuana compounds.”

In addition, a high percentage of PCPs from both the groups disagreed or somewhat disagreed that “discomfort asking patients about CBD use” and “concerns about legality” were barriers to initiating CBD communication with their patients. In examining across barrier types, group comparisons identified one barrier that approached statistical significance between ML and no-ML groups. PCPs in ML states were more likely to report “social stigma associated with marijuana compounds” as a barrier to CBD discussions compared with PCPs in no-ML states (no-ML: 15.9% vs. ML: 33.6%, p=0.0533).

Sensitivity tests

Two sensitivity tests were conducted to investigate variance in outcomes related to race as survey completers and non-completers differed on this sociodemographic variable. In CMH test analyses examining the effects of ML group while accounting for race in the total sample (Supplementary Table S4), results paralleled the findings from our main analysis, with differences being negligible between analyses ignoring and covarying for race, with one exception. For item 10.3, the distribution between question responses and ML status remained significant after adjusting for race (χ²=5.21, p=0.02), but the CMH test showed that ML effects, while present, were not independent at each level of the race variable. This suggests that race may have impacted the main outcome for this item.

The distribution of item responses as a function of ML status was shown to be independent of race for all other survey items. In propensity matched subgroup analyses comparing differences in CBD-related attitudes and practice behaviors between race-matched PCPs practicing in ML and no-ML states (Supplementary Table S5), the proportions/percentages of respondents agreeing or disagreeing with the question stem for each item showed minimal change from the main to the subgroup analysis, but there were three items (Q10.3, Q12, and Q18.1) for which significant results in the main analysis were no longer significant in the subgroup analyses. Parallel results from both the CMH and race-matched subgroup analyses for item 10.3 (“physicians should recommend continuing CBD supplements if they feel they are beneficial to their patient”) suggest that racial differences may be present and could confound our main analytic result for this item. For all other items, these test results suggest minimal impact of race on study outcomes.

Discussion

The main goal of the present study was to characterize attitudes/beliefs, clinical practice behaviors, and patient–provider communication barriers related to CBD supplements among PCPs practicing in outpatient primary care settings across a multistate health care system in the U.S. and determine if these attitudes, beliefs, and behaviors varied as a function of marijuana laws in the PCPs states of practice. To achieve this goal, we used data from qualitative interviews to develop an anonymous online survey querying CBD supplement-related attitudes, beliefs, and behaviors and administered it to 508 PCPs who provide medical care in outpatient primary care clinics within the multistate Mayo Clinic Health Network. Our survey results show that patient–provider discussions about CBD use are relatively common in PC settings, and almost always initiated by patients.

PCP who completed the surveys reported rarely screening for or initiating discussions about CBD use and identified several practice-level, provider-level, and patient-level barriers to open patient–provider dialogue about CBD. Regarding attitudes/beliefs, most PCPs were skeptical of the medical usefulness of CBD and did not feel that they should be recommending CBD supplements to their patients. A majority of PCP respondents felt that they had insufficient knowledge/training around CBD. Some attitudes/beliefs, perceived barriers, and practice behaviors varied in relation to the ML in the state of practice. These findings carry implications for clinical management of patients in PC settings and for medical education and PCP workforce training.

A similar percentage of PCPs from ML and no-ML noted that patients generally initiate CBD-related conversation in a clinical space, and CBD could be more helpful than unhelpful for “non-cancer-related chronic pain,” and “relieving or treating anxiety/stress.” The PCPs from ML and no-ML states differ in perceived SEs associated with CBD use. All concerns about SEs were noted higher in no-ML versus ML states. In addition, a higher percentage of PCPs from ML than no-ML states recommended continuing CBD if patients find it helpful, discussing risk and benefit analysis of CBD supplements with patient, and expressed need for more training about CBD. We found that PCPs were hesitant to screen for or initiate CBD discussions with their patients and that this hesitancy did not vary as a function of state marijuana legislation status.

This could be a missed opportunity with clinical implications for patients who choose to use and not actively disclose CBD to their providers, especially given rising rates of CBD use in the general population. Undisclosed CBD supplement use by patients could place them at risk for AEs associated with CBD use, particularly related to drug interactions. It is important to identify provider-level characteristics and system-level barriers that drive low screening rates and PCPs hesitancy to discuss CBD use with patients as identifying these variables may inform targeted education and system-level changes.

Common barriers to PCP initiated CBD discussions identified by our study include lack of time and competing demands of busy clinical practice. Addressing these barriers may require shifts to more technology-facilitated screening approaches at the system level. Relatedly, the efficient and innovative use of electronic health records (EHR) could be an answer. For example, patients who use the remote patient portal service could “self-screen” for CBD, which may prompt clinicians to discuss CBD use by patients in clinical space. Entering CBD supplement use in the medication list could also prompt clinicians for potential drug–drug interaction. However, these features may vary based on the type of EHR system.

Based on the results of our study and evidence drawn from marijuana literature.^20,21 PCPs perception of the benefit and harm of marijuana products (including CBD) are more knowledge dependent and are limitedly impacted by the marijuana legality of the state they practice. Our study also indicated that CBD-related topics are prevalent within the primary care practice. Prior studies have shown that patients living in MML states have a favorable opinion of marijuana for treating medical and psychiatric symptoms. In addition, states that have legalized marijuana have higher rates of marijuana use among the population even before the passage of ML.

Contrary to our hypothesis about the effects of ML on the frequency of CBD-related discussions, we found no difference in the frequency of these topics based on state legalization. Rather we found that CBD discussions, content, and barriers were largely similar in states that have and have not legalized marijuana, according to providers. As highlighted by the PCPs in our survey from both ML and no-ML states, we believe that patients may think that CBD is an over-the-counter product, which could have prevented them from frequently bringing the CBD-related topic with their PCPs during their clinical visits.

The study results confirmed our a priori hypotheses that the PCPs practicing in states that had passed marijuana legislation (compared with those practicing in states with no-ML) would find patient CBD supplement use more acceptable and have some differences in perceptions about harmfulness, SE, and provider behaviors related to discussing risk and benefit (rather than only risk) associated with CBD use in their clinical practice. Studies show that patients often use marijuana without input from PCPs, potentially placing them and their families at risk of adverse outcomes.^22,23 We believe that principles noted by the National Academy of Sciences for the use of complementary and alternative therapies in medical practice can be applied to CBD-related practice and that PCPs from ML states in our study seem to have a suitable approach by locating an equilibrium between patients' autonomy and concerns for patients' safety by protecting them from preventable harm.²⁴

There are several relevant limitations and some notable strengths to the present study. The findings are from an anonymous online survey of PCPs from the Mayo Clinic and MCHS. While self-report is the main way of characterizing attitudes, beliefs, and expectancies about drug effects, this approach carries inherent limitations. Respondents may have given answers that they perceived to be more acceptable to themselves, their peers, and health system, reflecting a social desirability bias. Respondents could have also experienced recall bias. Self-report measures may not be the best way to characterize practice behaviors, which may be more accurately assessed by provider observation or multi-informant reports from providers and patients. Our use of both qualitative and quantitative methods in the larger study mitigates some of these limitations.

Another notable limitation of this study is related to the sample's generalizability. Our study sample was entirely from a single Health System affiliate with PCPs drawn from four U.S. states. In designing the study, we chose to focus on PCPs based on the rationale that this group of health care providers represents the primary gatekeepers to population health in the U.S. CBD-related attitudes and practice behaviors may not be generalizable to physicians not working in primary care settings, physicians with specialty training, other provider types (e.g., nurse practitioners), or PCPs practicing in other health systems or practice settings. Our sample had other limitations: While the sample showed a composition similar to the general PCP workforce in the U.S.,²⁵ it was still mostly White (76%) and male (54%). PCPs who completed the survey and included in the main analysis were more likely to be White/Caucasian.

To address this limitation, we conducted a series of post hoc sensitivity tests to examine the effects of race on our study outcomes, the results of which showed minimal impact of race on study results, strengthening our main findings. Still, given the sample characteristics, our findings may not generalize to providers from other races and backgrounds. Additional research is needed in this area. Our secondary aim of characterizing differences in provider attitudes and practice behaviors as a function of state ML status was dependent on the natural experiment of state ML passage in U.S. states in which Mayo Clinic has PCP practice sites. Related to this aim, our sample, which surveyed PCPs from four U.S. states only one of which had passed RML, is a convenience sample. As a result of this sampling frame, too few respondents practicing in RML states were available for appropriately powered three-way analyses comparing RML, MML, and no-ML differences, and our main analyses were limited to focusing on larger ML versus no-ML comparisons.

A growing body of literature shows significant heterogeneity across states with MML and RML, with variability in provisions and implementation practices likely contributing to differential effects of ML passage on consumer and provider attitudes, behaviors, and health outcomes.²⁶ Future studies characterizing effects of ML on changes in provider attitudes and behaviors about CBD and medical marijuana should seek to draw larger samples with greater diversity and consider oversampling from MML and RML states to allow for comparisons within and across ML types. Finally, we did not gather clinic-site-specific information such as clinic location and zip code of physician practice, which could have helped assess the effect of rurality on explored associations.

Despite these limitations, this study has several strengths. Our mixed method design is a major strength, with the alignment between our survey results (presented here) and the semi-structured interview results from the qualitative portion of our study (presented in Sharma et al.¹³) bolstering the strength of our findings. Although the sample is a convenience sample from a single health care system, its large size and sample features, namely that it is a sample of PCPs that is generally representative of the U.S. PCP workforce and that included PCPs recruited from multiple U.S. states with diverse ML can be viewed as a strength.

To the best of our knowledge, our study is one of the first studies to examine provider–patient CBD-related discussion characteristics from the vantage point of PCPs. Given the sample characteristics and size, it is also one of the largest studies conducted to date in health care providers characterizing CBD-related attitudes and practice behaviors and the only one to examine differences in provider attitudes and behaviors related to CBD supplement use as a function of ML status.

Conclusions

In conclusion, the results of this anonymous survey of U.S. PCPs practicing in a single health system across four U.S. states showed that CBD supplement use is frequently brought up in primary care practice settings, typically by patients as a result of provider-reported barriers and that some provider attitudes and practice behaviors may vary as a function of state ML status. Regardless of state ML status, most PCPs did not feel that they should be recommending CBD supplements to their patients and generally believed that CBD was unhelpful for most conditions for which it is marketed. Additionally, most PCPs reported feeling that they had insufficient knowledge/training and low comfort discussing CBD with their patients. Our study results provide early evidence that state ML status may influence health care provider attitudes and beliefs.

We found that PCPs from ML states were more receptive to patients using CBD supplements, whereas PCPs from states without ML expressed more concerns about CBD-related SE and AE. Whether differences in attitudes and beliefs generalize to other samples and lead to differences in PCP practice behaviors warrant further investigation and may carry significant clinical implications. The results from this study may be used to optimize education and dissemination efforts around evidence-based best practices related to cannabinoid products for medical and psychiatric conditions. Addressing this knowledge and training gap carries the potential to improve gains related to novel cannabinoid-based treatments and mitigate risks for cannabinoid-related adverse outcomes in vulnerable subgroups who are more likely to experience adverse reactions to these products.

Footnotes

Acknowledgments

The authors would like to thank Dr. Rheanna Platt, Dr. Ashley Holland, and Tahart Sheikh for their assistance during the qualitative phase of this study. We also thank Mayo Clinic SRC for designing and disseminating the survey and Quantitative Health Sciences, particularly Dustin Duffy for statistical support.

Authors' Contributions

P.S.: Conceptualization, methodology, writing—original draft preparation, writing—review and editing, and funding acquisition; R.P.W.: Writing—review and editing; K.L.: Formal analysis; C.J.H.: Supervision and writing—review and editing.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by Mayo Clinic Health System, Office of Research.

Supplementary Material

Abbreviations Used

References

VanDolah

, Bauer

, Mauck

. Clinicians' guide to cannabidiol and hemp oils. Mayo Clin Proc, 2019; 94(9):1840–1851; doi: 10.1016/j.mayocp.2019.01.003

Gill

LL.

CBD goes mainstream. Consumer Reports April 11, 2019.

Brenan

14% of Americans say they use CBD products. Gallup August 7, 2019.

Walker

, Koturbash

, Kingston

, et al. Cannabidiol (CBD) in dietary supplements: Perspectives on science, safety, and potential regulatory approaches. J Diet Suppl, 2020; 17(5):493–502; doi: 10.1080/19390211.2020.1777244

Huestis

, Solimini

, Pichini

, et al. Cannabidiol adverse effects and toxicity. Curr Neuropharmacol, 2019; 17(10):974–989; doi: 10.2174/1570159X17666190603171901

Madeo

, Kapoor

, Giorgetti

, et al. Update on cannabidiol clinical toxicity and adverse effects: A systematic review. Curr Neuropharmacol, 2023; doi: 10.2174/1570159x21666230322143401

Bonn-Miller

, Loflin

MJE

, Thomas

, et al. Labeling accuracy of cannabidiol extracts sold online. JAMA, 2017; 318(17):1708–1709; doi: 10.1001/jama.2017.11909

Freedman

, Patel

. Inadequate regulation contributes to mislabeled online cannabidiol products. Pediatr Neurol Briefs, 2018; 32:3; doi: 10.15844/pedneurbriefs-32-3

Health

NI.

Dietary supplements: What you need to know. National Institutes of Health (NIH) Office of Dietary Supplements (ODS): Bethesda, MD; 2023. Available from: https://ods.od.nih.gov/factsheets/WYNTK-Consumer/ [Last accessed: April 18 ].

10.

Schilling

, Hughes

, Wallace

, et al. Cannabidiol as a treatment for chronic pain: A survey of patients' perspectives and attitudes. J Pain Res, 2021; 14:1241–1250; doi: 10.2147/jpr.S278718

11.

Administration. Cannabidiol (CBD)—potential harms, side effects, and unknowns. Substance Abuse and Mental Health Services Administration (SAMHSA): Rockville, MD; 2023. Available from: https://store.samhsa.gov/sites/default/files/pep22-06-04-003.pdf [Last accessed: April 18 ].

12.

Ryan

, McCabe

, Boyd

. Medicinal cannabis: Policy, patients, and providers. Policy Polit Nurs Pract, 2021; 22(2):126–133; doi: 10.1177/1527154421989609

13.

Sharma

, Holland

, Sheikh

, et al. Primary care provider attitudes, experiences and practices about cannabidiol (CBD) and barriers to patient-provider communication about CBD use: A qualitative study. PEC Innovation, 2022; 1:100044; doi: 10.1016/j.pecinn.2022.100044

14.

Ebbert

, Scharf

, Hurt

. Medical cannabis. Mayo Clin Proc, 2018; 93(12):1842–1847; doi: 10.1016/j.mayocp.2018.09.005

15.

Pacula

, Smart

. Medical marijuana and marijuana legalization. Annu Rev Clin Psychol, 2017; 13:397–419; doi: 10.1146/annurev-clinpsy-032816-045128

16.

Hammond

, Chaney

, Hendrickson

, et al. Cannabis use among U.S. adolescents in the era of marijuana legalization: A review of changing use patterns, comorbidity, and health correlates. Int Rev Psychiatry, 2020; 32(3):221–234; doi: 10.1080/09540261.2020.1713056

17.

Kondrad

, Reed

. Colorado family physicians and medical marijuana: Has recreational marijuana changed physician attitudes and behaviors?. J Am Board Fam Med, 2022; 35(1):102–114; doi: 10.3122/jabfm.2022.01.210116

18.

Sideris

, Khan

, Boltunova

, et al. New York physicians' perspectives and knowledge of the state medical marijuana program. Cannabis Cannabinoid Res, 2018; 3(1):74–84; doi: 10.1089/can.2017.0046

19.

Kondrad

, Reid

. Colorado family physicians' attitudes toward medical marijuana. J Am Board Fam Med, 2013; 26(1):52–60; doi: 10.3122/jabfm.2013.01.120089

20.

Holman

, Kruger

, Lucas

, et al. Healthcare provider and medical cannabis patient communication regarding referral and medication substitution: The Canadian context. J Cannabis Res, 2022; 4(1):32; doi: 10.1186/s42238-022-00141-0

21.

Wilson

, Klein

, Bindler

, et al. Shared decision-making for patients using cannabis for pain symptom management in the United States. Pain Manag Nurs, 2021; 22(1):15–20; doi: 10.1016/j.pmn.2020.09.009

22.

Boehnke

, Yakas

, Scott

, et al. A mixed methods analysis of cannabis use routines for chronic pain management. J Cannabis Res, 2022; 4(1):7; doi: 10.1186/s42238-021-00116-7

23.

Kruger

, Kruger

, Collins

. Cannabis enthusiasts' knowledge of medical treatment effectiveness and increased risks from cannabis use. Am J Health Promot, 2020; 34(4):436–439; doi: 10.1177/0890117119899218

24.

No authors listed. New model guidelines for the use of complementary and alternative therapies in medical practice. Altern Ther Health Med, 2002; 8(4):44–47.

25.

AAMC. Diversity in Medicine: Facts and Figures 2019. Association of American Medical Colleges: Washington, DC; 2019.

26.

Johnson

, Hodgkin

, Harris

. The design of medical marijuana laws and adolescent use and heavy use of marijuana: Analysis of 45 states from 1991 to 2011. Drug Alcohol Depend, 2017; 170:1–8; doi: 10.1016/j.drugalcdep.2016.10.028

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.70 MB

0.41 MB

1.05 MB

0.42 MB

1.04 MB