January 2018, VOLUME141 /ISSUE 1

Provider Perspectives on Use of Medical Marijuana in Children With Cancer

  1. Prasanna Ananth, MD, MPHa,
  2. Clement Ma, PhDb,c,
  3. Hasan Al-Sayegh, MBChB, MPHb,
  4. Leah Kroon, MN, RNd,e,
  5. Victoria Klein, BSd,
  6. Claire Wharton, BSd,
  7. Elise Hallez, PNPf,
  8. Ilana Braun, MDc,g,
  9. Kelly Michelson, MD, MPHf,h,
  10. Abby R. Rosenberg, MD, MSd,e,i,
  11. Wendy London, PhDb,c, and
  12. Joanne Wolfe, MD, MPHb,c,g
  1. aSection of Pediatric Hematology/Oncology, Department of Pediatrics, Yale School of Medicine, New Haven, Connecticut;
  2. bDepartment of Pediatric Hematology/Oncology, Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Boston, Massachusetts;
  3. cHarvard Medical School, Boston, Massachusetts;
  4. dSeattle Children's Cancer and Blood Disorders Center, Seattle, Washington;
  5. eTreuman Katz Center for Pediatric Bioethics, Seattle, Washington;
  6. fAnn & Robert H. Lurie Children's Hospital of Chicago, Chicago, Illinois;
  7. gDepartment of Psychosocial Oncology and Palliative Care, Dana-Farber Cancer Institute, Boston, Massachusetts;
  8. hCenter for Bioethics and Medical Humanities, Northwestern University Feinberg School of Medicine, Chicago, Illinois; and
  9. iUniversity of Washington School of Medicine, Seattle, Washington
  1. Dr Ananth conceptualized and designed the study, oversaw study execution and data collection, reviewed analyses, and drafted and revised the manuscript; Drs Ma, Al-Sayegh, and London planned and conducted the analyses and reviewed and revised the manuscript; Dr Braun provided guidance on the study design and critically reviewed the draft of the manuscript; Drs Michelson and Rosenberg oversaw study execution at their respective sites and reviewed and revised the manuscript; Ms Kroon, Ms Klein, Ms Wharton, and Ms Hallez assisted with study procedures at their respective sites and critically reviewed the draft of the manuscript; Dr Wolfe conceptualized and designed the study, assisted with survey design, reviewed analyses, and critically reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.

  2. We conducted a sensitivity analysis comparing the results of 2 multivariable models adjusted for provider age, sex, race, and location of practice: (1) comparing attitudes of providers legally ETC for MM with providers n-ETC (reference group), and (2) comparing attitudes of physicians to nonphysicians (reference group). We observed that adjusted ORs and 95% CIs do not substantially differ between models 1 and 2. Results from model 1 are presented in the article (Table 3).


BACKGROUND: Although medical marijuana (MM) may have utility in the supportive care of children with serious illness, it remains controversial. We investigated interdisciplinary provider perspectives on legal MM use in children with cancer.

METHODS: We sent a 32-item, cross-sectional survey to 654 pediatric oncology providers in Illinois, Massachusetts, and Washington characterizing MM practices, knowledge, attitudes, and barriers. Forty-eight percent responded; 44% (n = 288) were included in analyses. Providers were stratified by status as legally eligible to certify (ETC) for MM. We used Fisher’s exact and Wilcoxon rank tests and univariate and multivariate logistic regression models for group comparisons.

RESULTS: The provider median age was 35 years (range 22–70 years); 33% were ETC (83 physicians; 13 Washington state advance practice providers). Thirty percent of providers received ≥1 request for MM in the previous month. Notably, only 5% of all providers knew state-specific regulations. ETC providers were more likely to know that MM is against federal laws (P < .0001). Whereas most providers (92%) reported willingness to help children with cancer access MM, in adjusted models, ETC providers were less likely to indicate approval of patient MM use by smoking, oral formulations, as cancer-directed therapy, or to manage symptoms (P < .005 for all). Forty-six percent of all providers cited the absence of standards around formulations, potency, or dosing to be the greatest barrier to recommending MM.

CONCLUSIONS: Most pediatric oncology providers are willing to consider MM use in children with cancer and receive frequent inquiries. However, ETC providers endorse less favorable attitudes overall. The absence of standards is an important barrier to recommending MM.

  • Abbreviations:
    American Academy of Pediatrics
    confidence interval
    eligible to certify for medical marijuana
    Food and Drug Administration
    medical marijuana
    not eligible to certify for medical marijuana
    odds ratio
    Research Electronic Data Capture
  • What’s Known on This Subject:

    Existing studies have investigated the impact of widespread marijuana legalization on healthy children. Yet, children with serious illness now have unprecedented access to medical marijuana (MM). This necessitates inquiry into how MM may be perceived by frontline, interdisciplinary providers.

    What This Study Adds:

    Most pediatric oncology providers are willing to consider MM use in children with cancer and receive frequent inquiries. However, providers reveal limited knowledge about regulations, and those with legal eligibility to certify endorse less favorable attitudes overall.

    Medical marijuana (MM) has risen to the forefront of controversy throughout the United States.14 MM refers to marijuana use, in plant or extract form, to treat an illness or its symptoms. Twenty-nine states and the District of Columbia have legalized MM for those with serious health conditions, with 8 states legalizing recreational marijuana as well.5,6 Most jurisdictions consider cancer to be a qualifying condition and include provisions for children with life-threatening illnesses.

    Cannabinoids, the chemical components of marijuana, may relieve nausea, anorexia, and neuropathic pain.79 In pediatric cancer care, synthetic cannabinoid, or dronabinol, is routinely prescribed to manage nausea and anorexia.10 Cannabinoids may also carry an antineoplastic effect, although the evidence is limited to preclinical studies.11 The American Academy of Pediatrics (AAP) acknowledges MM as a potential supportive measure for children with serious illness.12,13 Indeed, as children with cancer experience substantial symptom-related suffering,14 there may be a role for MM in their care.

    Nevertheless, MM has not been adopted in pediatrics. First, there is concern for adverse psychiatric and cognitive effects in developing children.1,15,16 Second, appropriate formulations and dosing are unknown.17 Third, critics cite concern that MM access may promote illicit use or toxic ingestion.18,19 Finally, marijuana is a schedule I controlled substance, signifying that there are no currently accepted medical indications and that it carries high potential for abuse; schedule I classification designates procuring, using, or recommending MM as federal offenses.20

    New state MM policies may influence pediatric oncology practice, particularly for children with intractable symptoms for whom conventional therapies have been exhausted. Previous researchers investigated MM attitudes of adult oncology and palliative care providers.2124 Notably, approximately half of adult oncologists espoused favorable attitudes toward MM even when MM was not available legally.21 Moreover, public opinions are shifting, with >70% of Americans supporting adult MM use if recommended by a medical provider.20,2427

    Despite evolving legality and mounting public interest in MM, no studies have explored interdisciplinary provider perspectives on MM use in children with cancer. In this multicenter study, we describe pediatric oncology provider practices, knowledge, attitudes, and barriers regarding legal MM use in children with cancer. We hypothesized a priori that MM knowledge and attitudes might differ depending on whether a pediatric oncology provider is legally permitted to facilitate MM access through a formal certification process. Hence, this study compares the perspectives of providers who are eligible to certify (ETC) versus not eligible to certify (n-ETC) for MM.


    Study Design and Population

    From July 2015 to November 2015, a cross-sectional survey was sent electronically to 654 pediatric oncology providers identified by site-specific principal investigators at 3 National Cancer Institute–designated cancer centers: Dana-Farber/Boston Children’s Cancer and Blood Disorders Center (Boston, MA), Ann & Robert H. Lurie Children’s Hospital of Chicago (Chicago, IL), and Seattle Children’s Hospital Cancer and Blood Disorders Center (Seattle, WA). Eligible participants included all physicians, nurse practitioners, physician assistants, psychologists, social workers, and registered nurses who longitudinally care for children with cancer in inpatient or outpatient settings. This study was approved by the respective institutional review boards.

    Of 654 eligible participants, 313 (48%) responded. We excluded 12 (2%) participants who completed less than half of the survey and 13 (2%) additional participants who did not specify provider type. Ultimately, we included 288 (44%) participants in analyses.

    Survey Instrument

    We created a 32-item survey for this study to address the following domains: practices, including the frequency with which patients and families requested MM, providers recommended MM, and providers facilitated its procurement; knowledge of state-specific and federal MM regulations, mechanisms of access, cultivation, and possession; MM attitudes, including perceived risk of harm, perceived benefits, and opinions on its utility in symptom control; and barriers to recommending MM. Demographic items were incorporated from an existing instrument, the Survey about Caring for Children with Cancer.28 Preliminary survey items were revised by a senior research scientist with survey design expertise. Items were also reviewed for face and content validity by national experts in palliative care. The composite survey was piloted with 5 health care providers outside of pediatric oncology who engaged in cognitive debriefing to help refine items.

    Four questions pertained to practices. Knowledge was measured with 5 true–false and multiple-choice questions. Responses were coded as correct or incorrect. Attitudes were quantified with 6 questions by using a 5-point (0–4) Likert scale. Likert scales ranged from 1 for “strongly disagree or disapprove” to 4 for “strongly agree or approve,” with 0 representing “not sure.” Scores were dichotomized into favorable (3–4) and unfavorable (1–2) attitudes. Two questions addressed barriers to MM recommendation with Likert scale responses as previously stated; a third question on barriers asked respondents to rank potential concerns about MM from greatest to least. Finally, 1 question asked respondents to characterize the appropriate timing of MM use in the course of cancer treatment. Missing responses and responses of “I don’t know” or “not sure” were excluded from analyses.

    We entered the survey instrument into Research Electronic Data Capture (REDCap), a Health Insurance Portability and Accountability Act–compliant, encrypted, Web-based platform.29 We invited eligible providers to participate via e-mail, which included a unique survey hyperlink. Two weeks later, eligible participants who had not yet completed the survey received 2 reminders in 2-week intervals. Participants were requested to complete the survey independently, which took ∼10 minutes. Data were automatically entered into a deidentified database in REDCap, optimizing confidentiality. After a 6-week study period, the database was exported for analyses. Providers who completed the survey were entered into a drawing to receive 1 of 8 $25 gift cards for participation.

    Study Outcomes

    Primary outcomes included knowledge of MM regulations and attitudes about its use in pediatric oncology. Secondary outcomes included practice patterns, barriers to recommending MM, and perspectives on the timing of MM use.

    Study Groups

    For a segment of analyses, providers were categorized into 2 groups to contrast those who were ETC versus n-ETC for MM. In Massachusetts and Illinois, only physicians are legally ETC. In Washington, physicians, nurse practitioners, and physician assistants are ETC. All other providers are n-ETC in any state. To ensure fidelity in our approach to categorizing providers, we conducted a sensitivity analysis and examined attitudes by provider discipline.

    Site Policies

    Although MM is legal in states where this study was conducted, the Chicago site officially prohibits pediatric providers from facilitating MM access, which is in accordance with federal law. The Seattle site permits providers to recommend MM under certain circumstances. No formal policy has been established at the Boston site, although some providers have sought state-mandated training to certify for MM.

    Statistical Analyses

    Survey responses were summarized using descriptive statistics. Provider demographics and MM knowledge were compared among provider types by using the Fisher’s exact test for categorical variables and the Wilcoxon rank test for continuous variables. We used univariate logistic regression to characterize whether attitudes differed by provider type. We included significantly associated attitudes in a multivariate logistic regression model, controlling for age, sex, race (white versus any other race), and location of practice. We also constructed a multivariate logistic regression model to analyze differences in practice patterns among provider types, adjusting for state.

    To account for multiple comparisons in this study, P values ≤ .005 were considered statistically significant. Analyses were performed by using SAS 9.4 software (‎SAS Institute, Inc, Cary, NC).



    Across the 288 providers, the median age was 35 years (range 22–70 years); 85% (n = 246) were women; 92% (260 of 282) self-identified as white; and 99% (n = 283) were non-Hispanic (Table 1). Forty-nine percent (n = 129) practiced in Massachusetts, 40% (n = 116) practiced in Washington, and 15% (n = 43) practiced in Illinois. Thirty-three percent (n = 96) of providers were ETC, whereas 67% (n = 192) were n-ETC. Of ETC providers (n = 96), 86% (n = 83) were physicians; 14% (n = 13) were nurse practitioners or physician assistants. Of n-ETC providers (n = 192), 89% (n = 170) were nurses, 8% (n = 16) were nurse practitioners or physician assistants, 2% (n = 3) were psychosocial providers, and 2% (n = 3) were other providers. n-ETC providers were significantly younger, mostly women, and more predominantly of white race compared with ETC providers (P < .0001).

    TABLE 1

    Baseline Characteristics of Pediatric Oncology Providers


    Overall, 30% (85 of 284) of providers reported 1 or more inquiries for MM, including 40% (38 of 95) of ETC providers and 25% (47 of 189) of n-ETC providers (P = .013) (Fig 1). Of those providers receiving MM inquiries, 14% (12 of 83) facilitated access 1 or more times (facilitation occurred in Massachusetts and Washington only). Seventy-nine percent (70 of 85) of all inquiries were for relief of nausea and/or vomiting, 52% (44 of 85) were for anorexia, 26% (22 of 85) were for pain, and 24% (20 of 85) were for depression and/or anxiety. Only 8% (23 of 283) reported sometimes or frequently recommending MM to patients, with no significant differences between ETC and n-ETC providers (unadjusted, P = .8; adjusted by state, P = .79).

    FIGURE 1

    Proportions of pediatric oncology providers receiving MM inquiries, making recommendations, and facilitating access, comparing ETC versus n-ETC providers. Of note, only providers who reported receiving inquiries in the previous month were asked whether they facilitated access.


    Overall, 86% (245 of 285) of providers knew that their state had legalized MM; 76% (217 of 285) knew that marijuana is considered a controlled substance by the Food and Drug Administration (FDA) (Table 2). Fewer (59%; 165 of 281) knew that MM is against federal laws, and only 5% (14 of 282) accurately identified state-specific regulations concerning MM access, possession, and cultivation. ETC providers were more likely to know that MM is federally prohibited (75%; 70 of 93) compared with n-ETC providers (51%; 95 of 188; P < .0001). Otherwise, we found no significant differences in knowledge by provider type.

    TABLE 2

    Knowledge of MM Regulations


    Most providers (92%; 240 of 261) were willing to help children with cancer access MM (Table 3). Similarly, most indicated approval of oral MM formulations (89%; 223 of 250) and of using MM to manage symptoms (92%; 236 of 258). Many providers indicated approval of patients smoking MM (57%; 133 of 234) and of using MM as cancer-directed therapy (67%; 158 of 236). A vast majority (93%; 237 of 256) of providers responded favorably to the conduct of clinical trials investigating MM use in children.

    TABLE 3

    Attitudes and Barriers Regarding MM Use

    In univariate models comparing MM attitudes by provider type, ETC providers were significantly less likely than n-ETC providers to report willingness to help pediatric patients access MM (odds ratio [OR] = 0.26; 95% confidence interval [CI] = 0.1–0.65; P = .0041) (Table 3). ETC providers were also less likely to indicate approval of patient MM use by smoking (OR = 0.33; 95% CI = 0.19–0.59; P = .0002), oral formulations (OR = 0.12; 95% CI = 0.05–0.29; P < .0001), as cancer-directed therapy (OR = 0.04; 95% CI = 0.02–0.08; P < .0001), or to manage symptoms (OR = 0.19; 95% CI = 0.07–0.47; P = .0004). Adjusting for provider age, sex, race, and location of practice, ETC providers continued to have significantly lower odds of indicating approval of patients MM use by smoking (adjusted OR = 0.25; 95% CI = 0.11–0.55; P = .0005), oral formulations (adjusted OR = 0.17; 95% CI = 0.05–0.53; P = .0025), as cancer-directed therapy (adjusted OR = 0.04; 95% CI = 0.01–0.09; P < .0001), or to manage symptoms (adjusted OR = 0.16; 95% CI = 0.05–0.52; P = .0026). A sensitivity analysis comparing attitudes by provider discipline yielded comparable results (Supplemental Tables 4 and 5).


    Many providers (63%; 162 of 259) were not concerned about substance abuse in children who receive MM. Likewise, most providers (80%; 208 of 260) were not concerned about being prosecuted by the federal government for facilitating MM access. These attitudes did not significantly differ by provider type (Table 3). Forty-six percent (n = 136) of providers reported their greatest concern to be the absence of standards around MM formulations, potency, and dosing.

    Timing of Use

    When asked about circumstances in which providers would consider MM appropriate for children with cancer, most (89%; 255 of 288) responded that MM would be appropriate near the end of life, in treatment of cancer with primarily palliative intent (89%; 255 of 288), or in treatment of progressive or relapsed cancer (76%; 220 of 288) (Fig 2). Fewer providers felt that MM would be appropriate when treating patients with curative intent (52%; 150 of 288) or in the early stages of cancer treatment (35%; 100 of 288). Only 2% (5 of 288) of providers felt that MM was never appropriate. N-ETC providers were more likely to consider MM appropriate in treating cancer from early stages through treatment with primarily palliative intent (P ≤ .004 for all).

    FIGURE 2

    Proportions of pediatric oncology providers who would consider MM appropriate at various time points in the course of caring for a child with cancer, comparing ETC versus n-ETC providers.


    Our study is among the first to describe MM perspectives of interdisciplinary providers caring for children with cancer. We found that pediatric oncology providers were commonly willing to consider MM use in children with cancer despite limited knowledge of state and federal regulations. MM inquiries arose frequently, with facilitation occurring at a nonnegligible rate regardless of institutional barriers to MM provision. Interestingly, more favorable attitudes were endorsed by n-ETC providers. N-ETC providers were also significantly more open to MM use in early cancer treatment. ETC providers were increasingly amenable to MM use in advanced stages of cancer or near the end of life. The greatest barrier to recommending MM was reported to be the absence of standards regarding formulations, dosing, and potency.

    Several studies over the past decade have ascertained that physicians are apprehensive about adult use of MM.22,24,26,3032 This reluctance appears to be driven by the potential for side effects,33 scant high-quality scientific data,22,34 unclear dosage guidelines,17,26 and a lack of regulatory oversight by the FDA, unlike other therapeutic and supportive care drugs.35 Such concerns are magnified when pediatric clinicians must consider MM use by children and adolescents, particularly because habitual marijuana use is associated with dependence, impaired neurocognitive development, and poor academic achievement in children.12,15,36,37 Recommending MM may thus be fundamentally problematic for physicians who are accustomed to evidence-based practice, as they cannot be assured by empirical data that benefits outweigh possible harm.37

    Few prior studies have explored how interdisciplinary providers, including those without legal authority to recommend MM, perceive its use by patients.24,30,32 Notably, some states enable advance practice providers to certify for MM, and existing literature suggests that the degree to which providers endorse MM may depend more on whether they can legally certify, rather than on particular discipline.24,30 Consistent with these findings, n-ETC providers in this study reported more permissive attitudes toward MM use in children with cancer. Specifically, n-ETC providers were more likely to convey approval of pediatric patients using MM in smoked or oral formulations, as cancer-directed therapy, or for symptom management, despite scarce evidence to corroborate this use.35,37 To place these findings in context, n-ETC provider attitudes observed in this study echo those of the general public,27,38 whereas ETC provider attitudes align more closely with physician-based data.22,31,33 We surmise that ETC providers, whose licensure and clinical reputation could theoretically be jeopardized by MM recommendation in the face of federal prohibition,20,39 may be less willing to endorse MM use in children than providers who face fewer legal ramifications.40 Additionally, responses may have been influenced by institutional policies at study sites that discourage ETC providers from actively recommending or procuring MM for patients.30

    The paradigm shifts when providers consider patients who are seriously ill. Half of interdisciplinary hospice providers in a previous study reported feeling comfortable recommending MM to adult patients.24 Similarly, we identified that pediatric providers found MM use for children with advanced cancer or near the end of life to be more permissible than in earlier stages of cancer treatment. This sentiment is reflected in a current AAP position statement, which sanctions the use of MM for “children with life-limiting or seriously debilitating conditions.”12 Although the AAP opposes all other pediatric marijuana use, it recognizes that there may be circumstances in which MM could have benefit. Accordingly, in our study, only 2% of providers stated that MM was never appropriate for a child with cancer. Emerging research indicates that cannabidiol, a nonpsychoactive component of marijuana, can be effective for children with refractory seizures and has a tolerable safety profile.41 Randomized clinical trials using such MM formulations for supportive care in children with cancer are needed to better understand the therapeutic potential of this agent. It is evident from our study that providers are strongly in support of trials to this end.

    A striking aspect of the current study is the frequency with which pediatric oncology providers receive requests for MM. Nearly one-third of all providers fielded inquiries in the previous month alone, including one-quarter of n-ETC providers; 14% of those who received inquiries facilitated access. This frequency of inquiries is comparable to published adult literature.24,30 Given burgeoning interest in MM, especially in oncology care,3,35 it is critical that providers who are routinely approached for access to MM possess baseline knowledge on regulations, known benefits, and harm. However, providers in this study were largely unaware of state-specific regulations around MM, and approximately one-quarter did not know that MM is federally illegal. These findings confirm the need for greater education, perhaps through state-based initiatives.3032

    One limitation of this study is its cross-sectional design; we are consequently unable to trend attitudes as laws change. In addition, the response rate to this survey (48%) may be deemed low, albeit comparable to existing provider studies of similar design.31,42,43 If there were some degree of response bias, respondents might be expected to have more knowledge than nonresponders, and yet, respondents to this survey demonstrated knowledge gaps in several domains. Although this sample is not nationally representative, we reveal confluent attitudes and practices of providers despite differential approaches to legal MM access. In a scenario in which uniform practice guidelines are needed amid variable institutional and state policies, the attitudes of frontline, interdisciplinary providers caring for children with serious illnesses are therefore critically important to report.


    This study identifies that interdisciplinary pediatric oncology providers are overall open to considering MM in children, but ETC providers are more circumspect in its use as a therapeutic or supportive care agent. We reveal high rates of MM inquiries and a need for increased provider education. Future researchers should explore patient and family perspectives on MM, considering substantial interest in and increasing availability of MM. Additionally, this study calls for rigorously designed clinical trials in which researchers investigate the use of MM in children with cancer. Although marijuana policy has, to date, outpaced scientific discovery, the US Drug Enforcement Administration recently agreed to enable broader access to marijuana for investigators conducting medical research.44 This decree holds promise that a stronger evidence base may soon be available to help providers, patients, and families make informed decisions about using MM in cancer care.


    We acknowledge the Pediatric Palliative Care Research Network for its members’ contributions to developing the survey instrument. We also thank Madhumitha Sridharan for assistance with REDCap survey implementation.


      • Accepted October 23, 2017.
    • Address correspondence to Prasanna Ananth, MD, MPH, Section of Pediatric Hematology/Oncology, Department of Pediatrics, Yale School of Medicine, 330 Cedar Street, LMP 2082C, New Haven, CT 06520-8064. E-mail: prasanna.ananth{at}
    • FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

    • FUNDING: Dr Ananth was supported by the David A. Abraham Fund at the Dana-Farber Cancer Institute and the Ruth L. Kirschstein National Research Service Award (National Institutes of Health 5-T32-CA-136432-05). Dr Michelson was supported by grants from the Patient-Centered Outcomes Research Institute (grant 1304-6449) and the American Cancer Society. Dr Rosenberg was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health (award KL2TR000421). Funded by the National Institutes of Health (NIH).

    • POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.