Quality of Care in the Last Two Years of Life for Children With Complex Chronic Conditions.

CONTEXT: Limited data exists about care received by children with complex chronic conditions (CCCs) in the final years of their disease and end-of-life (EOL). OBJECTIVE: To examine hospital performance on EOL quality measures and to describe healthcare services during the last two years of life for children with CCCs who died in-hospital. METHODS: Retrospective automated electronic health record review of children with ≥1 CCC ICD-10 diagnosis code, who died inpatient between October 2020 and March 2023 at a single quaternary U.S. children's hospital. Quality was assessed based on performance on 15 measures across five domains: healthcare utilization, interprofessional supports, medical intensity, symptom management, and communication. Quality EOL care and healthcare services in the last two years of life were determined overall by age group and per patient. Descriptive statistics were used to evaluate demographic differences by age. RESULTS: 266 children with CCCs died in the study timeframe; 45% were infants (n = 120), 52% (n = 137) were male, 42% (n = 113) were white, 64% (n = 170) were non-Hispanic, and 59% (n=156) had public insurance. Children had a median of three CCCs (IQR 2.4; range 1-8). On average, children met 69% (SD 13%) of EOL quality measures for which they were eligible. In the two years prior to death, 98% (n = 261) had an ICU admission, 75% (n = 200) had a procedure requiring sedation, and 29% (n = 79) had received cardiopulmonary resuscitation. 86% (n = 229) died in the ICU. CONCLUSION: In this study, children with CCCs met 69% of quality measures and received high-intensity healthcare in the last two years of life.

Parent-Clinician Communication and Prolonged Grief in Parents Whose Child Died From Cancer.

CONTEXT: Parent-clinician communication is essential for high-quality end-of-life (EOL) care in children with cancer. However, it is unknown how parent-clinician communication affects parents' experience in the first two years after their child's death. OBJECTIVES: To examine the association between communication and prolonged grief among parents whose child died from cancer and to explore the mediation effect of preparation for EOL care. METHODS: We analyzed data from a cross-sectional survey of parents of children who died from cancer in the prior 6-24 months. We used multiple linear regression to examine the association between communication variables and prolonged grief symptoms. We also examined how preparation for EOL mediates these associations. RESULTS: Across N = 124 parents, the mean age was 46 years, 82% were White, and 64% were mothers. The average PG-13 sum score was 32.7 ± 10.6 (range 11-55, with higher scores indicating greater symptom severity). Most parents reported "very good/excellent" communication with clinicians (80%), adequate prognostic information (64%), and high levels of trust (90%). Nearly 39% of parents reported feeling "not at all prepared" for their child's EOL. Compared to parents who wanted more prognostic information, parents who perceived prognostic information to be adequate had significantly lower PG-13 sum scores (36.4 ± 10.8 vs. 30.5 ± 10.1, F = 9.26, P = 0.003). Preparation for EOL fully mediated this association. CONCLUSION: Early bereaved parents report severe prolonged grief symptoms. Interventions focused on providing adequate prognostic information and improving preparation for EOL may mitigate parental prolonged grief symptoms in the first two years of their bereavement.

Defining the Denominator for Measuring Quality of End-of-Life Care in Children with Cancer: Results of a Nominal Group Technique.

OBJECTIVE: To determine which groups of children with cancer for whom to apply the newly developed quality measures (QMs) for end-of-life (EOL) care. STUDY DESIGN: In a series of nominal groups, panelists answered the question: "Which children, diagnoses, conditions, or prognoses should be included when examining the quality of EOL care for children with cancer?" In each group, individual panelists proposed answers to the question. After collating individual responses, each panelist ranked their 5 top answers and points were assigned (5 pts for the best answer, 4 pts the second best, etc.). A team of pediatric oncology and palliative care clinician-scientists developed and applied a coding structure for responses and associated themes and subthemes for responses. RESULTS: We conducted 5 nominal groups with a total of 44 participants. Most participants identified as female (88%) and non-Hispanic White (86%). Seventy-nine percent were clinicians, mainly in pediatric palliative care, pediatric oncology, or hospice; 40% were researchers and 12% were bereaved parents. Responses fell into 5 themes: (1) poor prognosis cancer; (2) specific treatment scenarios; (3) certain populations; (4) certain symptoms; and (5) specific utilization scenarios. Poor prognosis cancer and specific treatment scenarios received the most points (320 pts [49%] and 147 pts [23%], respectively). CONCLUSIONS: Participants developed a framework to identify which children should be included in EOL QMs for children with cancer. The deliberate identification of the denominator for pediatric QMs serves as a potent tool for enhancing quality, conducting research, and developing clinical programs.

Parent Priorities in End-of-Life Care for Children With Cancer.

Importance: Robust quality measures to benchmark end-of-life care for children with cancer do not currently exist; 28 candidate patient-centered quality measures were previously developed. Objective: To prioritize quality measures among parents who lost a child to cancer. Design, Setting, and Participants: This survey study was conducted using an electronic, cross-sectional discrete choice experiment (DCE) with maximum difference scaling from January to June 2021 in the US. In each of 21 questions in the DCE, participants were presented with a set of 4 quality measures and were asked to select the most and least important measures within each set. All 28 quality measures were presented an equal number of times in different permutations. In the volunteer sample, 69 eligible bereaved parents enrolled in the study; 61 parents completed the DCE (participation rate, 88.4%). Main Outcomes and Measures: Using choices participants made, a hierarchical bayesian multinomial logistic regression was fit to derive mean importance scores with 95% credible intervals (95% Crs) for each quality measure, representing the overall probability of a quality measure being selected as most important. Importance scores were rescaled proportionally from 0 to 100, with the sum of scores for all quality measures adding up to 100. This enabled interpretation of scores as the relative importance of quality measures. Results: Participants included 61 bereaved parents (median [range] age, 48 [24-74] years; 55 individuals self-identified as women [90.2%]; 1 American Indian or Alaska Native [1.6%], 1 Asian [1.6%], 2 Black or African American [3.3%], 1 Native Hawaiian or Pacific Islander, and 58 White [91.8%]; 58 not Hispanic or Latinx [95.1%]). Highest-priority quality measures by mean importance score included having a child's symptoms treated well (9.25 [95% Cr, 9.06-9.45]), feeling that a child's needs were heard by the health care team (8.39 [95% Cr, 8.05-8.73]), and having a goal-concordant end-of-life experience (7.45 [95% Cr, 6.84-8.05]). Lowest-priority quality measures included avoiding chemotherapy (0.33 [95% Cr, 0.21-0.45]), provision of psychosocial support for parents (1.01 [95% Cr, 0.57-1.45]), and avoiding the intensive care unit (1.09 [95% Cr, 0.74-1.43]). Rank-ordering measures by mean importance revealed that symptom management was 9 times more important to parents than psychosocial support for themselves. Conclusions and Relevance: This study found that bereaved parents prioritized end-of-life quality measures focused on symptom management and goal-concordant care while characterizing quality measures assessing their own psychosocial support and their child's hospital resource use as substantially less important. These findings suggest that future research should explore innovative strategies to measure care attributes that matter most to families of children with advanced cancer.

Measuring Pediatric Palliative Care Quality: Challenges and Opportunities.

Pediatric palliative care (PPC) programs vary widely in structure, staffing, funding, and patient census, resulting in inconsistency in service provision. Improving the quality of palliative care for children living with serious illness and their families requires measuring care quality, ensuring that quality measurement is embedded into day-to-day clinical practice, and aligning quality measurement with healthcare policy priorities. Yet, numerous challenges exist in measuring PPC quality. This paper provides an overview of PPC quality measurement, including challenges, current initiatives, and future opportunities. While important strides toward addressing quality measurement challenges in PPC have been made, including ongoing quality measurement initiatives like the Cambia Metrics Project, the PPC What Matters Most study, and collaborative learning networks, more work remains. Providing high-quality PPC to all children and families will require a multi-pronged approach. In this paper, we suggest several strategies for advancing high-quality PPC, which includes 1) considering how and by whom success is defined, 2) evaluating, adapting, and developing PPC measures, including those that address care disparities within PPC for historically marginalized and excluded communities, 3) improving the infrastructure with which to routinely and prospectively measure, monitor, and report clinical and administrative quality measures, 4) increasing endorsement of PPC quality measures by prominent quality organizations to facilitate accountability and possible reimbursement, and 5) integrating PPC-specific quality measures into the administrative, funding, and policy landscape of pediatric healthcare.

Development of Primary Palliative Care End-of-Life Quality Measures: A Modified Delphi Process.

BACKGROUND AND OBJECTIVES: Children with complex chronic conditions (CCCs) and their families deserve high-quality pediatric palliative care (PPC) throughout their illness trajectory, including at end of life (EOL). Standard EOL quality measures (QM) have only recently been proposed, require surveys and/or manual chart review, and focus on children with cancer. Therefore, we aimed to develop expert-endorsed, hospital-based, primary PPC quality measures for EOL care for all children with CCCs that could be automatically abstracted from the electronic health record (EHR). METHODS: We followed a modified Delphi approach for expert opinion gathering, including: (1) a comprehensive literature review of existing adult and pediatric measures (>200 measures); (2) formation of a multidisciplinary expert panel (n = 9); (3) development of a list of candidate measures (20 measures); (4) national survey to assess each QM's importance and abstraction feasibility and propose new measures (respondents = 95); and (5) final expert panel endorsement. RESULTS: Seventeen EHR-abstractable QM were endorsed in 5 domains: (1) health care utilization: 4 measures (eg, <2 emergency department visits in the last 30 days of life); (2) interprofessional services: 4 measures (eg, PPC in the last 30 days of life); (3) medical intensity: 5 measures (eg, death outside the ICU); (4) symptom management: 2 measures (eg, documented pain score within 24 hours of admission); and (5) communication: 2 measures (eg, code status documentation). CONCLUSIONS: This study developed a list of EHR-abstractable, hospital-based primary PPC EOL QM, providing a foundation for quality improvement initiatives and further measure development in the future.

End-of-life care quality for children with cancer who receive palliative care.

BACKGROUND: We previously developed stakeholder-informed quality measures to assess end-of-life care quality for children with cancer. We sought to implement a subset of these quality measures in the multi-center pediatric palliative care (PPC) database. PROCEDURES: We utilized the Shared Data and Research database to evaluate the proportion of childhood cancer decedents from 2017-2021 who, in the last 30 days of life, avoided chemotherapy, mechanical ventilation, intensive care unit admissions, and > 1 hospital admission; were enrolled in hospice services, and reported ≤ 2 highly distressing symptoms. We then explored patient factors associated with the attainment of quality benchmarks. RESULTS: Across 79 decedents, 82% met ≥ 4 quality benchmarks. Most (76%) reported > 2 highly distressing symptoms; 17% were enrolled in hospice. In univariable analyses, patients with an annual household income ≤$50,000 had lower odds of hospice enrollment and avoidance of mechanical ventilation or intensive care unit admissions near end of life (odds ratio [OR] 0.10 [95% confidence interval (C.I.) 0.01, 0.86], p = 0.04; OR 0.13 [0.02, 0.64], p = 0.01; OR 0.36 [0.13, 0.98], p = 0.04, respectively). In multivariable analyses, patients with an income ≤$50,000 remained less likely to enroll in hospice, after adjusting for cancer type (OR 0.10 [0.01, 0.87]; p = 0.04). CONCLUSIONS: Childhood cancer decedents who received PPC met a large proportion of quality measures near the end of their life. Yet, many reported highly distressing symptoms. Moreover, patients with lower household incomes appeared less likely to enroll in hospice and more likely to receive intensive hospital services near the end of life. This study identifies opportunities for palliative oncology quality improvement.

Survey of Pediatric Palliative Care Quality Improvement Training, Activities, and Barriers.

CONTEXT: Children with serious illness deserve high-quality pediatric palliative care (PPC). With expansion of PPC provision, it is important to understand the quality improvement (QI) activities of PPC clinicians and programs. OBJECTIVES: To describe the 1) background, education/training, and activities in QI and 2) perceived barriers to QI efforts among PPC clinicians and programs nationally. METHODS: An electronic survey was sent to members of the Pediatric Palliative Improvement Network, the National Coalition for Hospice and Palliative Care Pediatric Task Force, and the PPC Research Network as part of a study to develop hospital-based, primary PPC quality measures. Surveys queried participants' background, education/training, individual/team QI efforts, and barriers to QI work. Results were summarized descriptively. RESULTS: Of the 95 respondents; most were female (84 [88%]) and/or white (84 [88%]). The majority (57 [54%]) were physicians, although participants represented a variety of clinical disciplines, researchers (10 [9%]), and administrators (6 [6%]). One-quarter (25 [26%]) reported having <10 hours total of training in QI, yet two-third (63 [66%]) participated in QI work. About one-third (35 [37%]) reported that their program had no dedicated staff for QI activities, yet over half (56 [59%]) of participants reported that their team participated in QI work. Participants reported that lack of personnel/time, standardized measures/tools, education/training/mentoring in QI, systems to promote QI work, and financing/grants were barriers. CONCLUSION: Over half of PPC participants in this study reported involvement in QI activities despite limited staffing/time, QI training, and standardized measures, which presents challenges to this work.

Perspectives of pediatric oncologists and palliative care physicians on the therapeutic use of cannabis in children with cancer.

BACKGROUND: Children with cancer are increasingly using cannabis therapeutically. AIM: The purpose of this study was to determine the perspectives and practices of pediatric oncologists and palliative care physicians regarding the use of cannabis for medical purposes among children with cancer. METHODS: A self-administered, voluntary, cross-sectional, deidentified online survey was sent to all pediatric oncologists and palliative care physicians in Canada between June and August 2020. Survey domains included education, knowledge, and concerns about cannabis, views on its effectiveness, and the importance of cannabis-related research. Data were analyzed using descriptive statistics. RESULTS: In total, 122/259 (47.1%) physicians completed the survey. Although 62.2% of the physicians completed some form of training about medical cannabis, nearly all (95.8%) desired to know more about the dosing, side effects, and safety of cannabis. Physicians identified a potential role of cannabis in the management of nausea and vomiting (85.7%), chronic pain (72.3%), cachexia/poor appetite (67.2%), and anxiety or depression (42.9%). Only four (0.3%) physicians recognized cannabis to be potentially useful as an anticancer agent. Nearly all physicians reported that cannabis-related research for symptom relief is essential (91.5%) in pediatric oncology, whereas 51.7% expressed that future studies are necessary to determine the anticancer effects of cannabis. CONCLUSIONS: Our findings indicate that most pediatric oncologists and palliative care physicians recognize a potential role for cannabis in symptom control in children with cancer. Well-conducted studies are required to create evidence for cannabis use and promote shared decision making with pediatric oncology patients and their caregivers.

Refining Patient-Centered Measures of End-of-Life Care Quality for Children With Cancer.

PURPOSE: There are no existing quality measures (QMs) to optimize end-of-life care for children with cancer. Previously, we developed a set of 26 candidate QMs. Our primary objective in this study was to achieve stakeholder consensus on priority measures. METHODS: We conducted an iterative, cross-sectional electronic survey, using a modified Delphi method to build consensus among clinician and family stakeholders. In each of the two rounds of surveys, stakeholders were asked to rate QMs on a 9-point Likert scale, on the basis of perceived importance. Health care professionals were additionally asked to rate measures on perceived feasibility. After each round, we computed median scores on importance and feasibility of measurement, retaining QMs with median importance scores ≥ 8. RESULTS: Twenty-five participants completed both rounds of the survey. In round 1, participants were asked to rate 26 QMs; nine QMs, including QMs pertaining to health care use, were removed because of median importance scores < 8. Two new measures were proposed for consideration in round 2, on the basis of participant feedback. Following round 2, 17 QMs were ultimately retained. QMs related to symptom screening and palliative care consultation were rated highly in importance and feasibility. QMs related to communication were rated highly important, yet less feasible. Measuring whether a patient's needs were heard by their health care team was rated among the least feasible. CONCLUSION: Childhood cancer stakeholders prioritized QMs pertaining to patient-reported outcomes, deeming measures of health care resource use less important. Future research should seek to develop novel tools for quality assessment to enhance feasibility of implementing priority measures.

Sociodemographic and hospital-based predictors of intense end-of-life care among children, adolescents, and young adults with hematologic malignancies.

BACKGROUND: Children, adolescents, and young adults with hematologic malignancies tend to receive high-intensity end-of-life care (HI-EOLC), but sociodemographic and hospital-based predictors of HI-EOLC remain unclear. METHODS: The authors conducted a population-based, retrospective cohort study with the Premier Healthcare Database. They identified individuals with hematologic malignancies who were 0 to 39 years old at death and died between 2010 and 2017. HI-EOLC was defined as experiencing 2 or more of the following: cardiopulmonary resuscitation, intravenous chemotherapy, hemodialysis, mechanical ventilation, tracheostomy placement, or an emergency department visit within the last 30 days of life and death in the intensive care unit. Multivariable logistic regression models were constructed to identify patient sociodemographic and hospital characteristics associated with HI-EOLC. RESULTS: Among 1454 decedents, more than half (55%) experienced HI-EOLC. In multivariable models, patients treated in medium (adjusted odds ratio [aOR], 1.63; 95% confidence interval [CI], 1.07-2.50) or large hospitals (aOR, 2.21; 95% CI, 1.45-3.39), insured by Medicaid (aOR, 1.40 ; 95% CI, 1.09-2.06), or receiving cancer-directed treatment in the Northeast (aOR, 1.50; 95% CI, 1.05-2.15) were more likely to receive HI-EOLC. CONCLUSIONS: A majority of children, adolescents, and young adults with hematologic malignancies experienced HI-EOLC, and the likelihood of HI-EOLC was influenced by the hospital size, type of insurance, and geographic region. Further research is needed to determine how to mitigate these risks.

A Stakeholder-Driven Qualitative Study to Define High Quality End-of-Life Care for Children With Cancer.

CONTEXT: Among adults with cancer, measures for high quality end-of-life care (EOLC) include avoidance of hospitalizations near end of life. For children with cancer, no measures exist to evaluate or improve EOLC, and adult quality measures may not apply. OBJECTIVE: We engaged key stakeholders to explore EOLC priorities for children with cancer and their families, and to examine relevance of existing adult EOLC quality measures for children with cancer. METHODS: In a multicenter qualitative study, we conducted interviews and focus groups with: adolescents and young adults (AYAs) with advanced cancer, parents of children with advanced cancer, bereaved parents, and interdisciplinary healthcare professionals. We transcribed, coded, and employed thematic analysis to summarize findings. RESULTS: We enrolled 54 stakeholders (25 parents [including 12 bereaved parents], 10 AYAs, and 19 healthcare professionals). Participants uniformly prioritized direct communication with children about preferences and prognosis, interdisciplinary care, symptom management, and honoring family preference for location of death. Many participants valued access to the emergency department or hospital for symptom management or supportive care, which diverges from measures for high quality EOLC in adults. Most wished to avoid mechanical ventilation and cardiopulmonary resuscitation. Notably, participants generally valued hospice; however, few understood hospice care or had utilized its services. CONCLUSION: Childhood cancer stakeholders define high quality EOLC primarily through person-centered measures, characterizing half of existing adult-focused measures as limited in relevance to children. Future research should focus on developing techniques for person-centered quality measurement to capture attributes of greatest importance to children with cancer and their families.

Patterns of medication use at end of life by pediatric inpatients with cancer.

OBJECTIVE: To describe medication utilization patterns by pediatric inpatients with cancer during their last week of life. METHODS: This retrospective study used data from the Vizient Clinical Database/Resource Manager, a national compilation of clinical and resource use data from over 100 academic medical centers and affiliates. Patients (0-21 years) with malignancy who died during hospitalization (2010-2017) were included (N = 1659). Medications were categorized as opioid, benzodiazepine, gastrointestinal related, chemotherapy, anti-infectives, or vasopressors. Exposure to each group was ascertained for all patients at 1 week and 1 day prior to death. Factors associated with exposure were examined using generalized estimating equations, and summarized using adjusted odds ratios (aORs). RESULTS: Over the last week of life, there was increased use of opioids (76% to 82%, aOR = 1.55, P < .001) and benzodiazepines (53% to 66%, aOR = 1.36, P = .02), while gastrointestinal-related medication use decreased (92% to 89%, aOR = 0.69, P = .001). Patients had decreased exposure to chemotherapy (10% to 5%, aOR = 0.46, P < .001) and anti-infectives (82% to 73%, aOR = 0.41, P = .002). Vasopressor use increased as death approached (15% to 28%, aOR = 1.67, P = .04). Factors significantly associated with exposure varied with medication category, and included age, race, length of stay, malignancy type, death in the intensive care unit, history of hematopoietic stem cell transplant, and do-not-resuscitate status. CONCLUSION: During the week preceding death, administration of symptom management medications increased for children with cancer, but use was not universal. Potentially life-sustaining medications were often continued. Variability in utilization suggests differences in provider/family decision making that warrant further study to develop an evidence-based approach to end-of-life care.

Parent and patient perceptions of medical marijuana in the childhood cancer context.

BACKGROUND: Medical marijuana (MM) is legal in 34 US jurisdictions. Yet, little is known about patient and parent perceptions of MM in pediatric cancer care. We examined attitudes, beliefs, and experiences regarding MM among parents of children with cancer and adolescent and young adult (AYA) patients, to help frame future research initiatives. PROCEDURE: In this qualitative study, we conducted semi-structured, one-on-one interviews with parents and AYAs at a comprehensive cancer center. Interviews were audio-recorded, transcribed, and coded using both descriptive and inductive coding approaches. We used content and framework analysis to identify key themes. RESULTS: Fifteen parents and 15 AYAs enrolled. Participants were generally receptive to MM use, concurrently weighing benefits and risks. Participants most often endorsed MM use for relief of nausea, anorexia, and pain. Simultaneously, participants identified concerns about MM, including potential physiologic and psychological effects on children and lack of research. However, concerns were frequently minimized, relative to chemotherapy or supportive care medications with perceived greater side effect profiles. Many participants expressed uncertainty regarding legal access, citing complex processes to obtain MM. Few participants had discussed MM with their oncologist, instead seeking guidance from the internet, family, or peers. Importantly, we elicited several misconceptions regarding MM, including its utility as cancer-directed therapy. CONCLUSION: Patients and families are receptive to using MM, motivated by potential for symptom relief and cancer-directed effects. Yet, lack of empiric evidence is a barrier, underscoring the need for robust clinical trial data to support MM recommendations and use.

Parent Perceptions of Team-Delivered Care for Children With Advanced Cancer: A Report From the PediQUEST Study.

CONTEXT: Childhood cancer care is delivered by interprofessional health care teams; however, little is known about how parents perceive overall team-delivered care (TDC). OBJECTIVES: We sought to describe parent perceptions of TDC and associated factors, including care rendered by individual clinicians, teamwork, information consistency, and patient and parent characteristics. METHODS: Cross-sectional surveys were distributed to parents of 104 children with recurrent/refractory cancer enrolled in a multisite symptom management trial. The primary outcome, TDC, was parent report of care quality delivered by the child's care team during the preceding three months. Likert-scaled items (excellent/very good/good/fair/poor) queried care quality delivered by individual clinicians, perceived teamwork, and other factors. Factors associated with parent perceptions of excellent TDC were identified using Fisher's exact test. RESULTS: Eighty-six parents (83%) responded. During the preceding three months, 63% (n = 54) of parents reported excellent TDC. However, only 47% (n = 40) described their care team's teamwork as excellent. Approximately one-quarter (24%) described care rendered by their child's oncologist as less-than-excellent. Among parents who reported psychosocial clinician involvement (71%; n = 60), only 43% described this care as excellent. Individually, excellent care from each clinician type (oncologist, psychosocial clinician, and primary nurse) was associated with excellent TDC (all P ≤ 0.001; no correction for multiple comparisons). CONCLUSION: Among parents of children with advanced cancer, more than one-third report less-than-excellent TDC. In addition, less than half report excellent teamwork, and ratings of care rendered by individual clinicians are highly variable. Findings suggest that interventions are needed to enhance interprofessional teamwork in the care of children with advanced cancer.

Influence of early phase clinical trial enrollment on patterns of end-of-life care for children with advanced cancer.

We conducted a retrospective cohort study of 125 pediatric oncology patients who died in 2010-2014 to explore how healthcare utilization, pediatric palliative care (PPC) receipt, and end-of-life care (EOLC) differed between patients enrolled in early phase clinical trials (EP) and those not enrolled (NEP). Baseline characteristics and healthcare utilization did not significantly differ between groups. EP patients received PPC consultation closer to death than NEP patients (median days before death = 58 [interquartile range = 16-84] vs. 85 [32-173]; P = 0.04). Our findings suggest that early phase trial enrollment does not substantially alter EOLC for children with advanced cancer but may contribute to later PPC engagement. Future studies should definitively assess the relationship between trial enrollment and PPC timing.

Provider Perspectives on Use of Medical Marijuana in Children With Cancer.

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.