Multilevel Barriers and Facilitators of Communication in Pediatric Oncology: A Systematic Review

Abstract

Multiple factors can facilitate or impede the fulfillment of communication functions in pediatric cancer. In this systematic review, we evaluated 109 studies from the preceding 20 years that presented qualitative or quantitative evidence of barriers or facilitators to communication in pediatric cancer. Using a multilevel framework developed in our prior study, we then analyzed and categorized the levels of barriers and facilitators identified in included studies. The vast majority of studies focused on individual-level barriers, rather than team, organization/system, collaborating hospital, community, or policy-level barriers. Future studies should explore the full range of factors that affect communication.

Introduction

Communication is a central component of family-centered care at all phases of illness in pediatric cancer.^{1, 2} We previously identified 8 core functions of communication, including building relationships, exchanging information, managing uncertainty, enabling family self-management, making decisions, responding to emotions, providing validation, and supporting hope.³ Fulfillment of these functions supports parental hopefulness,⁴ peace of mind,^{5, 6} trust,⁷ and feeling acknowledged⁸ and comforted.⁹ When communication fails, families can experience inaccurate prognostic understanding,^10–12 emotional distress,¹³ decisional regret,¹⁴ loss of trust,¹⁵ decreased hope,¹³ and even medical harm.¹³

Communication is a complex, multidimensional process that can be influenced by factors at multiple levels. Additionally, communication in pediatric oncology is particularly challenging because of the wide developmental spectrum of patients who range from infancy to young adulthood. Understanding which factors facilitate or impede functions can help us to promote high-quality communication in pediatric oncology and identify targets for communication interventions. In previous work, we developed a multilevel framework of factors that affect the fulfillment of communication functions, including individual, team, organization/system, collaborating hospital, community, and policy-level factors.¹⁶ This multilevel model was adapted from the field of organizational psychology¹⁷ and tailored to pediatric communication based on key stakeholder input. The field of organizational psychology maintains that multilevel and cross-level interactions across individual, group, organizational, and environmental factors can influence individuals’ behaviors within an organization.¹⁷ However, it is unclear whether the pediatric communication literature has evaluated barriers and facilitators at these multiple levels. This field lacks a systematic review that synthesizes and categorizes the multilevel facilitators and barriers that affect the full breadth of communication functions. Instead, recent review articles have explored barriers and facilitators related to narrow aspects of communication, such as clinical trial enrollment,^{18, 19} fertility preservation,²⁰ cultural differences,^{21, 22} incurable disease,²³ and shared decision making.²⁴ A broader review of barriers and facilitators of communication will inform clinical practice and highlight areas that have been insufficiently studied. In this systematic review, we evaluated all studies of barriers and facilitators of clinical communication between patients/families and clinicians in pediatric oncology published in the prior 20 years. We then applied the previously-developed multilevel framework¹⁶ to categorize each barrier or facilitator and identify gaps in the existing literature.

Methods

We conducted a systematic review following the guidelines provided by Khan et al.²⁵ We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for reporting results of reviews insofar as data permitted,²⁶ posing the question: “What factors facilitate or impede the fulfillment of communication functions in pediatric oncology?” Our review protocol was not registered, but is available upon request. Additionally, all data and forms from this study are available upon request. (See Appendix 1 for PRISMA Checklist)

Inclusion and Exclusion Criteria

This review was inclusive of all research articles presenting original data published between January 1, 2000 and May 11, 2020. For inclusion, outcomes of the study needed to be related to “communication” in general or specific communication functions in pediatric oncology that we identified in past research: exchanging information, building relationships, enabling family self-management, making decisions, managing uncertainty, responding to emotions, providing validation, and supporting hope.³ Furthermore, the studies had to report on factors associated with facilitating or impeding communication functions, include a focus on pediatric patients with cancer, and be published in English. Additionally, we included studies of any sample size that enrolled children 18 years or younger with cancer, parents of children, or clinicians of children. If a study included children and young adults, we also included these studies. We excluded intervention studies because their focus was not on examining barriers and facilitators, but on modifying these factors. We also excluded gray literature, commentaries, and review articles. We defined barriers and facilitators as factors that impeded or supported the fulfillment of at least one communication function or communication in general. For quantitative studies, we included studies in which the barrier or facilitator was associated with a communication outcome. For qualitative studies, we included studies in which major themes identified communication barriers or facilitators. We combined facilitators and barriers into a single category because they often represent a similar underlying concept. For example, one study might identify low-quality teamwork as a barrier and another study might identify high-quality teamwork as a facilitator. Yet, both studies are highlighting the same underlying concept of teamwork.

Search Strategy

A medical librarian (LHY) searched the literature for records including a combination of four concepts: oncology, communication, children, and physicians. The librarian created search strategies using a combination of keywords and controlled vocabulary in Embase.com 1947-, Ovid Medline 1946-, Scopus 1823-, Cochrane Central Register of Controlled Trials (CENTRAL), Cumulative Index of Nursing and Allied Health Literature (CINAHL Plus) 1937 -, APA PsycINFO 1800s -, and Clinicaltrials.gov1 1997-. Results were limited by date to 2000 – 2020 and conference abstracts filtered out where possible. All search strategies were completed May 11, 2020 and a total of 9,573 results were found. 4,427 duplicate records were deleted after using the de-duplication processes described in “De-duplication of database search results for systematic reviews in EndNote,”²⁷ resulting in a total of 5,146 unique citations included in the project library. Fully reproducible search strategies for each database can be found in Appendix 2.

Data Classification

We utilized Rayyan web-based systematic review software for initial screening of studies.²⁸ Three authors (BAS, ABF, and ALA) collaboratively screened studies for inclusion using a staged approach. First, the authors screened titles and abstracts, then reviewed full-text articles. At least 2 authors reviewed every article that was included in this final collection of studies. We indicated reasons for exclusions based on an ordered hierarchy, evaluating characteristics of articles in the following order: duplicates, foreign language, topic, population, publication type, study design, outcomes. We applied this same approach to each round of screening. After all exclusions, one author (KH) abstracted characteristics of each study into a spreadsheet. Another author (BAS) abstracted the barriers or facilitators identified in each study, using definitions of barriers, facilitators, and communication functions described in the inclusion/exclusion criteria. We further characterized the level of barrier or facilitator using a combined inductive and deductive coding approach. We used an evidence-informed framework of multilevel barriers and facilitators of communication, including individual, team, organization/system, collaborating hospital, community, and policy-level factors.¹⁶ We utilized content analysis²⁹ to analyze these data, using the multilevel model of barriers as an a priori framework for coding, while also remaining open to novel categories of barriers that emerged from the collection of studies. Through our analysis, we identified clinical factors related to characteristics of the disease as an additional factor. Two authors (BAS and ABF) subsequently consensus coded all included studies by iteratively coding articles independently, reviewing the other’s codes, and resolving disagreements through discussion.

Quality Assessment and Risk of Bias

To assess for quality and risk of bias in qualitative studies, we characterized fidelity to consolidated criteria for reporting qualitative research (COREQ) guidelines,³⁰ as well as the Critical Appraisal Skills Program (CASP) Qualitative Studies Checklist.³¹ For quantitative studies, we used the Appraisal tool for Cross-Sectional Studies (AXIS) tool.³² The lead author (BAS) trained KH to utilize these assessment tools and they collaboratively evaluated the first 30 studies. KH assessed the remainder of the studies independently.

Results

Our initial search yielded 5,146 articles, supplemented by 1 article we identified through manual checks of references lists for additional articles. (Figure 1) We excluded 4626 articles while reviewing titles/abstracts, and we excluded another 412 articles after reviewing the full text manuscripts. After exclusions, 109 articles were included in this analysis. Most studies were qualitative, employing interviews (61/109) or focus groups (13/109), and most studies (71/109) were performed at >1 site. Most studies were performed in either the US (43%) or Europe (28%). Approximately half of studies included perspectives of parents (59/109) and half included the perspectives of clinicians (55/109). A minority of studies included patients themselves (27/109). (Table 1) See Appendix 3 for detailed information about each study, as well as quality assessments. Barriers and facilitators at each level are described below, and in Figure 2 and 3. For cross-sectional survey studies, AXIS scores averaged 17/20, demonstrating moderate to high study quality. For qualitative studies, we identified several reporting deficiencies based on COREQ guidelines, with an average of 20/32 accomplished checklist items.

Figure 1.

PRISMA Flow Chart

We excluded 4626 articles while reviewing titles/abstracts, and we excluded another 412 articles after reviewing the full text manuscripts.

Table 1.

Characteristics of Studies

Variable	% Yes (n)	References
Country
United States	43% (47)	4, 10, 16, 34, 37, 38, 50, 51, 53, 55, 56, 58–60, 70–73, 77–79, 83, 89, 94–96, 98, 101, 102, 105, 106, 109, 110, 113–120, 124, 126, 128, 131–133
United States + Other Country	2% (2)	54, 66
Canada	6% (6)	45, 46, 64, 75, 103, 107
Europe	28% (31)	9, 36, 40–42, 44, 47, 49, 52, 65, 68, 69, 74, 76, 82, 84, 85, 87, 88, 90, 92, 97, 99, 104, 108, 125, 127, 130, 134, 147, 148
Australia & New Zealand	3% (3)	43, 48, 91
Asia	6% (7)	61, 62, 86, 112, 121–123
Middle East	7% (8)	7, 39, 67, 81, 93, 100, 111, 129
Africa	2% (2)	33, 35
Central and South America	3% (3)	57, 63, 80
Number of Sites
1	35% (38)	4, 7, 33–37, 42, 43, 45, 53, 55, 56, 63, 65, 67, 70, 77, 78, 80, 87, 88, 90, 93, 96, 100, 101, 111, 112, 118, 121–123, 127, 129–131, 133
2	16% (17)	10, 16, 39, 50, 59, 64, 68, 71, 82, 86, 94, 110, 113, 114, 119, 120, 132
3	4% (5)	57, 60, 66, 92, 128
4	3% (4)	38, 61, 103, 126
6	13% (14)	44, 46, 47, 73, 89, 105, 107, 109, 115–117, 124, 125, 147
≥7	14% (15)	9, 48, 51, 69, 76, 83–85, 95, 98, 99, 104, 106, 134, 148
Multiple not specified	15% (16)	40, 41, 49, 52, 54, 58, 62, 72, 74, 75, 79, 81, 91, 97, 102, 108
Number of Participants: Patients
≤50	17% (19)	36–38, 40, 43, 57, 67, 73, 76, 84–87, 97, 99, 102, 129, 131, 148
51–100	2% (2)	52, 105
101–150	3% (3)	122, 130, 133
>150	3% (3)	66, 77, 106
Number of Participants: Parents
≤50	30% (33)	4, 39, 40, 43–46, 57, 63, 67, 74, 80, 82, 84–87, 92–96, 99, 100, 107, 110, 111, 113, 128, 129, 131, 134, 148
51–100	12% (13)	9, 33, 35, 41, 47, 48, 73, 75, 78, 112, 115, 127, 147
101–150	3% (4)	79, 101, 109, 117
>150	8% (9)	7, 10, 34, 59, 114, 118–120, 132
Number of Participants: Clinicians
≤50	33% (36)	9, 40, 42, 47, 51, 53, 55, 57, 59–61, 63–65, 67, 68, 70, 74, 80, 82, 84, 85, 88, 92, 96, 98, 103, 108, 110, 111, 124–126, 131, 134, 148
51–100	8% (9)	16, 49, 50, 52, 56, 69, 81, 83, 91
101–150	4% (5)	62, 71, 72, 89, 123
>150	4% (5)	54, 58, 104, 121, 124
Number of Participants: Other *
≤50	4% (5)	86, 103, 110, 123, 148
Study Design **
Qualitative: Focus Groups	12% (13)	16, 40, 60, 61, 75, 76, 88, 95, 108, 124, 125, 131, 148
Qualitative: Interviews	56% (61)	4, 9, 36–39, 41–48, 51, 55, 57, 61–65, 67, 68, 70, 72–75, 78, 80, 82–88, 92–94, 97–100, 102, 103, 105, 107–109, 111, 113, 115, 116, 122, 127, 128, 134, 147, 148
Quantitative: Cross-Sectional Survey	39% (42)	7, 10, 33–35, 49, 52–54, 56, 58, 59, 66, 69, 71, 77, 79, 81, 89–91, 96, 101, 102, 104, 106, 110, 112–114, 117–124, 126, 130, 132, 133
Mixed Methods	6% (7)	44, 88, 102, 108, 122, 124, 148
Other	3% (3)	44, 50, 122
Barrier & Facilitators to Communication
Individual	90% (100)	4, 7, 9, 10, 16, 33–47, 49–75, 77–123, 125, 128, 130, 133, 147, 148
Team	9% (10)	16, 42, 44, 45, 55, 60, 70, 73, 124, 125
Organization/System	30% (33)	16, 39, 41, 42, 44–46, 48, 49, 51, 52, 54, 55, 57, 61, 64, 69, 73, 75, 76, 78, 79, 83, 96, 98, 99, 103, 106–108, 116, 124, 126
Collaborating Hospital	2% (2)	16, 70
Community	5% (6)	16, 61, 102, 118, 127, 128
Policy	4% (5)	16, 54, 75, 89, 100
Clinical Factors	16% (17)	46, 50, 57, 61, 66, 84, 88, 90, 97, 102, 106, 128, 130–134

^*Other participants included interpreters, research coordinators, ethics committee members, child life specialists, hematology technicians, patient siblings, and healthcare administrators.

^**Study designs were not mutually exclusive.

Figure 2.

Multilevel Barriers and Facilitators of Communication

Figure 3.

Percentage of Studies Identifying Levels of Barriers/Facilitators

Individual-level factors

Individual-level factors were identified in 100 studies, related to individual characteristics of the clinician, parents, or patients, or characteristics of the clinician-patient-family interaction. Clinician characteristics and behaviors that were examined as potential influences on communication included: complexity of language and terminology,^33–38 pacing of information,^39–43 demonstrations of competence,^{42, 44–48} level of comfort or training related to sensitive topics,^{4, 42, 49–62} perceived importance of certain topics,^{51, 61–68} concerns about harms of discussing certain topics, ^69–72 exploration of family’s values and taking them seriously,^{38, 40, 41, 73–77} and demonstrations of empathy and caring.^{7, 38, 39, 78, 79} Some studies identified a tension between a desire to show empathy and a need to create appropriate emotional distance to allow for effective communication and decision-making. ^{57, 64, 70, 80, 81}

For families, the following characteristics and behaviors were identified as potential barriers or facilitators of communication: parental preferences for disclosure,^{57, 61, 67, 74, 82–87} family’s expectations regarding disease course,^{10, 54, 61, 88–91} family’s emotional state,^{40, 42, 92–96} preparedness to receive difficult news,^{39, 41, 61} comfort with the topic,^{52, 69, 97–99} demonstrated priority/importance of the topic,^{36, 51, 52, 66} level of medical knowledge or education,^{35, 41, 43, 57, 61, 63, 91, 92, 100, 101} internal family functioning,^{54, 102, 103} age of the child, ^{51, 63, 72, 104–106} parental fears of harm to the child from communication,^{41, 92} and lack of fluency in the majority language.^{99, 107–110}

Several studies also identified characteristics of the clinician-family interaction that might facilitate or impede communication: impact of culture and cultural differences between clinicians and families,^{33, 50–52, 67, 108, 111} presence of additional people during sensitive conversations,^{41, 52} hierarchy or power differential between clinicians and parents,^{35, 43, 75, 112, 113} and contentious relationships and confrontational advocacy.^{96, 103} One study described clinicians feeling split loyalties to parents and children.⁴⁰

Factors related to race and/or SES were identified in 14 studies. US studies focused on race whereas non-US studies focused more on social status. Physicians in the US underestimated the parental desire for information for families from racial minority groups.¹¹⁴ Also, parents racial minority groups reported lower parental understanding of medical information,^115–117 decreased voluntariness in decision making,¹¹⁸ and greater decisional regret.^{119, 120} In non-US studies, social hierarchy potentially contributed to families of lower social status receiving less attention from doctors and having lower understanding of medical information.^{108, 112, 121–123} Additionally, families with lower financial/economic status seemed to receive less information and guidance from clinicians^{33, 63, 123} and demonstrated poorer understanding of medical information.^115–117

Team-level factors

Team-level factors were identified in 10 studies. Hierarchy within teams and perceived authority seemed to affect clinicians’ communication practices, especially related to prognostic discussions and introduction of palliative care.^{16, 55, 70} Clear delineation and delegation of this authority seemed to facilitate communication.^{16, 42, 55, 60, 124} Two studies described negative family experiences related to receiving contradictory information from different team members.^{16, 45} One study described how mutual support within the team supported communication.¹²⁵Another study described how the presence of multidisciplinary expertise supported communication.⁷³

Organization/System-level factors

Organization/system-level factors were identified in 33 studies, including time constraints,^{16, 42, 44, 64, 73, 78, 98, 108, 124} scheduling difficulties,⁷⁵ long wait times,⁴⁴ quality of transitions in phases of care,⁷⁸ and inflexible institutional routines.¹⁰⁸ Several studies also highlighted the need for appropriate physical space to hold conversations,^{39, 44, 52, 73, 76, 99, 103} educational resources,^{51, 52, 107, 126} and organizational staff and resources to provide sufficient care. ^{16, 46, 50, 51, 54, 61, 79, 98, 103} Lastly, 4 studies described the negative effects of clinicians frequently rotating and the presence of unfamiliar clinicians during difficult conversations. ^{16, 41, 45, 99}

Collaborating Hospital-level factors

Collaborating hospital-level factors were identified in 2 studies. One study described how medical professionals at outside institutions sometimes held negative views of palliative care, which created uncertainty and communication challenges for the primary medical team.⁷⁰ Another study identified barriers related to different standards of care and limited responsiveness of the collaborating hospital.¹⁶

Community-level factors

Community-level factors were identified in 6 studies. Four studies described the role of peer and social support in facilitating decision making.^{61, 102, 127, 128} Another study highlighted how these social influences can also impede communication by affecting voluntariness in decision making.¹¹⁸ Lastly, one study identified barriers related to perceptions of the hospital in the family’s community, social networks serving as alternative sources of medical information, and shortages of cultural representatives for certain patients.¹⁶

Policy-level factors

Policy-level factors were identified in 5 studies. Three studies described excessive medicolegal and logistical requirements, especially related to clinical trials and the complexity of informed consent documents.^{16, 45, 89} Three studies also identified barriers related to insurance and coverage/reimbursement.^{16, 54, 100}

Clinical factors

Factors related to disease characteristics were identified in 17 studies. The diagnosis of cancer created an urgency to start treatment. This urgency affected decision making, especially related to fertility preservation, in which patients had to balance the desire for fertility preservation with the risk of delaying initiation of therapy.^{46, 90, 97, 129–131} Notably, 26% (7/27) of studies that included pediatric and adolescent patient perspectives focused on communication around fertility preservation. Inherent uncertainty about prognosis and disease progression also impeded communication and decision making,^{61, 88, 102, 128} whereas clarity provided by disease progression supported discussions about end of life care.⁵⁷ The severity of disease also affected discussions of late effects, prognosis, and fertility preservation.^{50, 66, 84, 106, 132, 133} One paper identified limitations in effective and available treatments as a barrier to decision making. ¹³⁴

Discussion

This review reinforces that factors at multiple levels can facilitate or impede communication. Understanding this complexity is essential to developing interventions that improve communication outcomes for families in the future. However, 90% of these studies focused on individual-level factors, with fewer than 20% of studies addressing factors related to team, community, clinical factors, collaborating hospitals, or policy. Furthermore, few studies evaluated cross-level interactions of these factors. These individual-level factors are important, yet personal intention is only one contributor to how individuals behave within complex organizations.¹³⁵ A key tenet of implementation science is that implementing change within an organization requires an understanding of not only individual motivations, but also the institutional culture and the broader environment in which the organization is situated.^{135, 136} Individuals are situated in teams that function within organizations in the setting of the greater healthcare system. The multilevel framework of factors identified in this review helps to provide important context for understanding communication behaviors in pediatric oncology. The dearth of literature exploring these non-individual-level barriers must be addressed with high-quality studies. Future studies should also characterize the cross-level interactions of these factors to develop a more nuanced and complete understanding of the barriers and facilitators of communication.

Race and/or SES were associated with communication difficulties in 14 studies. Families from racial minority groups or who had lower social status received less attention, less information, and demonstrated more decisional regret. Families with low income similarly received disparate communication. These outcomes are likely the result of multiple factors related to race, status, and poverty. For example, physicians in the US underestimated the information needs and desires of parents from racial minority groups¹¹⁴ and their desire for involvement in decision making,¹²⁰ which could represent manifestations of physicians’ implicit or explicit biases. However, none of these studies explored the structural and societal barriers to communication and care for these families. Furthermore, race and ethnicity can have different meanings within and outside the US. In studies outside of the US, social status and hierarchy seemed to play an important role in communication. Future studies should examine the systemic effects of race, SES, and social status on communication with the goal of developing interventions to ensure that these families receive optimal communication and engagement in decision making.

Clinical factors also seemed to affect communication, especially communication about time sensitive issues. For example, decisions about fertility preservation were hindered because of the urgency to start treatment for cancer. In other studies, uncertainty about diagnosis or prognosis also impeded communication and decision making. These clinical factors are often not modifiable. Yet, clinicians should be aware of these factors to develop communication practices that can ensure patients and families have the support they need. For example, clinicians can address fertility preservation earlier and provide multimedia information to enhance understanding and decrease time pressure. Similarly, clinicians might address uncertainty and “what if” questions directly, even when the answers to these questions are unknown.¹³⁷

Ten studies of team-level factors identified challenges related to hierarchy within teams and unclear delegation of responsibility for key aspects of communication. Two other studies identified the negative effect on families of receiving contradictory information from different team members. These few studies point to the importance of effective teamwork in supporting communication. Supporting teamwork in the delivery of cancer care has been a major focus of national cancer organizations for the last decade.^{1, 138} One goal of supporting teamwork has been to decrease “tremendous waste and inefficiencies due to poor coordination of care.”¹ Communication is an essential part of this teamwork and requires additional study.¹³⁹ For example, how do teams develop a shared mental model about what has been communicated and what ought to be communicated with families? How do teams delegate responsibility for different aspects of communication and what is the role of hierarchy in supporting or hindering this communication? How can teams use structured approaches to support communication? For example, one group employed roadmaps during team meetings that tracked communication milestones and created accountability for fulfilling responsibilities.¹⁴⁰

The role of the family’s community was explored in 6 studies, primarily finding that social support and peer communities can affect decision making. Additionally, one study found that parents use social media communities as an alternative source of information, which might lead to the spread of misinformation. Given the growing role of social media in everyday life, it was surprising that so few studies explored the role of social media in communication. Other studies have shown that social media can help patients and families feel a sense of normalcy, support, and control despite their medical problems.^141–143 Furthermore, social supports are essential to helping families manage their child’s illness, which is a core function of communication.³ Future studies should further evaluate how the social and family structure affect communication.

Five studies of policy-level factors focused on medicolegal and logistical challenges, as well as insurance difficulties. These logistical requirements often focused on the length and ineffectiveness of the informed consent process. Given the broad reach of policy, it is imperative to understand the effect of new and established policies on communication. For example, the implementation of the 21^st Century Cures Act Final Rule in April 2021 mandates that healthcare organizations provide patients with access to clinical notes, laboratory results, medications, and problem lists, among others.¹⁴⁴ By 2022, organizations will likely be required to share all electronic health information (EHI) with patients. In addition to parents, adolescent patients will also have access to their EHI. This new policy could have positive and negative effects on clinician-family communication and relationships. To maximize the benefits and minimize the harms of this legislation, future studies should explore the perceived risks and benefits from key stakeholders in pediatric and adolescent oncology and develop best practices for implementation in oncology.

Lastly, interactions with collaborating hospitals were only evaluated by 2 studies, highlighting challenges related to second opinions, different standards of care, and limited responsiveness of the collaborating hospital. This dearth of studies is especially problematic for the care of rural-residing families. These families often live several hours from the academic center where they receive care. For these families, any acute illness will require visits to local hospitals for initial evaluation. These local hospitals then collaborate with the cancer center to determine the necessary course of action. In the broader literature, only one study has explicitly explored the care experience of rural-residing families in pediatric oncology.¹⁴⁵ Parents in this study described challenges communicating with physicians who provided emergent care at their local hospitals, including feeling that their concerns were not taken seriously. Parents also described limited local resources and deficiencies in the quality of care. According to this study, “Participants wanted physicians at their cancer center to be aware of limitations at community hospitals and, when unanticipated deficiencies developed, this was viewed as a lack of communication between local and cancer center providers.”¹⁴⁵ As such, collaborating hospitals play a key role in care and communication. Future studies should examine the factors that affect communication and care between cancer centers, collaborating hospitals, and families in pediatric oncology.

Almost no studies performed multivariable analyses with variables that represented factors from more than 1 level of barriers. Instead, most studies focused on individual characteristics of patients, parents, or clinicians as the drivers of communication behaviors. Structures and environmental characteristics, however, can strongly influence individuals’ behaviors.¹⁷ To understand the relative roles of these varying factors in influencing communication, researchers must consider these multilevel factors when designing studies and deciding which measures and variables to include in each study. Researchers might use the rich qualitative data from this collection of studies to identify which multilevel factors to include in future studies. To improve communication, we must move beyond a focus on the individual-level and account for the other factors that shape the way clinicians and families communicate. However, our team performed a scoping review of communication interventions in pediatric and adult oncology recently, finding that almost no intervention studies focused on pediatric cancer, and that the majority of interventions in adult oncology were educational sessions that taught communication skills.¹⁴⁶ Educational interventions necessarily target personal-level barriers in the hopes that building communication skills will overcome other barriers. In the future, researchers should develop interventions that target barriers at these multiple levels.

This systematic review should be interpreted in light of limitations. First, the majority of studies were qualitative in nature, which affects the generalizability of findings. Second, we excluded non-English articles, which could have resulted in missing data and decreased cultural diversity. Third, searching for studies related to communication is challenging because of the multiple functions of communication. We opted for the inclusion of broader search terms and screened through more than 5000 articles. However, our search strategy might have missed pertinent articles. Furthermore, many qualitative studies in this review had deficiencies in reporting methods and results, as evaluated by the COREQ checklist. As such, we are unable to fully determine the appropriateness of methods for these studies. Importantly, the pediatric and adolescent patient’s voice is also lacking in many studies. Future studies should strive to engage these patients in studies to understand their perspectives on communication.Few studies evaluated the cross-level interactions of these factors, which limits our understanding of the complexity of communication. Furthermore, our search strategy focused on “physicians” and “practitioners”, but we did not include “nurse” as a specific search term. Although our results identified studies with various clinicians, future work should explicitly include nurses in the search strategy, given their central role in communication. These limitations could have affected our aggregated impressions of the literature.

Multilevel factors can facilitate or impede communication in pediatric oncology. The vast majority of current studies focus on individual-level barriers and facilitators, with limited attention to factors related to teamwork, community, policy, collaborating hospitals, or race and SES. Although individual motivation is an important determinant of behavior, communication is a complex process that is influenced by these multilevel factors and their cross-level interactions. To improve communication experiences for patients and families in pediatric oncology, future studies should further characterize these barriers and facilitators of communication with the goal of developing interventions to ensure all families are optimally engaged in communication.

Supplementary Material

Appendix1

Appendix 1. PRISMA Checklist

Appendix2

Appendix 2. Full Search Strategy

Appendix3

Appendix 3. Detailed Study Information

Abbreviation Definition

AYA

Adolescent and Young Adult

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

COREQ

consolidated criteria for reporting qualitative research guidelines

CASP

Critical Appraisal Skills Program Qualitative Studies Checklist

AXIS

Appraisal tool for Cross-Sectional Studies tool