A total of 11 physicians participated in two focus groups. Demographic characteristics are listed in Table 1. The group was relatively evenly distributed by gender, and most participants were aged 36 to 50 (n = 10, 91%). Participants had at least 11 years of experience; the majority had completed formal fellowship training in pediatric hematology/oncology (n = 10, 91%), and more than one-third had completed training in hospice and palliative medicine (n = 4, 36%). Most participants worked in publicly funded hospitals (n = 7, 64%), and all participants practiced at centers with more than 100 new pediatric cancer cases per year. One participant (9%) practiced in a low-income setting, with the remainder practicing in a low-middle-income country (n = 7, 64%) or upper-middle-income country (n = 3, 27%). Participants worked in countries representing all WHO-defined regions (Supplemental Fig. 1).

Table 1 Demographic characteristics of study participantsFactors influencing treatment decision-making for advanced cancer

During the focus groups, physicians independently listed and then collaboratively discussed numerous factors impacting decision-making at diagnosis for children with advanced or poor prognosis cancer. Figure 1 stratifies factors by individual (patient, family, physician), hospital, health system, community, and country levels. Participant quotations are presented for factors influencing decision-making at each level, including factors related to prognostication (disease determinant), decision-making challenges (individual–physician determinant), resource limitations (hospital-level determinant), delayed diagnosis (health system determinant), influence of religion (community-level determinant), and geopolitical conflict (country-level determinant) (Figs. 1 and 2, Supplemental Table 1).

Fig. 1

Factors influencing participants’ treatment decision-making for children presenting with advanced cancer exist at several health system levels. Participants described factors that influence treatment decision-making at diagnosis for advanced childhood cancer in low- and middle-income countries existing at the level of the individual (patient, family, and physician), hospital, health system, community, and country. Participant quotations illustrate the factors influencing treatment decision-making at each health system level

Fig. 2

Factors described by participants organized by health system level, construct, and domain related to treatment decision-making at diagnosis for advanced childhood cancer in low- and middle-income countries. Factors described by participants organized by updated consolidated framework for implementation science (CFIR). Further factors are summarized in Supplemental Table 1

Overall, participants described 44 factors influencing treatment decision-making at five health system levels, presented in Fig. 3. Participants described physicians (i.e., themselves) and families as the key decision-makers for children presenting with advanced or poor prognosis cancer in LMIC settings.

Fig. 3

Multi-level determinants that influence treatment decision-making at diagnosis for advanced childhood cancer in low- and middle-income countries. Determinants or factors are influenced by the health care system level (i.e., community level) and by the domain (i.e., local conditions). These determinants are organized using a multi-level determinant framework (consolidated framework in implementation research; CFIR) to define, group, and map factors described in the focus groups

Individual-level factors

Individual-level factors were reported from the perspective of the physician participants. Physician-related factors encompassed the physician’s professional background, including prior training in palliative care, clinical experience, and individual beliefs. Participants suggested that junior and senior physicians may approach decision-making differently, with junior physicians more often recommending intensive therapy and senior physicians more likely to consider factors such as cost. Access to malpractice insurance was also noted as a variable that may enhance physician flexibility and proactivity in decision-making.

Patient and family factors influencing decision-making included health literacy, such as the family’s understanding or perceptions of cancer and cancer treatment. Physicians described how families may refuse treatment or express preference for traditional medicine due to lower cost, less toxicity, or cultural norms. Families may request intensive cure-oriented treatment even when discouraged by the physician, in part because the effort and cost of seeking care underpin their expectations for receiving treatment with curative intent. The financial capabilities of the family to support treatment costs were noted as an important factor, regardless of a family’s access to health insurance. A family’s health insurance could influence treatment decision-making if a specific center was not covered, or an in-network center did not have capacity to provide a given therapy. Additionally, physicians identified religious beliefs as an influence on treatment decision-making (e.g., attitudes and acceptance related to amputation, blood product transfusions, or burial processes).

Participants described examples of shared decision-making strategies when choosing treatment for children with advanced or poor prognosis cancer in partnership with the patient’s family. In some settings and cultures, however, they expected the physician to drive decision-making. Participants shared that older children and adolescents sometimes participated in decision-making, although this was influenced by local culture and practice and generally directed by the family.

Disease-related factors

Participants described disease-related factors considered in decision-making including the child’s diagnosis, disease stage, prognosis, comorbidities, treatment urgency, and quality of life. Notably, perceptions of prognosis and quality of life also influenced decision-making, with physicians balancing the likelihood of cure and disease-related toxicities without local data to guide decision-making. Comorbidities factored into treatment decision-making, including co-existing congenital syndromes or malnutrition that may increase the risk of the child experiencing treatment-related toxicities or infection. Physicians also described how they may initiate therapy due to poor clinical status or rapidly growing disease without completion of diagnostic work-up.

Hospital-level factors

At the hospital level, participants described several factors, including the influence of available resources (i.e., access to diagnostics, multimodal treatment, supportive care, clinical trials, trained personnel, housing, and hospital beds); access to local, regional, and international experts to aid in decision-making; and facilitators and challenges related to communication between the family and healthcare team. For example, physicians described restricting proposed treatments to align with available resources. More specifically, multimodal therapies and supportive care may be constrained by inadequate or inconsistent availability of medications or personnel. At times, physicians questioned medication quality in the context of the patient’s lack of expected treatment response. Additional decision-making challenges included a lack of diagnostic modalities that limit accurate diagnoses or delay treatment and, when available, are often cost-prohibitive for families or have suboptimal diagnostic turnaround times.

In the context of limited hospital capacity or staffing, physicians reported feeling an obligation to prioritize the care of patients with more favorable prognoses over those who presented with advanced disease. Participants also highlighted the variability of staff training in pediatric oncology or palliative care as a factor influencing their ability to recommend certain treatment options. Participants described the importance and influence of input from multidisciplinary teams locally, regionally, and internationally, especially for physicians with less clinical experience, those lacking formal subspecialty training, or in the setting of rare or complex cases. Time constraints and lack of available communication tools or educational resources in local languages further influenced treatment decision-making processes between physicians, patients, and families.

Health system level

Lack of cancer awareness in the community and fragmentation of the healthcare system were described as factors contributing to the presentation of advanced disease at diagnosis, complicating treatment decision-making. Physicians also described how organizational variance within the healthcare system (i.e., different centers within a city, country, or region that provide different treatment options at varying costs) may influence decision-making. Further, inequities in access to healthcare were perceived to shape treatment decision-making, as physicians felt unable to offer therapies that a family could not reliably or consistently access.

Community level

Participants shared that in some circumstances, spiritual leaders or community leaders held decision-making authority for children diagnosed with cancer living within their communities. Physicians also described the influence of non-family community members who may participate in decision-making. Local culture and religion were identified as variables that influenced decision-making in nuanced, varied, and profound ways. Lack of community education was reported as a driver for some families’ belief that cancer is incurable. Finally, physicians underscored the importance of local foundations to help subsidize treatment costs.

Country level

Treatment decisions were exacerbated by inadequate national policies related to treatment payment or delivery for childhood cancer. Participants discussed the importance of international collaboration, including the WHO Global Initiative for Childhood Cancer. Participants also shared the impact of the broader political climate, in particular for children who needed to cross borders to receive necessary care. Participants additionally reflected on the impact of unforeseen international crises, such as the recent COVID-19 pandemic that impacted access to healthcare resources.

Offering treatment with non-curative intent at the time of diagnosis

All participants agreed that it could be appropriate to offer therapy with non-curative intent at diagnosis for certain children presenting with advanced or poor prognosis cancer in LMICs. Participants offered contingencies for when they might consider a non-curative treatment strategy, such as when survival was unlikely or when cure-directed treatment would cause significant toxicities. In this setting, participants stressed the importance of involving families in the decision-making process. Some participants also acknowledged that they rarely offer non-curative therapy at diagnosis of advanced cancer or strive to avoid making this treatment recommendation. Salient themes, supported by participant quotations, are summarized in Table 2.

Table 2 Participant’s perspectives on offering treatment with non-curative intent at the time of diagnosisCharacterizing “poor prognosis” cancer

Notably, participants did not endorse a consistent definition of “poor prognosis,” and many physicians verbalized difficulty with defining this term. Various descriptions of poor prognosis referenced a low chance of cure or survival, physician perceptions of poor quality of life, and fragile patient status (i.e., advanced disease complicated by comorbidities). These concepts supported by participant quotations are summarized in Table 3.

Table 3 Participant descriptions of “poor prognosis” at diagnosis for children presenting with advanced cancer in low- and middle-income countries