Determinants of delayed childhood cancer care in low- and middle-income countries: A systematic review

Abstract

Early access to care is essential to improve survival rates for childhood cancer. This study evaluates the determinants of delays in childhood cancer care in low- and middle-income countries (LMICs) through a systematic review of the literature. We proposed a novel Three-Delay framework specific to childhood cancer in LMICs by summarizing 43 determinants and 24 risk factors of delayed cancer care from 95 studies. Traditional medicine, household income, lack of transportation, rural population, parental education, and travel distance influenced most domains of our framework. Our novel framework can be used as a policy tool toward improving cancer care and outcomes for children in LMICs.

1 |. INTRODUCTION

Cancer accounts for a large proportion of the global burden of disease in children, ranking as the ninth leading cause of disease for children.¹ The cancer burden in children is disproportionally concentrated in low- and middle-income countries (LMICs), where 85% of cancer cases occur.^2–4 There are wide disparities in survival rates in children with cancer around the world, ranging from 30% in LMICs compared to 80% in high-income countries (HICs).^5–7 The actual extent of disparities is likely underestimated due to data challenges such as the lack of cancer registries and vital registration systems in LMICs.^4,8–10 In this context, the World Health Organization’s (WHO) Global Initiative for Childhood Cancer (GICC) has aimed to reach at least a 60% survival rate for children with cancer around the world by 2030.¹¹

Many factors contribute to the global differences in cancer outcomes in children, including disparities in access to diagnostics or therapeutics, human resource limitations, financial barriers, lack of supportive care, and more advanced stages of disease when cancer is diagnosed in LMICs.¹² Unlike adult cancers, for which prevention and screening play a significant role, causative genetic and environmental factors of childhood cancers are less understood.^13–15 Early diagnosis and treatment constitute the most powerful approaches to improve survival for childhood cancers.¹⁶ However, children often face long delays in cancer diagnosis, with as low as 30% of children in LMICs receiving timely diagnosis and treatment.^17–19 The Three-Delay Model has been widely used in many areas of global health to evaluate delays in care, with delays described across three domains, including (a) deciding to seek health care; (b) reaching an appropriate health facility; and (c) receiving adequate care when a health facility is reached.^20–23 Understanding how the determinants of delays in care contribute to childhood cancer-related mortality is essential to guide strategic interventions and policy development.

Systematic reviews are critical to guide policy decisions, inform research priorities, and identify gaps in knowledge and are highlighted as a need in oncology research.²⁴ In addition, rigorous systematic reviews specific to pediatric oncology in LMICs are even further lacking.²⁵ Our objective was to identify determinants and risk factors of delays in childhood cancer care in LMICs using a systematic review. We used these determinants to propose a Three-Delay framework tailored to childhood cancer across the continuum of care.

2 |. METHODS

2.1 |. Conceptual framework

All determinants and risk factors of delayed cancer care in this systematic review were organized by adapting several theoretical models (Figure 1).²⁶ We organized the determinants and factors of delayed care into domains through the Three-Delay Model, a widely used framework in many areas of global health, which summarizes barriers to care associated with seeking, reaching, and receiving health care, depicting the patient’s journey from home to the primary health center, all the way up to higher level hospitals.^20–23 The subdomains were organized based on the WHO GICC framework, which longitudinally outlines the childhood cancer continuum of care from detection of symptoms to diagnosis, treatment, and survivorship.^26,27 We intersected the previous domains and subdomains with the Socioecological Model (SEM), a comprehensive framework used in public health interventions. This model is divided into four layers including the following levels: individual (behaviors, perceptions, demographics, etc.), interpersonal and family (socioeconomic factors, social support, etc.), community and organizations (infrastructure, workforce, referral networks, etc.), and policy and environment (health financing schemes, political agenda, etc.).²⁸ Finally, our framework was aligned with pediatric-specific cancer control plans, including strengthening health systems through an evidence-based, culturally specific implementation framework such as the CureAll program.^26,27

FIGURE 1.

Conceptual framework guiding the systematic review and adaptation of the final Three-Delay framework

2.2 |. Literature search

We followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) 2020 guidelines for the systematic review (Appendix SA).²⁹ Our detailed protocol, search strategy, and methodology were registered in PROSPERO (CRD42021256128) and were previously published.^30,31 We searched 10 electronic databases and three websites for peer-reviewed studies and grey literature from inception (Appendix SB). Search strings were constructed in compliance with the PICO (Patient, Intervention, Comparison, Outcome) framework,³² including (a) the Population: children (aged 0–18 years) from LMICs (based on the World Bank classification updated to June 2020),³³ (b) the [I]Exposure: factors contributing to timely childhood cancer care, and (c) the Outcome: delays in childhood cancer care (Appendix SC).

2.3 |. Inclusion criteria

Inclusion criteria were determined by compliance with the constructs in the PICO statement. No restrictions regarding language, publication date, outcome effect measure, or quality were applied. Evidence-based studies, including observational studies, qualitative studies, interventions, abstracts, conference papers, reports, and theses and dissertations, were eligible for inclusion. We included grey literature (i.e., WHO Global Index Medicus) to help mitigate the risks of publication bias. Childhood cancer was defined as all-inclusive cancers within this age category according to the International Classification of Childhood Cancer 3rd edition (ICCC-3).³⁴ Pediatric cancer care was defined as any step across the entire childhood cancer continuum of care. Studies were excluded if they had a sample population over 18 years old, included data from both LMICs and HICs, or examined both adults and children, and did not have a separate analysis for children.

2.4 |. Study screening, eligibility, and data extraction

The studies identified through the electronic databases were screened and assessed for eligibility in EPPI reviewer (version 4.12.0.0).³⁵ Two groups of reviewers (Cesia Cotache-Condor, Andie Grimm, Kelsey R. Landrum, Vinootna Kantety, and Jahsarah Williamson) independently screened in duplicate, with titles, abstracts, and full-text studies assessed against the eligibility criteria. Articles in languages different than English were assessed by a reviewer fluent in that language or translated with Google Translator and verified by a person fluent in that language. The group discussed and resolved by consensus any issues raised during the screening and eligibility processes. If the discrepancies persisted, the final decision was made by another senior coauthor.

Data from all studies meeting the inclusion criteria were extracted by two reviewers (Vinootna Kantety and Jahsarah Williamson). A third reviewer (Cesia Cotache-Condor) independently assessed accuracy of the data extracted on a random subsample of 15% of the studies. Reviewers used a predefined spreadsheet based on the conceptual framework described above to enter information, and any differences between reviewers were resolved by consensus. The retrieved information included title, authors, study design, sample size, age, location, year of publication, outcome, outcome measure effects, exposures (determinants and risk factors of delayed care), study time-frame, measure of delays in receiving care, workforce, infrastructure, out-of-pocket, catastrophic, and impoverishing expenditure. Determinants and risk factors of delayed care were differentiated based on whether they reported effect measures of association (RR = risk ratio, OR = odds ratio, HR = hazard ratio, and aPR = adjusted prevalence ratio). Exposures reporting these measures of association were defined as risk factors. Descriptive statistics were generated using SAS 9⋅4 (SAS Institute Inc., Cary, NC, USA) and ArcMap 10⋅3 (ESRI, Redlands, CA). Cancer diagnoses were classified into 13 categories, including general (all cancers), and the 12 categories based on the ICCC-3.³⁴

2.5 |. Methodological quality appraisal and bias assessment

A total of 61 full-text articles were independently assessed for quality by two groups of reviewers (Vinootna Kantety, Andie Grimm, and Kelsey R. Landrum). A third reviewer (Cesia Cotache-Condor) performed the reconciliation process. Any issues that were raised during the quality assessment process were discussed by the group and resolved by consensus. If discrepancies persisted, the final decision was made by a senior coauthor. An assessment of quantitative studies was performed by using the National Institutes of Health (NIH) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies.^36,37 The Critical Appraisal Skills Program (CASP) checklist was used to assess qualitative studies.³⁸ The Authority, Accuracy, Coverage, Objectivity, Date, Significance (AACODS) checklist was used to evaluate the grey literature.³⁹ All full-text studies (n = 61) were given scores based on their performance in their respective assessment tool. Each tool had a different number of questions (AACODS = 6 sets of questions, CASP = 10 questions, NIH = 14 questions). For all studies, a positive answer (yes) was given the score of “1,” a negative answer was given the score of “−1,” and an unknown answer was given the score of “0.”

For the AACODS tool, the scores were first calculated based on the questions within each set. For each set, if at least 50% of the answers were positive, then the entire set was marked as “yes,” and it was otherwise marked as “no.” Then, we considered every set as one question and proceeded with the methodology explained above. Subscores from each tool item were added to calculate an overall score for each study. Overall scores equal or less than “0” were classified as “low quality.” Overall scores greater than “0” but lower than “5” (qualitative tool) and “7” (quantitative tool) were classified as medium quality. Finally, overall scores equal or greater than “5” (qualitative tool) and “7” (quantitative tool) were classified as high quality.

3 |. RESULTS

The systematic review yielded a total of 95 studies that met inclusion criteria (Figure 2). We summarized determinants and risk factors of delayed childhood cancer care from a pooled sample of 39,636 participants among children, caregivers, and healthcare professionals (IHME), the WHO International Agency for Research on Cancer (IARC) (Table 1). Most studies were cross-sectional from hospital settings and evaluated determinants and risk factors of delays to diagnosis. Cancer types varied across studies, including general, site group I (leukemias, myeloproliferative diseases, and myelodysplastic diseases), site group II (lymphomas and reticuloendothelial neoplasms), and site group VIII (malignant bone tumors) representing the most common cancer types and making up to 27%, 9%, and 9% of the total, respectively. Additional details, including population, delays in weeks, time frame, publication, workforce, infrastructure, OPP expenditure, catastrophic expenditure, and impoverishment expenditure, can be found in Appendix SD.

FIGURE 2.

Flowchart of systematic review. Note: Other sources include the World Bank Group, Institute for Health Metrics and Evaluation (IHME), theWHOInternational Agency for Research on Cancer (IARC)

TABLE 1.

Summary of characteristics of studies included in the systematic review investigating the association between risk factors of delayed childhood cancer care in low- and middle-income countries (n = 95)

Author	Design	Country	Setting	Determinant/factor	Outcome	Type of cancer
Brown40	CS	Nigeria	Hospital	Age	Diag.	General
Walubita41	Qual	Zambia	Hospital	Parents’ or primary health workers’ lack of knowledge; household poverty; inadequate counseling services; shortage of supplies (i.e., blood); inadequate infrastructure (i.e., beds); traditional medicine; religion	Diag.	General
Fajardo-Gutierrez42	CS	Mexico	Hospitals	Maternal education low; no social security; travel distance	Diag.	General
Ekenze43	CS	Nigeria	Hospital	Education; referral; traditional medicine	Diag.	Wilms tumor
Stefan44	CS	South Africa	Hospital	Physician delay	Diag.	General
Mendes Lins45	CS	Brazil	Hospital	Distance; public outpatient clinic; order of birth; parental age; number of children; parental education	Diag.	ALL AML
Abdelmabood46	CS	Egypt	Hospital	None	Diag.	General
Handayani47	CS	Indonesia	Hospital	Alternative medicine	Diag.	General
Zahra Boutahar48	CS	Morocco	Hospital	None	Diag.	Brain tumors
Col Araz49	CS	Turkey	Hospital	Physician delay; rural; first diagnosis by general physician; outpatient clinic	Diag.	Leukemia, lymphomas, solid tumors
Berhane50	CS	Ethiopia	Hospital	Age >10; rural; parental education; low income; health insurance; cancer beliefs; holy water	Diag.	General
Ramirez-Ortiz51	CS	Mexico	Hospital	Dirtflooring; maternal education	Diag.	Retinoblastoma
Chukwu52	Coh	Nigeria	Hospital	Patient delay; first center visited	Diag.	General
Begum53	CS	Bangladesh	Hospital	Age >2; father’s low education; income	Diag.	General
Abdelkhalek54	CS	Egypt	2 Hospitals	Age; lower parental education; socioeconomic level	Diag.	General
Vasquez55	RC	Peru	Hospital	Older age; general physician	Diag.	Lymphoma, solid tumors
Carpenter56	RC	Botswana	Hospital	None	Diag.	General
Buckle57	CS	Uganda, Kenya	2 Hospitals	Financial; transportation; household duties; income; cancer perception; number of children; primary caretakers	Diag.	Burkitt lymphoma
Gilli58	CS	Brazil	Hospital	Urban residence; gait changes and paresis; father education level; gender; mother’s occupation	Diag.	Central nervous system
Brown59	CS	Nigeria	Hospital	Health system delay	Diag.	General
Grynszpancholc60	CS	Argentina	ONG	Private clinics; primary level hospital	Diag.	General
Chantada61	CS	Argentina	Hospital	Elementary education	Diag.	Retinoblastoma
Zapata-Tarres62	CS	Mexico	13 Hospitals	Lowest socioeconomic; highest socioeconomic; waiting time perception	Diag.	General
Essuman63	CS	Ghana	Hospital	Lack of awareness of cancer and financial barriers	Diag.	Retinoblastoma
Fabian64	CS	Multiple*	278 Health centers	Low-income level; lower middle-income level; age ≥22 months	Diag.	Retinoblastoma
Masika65	Qual	Tanzania	Hospital	Financial concerns; emotional concerns (parental guilt); need for information; need for tangible support; improvements in care practices; hospital living conditions; government assistance	Diag.	General
Leal-Cavazos66	CS	Mexico	n/d	Misdiagnosis; attempting globe salvage	Diag.	Retinoblastoma
Mehrvar67	CS	Iran	Hospital	Extraocular retinoblastoma	Diag.	Retinoblastoma
Sherief68	CS	Egypt	Hospital	Child age; parental age; education; socioeconomic status; maternal work; disease type	Diag.	General
Zhang69	RC	China	Hospital	Germinoma; slow growth	Diag.	Sellar germ cell tumors
Tagoe70	CS	Ghana	Hospital	Age; distance; malignancy type (solid tumor)	Diag.	Lymphomas, leukemia, retinoblastoma, Wilms tumor
Oscanoa71	RC	Peru	Hospital	None	Diag.	Non-Hodgkin lymphoma
Vasquez72	CS	Peru	n/d	Diagnosis by physician and not a pediatrician; advanced parental age; low level of education; older child	Diag.	Lymphoma, solid tumors
Abdelkhalek73	CS	Egypt	2 Hospitals	Low parental education; misdiagnosis; socioeconomic status	Diag.	General
Kilicarslan-Toruner74	Qual	Turkey	Hospital	Information sources; parental expectations of healthcare team; parental information needs/services; economic factors	Diag.	General
Pulivadula75	CS	India	n/d	Healthcare system delay; parental delay; rural; parental education; distance; diagnosis by general physician	Diag.	Hematological
Pilkington76	Qual	Uganda	Hospital	Home treatments; other responsibilities	Diag.	General
De Angelis77	CS	Nicaragua, Italy	2 Hospitals	Physician delay; patient delay; education/training programs on childhood oncological diseases	Diag.	ALL and AML
Papyan78	CS	Armenia	Hospital	Unavailability of registry; lack of essential services, multidisciplinary care, palliative support; costs	Diag.	Solid tumors, hematological
Cecen79	CS	Turkey	Hospital	Age; type of cancer/tumor; point of medical contact; metastasis at diagnosis	Diag.	Lymphoma, solid tumors
Afungchwi80	CS	Cameroon	3 Hospitals	Use of traditional medicine	Diag.	Burkitt lymphoma
Gonzalez81	CS	Mexico	15 Hospitals	Financial constraint; miss work; misdiagnosis	Diag.	Leukemia, lymphomas, solid tumors
Ocak82	CS	Turkey	Hospital	Age; paternal education; localization; tumor type	Diag.	Lymphoma, solid tumors
Bukha83	CS	Botswana	n/d	Brain tumors; age	Diag.	General
Nevarez84	CS	Mexico	Institute of pediatrics	Tumors of central nervous system	Diag., Treat. Init.	General
Renner85	Qual	Ghana	Hospital	Parent’s lack of knowledge; financial; traditional medicine; fear of Treat.; past experience in hospital	Diag., Treat. Init.	General
Njuguna86	CS	Kenya	Hospital	Healthcare delay; health insurance; alternative medicine; first center visited; hospital detention	Diag., Treat. Init.	General
Mulyowa87	CS	Uganda	Outpatient clinics	Poor adherence to treatment; missed appointment; failure or delay in initiation of necessary treatment	Diag., Treat. Init.	General
Schulze Schwering88	Qual	Malawi	Hospital	Parents’ lack of knowledge; primary health workers’ lack of knowledge of childhood cancers; referral was recommended with delay; traditional medicine; inadequate transport system; lack of free transport system	Diag., Treat. Init.	Retinoblastoma
Offor89	CS	Ghana	Hospital	Financial barriers	Diag., Treat. Init.	Burkitt lymphoma
Faruqui90	Qual	India	Hospital	Parents’ or primary health workers’ lack of knowledge; socioeconomic status; waiting times; lack of social support; multiple referrals; distance; unreliable public transport; traditional medicine; mistrust of health system; bureaucracy	Diag., Treat. Init.	General
Pourfeizi91	Mixed	Iran	Hospital	Perception of importance of primary symptoms; habitual delay in physician referral; low economic status; lack of family support; family problems; uncertainty about proposed treatment; high treatment costs	Diag., Treat. Init.	General
Balmant92	RC	Brazil	161 HOSPITALS	No past diagnosis; type of cancer; region of residence; age	Diag., Treat. Init.	Osteosarcoma, Ewing sarcoma
Junquiera93	Inter.	Brazil	Hospitals	Knowledge of childhood cancer among primary care providers; partnerships with specialized care services; existence of cancer care network	Diag., Treat. Init.	General
Baskin94	CS	Multiple*	Hospitals	Lacking adequate equipment; timely radiological tests; optimal radiation therapy; lack of multidisciplinary care	Diag., Treat. Init.	Brain tumors
Ashraf95	CS	Pakistan	Hospital	Age; type of malignancy; financial problems; distance; parental education status; perceptions about usefulness of treatment; use of alternative therapies	Diag., Treat. Init.	General
Gavidia96	PC	El Salvador	Hospital	Maternal illiteracy; anticipated travel time to hospital; low income; having a central line; belief about weather as a cause of illness	Diag., Treat. Init.	ALL and AML
Osifo97	RC	Nigeria	Hospital	Traditional medicine	Diag., Treat. Init.	General
Schroeder98	CS	Tanzania	Hospital	Treatment knowledge; complexity of care; limited provider capacity; poor care coordination	Diag., Treat. Init.	General
Hampejskova99	CS	Multiple*	Hospital	Healthcare professionals lack training	Diag., Treat. Init.	Retinoblastoma
Pondy100	CS	Cameroon	Hospital	Lack of financial means; use of other therapies	Diag., Treat. Init.	General
Adefehinti101	CS	Nigeria	Hospital	Financial constraint	Treat. Init.	Retinoblastoma
Resham102	CS	Pakistan	2 Hospitals	Insufficient resources; lack of access to medical care	Treat. Init.	Ewing sarcoma
Mills103	Coh	Gaza Strip	Hospital	Chemotherapy shortages; methotrexate monitoring; relapse; induction failure	Treat. Init.	ALL
Lowe104	Mixed	India	7 Hospitals	Contact with multiple physicians before diagnosis; distance	Treat. Init.	General
Azad105	RC	Nepal	Hospital	Presenting symptoms; type of cancer	Treat. Init.	Central nervous system
Grynszpancholc106	CS	Argentina	Hospital	Lack of pharmacy stock; lack in the country/drug bank; unauthorized medicine	Treat. Init.	Leukemia, lymphomas, solid tumors
Kashif107	RC	Pakistan	Hospital	Education of father/guardian of the patient; primary diagnosis; medical professional delays	Treat. Init.	Malignant mediastinal
Asirwa108	CS	Kenya	2 Hospitals	HIV; early relapse; high cost of cancer treatment; lack of social support; long distances traveled to the hospital	Treat. Aban.	Burkitt lymphoma
Salaverria109	Inter.	El Salvador	Hospital	Cultural and social factors	Treat. Aban.	ALL
Rossell110	Mixed	El Salvador	Community centers	Lack or irregular access to institutional contact persons; lack of transportation; lack of financial resources	Treat. Aban.	General
Alvarez111	RC	Guatemala	Hospitals	Distance; age; time since diagnosis; race/ethnicity	Treat. Aban.	General
Salaverria112	Inter.	El Salvador	Hospital	Weather; transportation; financial instability; caregiver ill or unable to leave work; caregiver error (forgetting appointment date); palliative care as an alternative; fear of treatment effects	Treat. Aban.	General
Ferman113	Inter.	Brazil	Hospital	Retinoblastoma diagnosis	Treat. Aban.	Solid tumors
Kumar114	RC	India	Hospital	Illiteracy; financial constraints;false perceptions about cure	Treat. Aban.	ALL
Meremikwu115	RC	Nigeria	Hospital	Costs of testing; financial constraints	Treat. Aban.	Burkitt lymphoma
Ishaya116	RC	Nigeria	Hospital	Large household size; mother as caregiver; travel time to hospital >2 hours; age	Treat. Aban.	General
Atwiine117	Qual	Uganda	Household	Financial; competing commitments; child looked cured; alternative treatment; cancer is incurable; side effects	Treat. Aban.	General
Borker118	CS	India	Hospital	Socioeconomic constraints; religious beliefs; fatalistic attitude; alternative therapies	Treat. Aban.	Leukemia and solid tumors
Slone119	RC	Zambia	Hospital	Region of residence; maternal education less than secondary school; negative hiv status	Treat. Aban.	General
Fadoo120	PC	Pakistan	3 Hospitals	Age; residence; paternal education; maternal education; ethnicity and language	Treat. Aban.	ALL
Stanley121	Mixed	Malawi	Hospital	Lack of guardian education; travel times; community influence; costs; suboptimal clinical logistical challenges	Treat. Aban.	Lymphoma
Libes122	Inter.	Kenya	4 Hospitals	Financial; misunderstanding of treatment plan	Treat. Aban.	Wilms tumor
Cai123	CS	China	20 Hospitals	Low income; money; education; rural; farmer	Treat. Aban.	ALL
Mostert124	CS	Kenya	Hospital	Alternative medicine; cancer perception; forced hospital stays; financial; health insurance	Treat. Aban.	General
Ngoc125	PC	Vietnam	Hospital	Poor prognosis; travel distance; gender; poverty; traditional medicine	Treat. Aban.	General
Kumar126	PC	India	Hospital	Rural; financial; not willing to enucleate	Treat. Aban.	Retinoblastoma
Friedrich127	Mixed	Multiple*	n/d	Socioeconomic status; education; travel time; treatment-related concerns; perceived prognosis; fear; awareness	Treat. Aban.	General
Fluchel128	CS	Ghana	Hospital	Financial burden for caregivers	Treat. Aban.	General
Mostert129	CS	Kenya	Hospital	Health insurance access; fee waiver procedures when payment is not made; hospital detention practices	Treat. Aban.	General
Bonilla130	RC	El Salvador	Hospital	Paternal illiteracy; maternal illiteracy; increasing number of household members; low monthly household income	Treat. Aban.	General
Ehrlich131	Mixed	Multiple*	n/d	Lack of physician palliative care training; lack of access to consultation; lack of home-based services; family resistance to palliative care; lack of healthcare workforce/resources	Palliative care	General
Ngwang132	CS	Cameroon	n/d	Cultural views and beliefs; lack of health units; inaccessibility of treatment products	Palliative care	Burkitt lymphoma

Abbreviations: Coh, cohort; CS, cross-sectional; Diag., diagnosis; Inter., intervention; Mixed, mixed methods; n/d, not detailed; PC, prospective cohort; QL, qualitative; RC, retrospective cohort; Treat. Aban., treatment abandonment; Treat. Init., treatment initiation.

Studies were distributed across 97 LMICs, with the largest number of studies from Africa (n = 104) and Nigeria (n = 11) at the regional and national levels, respectively (Figure 3). The years with the highest number of publications were 2018 and 2019. The cross-sectional design was most frequently used (59% of reports), while 5%–8% of the studies used a mixed-methods, qualitative, or intervention design. The main outcome reported by the studies was delay in diagnosis with 55% of the total number of studies, followed by treatment initiation (23%), and treatment abandonment (20%). Only 2% of studies discussed delays in receiving palliative care.

FIGURE 3.

Summary of descriptive statistics from 95 studies included in the systematic review by geographic location (A), publication year (B), childhood cancer type (C), outcomes (D), and study design (E). Note: Type of cancer was collapsed into 13 categories, including ICCC-3 categories (I = leukemias, myeloproliferative diseases, and myelodysplastic diseases; II = lymphomas and reticuloendothelial neoplasms; III = central nervous system and miscellaneous intracranial and intraspinal neoplasms; V = retinoblastoma; VI = renal tumors; VII = hepatic tumors; VIII = malignant bone tumors; IX = soft tissue and other extraosseous sarcomas; X = germ cell tumors, trophoblastic tumors, and neoplasms of gonads; XI = other malignant epithelial neoplasms and malignant melanomas) and one additional category for studies that addresses childhood cancers in general (G = general)

From all 95 studies included in this review, we found a total of 43 determinants and 24 risk factors that were associated with delayed childhood cancer care. From a subsample of 61 full-text studies evaluated for quality, half were classified as high quality (50.8%), and half were classified as medium quality (49.2%). No studies were classified as low quality. Further details from the quality appraisal can be found in Appendix SE.

The determinants and risk factors for delays in cancer care were summarized within the Three-Delay framework in three domains (D1: Seeking care, D2: Reaching care, and D3: Receiving care), one subdomain (onset of symptoms) under both “D1: Seeking care” and “D2: Reaching care,” four subdomains under the “D3: Receiving care” (Diagnosis, Referral, Treatment, and Palliative care), and four strata (individual, interpersonal and family, community and organization, and environment and policy) (Figure 4). Traditional medicine, household income, lack of transportation, rural population, parental education, and travel distance influenced most domains and subdomains. Palliative care was the domain with the most significant lack of data.

FIGURE 4.

Adaptation of the Three-Delay framework specific for childhood cancer care across the care continuum. Note: D1 = delay 1. D2 = delay 2. D3 = delay 3. The asterisk (*) and bold font indicates risk factors (RR, OR, HR, and aPR) of delayed childhood cancer as reported in the original studies. Determinants as risk factors in red font influence most domains in the framework

When we assessed determinants and risk factors by time point domain along the continuum, seeking care was mainly impacted by individual and family strata, while reaching care was mainly impacted by community and policy strata. Receiving care was the most often reported domain across the continuum. Determinants and risk factors impacting receiving care, specifically treatment, were widely spread between domains. Determinants and risk factors impacting access to diagnosis included lack of cancer knowledge on both individual and the community levels, health system variables, such as multiple referrals and waiting times, and lack of health insurance at the policy level. Referral care was impacted by a household’s income, travel distance, rurality, and having a dedicated referral communication contact. The treatment subdomain had a wider range of reported determinants and risk factors of delay compared to the rest of subdomains and domains, with presence across all strata (individual, family, community, and policy). The least studied point in the care continuum was palliative care, with only four predictors, including family resistance, lack of social support, lack of home-based services, and lack of physician palliative care training. Individual and family determinants and risk factors mainly included religious and cancer beliefs, traditional medicine, socioeconomic variables (income, marital status, household duties, parental education), lack of social support, and absence from work. While language and income barriers mainly impacted seeking care, across the community stratum, rurality, travel distance, and lack of transportation impacted both reaching and receiving care, including both referral to a health center and treatment. Barriers in cancer care infrastructure were consistent across the entire domain of receiving care. Across the policy and environmental strata, a country’s income level impacted seeking care, neither determinants nor risk factors were found to impact reaching care, and lack of health insurance and a country’s income level impacted receiving care.

4 |. DISCUSSION

There is an increasing recognition in the global health community of the pressing need to improve cancer care for children, particularly in LMICs that face the highest burden of cancer and the lowest survival rates. The WHO GICC has set a target to improve cancer survival in children to at least 60% by 2030.¹¹ As reducing delays in cancer diagnosis and initiation of care are among the most effective tools to improve cancer survival for children, a better understanding of how countries are performing in addressing delays in care is essential. Our systematic review suggested that traditional medicine, household income, lack of transportation, rural population, parental education, and travel distance were the leading predictors of delays in childhood cancer care in LMICs. Based on these findings, we propose a Three-Delay Model that can potentially serve as a decision policy tool ready to guide national and regional efforts toward timely care and increased survival among children living with cancer in LMICs.

Cancers are often underdiagnosed or diagnosed at a late stage in LMICs,^4,133 leading to increased mortality rates, higher costs of treatment, abandonment of care, and increased risks of household impoverishment.^20,134 Improved access to accurate data and data-driven decision tools are essential for government and other policymakers to reduce delays in cancer care for children.^8,135,136 A better understanding of the barriers to timely cancer care is a foundational step toward developing action plans to improve childhood cancer care along the entire continuum of care. The Three-Delay Model offers a comprehensive view that depicts the journey from the patient’s home to the healthcare network and the hurdles at each point in time across the continuum of care. Based on examples from other global health fields,^22,23 we adapted this well-known model to understand the trajectory of cancer care and the specific challenges faced by children in LMICs. Unlike previous applications of this framework,^20–23 the Three-Delay Model for childhood cancer care includes three domains (seeking care, reaching care, and receiving care), five subdomains, including onset of symptoms, diagnosis, referral, treatment, and palliative care. These domains and subdomains intersect with the four levels of the SEM, allowing us to dissect and evaluate how specific points of delay relate to endogenous and exogenous variables from a structural point of view.²⁸

According to our model, the delay in seeking care starts with the onset of symptoms, and takes place at the family and community level. At this point in time, individual and community factors influence the decision to seek care, and the role of the health system is limited to education, surveillance, and outreach. The delay in reaching care is mainly influenced by factors at the community level and the role of the health system falls on ensuring the patient has a timely and ease of journey to the first healthcare facility, usually a primary health center or a district/first-level hospital. The delay in receiving care is heavily impacted by the health system capacity and takes place from primary health centers all the way up to higher level and specialized hospitals. At this stage, strong referral networks, infrastructure, and workforce capacity ease the patient’s journey from a preliminary diagnosis of cancer at the primary health center or district/first-level hospital, to an evidence-based diagnosis, appropriate treatment, and palliative care at higher level and specialized hospitals.

Our study confirmed the existence of multidimensional and multifactorial barriers, preventing timely cancer care for children in LMICs. Therefore, future initiatives need to address the multiple and interacting barriers leading to delays in care, with isolated interventions to address single barriers unlikely to move the needle much to improve outcomes. Comprehensive cancer care packages within the universal health coverage (UHC) schemes should be designed to protect families from financial constraints, develop health system capacity, and enhance necessary support networks for patients, families, and health professionals across the entire continuum of care, with an increased attention to palliative care, the most neglected area across the continuum of care.

The need for capacity-building for childhood cancer systems is frequently inadequately prioritized in national health agendas. The mean health expenditure for cancer care in LMICs is only 6.2% of the global cancer expenditures, despite the high global burden of cancer located in LMICs.¹³⁷ Access to diagnostic and treatment services such as adequate number of workforce, medication, and ancillary services such as pathology and laboratory support are all essential to achieve timely and quality cancer care.^138,139 However, our data suggest that increased financial protection is also a necessary step in the roadmap toward decreasing delays in childhood cancer care in LMICs, to ensure swift seeking, reaching, and receiving care when needed. Caregivers often experience significant out-of-pocket expenses to ensure access to cancer diagnosis and treatment for their children, with financial challenges often leading to delayed care and abandonment of care.^123,126

Effective scale-up of childhood cancer systems in LMICs requires evidence-based financing streams accord to specific national priorities and sustainable financing mechanisms aligned under the principles of UHC.¹⁴⁰ Innovative financing strategies can include multisectoral public and private partnerships, pooling resources at regional or global levels, or leveraging global health financing facilities.^141,142 These strategies have been successfully implemented in other global health areas, such as malaria, tuberculosis, HIV, and child immunizations, and might be promising for childhood cancer. Specific hospital-level financing systems could include strategies such as giving travel, lodging, and meal vouchers to families from referral hospitals to travel to tertiary hospitals where specialized cancer care is often given, thereby reducing out-of-pocket expenses for the family.

Our study has several limitations. First, limited national-level data were available to measure all barriers of delayed childhood cancer care. This study only accounts for the exogenous (observable) variables related to delayed childhood cancer care. Endogenous (unobservable) variables such as beliefs and attitudes were not included because of the lack of data. Further studies exploring these endogenous barriers to timely care are needed. Second, although our search strategy is optimally inclusive, we might have missed some studies, particularly grey literature from countries where national cancer programs are led by private organizations. Only three international organizations were searched by our team. Third, the heterogenous definition of childhood cancer across the literature is a persistent challenge. In this review, children were defined as persons up to 18 years old to be as inclusive as possible. Finally, we found considerable clinical heterogeneity to measure the timing of delays across the continuum of care, preventing us from performing a meta-analysis. While uniformity in definitions and criteria is ideal, challenges in performing clinical research in LMIC settings often means that studies must be adapted around local resources, customs, regulations, and data availability. Despite these limitations, this study is the most comprehensive review of predictors of delayed care among children living with cancer in LMICs.

5 |. CONCLUSIONS

In LMICs, children with cancer face multifactorial barriers to access timely care across the entire continuum of care. We reviewed the current evidence of the key drivers of delays in care and found geographic, social, financial, health system, and health policy barriers that limit access to cancer care in LMICs. Based on these findings, we proposed a Three-Delay framework that can be used as a policy decision tool to guide financing streams and interventions to improve timely care and survival rates for children with cancer in LMICs. We encourage national governments and other public health leaders to use this tool as a template to explore their unique barriers and limitations.

Abbreviations:

AACODS

Authority, Accuracy, Coverage, Objectivity, Date, Significance

CASP

Critical Appraisal Skills Program

GICC

Global Initiative for Childhood Cancer

HIC

high-income country

LMIC

low- and middle-income country

NIH

National Institutes of Health

PICO

Patient, Intervention, Comparison, Outcome

WHO

World Health Organization

DATA AVAILABILITY STATEMENT

The review protocol is publicly available on PROSPERO and Open Science Framework. All data used in the systematic review are available in the Supporting Information Appendix.