Healthcare costs of cancer among children, adolescents, and young adults: A scoping review

Abstract Objective To collate and critically review international evidence on the direct health system costs of children and adolescents and young adults (AYA) with cancer. Methods We conducted searches in PubMed, MEDLINE, CINAHL, and Scopus. Articles were limited to studies involving people aged 0–39 years at cancer diagnosis and published from 2012 to 2022. Two reviewers screened the articles and evaluated the studies using the Consolidated Health Economic Evaluation Reporting Standards checklist. The reviewers synthesized the findings using a narrative approach and presented the costs in 2022 US dollars for comparability. Results Overall, the mean healthcare costs for all cancers in the 5 years post diagnosis ranged from US$36,670 among children in Korea to US$127,946 among AYA in the USA. During the first year, the mean costs among children 0–14 years ranged from US$34,953 in Chile to over US$130,000 in Canada. These were higher than the costs for AYA, estimated at US$61,855 in Canada. At the end of life, the mean costs were estimated at over US$300,000 among children and US$235,265 among adolescents in Canada. Leukemia was the most expensive cancer type, estimated at US$50,133 in Chile, to US$152,533 among children in Canada. Overall, more than a third of the total cost is related to hospitalizations. All the included studies were of good quality. Conclusions Healthcare costs associated with cancer are substantial among children, and AYA. More research is needed on the cost of cancer in low‐ and middle‐income countries and harmonization of costs across countries.


| BACKGROUND
Cancer incidence is rising among children 1,2 and adolescents and young adults (AYA), 3 which presents a significant disease and economic burden across the world. 4lobally, in 2020, there were an estimated 204,665 new cases of childhood cancers in children aged 0-14 years and 1,233,225 new cases in AYA aged 15-39 years. 5These were mainly leukemias, and brain and central nervous system (CNS) tumors among children, and breast and thyroid cancers among AYA. 5 In high-income countries, most of these cancers are treated successfully and yield a high 5-year survival rate of over 80% 6,7 for children and AYA with cancer.
A significant amount of healthcare resources, including medication, imaging, pathology, radiology, and specialists, [8][9][10][11] are needed for diagnosis and treatment of cancer, and to manage its side effects.In addition, the need for surveillance of secondary cancers or cancer relapse, and psychological care creates a significant demand for services after active treatment. 12he costs of providing healthcare for children and AYA with cancer are not well understood.Several studies and reviews on the costs of cancers in adult populations have been conducted [13][14][15][16][17][18][19][20] and highlight the high direct costs in adult cancer populations compared to those without cancer.However, parallel evidence is lacking for those diagnosed with cancer as children and AYA.
In this study, we performed a scoping review to evaluate the current literature on the healthcare costs of all cancers in children and AYA from a healthcare provider perspective.The evidence generated is important in understanding the health system cost implications of cancer, which is a key factor in informing government health policy, resource allocation and health services management.In addition, this evidence can contribute to economic evaluations concerned with allocating healthcare resources for childhood and AYA cancers.

| METHODS
This review followed the Joanna Briggs Institute (JBI) methodological guidance for conducting scoping reviews 21 and used dedicated JBI software.We developed a protocol outlining a detailed search strategy and inclusion criteria.The findings of this review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR). 22

| Eligibility criteria
English-language studies published from January 2012 to September 2022 were included and confined to this time frame to capture healthcare costs that reflected current clinical practices.The studies had to be quantitative and focused on cancer in children, adolescents, or young adults (defined as being aged 0-39 years).Healthcare costs were a primary or secondary outcome.We defined direct healthcare costs as expenses related to cancer treatment and care paid by the government and providers.Qualitative studies, protocols, reviews, commentaries, and conference abstracts were not included.We did not search for articles in the gray literature.

| Information sources and literature search strategy
The search for relevant articles was conducted in four databases: MEDLINE, PubMed, CINAHL, and Scopus.First, an initial limited search of PubMed was undertaken.We created a search strategy (Data S1) with the help of a medical librarian using Medical Subject Headings (MeSH) and article keywords.The strategy was then adapted to three other databases, including MEDLINE (Web of Science), CINAHL (EBSCO), and Scopus.

| Selection of sources of evidence
We collected and saved all the citations from various databases in EndNote X9 (Clarivate Analytics, PA, USA).We removed duplicate citations and imported the remainder into the JBI System for the Unified Management, Assessment, and Review. 21Two reviewers (DN and KM) independently screened the titles and abstracts of the retrieved articles.We checked all references of selected articles for other relevant ones, which we retrieved and added for full-text review.The reasons for exclusion were recorded in the JBI system.The two reviewers discussed any disagreements that arose over the eligibility of the article and came to a consensus for inclusion or exclusion.

| Data charting process and data items
The Excel™ database recorded study information, such as authors, year of publication, country, sample size, perspective, cost estimates, cost components (such as hospitalization, emergency department presentations, pharmaceutical use.), and study funding sources.
Data on the cost outcomes were charted in the currency reported in the respective studies and using the web-based CCEMG-EPPI-Centre Cost Converter, 23 adjusted to 2022 prices and converted to a common currency, United States Dollars (USD).

| Critical appraisal and quality assessment of sources of evidence
We appraised the quality of study reporting using the standardized Consolidated Health Economic Evaluation Reporting Standards (CHEERs) checklist. 24Given most of our studies were cost-of-illness studies, and CHEERs covers full economic evaluations, we limited our appraisal to 13 out of 23 items of the CHEERs checklist that were deemed relevant.

| Synthesis of results
Studies in this review had a diverse range of cost components or categories and methods across the different age groups which precluded performing a meta-analysis.Therefore, we performed a narrative synthesis and summarized the study characteristics, costing methods, study perspective, healthcare cost outcomes, and cost components.

| Selection of studies
Our search yielded 141 articles after the removal of duplicates.After screening the titles, abstracts, and full texts against the inclusion criteria, 30 articles were eligible for inclusion (Figure 1).

| Sample size
The study sample sizes referred to either the number of patients or episodes of care.Patient sample sizes ranged from 50 49 to 88,329 30 with a mid-range of 1376.The number of episodes of care, such as hospital admissions, ranged from 3853 50 to 6,675,222 35 in four studies. 29,32,34,50Two studies reported both the numbers of patients and episodes of care in the sample size description 42,48 (Table 2, Table S1).

| Data sources
Electronic administrative data and billing records were the predominant sources of resource and cost data for 26 studies (n = 26, 87%) (Table 1, Table S1).Three studies (n = 3, 10%) used data from the USA based Medical Expenditure Panel Survey. 25,26,30One study used data from the Healthcare Cost and Utilization Project 34 and one study used prospective cross-sectional survey data 49 (Table 1, Table S1).

| Quality of reporting of the costing methods
The quality of the studies was good overall, with sufficient information on the population and context of the study, clear descriptions of how costs were valued, and suitable measures used to summarize the study outcomes.Most studies (n = 26, 86%) reported over 80% of the items on the CHEERS checklist (Table S6).][53] However, based on the composition of the cost units, half of the studies (n = 12, 50%) employed a health system perspective, 12 studies used a payer's perspective and three studies had a societal perspective (Table 1).50% of the studies (n = 15) were funded by national health funding bodies, one 46 by a pharmaceutical company Novartis Farmacéutica, three studies had no funding and the remainder (n = 12, 40%) did not mention their funding sources.

| Direct health system costs
3 Five studies were exclusive to long-term cancer survivors, two of which were 3-and 5-year post cancer diagnosis.25,26,28,30,54 In the rest of the studies, participants either were still with cancer, 27,29,32,34,35,42,[44][45][46][47]55 within 1-year post diagnosis 37,49,50 or up to 3-year post diagnosis.31,33,53 Cost estimates were limited in time or specific to a particular care phase, cancer type, type of cost (e.g., hospital use and medication costs). High osts were seen in pediatric patients, those with leukemia, and during the first year of cancer diagnosis and end of life.
Based on selected studies, total healthcare costs for cancer per person varied from US$36,670 in Korea 51 to US$127,946 in the USA. 36

| Costs related to the first year of cancer diagnosis
The mean healthcare cost for children in the first year of cancer diagnosis ranged between US$34,953 in Chile 50 to over US$130,000 in Canada 37,39,41 (Table 2).Among adolescents in Canada, the highest mean healthcare cost was US$61,855 in the first year of diagnosis. 41n Australia, hospitalization costs followed a similar pattern, with children costing a median annual cost US$86,799 and adolescents costing US$47,487 per patient in their first year of cancer diagnosis 43 (Table 2).For AYA, a study in the USA estimated the median annual expenditure post-diagnosis was US$12,4091 postdiagnosis 36 (Table 2).

| Costs related to ongoing care and survivorship care
Like the first year of cancer diagnosis, total healthcare costs and hospitalizations costs during continuing care were higher for children compared to adolescents in both Canada and Australia.In Canada, mean annual costs for children and adolescents were US$16,253 and US$7749, respectively, while hospitalization costs during continuing care were US$13,072 for children and US$9427 for adolescents in Australia.
Three studies estimated the yearly costs of cancer survivorship beyond 5 years.On average, these ranged from US$1448 in the USA 28 to US$5914 among long-term childhood cancer survivors in France annually 54 (Table 2).Costs for cancer survivors were higher for AYA in the USA, estimated at US$9033 per year. 54 3.12 | Costs related to end-of-life care and palliative care Compared to other phases of care, costs were generally higher during the end-of-life care phase, commonly referred to as the last 12 months of life. Th cost of endof-life care is higher for children than for adolescents in Canada, at over US$300,000 39,41 and US$235,265, 41 respectively.In a US study of children with cancer and a high in hospital mortality, those receiving palliative care had lower hospital costs compared to those not receiving palliative care estimated at US$83,403 and S$152,179, respectively 29 (Table 2).

| Cancer-specific costs
In all studies, leukemia had the highest cost compared to other cancers for children and adolescents.For children, costs specific to leukemia were estimated at US$50,133 in Chile 50 to US$152,533 in Canada. 37,41Studies in Canada estimated the cost of leukemia ranging from US$149,045 to US$166,670 among adolescents. 40,41Other costly cancers included brain and central nervous system (CNS) tumors, bone and articular cartilage, and non-Hodgkin lymphoma (Table S3).
6][47] These included treatment with asparaginase preparations in the Netherlands with mean costs ranging from US$66,019 to US$129,497 45 and other treatment protocols in Canada for chemotherapy ranging from US$214,113 to US$299,354 38 (Table 2).

| Summary of evidence
Most of the studies were from high-income countries, predominantly the USA, with a broad range and mix of cost outcomes.In most studies, total costs were for children with cancer, related to hospital use, or specific to ALL.Findings in this review show healthcare costs were high for those diagnosed as children, those with leukemia, in the first year, and at end of life.Overall, costs related to hospitalization contributed to more than a third of the total healthcare costs.
Healthcare costs varied considerably across countries, according to this review.Since healthcare costs are a factor of volume and prices of health services, the variations from country-specific policies for setting and regulating healthcare prices largely influence the healthcare costs in different settings. 56,57In addition, factors such as the use of medical technologies, 58 healthcare administration costs, and the lifestyle of the country's population influence health service use and costs. 570][61] In the USA, for example, there is no Universal Health Coverage.Large segments of the population need private health insurance with government programs like Children Health Insurance Program (CHIPS) and Medicare limited to low-income families.Unlike the USA, other countries have a public provision of essential services and medicines through tax-funded universal healthcare or mandatory private health insurance.These systems increase access to essential health care services, protect individuals against financial risk, and increasing health service utilization and cost.Lastly, inconsistent conduct and study design elements, such as cost structures, cost perspectives, and other variables, could influence the magnitude of healthcare costs across studies, including those in a similar setting. 62,63ur findings show healthcare costs during cancer care follow a U-shape trend, highest during the initial year and end-of-life care, and lowest during the continuing phase.This is consistent with previous studies in the adult population which show a similar trend both during formal care [64][65][66][67][68][69] and informal care. 14In the study by Cheng Brian & Wangmo, provision of specialist palliative care was associated with lower cost compared to regular endof-life care.1][72] In the meta-analysis by May et al, it was reported that palliative care is associated with reduced time in the hospital, which subsequently reduces the costs of healthcare. 71e noticed limited data on the healthcare costs for those who survive beyond 5-year post cancer diagnosis.Previous research has highlighted a high morbidity burden caused by late effects of cancer, such as infectious complications, psychological conditions, chronic health issues, and secondary neoplasms [73][74][75][76] among the young cancer survivors.This information gap can therefore hinder effective healthcare planning, resource allocation and priority setting of health interventions for long-term cancer survivors.
In line with our findings, national estimates from Australia on health system costs of cancer show leukemia as one of the costliest cancers among those aged below 20 years. 66,77Literature for adult populations similarly show leukemia with the highest per-person costs during follow-up care 64,66 and was projected to cost 6.3% of the total global cost of cancers of 2020 to 2050 78 lung, breast, prostate, and colorectal cancers being the most expensive in terms of overall cost burden.We can attribute this to the high cost of chemotherapy 79 and high hospitalization resulting from the cytotoxic effect of leukemia treatment agents.Our review reveals that hospitalization is the leading contributor to healthcare costs among children and AYA, which is not surprising.We can attribute this to several factors, including centralization of cancer treatment at hospital-based centers for children and AYA due to specialist team needs which leads to longer hospital stays and increased costs.][82][83][84] Although over 50% of new cancer cases are from lowand middle-income (LMICs), 85 there are several obstacles to the conduct of quality research in these countries such as limited resources, poor research infrastructure and general lack of comprehensive cancer registration and surveillance systems that generate reliable data. 85

| Study Limitations
Our study had limitations.First, we mainly included the direct cost of cancer to the healthcare system, which limited a comprehensive discussion of the cost burden of cancer to the individual and society.Second, because of our provider-focused approach, we did not fully explore how diverse health insurance systems influence health care costs.
Due to the broad research question, the retrieved studies had heterogeneous outcomes, which restricted our capability to discuss costs across various study settings.Where possible, healthcare cost data collection and valuation methods should be standardized to support the comparability of cost outcomes across different settings. 86A national-level healthcare cost data warehouse, suggested by Visscher et al., can facilitate data collection and linkage to administrative databases for health economic research at the country level. 87In addition, ensuring uniform methods for cost calculation, consistent discounting, presentation of cost outcomes and conduct of sensitivity analysis will reduce healthcare cost variations as recommended in a previous review. 63egarding the cost of cancer, there is a need to put in place mechanisms that maximize value for money for cancer services to ensure sustainability of essential healthcare for patients with cancer.These include reevaluating the cost effectiveness of the current models of care and promote integration of palliative care at the end of life.This will facilitate prioritization of evidencebased care, reduce on the wastage of health resources and consequently reduce the cost burden of children and AYA.Furthermore, provision of ongoing psychosocial care, promotion of healthy lifestyles could support patients to navigate through the challenging effects of cancer. 86,88tter funding is necessary to improve childhood and AYA cancer registration in poorly monitored countries that will enable a more accurate and detailed assessment of healthcare costs incurred by health services and families and across different countries.This will lead to further research into the causes of cancers affecting children and young adults and determine whether there are modifiable risk factors that might be targeted to prevent these burdensome and costly cancers.

| CONCLUSION
This review summarizes the current literature on healthcare costs for cancer among children, adolescents, and young adults.It reveals that cancer health resource costs vary depending on age, phase of care, cancer site, and types of resources.We know very little about the healthcare costs of cancer in low-and middle-income countries and those attributed to the less common cancers.

F I G U R E 1
Flow diagram of search and study selection process.
Characteristics of the N = 30 included studies.
T A B L E 1 Health Care Costs per Phase of Diagnosis in 2022 US Dollars.