Inequities in childhood cancer research: A scoping review

SUMMARY

An integral part of understanding and then designing programs to reduce childhood cancer inequities includes adequate representation of people with cancer in research, including children. A scoping review was carried out to understand how cancer research is oriented toward inequities and to identify who has participated in childhood qualitative cancer research. A systematic search identified 119 qualitative studies that met inclusion criteria, with most studies taking place in high-income countries (n=84). Overall, data were lacking on social determinants of health at multiple levels—structural, household, child, and guardian. Only 29 studies reported on race and/or ethnicity, with the majority of those including predominantly or all white children. Six articles included socioeconomic information, and across most articles, attention was absent to the financial ramifications of cancer care. Limited reporting of sociodemographics highlights a broader issue of neglecting key demographics and social factors that contribute to inequities.

INTRODUCTION

Childhood cancer is a leading cause of childhood disease burden globally, with an estimated 400,000 children developing cancer each year.^1,2 Childhood cancer mortality is declining in high-income countries (HICs), though these improvements are not uniform across children in HICs. Within the United States and other HICs, for example, substantial disparities exist by race, ethnicity, disability, health insurance status, and parental and area-level socioeconomic indicators.^3–13 Further, more than 80% of children with cancer live in low- and middle-income countries (LMICs). A negative correlation exists between country GDP per capita and mortality rates for childhood cancer,¹⁴ with survival estimates ranging from more than 85% in HICs to less than 20% in many LMICs.^1,12,15–18 However, these inequities are not inevitable; clinician-researchers have demonstrated that it is possible to deliver effective treatment regimens in low-resource settings.^16,19

Policy and programmatic attention have moved toward identifying and addressing determinants and mechanisms creating and exacerbating childhood cancer inequities. By inequities, we mean the unjust differences in access to diagnosis and treatment, quality of care throughout the cancer care continuum, and outcomes that are socially, politically, and economically produced and systematic (see^20–22). As researchers develop frameworks,²³ strategies, and evidence-based interventions to address inequities in childhood cancers,²⁴ it is necessary that they engage an important group of social actors—the children themselves. This is important because children’s experiences and perspectives matter, and also because the structural determinants and processes that create inequities likely also shape children’s experiences, needs, and responses to their cancer. In effect, responding to the contexts in which children face cancer and addressing the processes that shape cancer burden and outcomes requires that researchers listen to children, hear what they are saying and showing, and appropriately respond. As just one example of how this might happen, Myra Bluebond-Langner’s foundational ethnographic study of children dying from leukemia in the 1970s showed both how children came to know and then conceal knowledge of their prognosis from their parents and clinical teams. In effect, what children with cancer know, feel, need, and want are not always transparent to adults, including their parents,^25,26 and meaningful, ethical engagement with children is needed.^27,28

Recent scholarship, however, suggests that children already impacted by systemic inequities may also be inequitably represented in cancer research. Two recent scoping reviews on the inclusion of children in qualitative health research identified that research with children—in general and in a specific populations (i.e., children with disabilities)— is more predominant in HICs.^29,30 The authors further found that children with disabilities were being excluded from participation in the design and implementation of health research. This offers further impetus for our review given that certain populations of children with disabilities have a higher likelihood of developing cancer, such as those with Down syndrome, congenital neurodevelopmental abnormalities, and certain birth defects.³¹ Further, cancer disease processes and treatment commonly lead to physical and cognitive changes that could also be classified as a disability.³² While these two reviews hint at potential gaps in research with children with cancer, they do not provide definitive answers. We therefore identify the need to assess the question: how much of the childhood cancer research carried out with children has engaged children from the groups and settings where social, political, and economic processes have generated unjust access to diagnostics, treatments, and resources? We carried out a scoping review to understand if and how qualitative cancer research with children aged 6 to 11 years addresses inequities in childhood cancer. Our scoping review is justified by the reality that (1) childhood cancer inequities exist, and (2) children with cancer offer valuable insights into their experiences that need to inform interventions, programs, and policies that affect their treatment and care. We focused on children in middle childhood because this is an age that is often missed in health research in general and cancer research in particular. In terms of cancer, children in middle childhood can get subsumed within adolescent and young adult (AYA) qualitative research studies, a category defined by the National Cancer Institute (NCI) in the United States as ages 15 to 39 years old. But their experiences and needs differ, not only because of their younger age, but because certain tumor incidence rates differ in middle childhood compared to younger ages and AYA, with inherently different treatment and survivorship experiences.³³

Our review focuses on qualitative studies with children with cancer—defined as a range of visual, verbal, participatory, and observational techniques – because of the potential of such qualitative methodologies to provide the in-depth understandings of children’s experiences from their perspectives.^29,34–38 While quantitative methods have value in childhood cancer, they are not positioned to bring out children’s voices the way qualitative methods can. In cancer disparities research, qualitative studies have gained importance for how they contextualize adult cancers, such as breast, colon, and cervix.^39–41 Identifying gaps in children’s inclusion is the first step to ensuring that children with cancer have opportunities to participate. We further anticipate that this review might direct childhood cancer disparities researchers unfamiliar with methods for doing research with children to the field of childhood studies and to those who have been doing this research already, so that they can better understand how to engage children and learn from them.

METHODS

We report this scoping review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) reporting guidelines. We chose to carry out a scoping review to identify the breadth and depth of qualitative literature on childhood cancer from children’s perspectives. A protocol for the review has not been published.

Information Sources and Search Strategy

A medical librarian in collaboration with the research team developed the search strategy. Three electronic databases were used (i.e., Ovid Medline, Embase, and CINAHL (the Cumulative Index of Nursing and Allied Health Literature)) to search the literature. The search terms were ‘Children’, ‘Cancer’, ‘Focus Groups’, ‘Participation’, ‘Health Intervention’, ‘Ethnography’, and ‘Qualitative Research’. We last updated the search in December 2021 to cover a 15-year time span (January 1, 2007 – December 28, 2021).

Eligibility criteria and screening

Our eligibility criteria included published qualitative studies involving pediatric cancer patients between the ages of 6 and 11. Studies that included parents, siblings, or others in addition to children with cancer in our target age category were also eligible for inclusion. We excluded studies if all participants were older or younger than the identified age range, if the ages of the children were not specified, or if participants were not diagnosed with pediatric cancer, or if children with cancer were not participants in qualitative methods. We further excluded non-English articles, conference abstracts, reviews, dissertations, and commentaries. We considered a range of qualitative, ethnographic, and participatory methods (referred to as “qualitative” in this paper), including interviews, focus groups, arts-based and participatory methods (e.g., drawing, photography), and participant observation.

The medical librarian imported search results into Microsoft EndNote and used the automatic duplicate finder to identify duplicate citations. Two reviewers screened each title and abstract independently, and resolved discrepancies by consensus. The team reviewed the full text of all articles that did not explicitly list participants or methods in the abstract. All included articles then underwent full-text review by two reviewers, with any discrepancies resolved by a third reviewer. We applied the same inclusion and exclusion criteria at each stage of screening.

Data extraction, charting, and collation

The team extracted data on year of publication, author(s), author(s)’ location of affiliated institution, study location, languages used in study, participant age range, gender, race and ethnicity, religion, education, and disability, and parental employment, education, financial status, insurance status, and housing and neighborhood factors. We based this extraction codebook on Davison et al.’s family ecological model and Sorenson et al.’s model for incorporating social determinants of health (SDoH).^42,43 We further aligned these with the World Health Organization’s (WHO) CureAll framework¹ and the Inaugural Symposium of Childhood Cancer Health Disparities⁴⁴ to ensure that our codes matched existing work on childhood cancer disparities work.

Three reviewers independently extracted data on sociodemographic factors and manually charted each factor using a spreadsheet developed in Microsoft Excel. A sampling of 30% of this data was checked for 100% agreement, with any uncertainties in extraction resolved by the first author.

RESULTS

The search identified 25,558 unique records. 25,159 did not meet inclusion criteria based on review of title and abstract and 274 were articles excluded after full-text review. Three additional records were identified from citation searching, and all were included after being assessed for eligibility. In total, 119 articles were included in this review, with half of these articles published since 2017 (n=59, 50%) and more than one-third since 2019 (n=47, 39%).

Where was qualitative childhood cancer research with children taking place?

Part of understanding who has and has not had opportunities to participate in childhood cancer research includes identifying where studies took place. The studies represented research that took place in 32 unique countries. Top country locations were Sweden (n=18), the United States (n=18), Canada (n=15), Brazil (n=10), the UK (n= 7), and China (n=9).ⁱ We further examined studies by region, with most studies taking place in Europe (n=45) and North America (n=33). Few studies took place on the African continent (n=2 articles representing 5 countries because Namisango et al.⁴⁵ carried out a 4-country study). While 11 studies were set in Central and South America, 10 of these studies were carried out in one country, Brazil. Finally, we categorized articles by World Bank income designation; most studies (n=84) took place in countries designated as high-income. Of the remaining studies, 26 took place in upper-middle-income countries, 11 took place in lower-middle-income countries, and only one took place in a low-income country (see Table 1).

TABLE 1.

Childhood Cancer Articles that include Qualitative Research with Children (n=119)

Study Characteristics	No. Studies	Total Studies, %
Year of Publication
2007–2009	10	8.4%
2010–2012	18	15.1%
2013–2015	19	16.0%
2016–2018	25	21.0%
2019–20211	47	39.5%
Study Locations2
Europe (including UK)	45	37.8%
North America	33	27.7%
Asia3	24	20.2%
Central and South America	11	9.2%
Africa	5	4.2%
Australia and New Zealand	4	3.4%
World Bank Categories2
High Income	84	70.6%
Upper Middle Income	26	21.8%
Lower Middle Income	11	9.2%
Low Income	1	0.8%
Youngest child participant4
Under-6	27	22.7%
6 to 8	69	58.0%
9 to 11	22	18.5%
Oldest child participant4
6 to 8	5	4.2%
9 to 11	9	7.6%
12 to 14	37	31.1%
15 to 17	35	29.4%
Over-17	32	26.9%

¹One study was published online in 2021 and then officially published in 2022

²One study was located in Kenya, Namibia, South Africa, and Uganda. Thus, the total of studies is n=122, while the number of papers in the review is n=119.

³One study was located in Turkey. For this table, we have grouped Turkey with Asia as the study took place in Kocaeli.

⁴One study used only the average to describe participant ages. Thus, the total number of studies is n=118, while the number of papers in the review is n=119.

Who were the children participating in qualitative childhood cancer research?

We assessed who participated and how the researchers wrote about sociodemographics to better understand whose experiences were documented. All articles in our review reported child age (n=119 articles) and all included children between 6 and 11 years old, our inclusion criteria. However, most studies also included children older than 11 (n=104), while 27 articles included children under 6 years in addition to the 6 to 11-year age range. Almost all articles also reported participants’ gender (n=101). However, these articles only used binary classifications to do so (i.e., girl, boy; female, male).

Twenty-nine studies reported on race and/or ethnicity of the children participating. These studies were primarily located in the US (n=13) and Canada (n=5), with the remaining in European countries (n=10) and Singapore (n=1). Of those identified, 28 articles reported on race. Sixteen of these 28 studies (57%) were comprised of either predominantly or all white children (greater than 75% of the study sample). Only five studies were comprised of less than 50% white participants, with four of these studies in the United States and one study in Singapore. Notably, nearly half of the articles reporting on race (n=13) used the problematic classification, “Caucasian.” See supplementary table 1.

Exclusion criteria for participating also offers insight into participation. Several studies excluded children with disabilities based on their physical (n=21) or cognitive (n=40) impairments. Language was also a limiting factor. Only 16 of the 61 (26%) studies reporting on language carried out research in more than one language.

What household resources shaped the lives of children participating in qualitative research on cancer?

Our review found limited information on the households and families in which children’s cancer experiences are embedded. As mentioned previously, such limited information constrained our initial approach to data extraction, which had included a number of family, neighborhood, community, and other social factors. We identified and then charted two broadly defined, but interconnected domains— household socioeconomic status/finances and parent/guardian employment—both in terms of reporting and qualitative findings.

Reporting of the study sample characteristics in each of these domains was limited. Six articles included socioeconomic information, with five reporting household income (all had wide distribution of participants across income levels),^46–50 and one reporting “financial difficulty” (two out of 10 participants reported difficulty).⁵¹ Only three reported on parental employment/occupations. Bicakli et al.⁵² reported mothers’ occupations as “Nine mothers were housewives and five of them had professional jobs” (p. 288), Coyne et al.⁵³ reported a sample with a high number of stay-at-home parents (64%), and De Vries et al.⁵⁴ identified that, in their sample, “parents’ occupations varied, indicating social diversity.” Across these nine articles, only one article reported financial information when a parent was not also participating in the study.⁵¹ This lack of financial information for participants was identified as a limitation in several articles, such as Gibson et al.,⁵⁵ who wrote: “We also did not record any information to enable socioeconomic status to be calculated. This could have had an influence on the types of food given and the financial implications” (p. 274).

The qualitative findings around finances and employment were revealing of structural determinants. Six articles discussed parents taking time off work without reference to financial ramifications, or with the ramifications implied (e.g., when reporting about a parent that could not take any more time off work after being off for a year^56,57 or could work from home during home-based cancer care⁴⁸). This absence of attention to finances in the articles could be due to healthcare system factors that, for example, enabled both parents to stay home with a child full-time with income compensation.^58,59 However, in Zhukovsky et al. a 9-year-old participant described that only his mother was with him in the hospital “cause dad’s home trying to make money...I wish he would like come out here and stay with us” (p. 5).⁵⁰

When oriented toward financial concerns, the qualitative findings revealed how children experienced and responded to the expense of cancer treatment. In contrast to the above examples, several articles more explicitly described financial challenges and job instability resulting from cancer treatment. Notably, most of these studies took place in LMICs. Vindrola-Padros⁶⁰ offered an example of the need to pay attention to effects of missing work on children with cancer. She quoted Martin,ⁱⁱ who had migrated from Bolivia with his mother for cancer treatment in Argentina: “I don’t like the place where we live here. We don’t have a lot of money because my mom can’t work because she has to be with me because of the treatment so we rent a room from this house. But I don’t like it. I hear gunshots and there are drunk people” (p. 438). As another example Namisango et al.⁴⁵ found that “Some adolescents had taken on the onus of meeting their own transport-related costs and increasingly found it difficult to meet such demands” (p.328). In addition to present financial concerns, they also demonstrated that children had future financial concerns as well, linking being out of school with their future opportunities.

The need to migrate or travel to treatment was a theme that ran across findings on financial strain and job instability. Several studies included above showed that such travel brought additional challenges beyond work and finances, such as difficulties communicating with health providers, language barriers, worry about family back home and the absence of a social support system at the hospital, the need to adjust to a new city and substandard living conditions and safety issues.⁶¹ This suggests more explicit attention is needed in childhood cancer studies with children not only to the experience of traveling, but also to how social determinants at multiple levels—structural, household, and child and guardian—intersect to shape experience and also exacerbate disparities.

DISCUSSION

Our review identified 119 articles in which children with cancer (middle childhood, ages six to 11 years) had participated in studies broadly defined as qualitative. While still small, this finding shows that childhood cancer researchers are taking seriously children’s participation in research and that interest in children’s participation might be increasing, given the steady rise in publications during our review years (Jan 1, 2007-Dec 28, 2021). This finding, alongside recent global attention to childhood cancer inequities,^1,44 suggests that now is the time to take stock of representation in the participation in such research. Our review identified multiple gaps and lessons, with three key takeaways needed for advancing a childhood cancer disparities research agenda that includes children.

1. The childhood cancer research with children in LMICs is limited but offers important lessons.

Our findings show a clear trend at a global scale; most articles reported on research in HICs (n=84; 71%). Even though the articles identified represented 32 unique countries, more than half of the studies took place in five countries (n=68; 57%): the United States, Sweden, Canada, Brazil, and the UK. As such, we are missing important perspectives on childhood cancer experiences in many places around the world, especially in LMICs, where only one low-income country was represented (Uganda) in this review. This is a pressing gap given evidence of global inequities. To this effect, a 2020 study reported that the highest childhood cancer mortality rates were in Eastern Africa, Central America, and South-Eastern Asia,⁶² yet our review included no studies in Central America, only two studies in South-Eastern Asia, and three studies in Eastern Africa.

Though limited in number, the studies carried out in LMICs offer important insights from children. “Children,” Vindrola-Pardros⁶⁰ wrote, “went beyond descriptions of the disease and treatment” to offer in-depth insight into a number of social determinants of health, including “travelling to a new city, leaving behind family members and friends, interrupting school and producing changes in the economic capacity of the family” (p. 439). The articles also demonstrated how children were responding to these challenges, not as passive recipients of cancer treatment but as actors in their treatment, in their families, and in their environments. This included trying to meet transport-related costs,⁴⁵ as described above, and also self-blame and wanting to leave treatment because of the effects of the treatment on their families.⁶⁰ These complex and multi-faceted impacts, and outcomes, of cancer in adults have been necessary for contextualizing factors commonly cited as contributing to disparities, such as fatalism and mistrust,³⁹ adherence to medical advice,⁴¹ rurality,⁶³ family support,⁶⁴ and distance and transportation.⁶⁵ We believe there is still much to learn from children around these impacts, but only when children are given the opportunity to express their experiences purposefully designed research to address inequities.

Given the growing research on childhood cancer disparities and health equity, contextualizing factors contributing to disparities is imperative to the development of more equitable policy and programmatic solutions. This includes high-income countries as well, where treatments for childhood cancer have led to high survival rates. However, survival rates tell little about the experience, and qualitative research can offer insight into the disparities experienced despite the “good result.”⁶⁶ To give just one example, survival rates say little about what it takes for children and families in the US to travel to cancer centers from the vast rural areas of the country or from the heavily segregated parts of US cities where decades of racist policies have contributed to a lack of access to public transportation, health facilities, and other basic needs for Black residents. The advancement of technologies and treatments for cancer holds the potential to exacerbate cancer disparities. This is evident in adult cancer in the US, with similar trends being shown in childhood cancer as well.⁴ As childhood cancer treatments become more available due to funding and policy attention, it is important to ensure that all children and families who need them have access, without creating more suffering. This will take a number of actions and structural solutions that include listening to children about their lived experiences with cancer. Importantly, we note that children with cancer are identified as important interest groups in the CureAll framework, a necessary inclusion that requires the upskilling of cancer researchers in critical childhood studies (see below).

2. Limited sociodemographic reporting is not just a problem for understanding who has participated, but also demonstrative of the limited critical engagement with social determinants of health that create cancer disparities.

Another finding of our review was the limited reporting on sociodemographic characteristics beyond age and gender of the child. Standards for reporting qualitative data identify the need to report relevant characteristics (SRQR, Standards for Reporting Qualitative Research) or demographic data (COREQ, Consolidated Criteria for Reporting Qualitative Research) of qualitative samples. We acknowledge the challenge researchers face of providing enough demographic data while also protecting participants’ identities due to small sample sizes in qualitative research and the small population of children with cancer at a given hospital, during a certain period. However, qualitative research differs from quantitative research as to the purpose of including such information. In qualitative research, it is important to understand why a sample was selected in the way they were selected and to link the demographics with theory development.⁶⁷ Long lists of demographics are not needed, as in quantitative research. Instead, a more purposeful focus on certain determinants of health that were not included (e.g., migrating for treatment) and usage of others that were (e.g., race) is needed.

Race as a demographic category offers an example of the need for more critically informed linkages between demographics and interpretations and theory-building. In countries like the US, racism is a main driver of health inequities and is a core determinant of health.⁶⁸ However, due to the current use of race and study findings, we were not able to identify how racism might affect childhood cancer experiences, with only 29 studies in our review reporting on race and/or ethnicity. Such reporting revealed trends; specifically, many studies reporting race had predominantly white study participants, even though they were located in places (e.g., US, Canada) with sizeable non-white populations. Further, race was frequently reported in the sample demographics without further interpretation or discussion. Notably, nearly half of the studies reporting on race used the term “Caucasian” as a gloss for white. Caucasian is not a neutral scientific category, nor is it biologically meaningful. It originated in the 18^th century to promote a racialized worldview and to justify slavery and discrimination.⁶⁹ This continued usage is indicative of an uncritical and acontextual approach to race and racism in countries like the US. This needs attention given the higher risk of death for racial and ethnic minority children with cancer than for non-Hispanic white children with cancer in the US.⁴ Qualitative research informed by critical race theory, has a substantial role to play in elucidating cancer diagnostic and treatment experiences for racial and ethnic minority children with cancer, as it does for adults with cancer.⁴⁰

While there were identified gaps in inclusion, there were also purposeful exclusions, namely of children with disabilities. This exclusion by researchers contributes to a lack of representation in cancer research, and aligns with prior scoping reviews that demonstrate the historical exclusion of children with disabilities from participation in research.³⁰ This exclusionary practice must also be considered at the intersection of disability, race, ethnicity, and socioeconomic status which compounds experiences of exclusion and has a domino effect, perpetuating marginalization by further limiting access to resources and influencing decision-making processes from research findings. To carry out better inclusive cancer research, partnering with disabled advocates, families, and children with cancer is necessary. Given our findings and the disabling effects of cancer, we recommend collaborating with disability studies scholars to apply theoretical perspectives such as disability critical race theory⁷⁰ to address systemic inequities related to disability, culture, and race in order to appropriately support inclusion in research and cancer care. By incorporating the insights of disability critical race theory into childhood cancer research, researchers can explore the overlooked sociocultural and structural aspects of the cancer experience in children with disabilities.

3. Childhood cancer health disparities research needs to center critical childhood studies approaches.

“Critical” aspects of critical childhood studies are needed to offer a lens through which to comprehend the inequities in childhood cancer that are socially, politically, and economically constructed and systemic.^71,72 Power dynamics, structural inequalities, and societal biases shape the experiences of children with cancer, and need critical examination in efforts to further equity in access to diagnosis, treatment, support, and participation in research. Through this critical lens, it becomes evident that addressing these inequities requires not just medical intervention, but also a comprehensive reevaluation of the broader systems that influence the experiences of children with cancer and their families. By critically examining these complex issues, researchers and practitioners can advocate for more equitable healthcare policies and practices that prioritize the well-being of all children affected by cancer.

There is a need to upskill researchers, not just in this work in childhood cancer, but in critical childhood studies in general. For example, age was reported across studies in our review, but most articles (87.4%) included children and young people with cancer who were older than 11 in their sample, and more than a quarter of the sample included young people above the age of 17 years. Previous reviews have shown that, when a wide age range participated, younger children’s experiences and voices often are minimized in the reporting.²⁹ Without critical attention to power and interpretation, research with children can tokenize or diminish younger children’s voices. (For more on this and a reflective guide on meaningful inclusion of children in health intervention research, see²⁹)

Limitations

A limitation of this review was our exclusion of articles not written in English, which likely affected country and regional findings. Articles published in English may often reach larger audiences due to the high representation of English speakers in global health research. Including more languages in future reviews can reduce this disparity and highlight the contributions of non-English scholarship that may otherwise be missed. Another limitation was the inclusion of children aged 6 to 11 years and this review many not be representative of trends in the inclusion of older children in the cancer literature. Further, a strength and limitation was that the articles covered a range of topics within childhood cancer from multiple different disciplinary perspectives, which we managed by focusing only on qualitative research.

Conclusion

Despite evidence of a growth of children’s participation in qualitative research on childhood cancer, significant gaps in diverse, global representation were identified as well as the need for researchers to increase efforts in analyzing the non-clinical mechanisms and social determinants that create disparities.

FIGURE 1.

Flow diagram for scoping review process to identify articles reporting qualitative research with children with cancer

From: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. doi: 10.1136/bmj.n71. For more information, visit: http://www.prisma-statement.org/

TABLE 2:

Studies included in scoping review by country and World Bank income designation

Country	Number Of Studies	Income Based on World Bank Designation for July 2022 to July 2023
Australia	4	High Income (n=84, 70.6%)
Canada	15
Denmark	4
France	1
Germany	1
Greece	1
Italy	1
Ireland	2
Japan	1
Netherlands	3
Norway	2
Singapore	1
Spain	1
Sweden	18
Switzerland	4
United Kingdom	7
United States	18
Argentina	1	Upper Middle Income (n=26, 21.8%)
Brazil	10
China ‡	9
Indonesia	1
Namibia†	1
South Africa†	1
Turkey	3
India	2	Lower Middle Income (n=11, 9.2%)
Iran, Islamic Rep	4
Kenya†	1
Pakistan	1
Philippines	1
Tanzania	1
Vietnam	1
Uganda†	1	Low Income (n=1, 0.8%)

^‡Articles that were located in Hong Kong (n=1) or Taiwan (n=2) were included in China due to World Bank classification.

^†One study was located in Kenya, Namibia, South Africa, and Uganda. Thus the total of studies is n=122, while the number of papers in the review is n=119.