1 INTRODUCTION

Establishing healthy nutrition and physical activity behaviors during a child's early years has lifelong benefits, including preventing obesity, at the individual, intergenerational, and societal levels.1-3 Worldwide trends in childhood obesity show rising prevalence over recent decades.4 Childhood obesity prevention is a global priority, including early childhood as a priority area of action.5

A recent Cochrane systematic review found that early childhood multicomponent interventions that include diet and physical activity behaviors can reduce obesity risk in children 0–5 years.6 Other reviews have also shown that early individual-level interventions for obesity prevention that involve parents, families, and health professionals effectively reduce obesity in the short term.7-10 Yet, to date, there has been limited focus on minority populations.

Children from culturally and linguistically diverse backgrounds in English-speaking countries, such as Australia, experience higher prevalence rates of overweight and obesity.11, 12 Families from culturally and linguistically diverse backgrounds may face challenges accessing early childhood interventions and services for various reasons, including language barriers and cultural differences.13-15 To reduce overweight and obesity inequities, there is a pressing need to ensure early childhood interventions supporting healthy growth are culturally relevant and accessible to diverse populations.

Cultural adaptation of interventions to reach new target populations—modifying an intervention to suit different cultures, languages, and contexts—is an established avenue to leverage existing effective evidence-based interventions.16 Culturally adapted interventions have been associated with better health outcomes among target populations.17 Over the past decade particularly, theoretical approaches and guidelines for cultural adaptations have advanced.17-21 In the implementation science field, where the concept of adaptation includes any context (not just cultural) modifications to interventions, reporting frameworks also consider when, who, and why an intervention was adapted.22

A 2012 review of culturally adapted health promotion interventions promoting healthy eating and physical activity23 found that interventions primarily focused on adults, with few targeting young children. To date, reviews of culturally adapted obesity prevention interventions targeting children have included interventions for African–American girls (aged 5–18 years)24; African–American youth (aged 6–18 years)25; and minority preschool children in the United States (aged 2–5 years).26 These reviews did not include literature outside of America and are now 5–10 years since publication. There have been no reviews that include culturally adapted obesity-related behavioral prevention interventions for children aged 0–2 years. This is of importance given the early years are a critical time for addressing modifiable obesity risk factors.27

This systematic review aims to identify culturally adapted health programs or interventions targeting obesity-related behaviors (including infant feeding, nutrition, physical activity, and/or sleep) among children 0–5 years, then analyze the cultural adaptation approaches and outcomes. Our specific research questions were as follows: (a) What processes are used for culturally adapting childhood obesity-related behavioral prevention interventions and what types of adaptations are made, and (b) how effective are these adaptations in achieving health-related outcomes?

2 METHODS

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines28 (see Table S1). The review protocol was registered with PROSPERO (CRD42018105596).29 Three amendments to the registered protocol include (1) no restrictions on search dates to expand search limits; (2) no searches of gray literature to restrict review scope; (3) use of a different quality appraisal tool with criteria for mixed methods studies necessary for this review.

2.1 Study inclusion and exclusion criteria

This review included peer-reviewed prevention intervention studies published in English that (i) were culturally adapted and explicitly accounted for participants' culture, ethnicity, or race; (ii) targeted healthy children not affected by obesity aged 0–5 years; (iii) included a component related to at least one obesity-risk related behavior, specifically nutrition, physical activity, sleep, or infant feeding practices (e.g., breastfeeding, introducing solid foods). These behaviors are recognized risk factors for obesity in childhood.30

Studies that solely undertook language translations of intervention materials or measures and studies that developed new interventions were excluded from this review. Interventions that primarily focused on treatment and/or management of obesity or other health conditions were also excluded, as these interventions recruit a different target group and have a different focus compared with prevention interventions. In addition, interventions targeting children aged over 5 years or children with a specific health condition were excluded. There were no exclusion criteria placed upon study design, study duration, or publication date.

2.2 Search strategy

A preliminary literature search was conducted in July 2018 using Ovid MEDLINE, followed by a comprehensive search using Ovid (MEDLINE(R) and Epub Ahead of Print, Embase, ERIC, Global Health, PsycINFO), CINAHL, Scopus, and Web of Science databases in October 2018, and again in March 2021. The lead author (SM) was a doctoral candidate and sought guidance from the University of Sydney librarians with subject expertise to refine the search terms and map them to appropriate subject headings in each database. The final search terms were limited to human subjects. See the supporting information for the search terms used in each database (Tables S2–S5). Key search terms included (1) infant OR child OR preschool, AND (2) cultur* AND (tailor* OR adapt*), AND (3) trial OR program OR intervention, AND (4) nutrition OR diet OR physical activity OR tummy time OR sleep. Hand searches of reference lists of relevant review articles were also conducted.

2.3 Selection process

All search results were uploaded to Covidence systematic review software31 to aid the independent screening process. After duplicates were removed, the title and abstract of records were screened by the lead author (SM) and independently by another reviewer (CR, LMW, ST, PL, and ME), with a third independent reviewer resolving any discrepancies (PL or YL). The full article was retrieved if the title and abstract did not provide enough information to inform a final decision. This process was repeated for the full-text screening (SM screened all records; CR, LMW, ST, PL, and ME independently double screened the records; YL resolved discrepancies). When the full text was required but not accessible (n = 3),32-34 the authors were contacted for further information. Two responses confirmed the abstracts were outside the review scope, and one did not receive a response, and no information was available in English, so it was also excluded.

2.4 Data extraction and synthesis

Data were extracted from each article using a template in Excel developed for this review. The template was drafted and piloted with six studies, then further refined. Data extracted included details of the study (e.g., design, country, and stated aims), the population (culture or ethnicity of the target group, description of the target group, sample size, sample characteristics, and target group prior to adaptation), the culturally adapted intervention (e.g., name, brief description, setting and mode of delivery, design theories, and target behaviors), a description of the cultural adaptation process, a summary of reported outcomes, and key author conclusions. For study designs that assessed effectiveness or efficacy, data on the health measures and outcomes were also extracted for synthesis.

The data extraction items related to the cultural adaptation process were informed by Barrera and colleagues' Stages of Cultural Adaptation theoretical model35 and the Framework for Reporting Adaptations and Modifications-Enhanced (FRAME).22 Barrera's model outlines the stages of the cultural adaptation process, which relates to study design. The FRAME assists with characterizing modifications to interventions, including cultural adaptations. The cultural sensitivity dimensions of surface or deep structure from Resnicow et al.36, 37 were used to classify the reported adaptations. These key cultural adaptation data extraction items are described in Table 1.

TABLE 1. Data extraction items related to cultural adaptation process and strategies Data item heading Description of item and definitions, where relevant Stage(s) of cultural adaptation (design) Which stages of adaptation are presented, according to the Stages of Cultural Adaptation process model?35 The stages include Stage 1: Information gathering; Stage 2: Preliminary adaptation design; Stage 3: Preliminary adaptation tests; Stage 4 Adaptation refinement; Stage 5: Cultural adaptation trial. Cultural adaptation theory or framework (design) Were there references to any cultural adaptation theories or frameworks? If yes, which one(s)? Description of process for adaptation (how) What process was undertaken to make cultural adaptations to the intervention? Was this described? Description of cultural adaptations made/cultural adaptation strategies (what) What cultural adaptations were made to the intervention content and/or contextual factors (such as delivery setting and mode)? What (If any) other changes were made to the original intervention? Surface or deep structure cultural adaptations (what) Were the cultural adaptations made at the surface or deep structure level according to Resnicow and colleagues' cultural sensitivity dimensions?36, 37 Surface structure adaptations include modifications to the observable characteristics of a target population, such as people, language, music, and foods. Deep structure adaptations involve incorporating relevant cultural, social, historical, and environmental factors that influence health behaviors. Who made the adaptations (who) Who determined that the adaptations should be made? Who led and who undertook the cultural adaptations? Involvement of the target group (who) To what extent were members from the target population group involved in making and/or informing the cultural adaptation process and strategies? What was the approach to involvement? When adaptations were made (when) When in the cultural adaptation process were the adaptations made? Were there multiple time points? The rationale for adaptation (why) What were the reasons provided for undertaking the cultural adaptation? What were the influences on the decision? Note: Informed by the Framework for Reporting Adaptations and Modifications-Enhanced (FRAME).22

The lead author (SM) extracted and coded the data using the data extraction template, and another author independently cross-checked and edited for accuracy and completeness (CR, LMW, ST, PL, and YL). For three articles, published reports and supplementary material related to the study were referred to when extracting data. Missing or unclear information in the articles was recorded as such in the data extraction template. To assist with data presentation and synthesis, the lead author coded the level of detail reported about the cultural adaptation process and strategies for each intervention. The level of detail reported was coded as “detailed description,” “some description,” “limited description,” or “not presented, not described” (see Table 3).

In this review, with multiple study designs included, we used a segregated design,38 where qualitative and quantitative data were considered separately but complementarily to answer the two key review research questions. Qualitative data and descriptions were used to understand cultural adaptation processes, and quantitative data were used to understand intervention effectiveness for achieving health-related outcomes. Data from this systematic review were synthesized and reported narratively to summarize and explain the findings.

2.5 Critical appraisals

As all study designs were included in this review, the quality appraisals were conducted using a tool for varied designs; the Mixed Method Appraisal Tool (MMAT).39 The MMAT is a reliable and valid tool for assessing the methodological quality.40, 41 MMAT focuses on core items representing the overall quality of evidence and risk of bias, which may impact the validity of the study findings. The MMAT includes two screening questions and five criteria according to the relevant study methodology category (qualitative research, quantitative randomized controlled trials, quantitative nonrandomized studies, quantitative descriptive, and mixed method studies). Each item is rated “yes,” “no,” or “can't tell.” For each of the included interventions, the peer-reviewed publication that included key behavioral outcomes was assessed by the lead author (SM) and independently by another author (CR, LMW, ST, YL, PL) using separate Excel files. Ratings were then compared, and discrepancies were resolved through author group discussion. The results of the assessments were considered during data synthesis but did not result in exclusions from this review.

3 RESULTS 3.1 Study selection

The searches generated 8,322 records, and after removing duplicates, 4,708 records were independently double screened by title and abstract (95.1% interrater agreement). The full texts of 107 records were evaluated against the inclusion and exclusion criteria. A total of 16 articles representing 12 unique interventions were included in this review (Figure 1).

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram of systematic search findings and study selection

Notable exclusions that might appear to meet the criteria of this review included interventions that referred to cultural adaptation literature; however, they were designed or developed specifically for a particular cultural population group rather than adapting an existing intervention (e.g.,42, 43). Another notable exclusion was an intervention for childhood obesity prevention in Europe,44 which was developed with the intention of context and cultural adaptations during implementation; however, the cultural adaptations have not been documented.45

3.2 Description of studies

Table 2 summarizes the included interventions (n = 12) and articles (n = 16).46-61 Detailed intervention characteristics are available in Table S6. Included interventions primarily aimed to prevent childhood obesity (n = 5) or improve obesity-related behaviors (n = 5), and most included content to promote multiple obesity-related behaviors (n = 7). Most were from the United States of America (n = 7), with others from Australia, Bangladesh, Malaysia, and the United Kingdom (n = 5).

TABLE 2. Main characteristics of included interventions (n = 12) Intervention characteristics n Reference numbers Stated aim Obesity prevention 5 47,49/50/51,55,56,57/58 Improving specific behavior(s) 5 48,53/54,59,60,61 Child development 2 46,52 Target obesity-related behavior(s) Multiple lifestyle behaviors 7 47,49/50/51,52,55,56,57/58,60 Activity and sedentary behaviors 3 53/54, 59, 61 Feeding and nutrition 2 46, 48 Target child age In utero until 2 years 4 46,56,57/58,61 2–5 years 7 47,48,49/50/51,52,53/54,55,59 0–5 years 1 60 Engaged in intervention Child 3 53/54,59,60 Parent/caregiver and child 9 46,47,48,49/50/51,52,55,56,57/58,61 Settingb Early education and care settings 5 47,49/50/51,53/54,59,60 Community venue 3 52, 59, 61 Home 3 48, 52, 56 Healthcare clinic or hospital 3 55,57/58,61 Community health center 2 46, 48 Mode of deliveryb Face-to-face, groups 10 46,47,48,49/50/51,53/54,55,59,60,57/58,61 Face-to-face, individual 2 48, 52 Telephone, individual 1 56 Online component 1 47 Provider/facilitator Health professional 4 46, 52, 55, 56 Childcare staff member 3 47,49/50/51,53/54 Parents or peers 3 48,60,57/58 Research assistant 2 59, 61 Theory of intervention Described/referred to 7 47,48,49/50/51,55,56,59,57/58 Not described 5