A systematic review of economic evidence for community‐based obesity prevention interventions in children

Summary Multicomponent community‐based obesity prevention interventions that engage multiple sectors have shown promise in preventing obesity in childhood; however, economic evaluations of such interventions are limited. This systematic review explores the methods used and summarizes current evidence of costs and cost‐effectiveness of complex obesity prevention interventions. A systematic search was conducted using 12 academic databases and grey literature from 2006 to April 2022. Studies were included if they reported methods of costing and/or economic evaluation of multicomponent, multisectoral, and community‐wide obesity prevention interventions. Results were reported narratively based on the Consolidated Health Economic Evaluation Reporting Standards. Seventeen studies were included, reporting costing or economic evaluation of 13 different interventions. Five interventions reported full economic evaluations, five interventions reported economic evaluation protocols, two interventions reported cost analysis, and one intervention reported a costing protocol. Five studies conducted cost‐utility analysis, three of which were cost‐effective. One study reported a cost‐saving return‐on‐investment ratio. The economic evidence for complex obesity prevention interventions is limited and therefore inconclusive. Challenges include accurate tracking of costs for interventions with multiple actors, and the limited incorporation of broader benefits into economic evaluation. Further methodological development is needed to find appropriate pragmatic methods to evaluate complex obesity prevention interventions.

39 million children aged under 5 years. 1 Overweight and obesity in childhood and adolescence are associated with lower health-related quality of life (HRQoL) and additional healthcare costs. [2][3][4][5][6] Obesity in childhood and adolescence substantially increases the risk of chronic disease later in life, including type 2 diabetes mellitus, cardiovascular disease, hypertension, and various types of cancers. 1 The high health and economic burden of obesity 7,8 clearly demonstrates the need for childhood obesity prevention interventions that are effective, scalable, and economically viable in terms of cost-effectiveness. 9,10 Tackling the issue of childhood obesity is challenging, given the wide range of complex, multifactorial determinants across multiple ecological levels. 11 Multisetting, multistrategy interventions that incorporate community engagement and systems change (communitybased interventions [CBIs]) have been identified as promising in reducing the prevalence of obesity at the population level. This in contrast with single-sector interventions, which cannot address the multifactorial root causes of obesity in a meaningful way. For instance, Shape-Up Somerville was a community-based environmental change intervention across households, schools, and communities conducted in Massachusetts, USA, that demonstrated reductions of body mass index z score (BMIz) among school-aged children. 12 Similarly, the Romp & Chomp intervention conducted in Victoria, Australia, targeted children aged from birth to 5 years through community capacity building and environmental changes in early childhood education and care settings and demonstrated a significant difference in BMI among children aged 3.5 years between the intervention and comparator groups. 13,14 Given scarce resources, such CBIs must be both effective and cost-effective, and economic evaluation of multisetting, multicomponent CBIs is important to inform decision-makers on which interventions represent good value for money. 9,15 However, the economic evaluation of complex interventions for obesity prevention presents several methodological challenges. 10 These challenges include limited evidence that interventions incorporating changes in behavior, settings, and environments result in a significant outcome, such as reduced BMI, within relatively limited study timeframes; and the complexity of rigorously attributing costs and effects across interventions involving multiple players at multiple levels of the system simultaneously. 10 Whilst systematic reviews of the economic evidence for obesity prevention interventions have been published, 15,16 to the best of our knowledge, there are no current systematic reviews of the economic evidence for multicomponent and multisectoral community-based interventions targeting obesity prevention. A broad 2019 systematic review of economic evaluations of obesity interventions in children and adolescents 16 included 28 prevention interventions. 16 The search strategy used did not however include terms specific to CBIs, as the aim of the paper was to summarize the economic evidence for childhood obesity prevention and treatment interventions more generally and to assess the economic methods applied. As a result, the review did not provide a comprehensive overview of the state of the economic evidence for complex CBIs for childhood obesity prevention specifically, and in fact only included economic evaluations of two CBIs 17,18 at that time. In 2014, Flego et al. conducted a narrative review to synthesize the evidence of the cost-effectiveness of CBIs for obesity prevention. 15 The review included 11 studies comprising full economic evaluations [17][18][19][20][21][22][23][24] and did not include partial economic evaluation (e.g., cost analysis) or protocol papers. This limits the evidence that could be synthesized on the cost of intervention, particularly if interventions may not have had a statistically significant effect on their primary outcome and so a full economic evaluation may never have been published. This also limits the evidence for contemporaneous methods to undertaken economic evaluations of CBIs, and this may slow the development of robust and rigorous methods for estimating the value for money of these complex interventions. In addition, because cost-effectiveness research in this area was in its infancy at the time of this review, 15 the definition of CBIs used was also broad. This resulted in the inclusion of seven studies that were predominantly conducted in the school setting, with minimal intervention engagement from other settings (e.g., households and community). [19][20][21][22] More recent developments in CBIs include an increasing focus on meaningfully engaging multiple community sectors using a participatory approach. 25 This is distinct from earlier studies, which may have focused on a single sector such as a school, with relatively minimal engagement or participation from the wider community. 25 Given the advances in CBI research made in the years since the review was published, it is likely that only four of the included economic evaluation studies would now be considered as evaluating multisectoral CBIs (rather than school-based interventions).
Therefore, a significant gap in synthesizing the economic evidence for CBIs exists, particularly when considering the increasing interest in CBIs and the new studies conducted within the last decade. To address these gaps, it is necessary to synthesize evidence by including a broader range of economic studies that investigate communitybased interventions for obesity prevention involving multiple stakeholders. As such, the objective of this review is to synthesize up-to-date evidence of the costs and cost-effectiveness of community-based obesity prevention interventions and to provide a comprehensive examination of the methods used in published studies. This analysis will generate evidence of the economic credentials of CBIs for use in resource allocation and decision-making and underpin discussion on the future research directions required to overcome the significant challenges in understanding the costs and benefits of multicomponent and multisectoral obesity prevention interventions.

| METHODS
The systematic review was registered with PROSPERO (CRD42021262817) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 26 The PRISMA Checklist is in Table S1 General search terms were based on those used in the academic database search, with screening of the first 10 pages of search hits.
General search terms were then combined with the names of identified CBIs from a recent systematic review 25 to conduct a targeted grey literature search. Full details of the search strategies are presented in Table S2.
Study inclusion criteria were as follows: • Intervention in the general child population (aged from birth to 18 years); • Multicomponent and multisetting community-based obesity prevention intervention, defined as targeting a whole community and engaging at least two community sectors (e.g., households/families, schools, media, businesses, health services, community/recreation centers, and local governments) through a participatory approach, focused on the prevention of childhood obesity 25 ; • Reported either the protocol for or the results from a full economic evaluation (cost-effectiveness analysis [CEA], cost-utility analysis [CUA], cost-benefit analysis [CBA]) or a costing study, of an obesity prevention CBI; • Published between January 2006 and April 2022. The timeframe was started at 2006 for the relative recency of most CBIs for obesity prevention 25,27,28 and the aim of providing a synthesis of upto-date economic evidence; • Published in the English language; • Original articles, study protocols, reports. Protocol papers were included in the review to include the most recent methodologies employed, although economic results may not be available for these interventions.

| Data extraction and synthesis
A data extraction template was developed in Microsoft Excel following the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist. 29 Extracted information for each study included study design and settings, target population, sample size, brief summary of intervention and comparator, study perspective, discount rate, time horizon, measurement of effectiveness, choice of health outcomes, measurement and valuation of preference based outcomes, resource use and cost categories, currency, reference year, model assumptions, characterization of uncertainty and heterogeneity, sensitivity analysis, results, limitations, source of funding, and conflicts of interest. Data extraction was completed by one reviewer (MS) and checked by another reviewer (VB) to ensure all relevant information was extracted.
Findings were narratively synthesized, with a detailed narrative summary of the relevant CHEERS items 29 for each included study available in Table S3. Reported costs of each included intervention were converted to 2022 US dollars (USD) to facilitate discussion of the major findings. Conversion was carried out using the EPPI-CCEMG cost calculator (https://eppi.ioe.ac.uk/costconversion/).  17,18,24,[30][31][32][33][34][35][36][37][38][39][40][41][42][43] From the 17 included studies, 15 discrete economic evaluations, protocols, or costing studies of 13 CBIs were identified ( Table 1). The economic evaluation of a CBI was reported in full in the publication by Ananthapavan et al. 34 and also included as part of a larger prioritysetting study. 35 The protocol for economic evaluation of Whole of

Systems Trial of Prevention Strategies (WHO STOPS) for Childhood
Obesity was reported in detail in the publication by Sweeney et al. 39 and a brief summary included in the wider study protocol. 33 Likewise, the protocol for the economic evaluation of Obesity Prevention in Communities (OPIC) was reported in two studies. 40

| Summary of intervention characteristics
Twelve CBIs were conducted exclusively in high-income countries (HICs) ( Table 1). 17,18,24,[30][31][32][33][34][35][36][37][38][39]42,43 The OPIC study was conducted in both high-income and upper-middle-income countries from the Western Pacific region (Fiji, Tonga, New Zealand, Australia). 40 (Table 1) were undertaken in school settings, with other sectoral engagement including early education and care, local health services, community and support organizations, households, and restaurants. 17,18,30,31,36,37,40,41,43 BAEW, for example, targeted the whole community through engaging local health agencies, community stakeholders, and primary schools. 18 The CORD CBI primarily focused on primary healthcare clinics along with other sectors, including education (e.g., education centers and schools). 32 The R&C intervention was a community-based and community-wide obesity prevention intervention, incorporating strong engagement with early education and care settings. 42 CBIs evaluated by Ananthapavan et al., 34,35 and the WHO STOPS intervention 33,39 utilized a systems approach through engaging a range of organizations and community settings (e.g., partnerships between researchers and community leaders, parents and community leaders from local government, health agencies, businesses, and clubs).

| Summary of full economic evaluations (n = 7)
Seven full economic evaluations of five different CBIs were identified 17,18,24,31,34,35,42 ( Table 2). One full economic evaluation was F I G U R E 1 PRISMA flow chart of study selection on economic evidence of obesity prevention intervention.
The APPLE intervention has been the subject of two full economic evaluations. 17,24 The study by McAuley et al. 17  Given the low incremental benefits from the intervention (0.007 QALYs gained), the modeled CUA was not cost-effective, with an estimated cost per QALY gained of USD173,062. 24 Similarly, the BAEW CBI has been the subject of two full economic evaluations. 18,24 The study by Moodie et al. 18 conducted a modeled CUA of BAEW for the population of children aged 4-12 years in the Australian regional town of Colac (n = 2184). Costs and effects of BAEW were also extrapolated to the national level (assumed 10% of Australian primary school children). 44,45 Health benefits were measured as change in BMI and disability adjusted life years (DALYs) saved over the lifetime of the cohort using the ACE-Obesity model, 45 with an assumption of 100% maintenance of the intervention effect. 18 DALYs averted were calculated as the T A B L E 1 Summary of included studies (n = 17 studies incorporating 15 discrete economic evaluations, protocols or costing studies of 13 CBIs). USD20,075).
The BAEW CBI was also the subject of a modeled CUA where the effect of BAEW intervention was applied to the New Zealand general, M aori, and Pacific populations of children aged 9 years. 24 The cost per participant applied to the model was calculated from the BAEW report. 46 The time horizon for the evaluation was 92 years (lifetime), and the model assumed a 1% decay in intervention effect per annum after the fifth year of the intervention. No utility difference was applied to BMI categories for the base-case analysis. 24 In this analysis, BAEW was not cost-effective (USD130,093; USD142,087; USD104,239 per QALY gained for Pacific, general and M aori populations within New Zealand, respectively). 24 Both analyses revealed that interventions became not cost-effective when incorporating intervention effect decay in sensitivity analyses (Tables 2 and S3).
Ananthapavan et al. 34 reported the results of a modeled CUA of a hypothetical complex CBI targeting Australian children aged from 5 to 18 years. The analysis formed part of a national priority setting study 47 and was conducted from a limited societal perspective. Costs were estimated using a variety of sources (e.g., trial data and the liter-

| Summary of protocols of full economic evaluations
Protocols for full economic evaluations of five CBIs, for which the results were not available, were included in this review. 32,33,36,37,[39][40][41] A detailed protocol for the economic evaluation of the WHO STOPS CBI was reported by Sweeney et al., 39 and briefly summarized in the study's main protocol paper. 33 The economic evaluation as proposed  The protocol for HENRY (OFTEN) (HENRY with additional components) also aimed to evaluate parent engagement to implement intervention components (e.g., healthy diet and lifestyles for children). 37 The proposed method included generating cost-effectiveness acceptability curves to determine cost-effectiveness of the intervention. 37 A willingness-to-pay survey from a health commissioner perspective will be explored and used to inform the cost-effectiveness thresholds for economic evaluation. 37 The CORD CBI aimed to improve healthy eating and physical activity among children aged 2-12 years and to estimate the benefits that can be gained per dollar invested. The protocol outlined that activity-based costing would be used to measure intervention cost longitudinally; however, additional details and methods for the economic evaluation were not clearly presented. 32 3.5 | Summary of cost analyses (n = 2) and protocols for cost-analysis (n = 1) Two cost analyses of CBIs were included in the review. 30 18 were included in previous reviews by Flego et al. 15 and Zanganeh et al., 16 and Flego's review also incorporated modeled economic evaluations for the APPLE and BAEW interventions from the grey literature. 24 Since then however, there has been a significant increase in the economic evidence for multicomponent and multisectoral CBIs, with a further four full economic evaluations (CORD, 32 a hypothetical CBI, 34,35 Romp & Chomp, 42 30 reported that measurement of HRQoL changes was not practical for the OPAL intervention and economic evaluation was therefore not feasible, and a cost-analysis was conducted instead.
In addition, the nature of complex CBIs means that several, concurrent and interacting policy and practice interventions may be in place at the same time, limiting investigators' ability to separately identify and rigorously attribute intervention effects. 53 Longitudinal studies could be a better solution to mitigate these challenges in conducting economic evaluations for such CBIs for obesity prevention because they would allow repeated measures from the same individuals for the examination of changes in health outcomes and costs over time.
However, longitudinal studies often face challenges in measuring long-term intervention effects due to attrition and changing environments. 54,55 Only one intervention (CORD) identified in this review adopted a longitudinal design to estimate the return on $1 invested.
More exploration of longitudinal design is recommended to assist policymakers to make informed decisions about which interventions are the most cost-effective or economically viable for preventing obesity in childhood.
Evidence of intervention effect on the primary outcome of body weight or BMI is also sometimes lacking. Results from our review demonstrate that the absence of intervention effectiveness in some cases has led to limited economic evidence being produced. For instance, the economic evidence is not available to date for multicountry OPIC project as planned 40,41 while several published studies from the project are available with mixed results (mostly not significant). [56][57][58][59] A recent paper from the WHO STOPs intervention also revealed that although the intervention was effective for the first 2 years, it became ineffective after 4 years. 60  to age 15 years, 42 and there is clearly scope for the inclusion of similar short-to-mid-term benefits to be included in more modeled economic evaluations.
To date, there is limited economic evidence that attempts to incorporate spillover effects, or the potential wider health benefits of CBIs for obesity prevention (e.g., productivity cost averted). 39,62  Clear decision rules are required to attribute these costs, 39,66 yet no guidance currently exists in terms of how these decision rules should be formulated and/or applied. Clearly, these are areas for future work.

| Future research and policy implications of findings
The review uncovered several challenges in synthesizing economic evidence from the included interventions. These challenges include a F I G U R E 2 Recommendations for future research scope and policy implications.
lack of long-term evidence on intervention effectiveness, reliance on model-based studies, difficulties in tracking resources across multiple sectors, and a shortage of longitudinal studies to track costs and outcomes over time. As a result of these challenges, cost evidence was either reported to a lesser extent (e.g., OPAL) or not available in publications (e.g., OPIC). Based on these findings, a list of future research and policy implications has been included in Figure 2. The identified challenges underscore the need for more rigorous economic evaluations of interventions, particularly those involving multiple sectors.
Additionally, greater attention should be paid to collecting long-term outcome and cost data, as well as exploring innovative approaches to tracking resource utilization across sectors. These steps will help to improve the quality of economic evidence available to inform decision-making in healthcare policy and practice.

| Strengths and limitations
The major strengths of this review include the application of a systematic method following PRISMA guidelines and the most recent CHEERS checklist. 29 Other significant strengths are the comprehensive search strategies, the use of a wide range of academic databases and the grey literature, and review by two independent reviewers.
However, this review is limited to complex obesity prevention CBIs.
The methods used to understand the cost-effectiveness of complex CBIs outside child obesity prevention fall outside the scope of this review but may offer additional methodological insights. In addition, quality assessment of the included studies was not undertaken given the lack of a well-validated and universally accepted tool for economic evaluation studies. 67

| CONCLUSIONS
The findings of this review indicate that there is ongoing interest in evaluating cost-effectiveness of CBIs for childhood obesity prevention; however, available evidence is limited. The studies demonstrated mixed results and used widely varying methods. Complex CBIs are important for preventing childhood obesity, and reliable economic analysis, using standardized or comparable outcomes, is critical for the efficient allocation of scarce resources. This review further demonstrated that more research and methodological development are required to overcome challenges in accurately tracking resources and ensuring that the full breadth of potential benefits generated by CBIs for obesity prevention can be captured.

CONFLICT OF INTEREST STATEMENT
No conflict of interest statement.

DATA AVAILABILITY STATEMENT
The data are available within the manuscript and supplementary files.