Introduction

Despite recommendations to prioritize obesity prevention efforts, [1–4] epidemiological data from the 2015–2016 National Health and Nutrition Examination Survey (NHANES) found that the prevalence of early childhood obesity remains at a 10-year high [5]. Furthermore, obesity rates among school-aged children aged 6–11 years are approximately 25% higher compared with preschool children aged 2–5 years [5]. In addition, even higher obesity rates are differentially associated with minorities and children from low-income families [6]. These prevalence patterns underscore the need to focus on early childhood obesity prevention efforts with the goal of meeting the Healthy People 2020 obesity rate target of 9.4% [2]

The evaluation of such efforts should be guided by framework models that consider the various levels that influence an individual’s health trajectory [7, 8]. For example, the National Institute on Minority Health and Health Disparities (NIMHD) Research Framework represents the multiple levels of modifiable and interacting determinants that contribute to health disparities (Fig 1) [9]. While the multi-level Life Course Health Development Framework perspective also recognizes that health-development unfolds over the life course, is sensitive to time and environment, adaptive, and requires a balance among all levels of health [10]. Together, these models are well-suited for evaluating and characterizing childhood obesity prevention efforts and informing future interventions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Framework used to characterize components of early childhood obesity prevention interventions across the early life course.

Adapted from The National Institute on Minority Health and Health Disparities (NIMHD) Research Framework [9].

https://doi.org/10.1371/journal.pone.0209787.g001

We conducted a scoping review to provide an overview of the current state of obesity prevention efforts in the United States (US) with children less than 6 years of age, and to answer two questions: “What types of interventions and policies are being used for obesity prevention across the early life course and at multiple levels of influence?” and “How effective are they?” The secondary aim was to describe the best available evidence on the cost and cost-effectiveness of implementing obesity prevention interventions and policies.

Methods

A scoping review was conducted to expand on previous systematic and narrative reviews of obesity prevention efforts in young children and to identify scientific evidence from a broad range of interventional studies, government and non-governmental programs, and local and national policies in the US. The scoping study design was chosen because it offers a framework to identify, “map”, merge evidence, and synthesize a broad range of evidence. Furthermore, the scoping review methodology is focused on providing conceptual clarity and allows researchers to focus on questions with relevance to target populations and locations [11–12].

The scoping review process is based on the Arksey and O’Malley’s 5-stage methodological framework [11]. The 5-stages that served as a roadmap for the present review are: 1. identifying the research question; 2. identifying relevant studies; 3. study selection; 4. charting the data, and 5. collating, summarizing and reporting the results. [11]. While the Joanna Briggs Institute’s Reviewer’s Manual of best research practice guidelines for conducting a systematic scoping review, served as a guide for the present review [12]. To ensure consistency, transparency, and reproducibility an a priori scoping review protocol was developed and directed the review process.

Search strategy

A search strategy was devised with the assistance of an Information & Education Librarian (MG) for PubMed using keywords from obesity prevention articles and modified for the additional electronic databases. Table 1 shows the search syntax and strategy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Search strategy and study selection criteria.

https://doi.org/10.1371/journal.pone.0209787.t001

Results

Collating, summarizing and reporting the results (Stage 5)

Fig 2 describes the literature search and study selection process. We identified a total of 2,467 records. After removing duplicate records, the titles and abstracts of 2,180 records were screened for inclusion. The full text of 73 articles were reviewed for eligibility and 34 studies were included in the review.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Flow diagram showing literature and study selection.

Adapted from Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ (Clinical research ed). 2009;339:b2535, [15].

https://doi.org/10.1371/journal.pone.0209787.g002

Study characteristics

The included studies examined the collective experiences of approximately 900 pregnant women, 1,600 infants and 10,000 preschool aged children across 16 states, in 10 urban centers, and a mix of suburban and rural communities (Tables 2–4). The interventions (n = 25) were initiated during three stages in the early life course: pregnancy (n = 3),[16–18], infancy (n = 3), [19–21] and preschool (n = 19), [22–40]. The majority (88%, 22/25) of the interventions used an experimental study design (19–40).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics of interventions during pregnancy.

https://doi.org/10.1371/journal.pone.0209787.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics of interventions during infancy.

https://doi.org/10.1371/journal.pone.0209787.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Characteristics of interventions during preschool.

https://doi.org/10.1371/journal.pone.0209787.t004

We identified 6 publications examining the impact of city, state and national obesity prevention policies [41–46]. Three additional studies calculated the net cost or cost-effectiveness of obesity prevention interventions [47–49].

In total, 11 (44%) interventions reported a positive benefit on a weight-based measure of growth or weight status (e.g., weight, weight/weight percentiles, BMI/ BMI z-score, and BMI categories) [17, 19, 20, 22–25, 28, 31, 36, 38]. The effectiveness of the interventions was inconsistent and contradictory across all stages of the early life course. Tables 2–4 provide descriptions of the characteristics and outcomes of the interventions.

Interventions in preschool aged children

Childcare center-based interventions.

Over half (63%, 12/19) of the preschool-aged interventions enrolled children from low-income, racially and ethnically diverse families from Head Start centers (n = 7), YMCA-affiliated childcare centers (n = 2) or other subsidized childcare programs (n = 3), (Table 4). The childcare center-based interventions included in this review were designed with interventional components that primarily focused on influencing individual- or interpersonal-level health behaviors of the children and preschool teachers. Only 42% (5/12) of these interventions reported a significant improvement in either BMI-z score [22], BMI [23–24, 28] and BMI percentile [36].

Of the interventions demonstrating a positive effect on BMI, two studies administered 30 minutes of moderate to vigorous physical activity (MVPA) for primarily African American (AA) children attending YMCA-affiliated preschools [23, 24]. Similarly, the “Hip-Hop to Health” efficacy trial [28] enrolled predominately AA preschoolers attending Head Start and used trained educators to deliver 20-minute healthy-lifestyle behavior themed lessons and 20 minutes of directed physical activity (PA). However, the same intervention did not have a beneficial effect on BMI in Latino preschoolers [29], nor did other similar effectiveness trial in Latino [30] and AA preschoolers [33, 50]. Another study found that five, 1-hour long healthy-lifestyle themed workshops presented by trained nurse childcare health consultant to parents, childcare teachers and staff, significantly decreased mean BMI-z scores in children from underserved minority families [22]. Finally, a multilevel, childcare-based intervention [36] showed a significantly smaller increase in the BMI percentile when intervention centers implemented early childcare center policies focused on modifying individual-level child, parent, and teacher behaviors with physical/built and sociocultural environment changes.

Primary care providers clinic-based.

Two primary care clinic-based studies [25, 37] reported contradictory results. Both studies targeted individual- and interpersonal-level behavioral changes, implemented in a healthcare environment. Only Cloutier and colleagues found a significantly greater reduction in BMI percentile in the intervention group that participated in bilingual, culturally adapted, motivational interviewing (MI) sessions during primary care provider (PCP) visits and phone-coaching session [25].

Community center-based.

Similarly, two interpersonal-level, family-based studies were conducted a community center-based environment with mixed results. Slusser and colleagues [38] randomized Latino mothers of preschoolers to receive nine culturally tailored, Spanish language, parent-training sessions or be in a Wait List Group (WLG). Despite reporting 33.3% attrition, the intervention group experienced a greater reduction in BMI percentile differences compared with the WLG at nine months. In contrast, Haines and colleagues [32] failed to demonstrate a significant improvement in BMI with a family-based, community health center intervention.

Other settings.

Additional preschool-aged interventions were conducted at a WIC site [40], online [39] and in the home and over the telephone [31]. Only one study reported a positive effect on BMI. In the “Healthy Habit, Healthy Home” [31] health educators used MI coaching techniques during home and phone coaching sessions, along with text messages to promote interpersonal-level changes in healthy family routines and PA, encourage family meals and beverage choices, adequate sleep and change the physical-built environment by asking families to remove TVs from the child’s bedroom. Although a WIC-based intervention within the community-wide Massachusetts Childhood Obesity Research Demonstration [51] found no effect on BMI-z scores, a post-hoc analysis excluding Asians (due to disproportionate distribution of Asian children in the comparison community) found a significant improvement in BMI-z scores [40].

Costs of obesity prevention interventions.

Table 5 includes a summary of three studies appraising the cost of preventing obesity. Cradock and colleagues [47] estimated the total annual cost per child associated with implementing and nationally disseminating the PA component of the childcare center-based “Hip-Hop Jr.” PA intervention was $22.65 yearly [33]. Also, Wright and colleagues [45] calculated the net cost of a primary-care based intervention [52], aimed at reducing obesity related behaviors and BMI in overweight and young, obese children at $196 per child [49]. In the third study, Ma and Frick [48] modeled the breakeven point of a hypothetical intervention producing a 1% reduction in the prevalence of obesity among children 0–6 years. Accounting for future medical costs, population-based interventions could cost up to $339 per child and still have a favorable health benefit/cost profile.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Cost and cost effectiveness of interventions and policies.

https://doi.org/10.1371/journal.pone.0209787.t005

Policy interventions.

Six studies [41–46] estimated the cost and health benefits of city-wide, and state and national policies aimed at preventing future weight gain and obesity (Table 5). Kuo and colleagues [41] developed a simulation model estimating the impact on annual weight gain in Los Angeles (LA) County of a California menu law mandating large restaurant chains to display the caloric content of menu items. The model assumed 10% of all customers would consume 100 calories less per meal and found the law was projected to avert as much as 500,000 lbs. of the estimated annual LA County population weight gain (1.25 million lbs.) in children 5 to 17 years old.

The Northeast Iowa Food and Fitness program enacted multilevel changes during a 6-year long program. The changes targeted schools and home meals and PA, established school gardens, and at the community-level provided access to outdoor recreational spaces and programs, local farmers markets and affordable healthy food. It was shown that children ages 4–12 years who had longer periods of program exposures (2 to 6 years) demonstrated a greater improvement in appropriate growth rate compared with children with shorter periods of program exposure (0 to 1 year) [46].

Dharmasena et al. utilized an economic demand model based on household purchasing habits to assess the impact of a 20% tax on SSB consumption, caloric intake and weight. Results using the most conservative estimate showed an overall reduction in SSB with corresponding increases in fruit juice and low-fat milk consumption. The interrelated changes in beverage consumption patterns was forecast to produce an average reduction of 449 calories per month resulting in a mean body weight reduction of 1.54 lbs/year.

Costs and cost effectiveness analysis.

Finally, three studies describe the economic impact and health consequences of obesity prevention policies using a Markov-based cohort model to estimate the cost effectiveness of: an excise tax on SSBs, [44] eliminating the tax subsidy for TV advertising, [42] and implementing a set of hypothetical childcare center-based policy changes [45], (Table 5). Sonneville and colleagues [42] found that eliminating the ability of food manufactures to deduct the cost of advertising unhealthy foods would result in a mean BMI reduction 0.028 per child [42]. Likewise, Wright and colleagues [45] estimated that a hypothetical childcare center-based policy (eliminating SSBs, limiting fruit juice, serving low-fat milk, limiting screen time and increasing MVPA) would result in a mean BMI reduction of 0.019 BMI units per child [45]. Finally, Long and colleagues [44] estimated that a tax of $0.01/ounce on SSBs would reduce the total calories consumed by children ages 2–4 and 5–9 years, by -1 to -13 kcal/day, respectively, and result in a mean reduction in BMI of approximately 0.16 units for children 2–19 years. The cost per unit BMI reduction based on these policies ranged from $1.16 for eliminating the advertising subsidy to $57.80 for the childcare center-based policy.

Discussion

This scoping review identified the characteristics and effectiveness of obesity prevention interventions, programs and policies across the early life course and at multiple levels of influence in the US. There were a number of key findings (S2 Appendix). We found slightly less than half of the interventions initiated during pregnancy, infancy or preschool were effective at improving a weight-based measure of growth or weight status in young children [17, 19, 20, 22–25, 28, 31, 36, 38]. All interventions included strategies to influence health behaviors at an individual or interpersonal level. However, few studies removed obstacles in the physical/built environment, sociocultural environment or healthcare system. The majority of the interventions were conducted in children at higher risk of obesity, in early childcare education settings. The impact of menu labeling laws, taxing SSBs and eliminating incentives for TV advertising of unhealthy foods, on a direct weight-based measure of growth or weight status in children under age 6 years remains unclear. Finally, this review confirmed the lack of available data on the cost of implementing obesity preventions efforts in the US.

We used the NIMHD Research framework [53] to guide our examination of obesity prevention efforts considering factors relevant to obesity and health equality research. We found that all interventions initiated during pregnancy and infancy focused on modifying biological risk factors of obesity by enhancing individual-level knowledge of healthy eating patterns, appropriate gestational weight gain, prolonged breastfeeding, delayed introduction of complementary feeding, responsive feeding techniques and appropriate infant growth patterns [19, 20]. For example, in the primary care setting, MI coaching techniques were applied to reduce obesogenic behaviors [25, 37], while in the childcare center setting, parents and teachers gained the necessary knowledge and skills to serve as role-models of healthy-lifestyle behaviors [19, 28, 36, 38]. Also, preschool children were encouraged to modify their behaviors using culturally adapted, nutrition and PA lessons [19, 28, 31, 36, 38], food group-themed, hand puppet activities [29], English and Spanish-language CDs [30], healthy snacks [22, 28, 36] and structured MVPA [23, 24, 28].

All effective interventions identified in this review incorporated interpersonal-level strategies affecting family behaviors and home routines. Parental and family participation was either a primary or secondary component of all successful interventions. Parents attended culturally tailored education sessions [38], home visits [17, 20, 31], or were assigned homework, received instructional handouts and newsletters promoting frequent family meals, adequate sleep, family PA and limiting screen time [20, 22, 25, 28, 31, 36, 38]. The inclusion of parent-direct strategies in successful interventions is consistent with the findings of a recent review of obesity prevention interventions in early childcare centers. Ward and colleagues [54] found that higher parent engagement in the early childcare settings enhanced the effectiveness of interventions by achieving a positive weight related outcome [54].

Recognizing the influence of physical environments on the risk of childhood obesity and health disparities is critical to the design of multilevel obesity prevention interventions. Yet, less than one-third of the studies aimed to modify a component in the physical/built environment. These interventions effectively removed barriers to facilitate healthier lifestyle behaviors. They included components such as removing TVs from bedrooms, providing alternative playtime activities, and implementing childcare centers polices limiting SSBs, serving water, low-fat milk, fruits and vegetables as snacks; increasing hours of PA, and limiting screen time [22, 31, 36]. At the community-level, Karanja and colleagues [17] enacted tribal-wide policies providing access to breast feeding rooms and reallocating resources by stocking vending machines with water and providing water coolers at community-sponsored activities. Collectively, these types of physical environment changes rendered healthy behaviors as the default behavior.

Approximately half of the interventions integrated strategies to change sociocultural environmental-level factors, with considerable variability in the intensity of the components. Few included primary objectives examining the effectiveness of culturally-tailored training programs [38, 39] or other high-intensity activities such as media campaigns encouraging drinking water and breastfeeding as cultural values [17] or establishing group sessions intended to support shared sociocultural values [19]. While the majority of studies included the following types of moderate- to low-intensity components: using bi-cultural/bilingual interventionalist [19, 25, 29, 30, 32, 36, 38, 39], culturally adapted curriculum [25, 26, 28–31, 33, 34, 36, 38, 39] and providing culturally relevant recipes [19, 25, 26, 35, 39].

Healthcare system-level changes facilitate access to healthcare resources, engage parents in healthcare decisions and improve parent-healthcare provider relationships. Furthermore, the early initiation of interventions in the healthcare setting might alter the course of health and disease, and reduce future health disparities [55]. Yet, of the 5 healthcare-based interventions [16, 19, 21, 25, 37], only two reported improvements in a measure of obesity. One was informed by the chronic care model and used brief MI format [25], while the other used a patient-centered approach and group sessions [19]. In contrast, others failed to improve childhood growth trajectories [16, 21, 37]. Such results suggest that even in a healthcare setting, the intensity of the program and study population are important considerations for the success of any intervention.

This review identified studies that estimated the health and economic impact of regional and national SSBs pricing strategies, labeling laws and food marketing policies. These strategies were shown to change purchasing behavior and improve the prevalence of obesity, and potentially to generate tax revenue, drive the reformulation of unhealthy foods, and change social norms [42, 44]. Although the population reach and societal-level impact of obesity prevention policies are high, the long-term outcome on the prevalence of early childhood obesity remains uncertain.

None of the included interventions reported on the costs or cost effectiveness of the study. The lack of economic evaluations is a surprising finding given that the annual cost of obesity-related medical spending was estimated to exceed $147 billion [56]. Consistent with our findings, Wolfenden and colleagues [57] noted that 88% of the systematic reviews of obesity prevention intervention in children did not report whether a cost analysis has been conducted. In the absence of CEA data, a reliable cost threshold could assist stakeholders to determine the amount of money to spend on obesity prevention interventions. Ma and Frick [48] established the break-even cost of $339 per child as a “good value” for interventions resulting in a 1% reduction in childhood obesity prevalence.

There were several limitations to our review. First, our scoping review presented a comprehensive overview of the quantity and context of current childhood obesity prevention efforts in the US, which limits the generalizability of our findings to other countries. Next, our review did not identify any obesity prevention interventions conducted during pre-pregnancy and identified very few studies conducted during pregnancy. It is likely that our search strategy may not have been sensitive enough to identify the full breadth of research activities during these early life stages. Since we followed standardized methodology for scoping reviews, we did not assess the effect size of the interventions or systematical evaluate the quality of the individual studies including the risk of bias quality. We acknowledge that many of the studies were of varying quality based on study design, sample size, intensity, analytical approaches, high attrition rates and low parent attendance rates. Although we excluded studies judged to be fatally flawed, our results are subject to a range of biases due to these many threats to the internal and external validity of the study results. Furthermore, given that studies reporting positive results are more likely to be published, conclusions may also be subject to publication bias [44]. Finally, our synthesis was based on interventions reporting a weight-based measure of growth or weight status in normal weight children. Because we did not consider favorable behavioral or PA outcomes the generalization of our results may be limited.

Conclusions

This review presents an overview of the current state of obesity prevention efforts across multiple levels of influence and the early life course in the US. The majority of efforts focused on individual and interpersonal-level health behavior changes in preschoolers. Thus, there is a need to intensify obesity preventive efforts during critical periods of health development and target multiple levels of influence, especially regarding physical, sociocultural and healthcare system-level obesity risk factors. Furthermore, there is considerable uncertainty around estimates of the economic impacts of obesity prevention interventions and policies. Future studies should estimate the feasibility, effectiveness, and cost-effectiveness of obesity prevention interventions and policies. Addressing these research gaps may provide government agencies, policy makers, and healthcare payers with the necessary scientific evidence to make informed decisions regarding the allocation of funds for initiatives aimed at decreasing the prevalence of early childhood obesity.

Supporting information

Acknowledgments

We thank Mina Ghajar (Librarian, George F. Smith Library of the Health Sciences, University Libraries Rutgers) for her assistance with the literature search strategy.