A systematic analysis of childhood obesity prevention interventions targeting Hispanic children: lessons learned from the previous decade
P Branscum, Graduate Assistant, Health Promotion & Education, University of Cincinnati, PO Box 210068, Cincinnati, OH 45221-0068, USA. E-mail: firstname.lastname@example.org
Hispanic children suffer from the highest overall rates of prevalence for overweight and obesity in the US. In the last decade some interventions for prevention of childhood obesity have been developed and tailored to target this subgroup. The purpose of this review is to systematically analyze and summarize findings for health education and promotion interventions aimed at the prevention of childhood overweight and obesity among primarily Hispanic children. A systematic review of PubMed, CINAHL, and ERIC was done for the time period 2000 to May 2010. A posteriori effect size for the primary outcome of each intervention was calculated using G*Power. A total of nine interventions were located; five randomized controlled trials and four were either quasi-experimental or pilot studies. Among these studies, only four had significant findings, and calculated effect sizes (Cohen's f) ranged from small to medium with the highest f = 0.26. Interventions were more likely to be successful when participants were at higher risk for obesity, a parental component was included, the intervention contained theoretical underpinnings, the intervention was delivered by a dedicated staff, the intervention served older children and the intervention was longer in duration. More interventions need to be developed for Hispanic children. Future interventions should also develop and utilize culturally appropriate and sensitive materials and approaches.
Childhood overweight and obesity continue to be an epidemic in the United States. While all children experience metabolic and psychological consequences because of this problem, there are notable disparities among minority groups. According to the recent National Health and Nutrition Examination Survey completed in 2006, 16.3% of all children between the ages of 2 to 19 years were found to be obese (≥95th percentile) and 31.9% were overweight (≥85th percentile). When compared across racial and ethnic groups, children with a Hispanic background suffer from the highest overall rates (20.9% obese and 38% overweight), especially among Hispanic boys (23.2% obese and 40.8% overweight) (1).
Hispanic Americans are the largest minority group in the US, comprising of 15% of the current population (46.9 Million). Hispanic Americans are also the fastest growing minority group and it is expected that by 2050 they will make up 30% of the US population (2). The term Hispanic does not refer to a racial category, rather it refers to an ethnicity that is made up of descendents from over 25 Spanish speaking countries. The most common countries of origin for US Hispanics include Mexico, Central and South America, Puerto Rico and Cuba (3).
Hispanics currently have higher rates of poverty, food insecurity and obesity compared with Caucasians (4). As reviewed by the Latino Consortium of the American Academy of Pediatrics Center for Child Health Research, compared with Caucasians children, Hispanic children experience disparities in mental health, oral health, overweight, obesity, type 2 diabetes, asthma, and school drop out rates (5). Hispanic children reportedly have higher rates of depression, anxiety, fears and phobias. They are also more likely to develop dental carries, and less likely to have them treated or filled. Type 2 diabetes is another problem Hispanic children face, with data suggesting Hispanics make up 45% of all newly diagnosed diabetes cases. Currently, approximately half a million Hispanic children have asthma. Notably, Puerto Rican children have higher rate of asthma (11%) than any other racial/ethnic group. Finally, Hispanic teens have a much higher high school dropout rate (29%), compared with Caucasians (7%) and African Americans (13%) (5).
There are also disparities for accessing health care, with adult Hispanic patients being less likely to receive mammograms, Papanicolaou test, influenza vaccinations, prenatal care and cardiovascular procedures (6). Hispanic children also experience inequities in accessing health care services. Compared with Caucasian children, Hispanic children are twice as likely to have no health insurance, are referred to specialist at half the rate (11%) than Caucasian children (22%), and Hispanic parents often report their primary care physician either does not usually understand or slightly understand their child's health needs (7). However, unlike Caucasians, there does not appear to be an Socio-economic status (SES) gradient for Hispanic health. Using a nationally representative sample of Caucasian and Hispanic children, Balistreri and colleagues (2009) demonstrated a clear negative relationship among Caucasian children's body mass index (BMI) and parents' education attainment and income; however, this relationship was not found among Hispanic children. BMI was lowest among higher income Hispanic children; however, it was highest among middle-income Hispanic children. This lack of SES gradient suggests that in upcoming years as Hispanics experience economic gains, improvements for health and BMI status may not accompany (8).
It is well known that overweight and obesity in childhood is problematic for many reasons. First, overweight children are more likely to remain overweight into adulthood, placing them at risk for chronic metabolic conditions later in their lives. Overweight children also experience negative metabolic and psychosocial problems at an early age, which are more likely to engage in more risky behaviours such as tobacco and alcohol use (9). Health education and promotion efforts have the potential to curb this problem, however many studies often focus on a primarily Caucasian sample. This purpose of this review is to systematically analyze and summarize findings for health education and promotion interventions in the last decade aimed at the prevention of childhood overweight and obesity among primarily Hispanic children.
An extensive literature search was conducted to collect studies for inclusion in this article using the databases PubMed, ERIC, and CINAHL. Keywords that were used included ‘Hispanic’, ‘Latino’, ‘childhood obesity’ and ‘intervention.’ Inclusion criteria for including studies in this review were: (1) publication in English language (2); a primary research article evaluating any form of intervention strategy for the treatment or prevention of childhood obesity (3) publications between 2000 and May 2010 and (4) the primary audience for intervention was Hispanic, Latino or Mexican American. Exclusion criteria were articles in languages other than English, and review articles. To evaluate the a posteriori effect size for the primary outcome of each intervention, Cohen's f was calculated using G*Power version 3.1.2 (10). Cohen's f-values will be interpreted using the following criteria: 0.1 to 0.24 represents a small effect size, 0.25 to 0.39 will represent a medium effect size, and greater than 0.4 will represent a large effect size.
A total of nine studies were found for inclusion in this review. Table 1 summarizes five of the nine interventions that were all randomized controlled trials and Table 2 summarizes the remaining four interventions that were either quasi-experimental or pilot studies. Elements of research design and study demographics included in each table include: intervention name, percentage of Hispanic children in the study, theory used, duration of the intervention and salient findings.
Table 1. Summary of prevention and treatment interventions used to combat obesity among Hispanic children/adolescents, using a randomized controlled design
|1a.||Intensive intervention (II) (13)|
All overweight 12.5 (0.6) years
|Behaviour theory||Randomized control trial|
n = 71
(tx = 46)
(cnt = 25)
|Weekly lessons on diet/nutrition (1 day) and physical activity (4 days)||24 weeks||Significant reduction in BMI z-score and percentage body fat at 3- and 6-month follow-up.|
|1b.||Intensive intervention (II) (14)|
12.4 (0.7) years
|Behaviour theory||Randomized control trial|
n = 60
(tx = 40)
(cnt = 20)
|Weekly lessons on diet/nutrition (1 day) and physical activity (4 days)|
Children received a healthy snack every day during the programme
|24 weeks||Significant reduction in BMI z-score at 3- and 6-month follow-up.|
Significant reduction in total cholesterol and LDL cholesterol compared at the 6 month follow-up.
|1c.||Intensive intervention (II) (15)|
|Behaviour theory||Randomized control trial|
n = 80
|Weekly lessons on diet/nutrition (1 day) and physical activity (4 days)||24 weeks||Total quality of life (QOL) significantly increased for both II and SH group, but only the II group significantly increased the ‘physical’ QOL subscale at the 6-month follow-up.|
Significant improvements in BMI z-score accounted for the improvement in physical QOL for the II group.
|2.||Hip hop to health Jr.(17)|
4.2 (0.6) years
|Social cognitive theory||Randomized control TRIAL|
n = 401
(tx = 202)
(cnt = 199)
|Three 40-min. sessions per week.|
20-min. on healthy eating behaviours.
20-min. engaged in a physical activity.
Parent sessions included weekly newsletters (12 weeks)
|14 weeks||No differences between treatment and control group for BMI, physical activity or dietary behaviour measures.|
Parent participating was relatively low (54% returned at least 1 activity)
|Social cognitive theory||Randomized control trial|
n = 896
(tx = 423)
(cnt = 473)
|Physical education and school cafeteria components of CATCH implemented||3 years||The rate of ‘overweight’ increased higher among non-CATCH schools.|
Physical fitness measures and times spent in MVPA and VPA were similar across for CATCH and non-CATCH children.
CATCH schools met the fat recommendations for lunches in the 4th grade only, but never met the sodium recommendations.
Table 2. Summary of prevention and treatment interventions used to combat obesity among Hispanic children/adolescents, using a quasi-experimental design
10.8 (1.2) years
38.9 (8.2) years
|Social cognitive theory||Quasi experimental|
n = 37 dyads
Daily sessions included an exercise, nutrition education and behavioral counselling component.
Sessions taught parents how to incorporate healthy lifestyles for daughters at home.
9:00 h–17:00 h
2 h per session
|Immediately following the intervention:|
Significant reduction of body weight, BMI, and waist circumference.
Significant improvement for physical fitness via 20-meter shuttle run.
|2.||Get Moving! (12)|
12.5 (0.6) years
|Self-determination theory & the theory of meanings of behaviour||Quasi experimental|
n = 459
(tx = 136)
(cnt = 323)
|A media based physical activity intervention through which classes developed PSA's to promote physical activity||Five to seven consecutive classroom session||Significant decrease in sedentary activities, but no changes for physical activity.|
Significant increase for intrinsic motivation.
Decrease in sedentary activities was not mediated by intrinsic motivation.
|3.||Increased 60 min of Recess (16)|
3.5 (0.5) years
n = 33
(tx = 18)
(cnt = 15)
|Children in treatment group were given 60 min of extra recess time for two consecutive days.||2 days||Increased recess time did not significantly impact the amount or intensity of physical activity.|
|4.||Unnamed physical activity programme (19)|
12.6 (1.0) years
|Self-perception theory||Pilot study|
n = 30
(tx = 30)
|Three physical activity sessions/week|
Included warm-up, stretching, cool down exercises and rest breaks.
|12 weeks||Significant improvements for fitness measures (i.e. 1-mile run, flexibility), BMI and triceps skinfold|
Significant improvements for all self-perception subscales, except physical appearance and global self-worth.
Tables 3 and 4 show the effect sizes from interventions on all primary outcome measures. All but one study evaluated some measure of weight status (BMI, BMI-percentile or BMI z-score) and among these studies, only four had significant findings. Calculated effect sizes (Cohen's f) ranged from small to medium with the highest f = 0.26.
Table 3. Calculated effect sizes for primary outcomes for reviewed randomized controlled interventions
|1a.||Intensive intervention||z-BMI||n = 71||Repeated measures anova||0.019||0.80||0.17|
|1b.||Intensive intervention (II)||z-BMI||n = 60||Repeated measures anova||0.001||0.80||0.24|
|1c.||Intensive intervention (II)||z-BMI||n = 80||Repeated measures anova||0.001||0.80||0.21|
|2.||Hip hop to health Jr.||BMI-percentile||n = 401||Repeated measures anova||–||–||–|
|3.||CATCH||BMI-percentile||n = 896||SAS Proc Glimmix mixed model||–||–||–|
Table 4. Calculated effect sizes for primary outcomes for reviewed quasi-experimental interventions
|1.||BOUNCE||BMI||n = 31||Repeated Measures anova||0.05||0.80||0.26|
|2.||Get moving!||BMI-percentile||n = 459||Repeated Measures anova||–||–||–|
|3.||Increased 60 min of recess||Physical activity||n = 33||Wilcoxon Rank Sum/Mann–Whitney U-Test||–||–||–|
|4.||Unnamed programme||BMI||n = 30||t-test||–||–||–|
Tables 5 and 6 show intervention characteristics, which suggests why these four studies were successful. As reviewed by Stice and colleagues (2006), childhood obesity interventions are typically more successful when participants are at higher risk (primarily overweight or obese), involve a parent component, are theory-based, have a dedicated intervention staff, have older children and have longer intervention periods (20). All four studies with significant findings were studies in which all children were overweight or obese. This is expected since overweight and obese children have more weight to lose than normal weight children. Also, these children would likely have more motivation to adopt healthier lifestyles, because their perceived susceptibility is likely higher than their normal weight peers.
Table 5. Design features of reviewed randomized controlled interventions
|1a.||Intensive intervention||X||X||X||X||X||12.5 (0.6)||66 sessions|
|1b||Intensive intervention||X||X||X||X||X||12.4 (0.7)||66 sessions|
|1c||Intensive intervention||X||X||X||X||X||12.1||60 sessions|
|2.||Hip hop to health Jr.|| || ||X||X|| ||4.2 (0.6)||42 sessions|
|3.||CATCH|| || ||X||X|| ||8.3||3 years|
Table 6. Design features of reviewed quasi-experimental interventions
|1.||BOUNCE||X||X||X||X||X||10.8 (1.2)||15 sessions|
|2.||Get moving!|| || || ||X||X||12.47 (0.63)||5 to 7 sessions|
|3.||Increased 60 min of Recess|| || || || || ||3.5 (0.5)||4 sessions|
|4.||Unnamed programme|| || || ||X||X||12.6 (1.0)||12 weeks|
The purpose of this study was to evaluate current health promotion and education interventions targeting the prevention of Hispanic childhood obesity. Few interventions have been implemented in the past decade with a primary focus on the prevention of Hispanic childhood obesity. Furthermore, the experimental rigor in the studies reviewed in this article greatly varied as some were multi year, large-scale randomized controlled trials and others were brief, small-scale pilot studies. Results from larger scale Randomized controlled trial (RCT)'s were generally more favourable, and conclusions drawn from non-RCT studies should be interpreted with caution.
Few studies showed significant improvements for changes in weight status among children. This is not surprising because less intensive approaches were commonly used and implemented. Studies also rarely employed follow-up evaluations, which would be ideal to show whether the effects of the intervention truly prevented an increase in the incidence of overweight and obesity in subsequent months or years. Two studies used absolute BMI (kg m−2) which is not recommended for children, given they are expected to grow to some extent as they mature. Indices such as BMI-percentiles, which can be collected from Centers for Disease Control and Prevention (CDC)'s growth charts, or BMI z-scores should always be used. Studies that aim to prevent overweight and obesity among children should also consider evaluating behaviours related to weight maintenance and antecedents of behaviours, such as self-efficacy or self-control. Overall, there were few measures other than weight status used throughout these studies, which included: physical activity (five studies), physical fitness (three studies), blood pressure (three studies), psychosocial measures (two studies) and dietary behaviours (one study).
All four studies with significant findings had a parent component as part of the intervention. A parent component in any childhood intervention is ideal, because children cannot make many important health related decisions for themselves. For example, children usually have little control on the restaurants they visit, school lunches that are made for them (in and out of the school cafeteria), foods available in their household, and opportunities to engage in physical activities. Children should also be taught how to interact with their parents on how to get them to act on their behalf with regards to these health issues. For example, children can learn negotiation skills and how to talk with their parents about the healthy foods they prefer to eat or physical activities they could engage in together.
Theory-driven interventions are also ideal, but not sufficient as evidenced in this review. While eight interventions were theory-based, only half had significant changes in weight status. Only two studies measures psychosocial constructs of the theory they utilized (12,19). Future studies would benefit by including psychosocial measures.
A dedicated intervention staff was another important characteristic for studies with significant outcomes. School-teachers are currently leaned upon to prepare their children for standardized tests, which could demotivate them to engage in extracurricular subjects, such as health and wellness. Furthermore, professionals that deliver health interventions typically have specialized training in this area, are motivated to do a good job, and can put more time and energy into refining the intervention throughout.
Interventions with significant outcomes also had older children compared with most of the other studies. Age ranges for these studies were approximately 10–12 years old children, and age ranges for the remaining studies included 3–12 years old children. As mentioned previously, children typically do not have much control over their immediate environment; however, as they get older they start to show more independence, including how they spend their free time and what foods they eat. They especially have more control over snack foods or foods not eaten in major meals they eat with their parents/guardians. Older children also have a better ability to understand more complex concepts and skills. No interventions were conducted with Hispanic high school adolescents. It is unclear why this is the case; however, as previously noted approximately a third of Hispanic teens drop out of high school, which may make it difficult to target this group at that age.
Interventions with significant results were also those that had a longer implementation period. Besides evaluating weeks, or months we decided to evaluate number of sessions and total contact hours during the intervention. This was done because while some studies were shorter, the total contact hours were actually high. For example, the Bounce programme only lasted 3 weeks, but the total intervention lasted 15 days for 8 h per day (11). It was also impossible to determine the actual length of intervention for some studies, including CATCH, because the programme was integrated into their normal school curriculum.
Finally, cultural influences may help explain why some studies failed to have a significant impact. For example, it has been noted that Hispanics do not perceive thinness as ideal and fatness as unhealthy as Caucasians do, and being overweight is looked upon as a sign of health and strength among many Hispanics (3). Hispanic parents could also want to indulge their children as they make economic gains. Studies suggest as Hispanics become more acculturated, their diet quality tends to suffer. It has been documented that as acculturation increase, Hispanics consume less fibre, fruits, vegetables and whole milk and Hispanic mothers are less likely to breast-feed (3). Acculturation may also explain why English speaking Hispanics report better health care than non-English speaking Hispanics. Fiscella and colleagues (2002) reported that English speaking Hispanics and Caucasians did not differ in terms of likelihood of having a physician visit, mental health visit or flu shot; however, non-English speaking Hispanics were significantly less likely to have any of these services (6). Parent acculturation was only measured in one study in this review (17). This could potentially be a barrier while working with Hispanic children. Future studies may benefit from evaluating acculturation as a possible confounding variable. However, this may be difficult, as acculturation among Hispanics is becoming difficult to measure and explain as more cities in Mexico and Latin America are urbanized and influenced by American culture.
Implications for future studies
As childhood overweight and obesity increases, especially among minority groups such as African Americans and Hispanics, so does the need for innovative intervention strategies to combat this problem. Few studies have targeted a primarily Hispanic audience in the past decade for childhood obesity prevention and among these few studies, only two interventions appeared to have a great impact. The Bounce programme was not a school-based programme, rather a summer programme that met eight hours per day, for three weeks (11). While this strategy is ideal for preventing already overweight children from further weight gain, it may not be feasible to implement on a larger scale. The other three studies with significant findings related to weight status were all done with an intensive intervention (II) during school hours (13–15). This intervention may serve as a model for future health promotion and education endeavours.
Future interventions should target both physical activity and dietary behaviours. Important dietary behaviours include increasing fruit and vegetable consumption, decreasing fat intake, decreasing the consumption of sugar-sweetened beverages, having adequate consumption of water and/or non-caloric beverages and restricting portion sizes of meals and snacks. Children should also be encouraged to participate in 60 min of moderate to vigorous physical activities on most days of the week. For studies that include these targets, evaluation should also be a priority. Dietary behaviours can be evaluated using comprehensive measures (i.e. food diaries, 24 h recalls) or brief instruments that have been previously validated (i.e. SPAN (21)). Physical activity is similar in that researchers can choose from a variety of comprehensive measures (i.e. accelerometer's, physical activity diaries) or brief validated instruments (i.e. PDPAR (22)).
Theories should also be better operationalized and evaluated for future studies. As previously mentioned, while most studies cited using a theory for their intervention, few effectively operationalized theoretical constructs to target during the course of the intervention, and few effectively evaluated these constructs. Popular theories that have been used in obesity prevention research include social cognitive theory and the theory of planned behaviour. Important constructs that overlap in these theories and would be important to target include self-efficacy (an individuals confidence in their ability to enact a behaviour or overcome a barrier), proxy-efficacy (an individuals confidence in their ability to get others to act on their behalf to reach a desired outcome), and self-control (an individuals ability to self-regulate their behaviours).
An important aspect that was present in studies was that intervention materials were usually culturally appropriate. This included having diet recommendations that included ethnically relevant foods and suggestions for physical activities. In fact, a limitation that Heath and Coleman (2002), noted was that some aspects of CATCH were not able to be implemented because CATCH materials were not yet translated and modified (23). This could have contributed to their non-significant findings. Future interventions for Hispanic children need to develop and utilize culturally appropriate and sensitive materials and approaches.
Conflict of interest statement
No conflict of interest was declared.
We are thankful to our programme and University for allowing us to work on this project.