Children today have fewer opportunities to be active in a safe and independent manner, especially in large cities in developing countries that are rapidly urbanizing (11). According to a birth cohort study conducted in Pelotas, Brazil, the prevalence of sedentary lifestyles, characterized by excessive TV viewing and lack of PA, among youth is as high as 58% (12). Factors that decrease energy expenditure, such as the declining time for physical education (PE) in schools, may play an important role in the prevalence of overweight among children. Because students spend large amounts of time in school, there is a great potential for increasing their level of PA through school-based interventions (13,14). In a systematic review of 14 evidence-based PA promotion strategies, the US Guide to Community Preventive Services (Community Guide) found strong supporting evidence that school-based PE programs increased PA among school children (15). The school-based PE interventions included modified PE classes, generally with more classroom time and more moderate-to-vigorous PA (16).

Relative effect sizes for the school-based PE intervention studies were calculated as the net percentage change from baseline for all reported measures of aerobic capacity and physical activity levels and for all time periods. The formula for calculating the net effect (Figure 1) varied depending on the study design – with or without a control group or pre-intervention outcome measurement (16). To identify common characteristics of the school-based PE interventions, the authors reviewed each of the school-based PE articles and abstraction forms in detail, categorizing and listing intervention components. These ‘core components’ were determined according to frequency of use by most of the studies and also by taking into account the intervention activities recognized by the Community Guide (16).

Two studies (23,24) reported minutes spent in moderate and vigorous PA, and one study (20) reported four different outcomes related to behavior: percentage of children biking or walking to school, estimated energy expenditure based on observed PA behavior of children, percentage of children who were very active during PE classes and percentage of time spent walking during PE classes. Further reported outcomes were aerobic capacity as estimated by maximal oxygen uptake (VO₂ max) (21), results from timed runs (23) or endurance testing (shuttle runs) (22).

Although the outcomes, ascertainment methods, and duration of interventions varied between studies, there were consistent positive increases in PA levels for all outcomes measured during PE classes (relative net effects ranged from 7% to 307%) in all three randomized studies (20,21,23) (Figure 1). The extreme magnitude of the 307% net effect is explained by the low baseline walking levels in the intervention group versus the control group (5.6 vs. 14.4%, respectively). At post-test the intervention group increased time spent walking during PE classes nearly 5–6 times that of baseline, while the control group increased walking time by only 1.5 times baseline. Likewise, endurance (23) and active transportation to school (20) outcomes showed small but positive net effects for the same group of studies. With the exception of one cohort in Heath (24), the non-randomized studies increased moderate and vigorous PA levels during PE classes. These increases ranged from 5% to 69% for moderate and 50% to 55% for vigorous PA levels during PE classes. The results from the shuttle run test (22) demonstrated increased endurance performances among boys (35%) and girls (37%).

After reviewing the details of each study (Table 1), four types of intervention activities were identified and summarized as the ‘core components’ presented in Table 2. Core components included the provision of information about the benefits and importance of health and active living (20,22); types of PA suitable for a particular age range (20); overall PE class planning (lesson plan development, class and environment management) and methods of PA practice (23,24); increases in the duration of the already scheduled PE classes (23,24); addition of extra PE classes per week (21,22); use of age-specific and enjoyable activities (20–24) (such as balance exercises, dancing, jumping, and running); availability of new materials for classes (balls, basketball boards) (20–22); and adaptable, flexible and culturally sensitive activities based on ethnicity or socioeconomic status (20,22–24).

High-quality school-based PE programs, combining all or some of the core components described here, increase PA among children in Latin America in the same way as they do in the USA (18). Similar to the GUIA literature review (18), the 13 US qualifying studies revealed consistent increases in time spent in PA at school (16), The net increase in the amount of PE class time spent in moderate and vigorous PA was 50%, ranging from 6% to 125%. The percentage of class time increase in moderate and vigorous PA was 10%, ranging from 3% to 15%, although one study reported a 762% increase from a very small baseline value. Two studies showed increases in energy expenditure, and 11 studies reported increases in aerobic capacity with a median of 8% (interquartile range 3% to 19%) (16).

When comparing intervention effects for the same outcomes, the five studies from the present study are generally consistent with the US CG results, both in direction and effect size. Some of the US CG net effects also had a wide range, particularly for the amount and percentage of PE class time spent in moderate and vigorous PA. The three studies that reported increased levels of PA during PE class (20,23,24) used the same observational method (16) to measure the amount of PE class time spent in moderate and vigorous PAⁱ, but for different outcomes, such as the time spent being very active and the amount of walking during PE classes (20). The wide range in net effect sizes reflects differences in the degree of change in the outcome measures for intervention and comparison (if applicable) groups before and after the intervention. They also reflect the magnitude of the pre-intervention measures (see Figure 1 note).

Insufficient evidence supported recommending classroom-based health education to increase PA, mainly because of an insufficient number of qualified studies (n = 3) (18). These interventions consist of multicomponent health education classes in elementary, middle, or high schools that aim to help students to develop the skills they need to make rational decisions about adopting healthier behaviors (16).

Among the three reviewed studies, the net effects for various outcomes (e.g. frequency of exercise-related behaviors, any moderate and vigorous PA in the previous week) ranged from 0.0 to 0.2; net effects for 11 of the 14 outcomes were between 0 and 0.1. These net effects tended to be much lower in magnitude than those observed for the school-based PE interventions, which suggests that merely providing health information and promoting behavioral skills for healthy decision-making may be insufficient to significantly change PA behavior. However, more research of classroom-based health education, both in the USA (16) and in Latin America, is needed to confirm this finding.

The impact of an intervention is determined not only by its reach and efficacy but also by the extent to which the intervention is adopted, implemented as intended and maintained at the system and individual level (25). An effective intervention needs to have broad reach and be feasible to implement in ‘real-world’ settings in order to make a public health impact (25). Some of the major challenges for interventions are finding creative and cost-effective ways to implement evidence-based programs and policies at the population level, while providing adequate information and resources (e.g. infrastructure, training, proper economic and social environments) (26). This is particularly important in a region where legislation to promote school-based PE is highly variable. In addition, it is key to consider differences in terms of social influences that may affect children's and adolescents’ behavior (e.g. family, social networks, culture) when adapting programs that are effective in a specific setting or population (27). Although the purpose of school-based PE is larger than enhancing physical fitness only, it also has great potential for PA promotion because of its effectiveness in increasing PA among children (18) and because school programs reach virtually all children at relatively low cost with preexisting infrastructure.

In a context of increased academic and political credibility for PA promotion and PE during the past decade (28), supportive school PE policies appear to exist in many Latin American countries (29,30). In these countries, even though PE is mandatory for students in public and private elementary and secondary schools, the requirements concerning frequency and duration of classes is generally low (31). An optimal PE curriculum should include a variety of learning and practice experiences that will provide the basis for active living in the adult years (32). However, according to experts in the area, PE programs in Latin America are often of low quality and quantity. Therefore, a comprehensive school policy to increase PA practice during school-based PE and enhance its quality is required. In addition, frequency and duration of classes is often less than optimal because of challenges related to salary, work environment, physical space and curriculum structure (31); consequently, barriers to implementation exist primarily within the school systems and need to be addressed as well.

According to Melo (28), there is a call for academic institutions where PE teachers are trained to push for an adequate transition from what is learned in the academic world to what is needed in everyday practice in the schools. Often, the fields of academic knowledge production (e.g. universities/research institutes) and policy formulation/implementation are very different, frequently with incompatible goals and methods (33). Therefore, collaborative partnerships between researchers and practitioners, along with regulatory and legislative enforcement and political commitment and support for PE, are particularly desirable to help bridge the gap between theory and practice and for successful and sustainable program implementation.

Some limitations of this review should be considered. The five identified studies applied diverse ways to measure similar or same outcomes (e.g. moderate and vigorous PA levels during PE classes), which limited the ability of the previous review to derive a summary measure for the effect of school-based PE across the five reviewed studies. The net effect plot is an effort to be transparent in illustrating the variability in measures and lack of a gold standard in PA measurement in kids. By computing the net intervention effect, we were able to standardize and compare the effects. Although the literature search was conducted to identify interventions delivered in Latin America, we discuss findings from only five studies, two of which were implemented on the US/Mexican border (23,24). However, because these interventions took place in schools attended mainly by Hispanic children of low socioeconomic status, and since information on quality or availability of equipments and facilities was not available in the reviewed papers, we considered this population and setting comparable to those from other Latin American countries. The three remaining studies were limited to Chile (21,22) and Brazil (20). This limitation should be taken into account when disseminating school-based PE interventions in Latin America. Another consideration is the variability in the interventions’ duration (4 months to 3 years). The two longest studies (23,24) are funded adaptations of a broad, well-designed, tested and documented intervention studyⁱⁱ, which could justify in part the length of these two interventions. The three remaining studies (20–22) were much less supported and were single initiatives largely developed by academic institutions (University of Chile, Catholic University of Chile and Federal University of Sao Paulo) in collaboration with local health or education sectors. Nevertheless, the two longest interventions (23,24) did not result in larger net effect for similar outcomes when compared with the study with shortest duration (20).

Identifying interventions to increase PA levels in school settings is particularly important in Latin America because of the rapid demographic, epidemiologic and nutritional transitions associated with the increase in risk factors for chronic disease, including obesity and physical inactivity (4). These problems are unequally distributed across populations, and they disproportionately affect the most impoverished Latin Americans. Given the limited literature on evidence-based interventions for PA promotion in Latin American countries, developing recommendations to address physical inactivity among youth is essential.