1. Introduction
Non-communicable diseases (NCDs) are the leading cause of mortality worldwide, including in low- and middle-income countries (LMICs) [
1,
2]. Physical inactivity and sedentary behavior are well-known risk factors for NCDs. It was estimated that lack of physical activity could contribute up to 6–10% of the NCD morbidity (e.g., coronary heart disease, type 2 diabetes, and breast and colon cancers), and 9% of NCD premature mortality [
3]. It has been estimated that physical inactivity in LMICs is responsible for 75.0% of the 13.4 million disability-adjusted life years globally [
4], and life expectancy could increase by 0.68 years if physical inactivity was eliminated [
3].
In youths, physical activity is beneficial to metabolic health in many ways, including the reduction of abdominal fat, blood pressure, blood glucose, and arterial stiffness, and increase in blood HDL cholesterol [
5]. In addition, regular physical activity during childhood has a favorable effect on several outcomes in adulthood, including reduction of the risk of cardiovascular disease [
6,
7]. One previous review of the Global School-based Health Survey (GSHS) conducted between 2003 and 2007 in 34 countries showed that only a minority of adolescents (23.8% of boys and 15.4% of girls) met the World Health Organization (WHO) physical activity target (≥60 min/day), and more than one third of children engaged in sedentary behavior for ≥2 h per day (e.g., watching television or playing computer games) [
8]. Based on the GSHS in 2001 to 2016, another study reported that ≥80% of students aged 11–17 years were physically inactive [
9]. However, physical inactivity and sedentary behavior can have detrimental effects on health independently of each other [
10], and information on sedentary time was not included in this study. A review of 130 surveillance studies showed ≥50% of children and adolescents engaged in sedentary behavior for ≥2 h/day in recent years [
11]. Based on the GSHS in 2006–2016, it was estimated that 26.4% of adolescents had sedentary behavior for ≥3 h/day during their leisure time [
12].
Member states of WHO agreed on a global target of a 10% relative reduction in physical inactivity between 2010 and 2025 for the prevention and control of NCDs [
13]. In addition, WHO has developed a new global plan for physical activity in youth, which includes enhancing physical education and school-based programs, walking and cycling to school programs, and improving access to public open spaces, among several other measures [
14]. Therefore, it is important to regularly update estimates on the prevalence of physical activity and sedentary behavior in children and adolescents in all countries in order to monitor and guide effective interventions to increase physical activity and decrease sedentary behavior.
There is a relationship between the socioeconomic status of a country and levels of physical activity [
15] and sedentary behavior [
16]. For example, Wang et al. found that people tended to engage in more physical activity in economically advanced regions in China, and the correlation coefficient between physical activity and gross domestic product (GDP) per capita was 0.23 for men and 0.15 for women [
15].A systematic review and meta-analysis including 39 countries found that the prevalence of physical inactivity differed according to country GDP per capita. For example, the prevalence increased with GDP per capita in LMICs, but decreased in high-income countries [
16]. However, Guthold et al. did not find a consistent pattern of physical inactivity according to country income (low, low-middle, upper-middle, and high level) [
9]. Furthermore, the socioeconomic status of a country was assessed based on the GDP per capita in most studies. However, the purchasing power parity (PPP) per capita may be a better indicator of a country’s economic development than GDP per capita, because PPP is adjusted for the living costs and inflation of a country [
17].
Therefore, using the most recent GSHS data, we assessed the prevalence of physical activity and sedentary behavior, and their associations with a country’s economic development (measured with PPP/capita) in young adolescents aged 12–15 years in 68 LMICs.
4. Discussion
Overall, the prevalence of both sufficient physical activity and low sedentary behavior was relatively low among young adolescents aged 12–15 years in LMICs. There was an upward trend for both physical activity and sedentary behavior along increasing country PPP/capita quintiles. In addition, the relation to country PPP/capita quintiles(i.e., analysis done with countries as unit) showed that the prevalence of sufficient physical activity (≥1 h per day) decreased between the first and second PPP/capita quintiles and then increased along increasing PPP/capita quintiles (i.e., a J-curve shape). The prevalence of low sedentary time decreased with the increase of PPP quintiles among both sexes.
Previous studies have reported a low prevalence of sufficient physical activity in young adolescents worldwide [
9]. In 2011, a study using data from 105 countries showed that only 19.7% adolescents met the recommended level of ≥60 min/day of physical activity [
25], which is similar to our results. Steene-Johannessen et.al. found that only 29% of children and adolescents aged 2–18 years were sufficiently physically active (using the same criterion of ≥60 min/day) among European countries [
26]. Another previous study using the GSHS data from 34 countries between 2003 and 2007 showed that, among adolescents aged 13–15 years, 23.8% of boys and 15.4% of girls met the recommendation of sufficient physical activity [
8], which is consistent with our findings that boys are more active than girls. The sex difference might be due to environmental factors and gender norms. For example, unlike boys, girls’ outdoor physical activity can be associated with street violence [
27], and passive road safety was significantly associated with the increase of physical activity among girls [
28]. It seemed that girls may receive more parental restrictions for engaging in exercise, and exercise by girls may be more easily affected by environmental factors [
29]. In addition, the road environment (e.g., traffic/pedestrian lights, residing on a cul-de-sac) could influence adolescents’ physical activity [
30]. These findings suggest that global gender specific strategies on environmental barriers could improve physical activity among adolescents.
Few studies have assessed the prevalence of sedentary behavior in young adolescents using global data, and little is known about this question in LMICs. Our findings suggest a high prevalence (35.4%) of sedentary behavior (>2 h/day) among young adolescents aged 12–15 years in LMICs, which is consistent with a review based on 130 surveillance studies [
11]. A recent meta-analysis showed that 3–4 h/day of TV reviewing was associated with a greater risk of all-cause, cardiovascular, and cancer mortality, as well as incident type 2 diabetes [
31]. Rees-Punia et al. reported that the risk of premature death could be reduced if the amount of sitting time was replaced by physical activity, even moderate, and the benefits would be larger if sedentary behavior was replaced with moderate to vigorous physical activity [
32].
We also found that less than 10% of adolescents overall met both recommendations on sufficient physical activity and low sedentary behavior. Because low physical activity and high sedentary behavior have independent adverse effects on health [
33,
34], adolescents with both insufficient physical activity and sedentary behavior may suffer from amplified adverse health outcomes. In addition, we found that over 50% of adolescents who had low sedentary behavior had insufficient physical activity, suggesting that adolescents who have low levels of sedentary behavior can have insufficient physical activity. One previous meta-analysis showed that there was only a small inverse association between physical activity and sedentary behavior among children and adolescents, emphasizing that the two behaviors do not directly necessarily correlate well [
35]. Our findings suggest the need to promote sufficient physical activity among adolescents, including for those who do not have sedentary behavior, and vice versa.
Several previous studies have examined the association between a country’s economic status and physical activity using a national GDP per capita with inconsistent findings [
36,
37,
38]. Data from European countries have demonstrated that country GDP/capita was positively associated with levels of physical activity [
36], although other studies showed that high-income countries have a large prevalence of adolescents with insufficient physical activity [
37]. Consistent with this latter study, we found that country economic level, assessed by the country’s PPP/capita, was positively associated with the number of days with sufficient physical activity in young adolescents in LMICs [
35].
When the distribution of physical activity is analyzed within a particular country (i.e., not between countries), based on the GSHS conducted from 2009 to 2016, Vancampfort et al. found that adolescents from poor families were less likely to meet the recommended sufficient physical activity [
39].This question was not assessed in the present study, as we did not have socio-economic indicators at the individual level for all students included in this study. However, it is known, for example, that students from families with high socio-economic status are more likely to attend gym memberships and to live in a favorable environment, like an area with green areas, sports facilities, bike trails, and adequately connected streets [
12].
We also observed in our study among LMICs that young adolescents were more likely to have high sedentary behavior in countries with high than low PPP/capita. Adolescents from the more developed countries (i.e., those with the highest PPP/capita) are likely to have easier access to the internet, computer games, and TV than those from less developed countries (lower PPP/capita). A recent systematic meta-analysis showed an inverse association between socioeconomic status and sedentary behavior in high-income countries but a direct association in LMICs [
16]. Moreover, a Brazilian study showed that about 57.3% of adolescents had high screen time, and screen time was positively associated with the socioeconomic status of the adolescents [
40]. Altogether, these observations in the literature and our findings suggest that, although the economic development of a country (among LMICs) maybe associated with a greater level of physical activity among adolescents, increased country development may also be associated with adverse health effects due to increased sedentary behavior according to an individual’s socioeconomic status. We also found that, although physical activity increased along the upper PPP/capita quintiles, the prevalence of adolescents with sufficient physical activity tended to decrease within the first and second PPP/capita quintiles, suggesting a J-shape relation.
Economic development at the country level may also alter people’s lifestyles. The weakly positive association of a country’s PPP/capita with the level of physical activity among adolescents and the inverse association between a country’s PPP and sedentary behavior suggest that LMICs with higher PPP/capita should specifically address interventions to reduce sedentary behavior, even though adolescents in these countries may have more physical activity. In addition, LMICs with low PPP/capita should develop interventions to promote physical activity despite adolescents in these countries tending to have less sedentary behavior. Over all, it is important to develop specific strategies to promote physical activity and reduce sedentary behavior among adolescents in LMICs.
Strategies to promote physical activity and reduce sedentary behavior should target people of all ages and genders, in different settings, and be performed through interventions involving multiple sectors [
41]. The global action plan on physical activity 2018–2030 set four strategic objectives to increase physical activity and reduce sedentary behavior, including creating active societies, creating active environments, creating active systems, and encouraging people to be active [
14]. This can include measures that promote active commuting to school (e.g., walking or cycling), setting a minimum number of hours of physical activity in school curricula, promoting sports for all at leisure time, programs of physical activity in communities and in other settings, and developing green areas and sport premises. Of note, many interventions promoting physical activity can have benefits that extend beyond health, such as improving social cohesion or benefiting the local economic actors, which may further accelerate their implementation. It is also important to regularly evaluate the impact of physical activity interventions and policy. A recent systematic review reported that only 69 of 292 intervention studies mentioned a proper evaluation framework, limiting their potential to be further sustained or implemented elsewhere [
42].
Study Strengths and Limitations
This study had two main strengths. First, it included a large sample of participants from many LMICs in several regions, which strengthens the generalizability of the findings to young adolescents. Second, the same questionnaire was used in all countries, making the results directly comparable across all countries. However, several limitations should also be noted. First, physical activity and sedentary behavior were self-reported, which is far less accurate than objective measurements, such as accelerometers or energy expenditure assessed with doubly labeled water. This may lead to biases toward both over- or underestimation. However, objective measurements of physical activity and sedentary behavior (e.g., accelerometers) are quite resource intensive, making such measurements difficult to use in large epidemiological surveys. Second, sedentary behavior (e.g., sitting time, standing time, and lying time) is difficult to assess precisely by a questionnaire, and could provide more accurate data and would be measured more accurately with objects such as accelerometers, including in smartphones. Cut-offs for sedentary time associated with detrimental health outcomes (e.g., ≥2 h in our study, excluding sitting time at school or for homework) are still quite arbitrary and less validated than physical activity cut-offs associated with detrimental health outcomes. It is also not fully clear to what extent sedentary behavior and insufficient physical activity are independent causes of health outcomes, and whether a person can compensate sedentary time by having more physical activity and vice versa. Third, the comparison of estimates between WHO regions and countries should be interpreted with caution, because surveys were not done at the same time (2009 to 2016). However, the majority of surveys (54 of 68 countries) were performed between 2011 and 2016. Fourth, although some potentially confounding factors were adjusted in a multivariate analysis, residual confounding or unmeasured factors might have influenced our findings. In addition, despite the same wording, questions on physical activity and sedentary behavior (both cutting across many social and behavioral dimensions) may be understood or interpreted differently according to a person’s country, culture, sex, or social background. Fifth, the ecological design of our study (the correlation between PPP/capita at the country level and physical activity and sedentary behavior levels at the individual level) might lead to ecological fallacy.