Artigos

Introduction
Low physical activity among adolescents is a global public health concern. In 2016, nearly 80% of the world’s adolescents aged 11 and 17 were considered insufficiently active(1). In Brazil, only 34.4% of 9th graders were classified as physically active in 2015(2).
Among school-age youth, physical inactivity is associated with increased cardiometabolic risk, higher levels of body fat, lower fitness and compromised bone health(3). Additionally, insufficient physical inactivity during adolescence is a strong predictor of continued inactivity in later life(4). Physically inactive children are more likely to exhibit biological risk factors such as high blood pressure, elevated cholesterol levels, and increased body fat(5). Conversely, the risk of developing cardiovascular disease decreases progressively with increased physical activity(6).
Brazilian studies have consistently shown a high prevalence of physical inactivity among adolescents. Despite variations in criteria and cutoff points, the prevalence of physical inactivity ranges from 55.9% to 82.7% across different Brazilian cities(7-10).
Overall, girls, older adolescents, and those with lower education attainment, and individuals from lower socioeconomic backgrounds have been identified as less active(7, 8, 11, 12). Physical inactivity in children and adolescents is also associated with risky health behaviors, including unhealthy eating, sedentary behavior, smoking, alcohol consumption, and drug use(8, 13).
Social support, defined as the resources provided individuals within one’s social network, has been linked with physical activity in adolescents(4). Longitudinal studies have shown that logistical support, parental role modeling and having active friends are positive factors that encourage young people remain active(15). However, as adolescents age, the influence of family tends to decrease while peer influence becomes more prominent(16). The physical activity levels of friends positively influence adolescents, with those having active friends being more likely to be active themselves(16). On the other hand, limited social networks can act as a barrier to engaging in physical activity(17). Research on adolescents has demonstrated that feelings of loneliness and having few friends are characteristics associated with physical inactivity(18-20). While several systematic reviews have explored the relationship between adolescent physical activity and the support of parents and friends, less than 1% of the studies have been conducted Brazil(14, 16, 21-24). Therefore, the aim of this study is to investigate the potential association between social networks, social isolation and physical inactivity among Brazilian adolescent students.

Methods
The present study utilized data from the third edition of the National Adolescent School-based Health Survey (PeNSE-2015), which monitors health risk factors among Brazilian students. The PeNSE-2015 is a cross-sectional, school-based study with data collected through self-administered electronic questionnaires from April to September 2015. Data are publicly available and can be accessed through the Brazilian Institute of Geography and Statistics (IBGE) website(2).
Sampling plan included 9th grade students from public and private schools across Brazil in 2015. The age range of the students was 13 to 15 years, a reference group for school surveys. This age group represents the the minimum age suitable for responding to self-administered questionnaires(2). According to IBGE, approximately 2.6 million students were enrolled in fundamental school in Brazil in 2015.
The 2015 PeNSE sample was randomly selected from 53 geographical strata: including 26 state capitals and Federal District, as well as the 26 states. Schools within each stratum were selected based on the 2013 School Census, and within each selected school, 9th-grade classes were randomly selected with probability proportional to size. Thus, the primary sampling units were the schools and the secondary units were the classes, resulting in an independent sample of students in each stratum. All students present in the selected classes were eligible and invited to participate in the study(2).
The inclusion criteria were to be enrolled in the 9th year of elementary school, during the 2015 academic year, and regularly attending public or private schools located in urban or rural areas of the National Territory. Night shift students and schools with fewer than 15 students enrolled in the 9th grade in 2013 were excluded(2).
The sample size for each geographic stratum was calculated to provide estimates of proportions (or prevalence) for selected characteristics of interest, with a maximum sampling error of approximately 3 percentage points and a 95% confidence level. A prevalence of 50% was used in the calculation. The number of schools was obtained by dividing the number of 9th grade classes in the sample by the average number of 9th grade classes per school in each stratum. Sample weights accounted for the sizes of schools, classes, and students. Based on this sample design, the analysis included 101.500 students from 3.040 schools.
To ensure quality and standardization of data collection, IBGE state supervisors were trained and, in turn, provided training to the interviewers.
Leisure-time physical activity levels among students over the past seven days were assessed using questions on the frequency and duration of extracurricular activities, such as sports, dance, gymnastics, weight training, and martial arts. These questions were adapted from the International Physical Activity Questionnaire (IPAQ), a tool with established reliability and validity (25). Based on weekly accumulated minutes, physical activity was categorized into four levels: ?300 minutes per week (m/w), 150–299m/w, 1–149m/w and zero m/w (26).
The main exposures were social isolation and physical social network. Social isolation was assessed with a question ont how often adolescents felt lonely (never/rarely/sometimes/most of the time/always), and the social network was assessed by asking about the number of close friends they had (no friends/one friend/two friends/three or more friends).
Confounding factors included demographic, socioeconomic and behavioral variables. Demographics were gender (male/female), skin color (white/black/brown/indigenous/yellow), and age (11 to 13 years/ 14 years/ 15 years/ 16 years/ 17 years or older). Socioeconomic status was determined based on maternal education (did not study / incomplete elementary school / complete elementary school / incomplete high school / complete high school / incomplete undergraduate education or more).
Behavioral characteristics included: frequency of vegetable consumption per week (zero / 1 to 3 days / 4 to 6 days / every day), smoking status (non-smoker / smoker), number of alcoholic drinks consumed in the last month ( none / less than 1 glass / 1 to 2 glasses / 3 to 4 glasses / 5 glasses or more), use of illicit drugs (cannabis, cocaine and inhalants) in the last month (did not use / used) and daily sitting time (? 2hs/ >2hs/day).
Data were analyzed using Stata version 15 (StataCorp, College Station, TX, USA). Descriptive analyses were conducted to calculate absolute and relative frequencies. Univariate associations between categorical variables were assessed using the chi-square test for linear trend. Polytomous (multinomial) logistic regression was employed instead of ordered logistic regression due to concerns regarding the proportional odds assumption. Although the outcome categories of physical activity levels have an inherent order (from high activity to complete inactivity), the Brant test revealed a statistically significant violation of the proportional odds assumption for both exposures, an essential requirement for the validity of ordered logistic models.
In contrast, multinomial logistic regression does not impose this assumption and allows for greater flexibility by estimating separate odds ratio for each level of the outcome relative to the reference category (?300 minutes/week). For the multivariable analysis, a confounding control strategy was adopted. Variables were considered confounders if they were significantly associated with both the exposure and the outcome (p < 0.05) and were not hypothesized to lie on the causal pathway. All analyses accounted for the complex survey design, incorporating primary sampling units (schools), geographic strata, and sampling weights. Analyses were performed for the total sample and stratified by sex. A significance level of 1% (p < 0.01) was adopted to identify statistically significant associations.
The PeNSE-2015 study was approved by the National Research Ethics Committee (CONEP) of the National Health Council (CNS), under approval number 1.006.467, on March 3, 2015.

Results
Our analysis included 101,927 students, comprising 49,218 boys and 52,709 girls. Among Brazilian adolescents, leisure-time physical activity levels varied considerably by sex and across sociodemographic, behavioral, and psychosocial characteristics. Overall, only 15.7% (95% CI: 15.3–16.2) of students engaged in ?300 minutes/week of extracurricular physical activity (Table 1), with a marked sex disparity: 22.7% (95% CI: 22.0–23.5) among boys (Table 2) and 9.1% (95% CI: 8.6–9.6) among girls (Table 3). In contrast, complete inactivity (0 minutes/week) was reported by 34.6% (95% CI: 33.9–35.3) of the overall sample, with significantly higher prevalence among girls (43.9%, 95% CI: 43.0–44.9) compared to boys (24.8%, 95% CI: 24.0–25.6). Higher levels of physical activity were generally associated with younger age, higher maternal education, healthier eating habits (e.g., fruit and vegetable consumption), and stronger social networks (i.e., having three or more close friends and low levels of social isolation). Conversely, lower physical activity levels were more common among adolescents who reported frequent social isolation, fewer close friends, greater screen time, and lower parental education (Tables 1, 2 and 3).
After adjusting for sociodemographic and behavioral variables, the odds for complete inactivity (zero min/week) of physical inactivity were 26% (OR 1.26; CI95%: 1.06, 1.51) higher in students with reporting feeling “always” isolated compared to those “never” isolated. Adolescents with “no friends” also had significantly higher odds of complete inactivity (OR: 1.69; 95% CI: 1.35–2.11) compared to those with “three or more friends”. In sex-stratified analyses, these associations remained significant among boys, especially for those who were always isolated (‘OR for 0 min/week: 1.71; 95% CI: 1.30–2.25), and among girls with limited social networks—girls with no friends had 1.53 times the odds of being completely inactive (95% CI: 1.07–2.18) compared to their more socially connected peers (Table 4).

Discussion
This study aimed to investigate the association between social networks and social isolation with physical inactivity levels in Brazilian schoolchildren. The findings revealed that the majority of students were either physically inactive or insufficiently active, a concerning pattern that has been consistently reported in previous studies.(1, 22, 27).
In the sex-stratified analysis, we observed significant differences in leisure-time physical inactivity levels, with girls being more inactive than boys, consistent with other research(7, 28, 29). These differences can be attributed to biological and sociocultural factors, particularly the gender roles assigned to boys and girls from a young age, which influence their engagement in physical activities. Girls are often encouraged to participate in lighter activities, based on the misconception about the the fragility of their bodies and aesthetical concerns, whereas boys are typically urged to engage in more vigorous physical activities, due to the false perception of greater physical strength and capability(30).
A restricted social network was significantly associated with higher levels of physical inactivity. These results are consistent with other studies indicating that friends play a crucial role in supporting adolescents’ physical activity(21, 24, 28, 30-32). A systematic review identified three key indicators of friendship that positively influence physical activity in children and adolescents: communication between friends about physical activity, role modeling and co-participation(23). Among Brazilian adolescents, the presence of friends was found to be the strongest association with physical activity(33).
Social isolation was linked with lower levels of physical activity only among male students. Previous studies in both high and middle-income countries have also reported an association between physical inactivity and feelings of loneliness in adolescents(20, 34). Our results suggest that social network may be more influential than social isolation in this age group and population, highlighting the importance of peer relationships in adolescents’ physical activity. While loneliness can be a transient experience, friendships tend to be more enduring, providing sustained support to overcome the barriers to regular physical activity. In Brazilian students, higher levels of social support from parents and friends have been associated with greater self-efficacy in overcoming physical activity barriers (35). It could be hypothesized that the self-efficacy fostered by friends’ support may better help adolescents to cope with loneliness.
The pathways that explain the relationship between friendship networks and physical activity behavior can be explained by two models: peer contagion and peer selection. Peer contagion suggests that once individual become friends, their behaviors tend to align, particularly through social influence. Peer selection posits that individuals choose friends with similar interests, attitudes and behaviors(36). Qualitative studies have shown that young people seek physical activity primarily for enjoyment, and the presence of friends is a key motivator for this “fun" factor(23). Conversely, individuals with a limited social network, who have fewer friends, are likely to experience less social influence to engage in physical activity, and peer influence in general may also be weaker.
In the sex-stratified analyzes, the adjusted results indicate a relatively stronger effect of both social networks and social isolation on the physical inactivity levels of male students, suggesting that boys may be more influenced by their social networks compared to girls. These findings are consistent with a systematic review of 13 studies involving 24.997 children and adolescents, which showed that boys were more influenced by their friendships’ networks in terms of physical activity than girls(16). This observation has been corroborated by studies on Brazilian adolescents, where the support of friends had a greater effect on boys’ physical activity than on girls’(35). Additonally, a cross-sectional study reported that self-efficacy, social norms, and community involvement in physical activity were important predictors of moderate-to-vigorous physical activity for boys; whereas for girls, only self-efficacy was a significant predictor(37).
Although the data analyzed in this study were collected prior to the COVID-19 pandemic, they remain relevant for understanding the association between social ties and leisure-time physical activity among adolescent students. First, the pandemic-related restrictions, such as school closures and social distancing, were temporary disruptions rather than long-term structural changes. As the transitory nature of these restrictions, it is plausible to assume that patterns of socialization and physical activity among adolescents are currently closer to the levels observed in the pre-pandemic period. Second, using pre-pandemic data allows for an examination of these associations under more normal social conditions, free from the confounding effects of emergency health policies and behavioral constraints. Thus, the data analyzed preserve relevance and validity and provide a valuable reference point for understanding the associations between social relationships and leisure-time physical activity in adolescent students.
Our study has limitations that should be acknowledged. Due to its cross-sectional nature, where outcomes and exposures were measured simultaneously, reverse causality cannot be ruled out. It could be argued that physical inactivity leads to a smaller network of friends and greater social isolation. For instance, using the same dataset, Werneck and collegues(38) found that physical activity was associated with lower prevalence of social isolation. Furthermore, our study utilized individual questions to measure the exposures, whereas well-established scales, such as the Social Support for Exercise Behavior Scale, the Survey of Children’s Social Support and the Adolescent Physical Activity Survey exist for assessing these constructs(39). We also did not evaluate family social support, can play a significant role in encouraging and sustaining physical activity in adolescents(14). Additionally, our study relied on subjective measures of physical activity, which are less precise and accurate. Another limitation is the weekly aggregation physical activity time on a weekly basis, which fails to differentiate between consistent daily engagement and sporadic patterns concentrated over fewer days. Although the cutoff point of 300 minutes per week is consistent with international recommendations for additional health benefits, it does not capture adherence to the recommendation of 60 minutes of daily physical activity for teenagers. This may result in the misclassification of activity patterns and limit more sophisticated interpretations of behavioral routines associated with physical activity. Future research should employ longitudinal designs, validated scales for assessing exposures, and objective methods for measuring the outcome. However, our study’s strength lies in its large, nationally representative sample, which provides valuable insights into the relationships between social networks, social isolation and physical inactivity levels in adolescents.
In conclusion, a lower number of friends was associated with reduced levels of physical activity among adolescent students. Given the growing concern regarding insufficient levels of physical activity and its implications for the physical and mental health of adolescents, it is essential to understand the role of psychosocial factors, such as social support networks and peer influence, in promoting healthy behaviors. This understanding is vital for the development of effective public health interventions aimed at improving adolescent health and well-being.

References
1. Guthold R, Stevens GA, Riley LM, Bull FC. Global trends in insufficient physical activity among adolescents: a pooled analysis of 298 population-based surveys with 1.6 million participants. Lancet Child Adolesc Health. 2020 Jan;4(1):23-35.
2. IBGE IBdGeE. Pesquisa Nacional de Saúde do Escolar. 2024 [08/08/2024]; Available from: https://www.ibge.gov.br/estatisticas/sociais/educacao/9134-pesquisa-nacional-de-saude-do-escolar.html?=&t=microdados.
3. Poitras VJ, Gray CE, Borghese MM, Carson V, Chaput JP, Janssen I, et al. Systematic review of the relationships between objectively measured physical activity and health indicators in school-aged children and youth. Appl Physiol Nutr Metab. 2016 Jun;41(6 Suppl 3):S197-239.
4. Belanger NMS, Patrick JH. The Influence of Source and Type of Support on College Students' Physical Activity Behavior. J Phys Act Health. 2018 Mar 1;15(3):183-90.
5. Ribeiro JC, Guerra S, Oliveira J, Teixeira-Pinto A, Twisk JW, Duarte JA, et al. Physical activity and biological risk factors clustering in pediatric population. Prev Med. 2004 Sep;39(3):596-601.
6. Andersen LB, Harro M, Sardinha LB, Froberg K, Ekelund U, Brage S, et al. Physical activity and clustered cardiovascular risk in children: a cross-sectional study (The European Youth Heart Study). Lancet. 2006 Jul 22;368(9532):299-304.
7. Bergmann GG, Bergmann ML, Marques AC, Hallal PC. Prevalence of physical inactivity and associated factors among adolescents from public schools in Uruguaiana, Rio Grande do Sul State, Brazil. Cad Saude Publica. 2013 Nov;29(11):2217-29.
8. Ceschini FL, Andrade DR, Oliveira LC, Araujo Junior JF, Matsudo VK. Prevalence of physical inactivity and associated factors among high school students from state's public schools. J Pediatr (Rio J). 2009 Jul-Aug;85(4):301-6.
9. Condessa LA, Soares CA, Mielke GI, Malta DC, Caiaffa WT. Prevalence of physically active adolescents in Brazilian capitals: National Adolescent School-based Health Survey 2012 and 2015. Rev Bras Epidemiol. 2018 Nov 29;21(suppl 1):e180012.
10. Moraes AC, Fernandes CA, Elias RG, Nakashima AT, Reichert FF, Falcão MC. [Prevalence of physical inactivity and associated factors in adolescents]. Rev Assoc Med Bras (1992). 2009 Sep-Oct;55(5):523-8.
11. Azevedo MR, Araujo CL, Reichert FF, Siqueira FV, da Silva MC, Hallal PC. Gender differences in leisure-time physical activity. Int J Public Health. 2007;52(1):8-15.
12. Varo JJ, Martinez-Gonzalez MA, De Irala-Estevez J, Kearney J, Gibney M, Martinez JA. Distribution and determinants of sedentary lifestyles in the European Union. Int J Epidemiol. 2003 Feb;32(1):138-46.
13. Pate RR, Davis MG, Robinson TN, Stone EJ, McKenzie TL, Young JC, et al. Promoting physical activity in children and youth: a leadership role for schools: a scientific statement from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism (Physical Activity Committee) in collaboration with the Councils on Cardiovascular Disease in the Young and Cardiovascular Nursing. Circulation. 2006 Sep 12;114(11):1214-24.
14. Laird Y, Fawkner S, Kelly P, McNamee L, Niven A. The role of social support on physical activity behaviour in adolescent girls: a systematic review and meta-analysis. Int J Behav Nutr Phys Act. 2016 Jul 7;13:79.
15. Davison KK, Jago R. Change in parent and peer support across ages 9 to 15 yr and adolescent girls' physical activity. Med Sci Sports Exerc. 2009 Sep;41(9):1816-25.
16. Sawka KJ, McCormack GR, Nettel-Aguirre A, Hawe P, Doyle-Baker PK. Friendship networks and physical activity and sedentary behavior among youth: a systematized review. Int J Behav Nutr Phys Act. 2013 Dec 1;10:130.
17. House JS, Landis KR, Umberson D. Social relationships and health. Science. 1988 Jul 29;241(4865):540-5.
18. Cheng LA, Mendonca G, Farias Junior JC. Physical activity in adolescents: analysis of the social influence of parents and friends. J Pediatr (Rio J). 2014 Jan-Feb;90(1):35-41.
19. Atkin AJ, Corder K, Goodyer I, Bamber D, Ekelund U, Brage S, et al. Perceived family functioning and friendship quality: cross-sectional associations with physical activity and sedentary behaviours. Int J Behav Nutr Phys Act. 2015 Feb 21;12:23.
20. Peltzer K, Pengpid S. Leisure Time Physical Inactivity and Sedentary Behaviour and Lifestyle Correlates among Students Aged 13-15 in the Association of Southeast Asian Nations (ASEAN) Member States, 2007-2013. Int J Environ Res Public Health. 2016 Feb 15;13(2):217.
21. Mendonca G, Cheng LA, Melo EN, de Farias Junior JC. Physical activity and social support in adolescents: a systematic review. Health Educ Res. 2014 Oct;29(5):822-39.
22. Jago R, Macdonald-Wallis K, Thompson JL, Page AS, Brockman R, Fox KR. Better with a buddy: influence of best friends on children's physical activity. Med Sci Sports Exerc. 2011 Feb;43(2):259-65.
23. Maturo CC, Cunningham SA. Influence of friends on children's physical activity: a review. Am J Public Health. 2013 Jul;103(7):e23-38.
24. Macdonald-Wallis K, Jago R, Sterne JA. Social network analysis of childhood and youth physical activity: a systematic review. Am J Prev Med. 2012 Dec;43(6):636-42.
25. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003 Aug;35(8):1381-95.
26. WHO WHO. WHO guidelines on physical activity and sedentary behaviour. Geneve: World Health Organization; 2020.
27. Coelho LG, Candido AP, Machado-Coelho GL, Freitas SN. Association between nutritional status, food habits and physical activity level in schoolchildren. J Pediatr (Rio J). 2012 Sep-Oct;88(5):406-12.
28. Hallal PC, Knuth AG, Cruz DK, Mendes MI, Malta DC. [Physical activity practice among Brazilian adolescents]. Cien Saude Colet. 2010 Oct;15 Suppl 2:3035-42.
29. Keating XD, Guan J, Pinero JC, Bridges DM. A meta-analysis of college students' physical activity behaviors. J Am Coll Health. 2005 Sep-Oct;54(2):116-25.
30. Farias Junior JC, Lopes Ada S, Mota J, Hallal PC. Physical activity practice and associated factors in adolescents in Northeastern Brazil. Rev Saude Publica. 2012 Jun;46(3):505-15.
31. Beets MW, Vogel R, Forlaw L, Pitetti KH, Cardinal BJ. Social support and youth physical activity: the role of provider and type. Am J Health Behav. 2006 May-Jun;30(3):278-89.
32. Fermino RC, Rech CR, Hino AA, Rodriguez Anez CR, Reis RS. Physical activity and associated factors in high-school adolescents in Southern Brazil. Rev Saude Publica. 2010 Dec;44(6):986-95.
33. Prado CV, Lima AV, Fermino RC, Anez CR, Reis RS. [Social support and physical activity in adolescents from public schools: the importance of family and friends]. Cad Saude Publica. 2014 Apr;30(4):827-38.
34. Richard A, Rohrmann S, Vandeleur CL, Schmid M, Barth J, Eichholzer M. Loneliness is adversely associated with physical and mental health and lifestyle factors: Results from a Swiss national survey. PLoS One. 2017;12(7):e0181442.
35. Farias Junior JC, Reis RS, Hallal PC. Physical activity, psychosocial and perceived environmental factors in adolescents from Northeast Brazil. Cad Saude Publica. 2014 May;30(5):941-51.
36. de la Haye K, Robins G, Mohr P, Wilson C. Obesity-related behaviors in adolescent friendship networks. Social Networks. 2010;32(3):161-7.
37. Trost SG, Pate RR, Ward DS, Saunders R, Riner W. Correlates of objectively measured physical activity in preadolescent youth. Am J Prev Med. 1999 Aug;17(2):120-6.
38. Werneck AO, Collings PJ, Barboza LL, Stubbs B, Silva DR. Associations of sedentary behaviors and physical activity with social isolation in 100,839 school students: The Brazilian Scholar Health Survey. Gen Hosp Psychiatry. 2019 Jul-Aug;59:7-13.
39. Sallis JF, Grossman RM, Pinski RB, Patterson TL, Nader PR. The development of scales to measure social support for diet and exercise behaviors. Prev Med. 1987 Nov;16(6):825-36.