INTRODUCTION

Overall estimates indicate that physical inactivity in adolescence is high, and about 80% of adolescents do not reach the minimum recommendations of physical activity for health (^{Sallis et al., 2016}), data similar to those found in the Brazil (^{Ferreira et al., 2018}). Evidence from longitudinal studies indicates that the level of physical activity reduces in adolescence (^{Dumith, Gigante, Domingues, & Kohl III, 2011}). However, this decline seems to be associated with biological maturation (^{Sherar, Cumming, Eisenmann, Baxter-Jones, & Malina, 2010}), since the literature indicates a larger decrease in early maturing adolescents (^{Bacil, Mazzardo Junior, Rech, Legnani, & Campos, 2015}).

In fact, adolescence is marked by a set of physical, hormonal, psychological and behavioral changes, among others, which have some influence on the involvement in physical and sports activities, differently between the sexes (^{Bacil et al., 2015}; ^{Seabra, Mendonca, Thomis, Anjos, & Maia, 2008}). For girls, the pubertal development is inversely associated with psychological well-being and self-esteem which, in turn, may negatively impact on the involvement and sensation of pleasure related to physical activities (^{Davison, Werder, Trost, Baker, & Birch, 2007}). In boys, on the other hand, physical changes associated with maturation, such as increased height and muscle mass, result in a physical size adequate for success in sports activities (^{Bacil et al., 2015}).

A number of individual, interpersonal, and environmental factors are associated with physical activity in adolescence, including social support (^{Bauman et al., 2012}; ^{Sallis et al., 2016}; ^{Van Der Horst, Paw, Twisk, & Van Mechelen, 2007}). Conceptually, social support can be understood as aid or care received through interpersonal relationships, which is one of the most important functions of social relations (^{Glanz, Rimer, & Viswanath, 2008}). In addition to the influence that maturation exerts on physical and psychological changes related to physical activity, evidences point to alterations in intrapersonal and interpersonal relationships (^{Sherar et al., 2009}). Recent reviews have identified a consistent positive association between different types and sources of social support and physical activity in adolescents (^{Beets, Cardinal, & Alderman, 2010}; ^{Laird, Fawkner, Kelly, McNamee, & Niven, 2016}; ^{G. Mendonça, Cheng, Melo, & Farias Junior, 2014}), without, however, considering the effect of maturation on most evidence.

The literature indicates differences in social support for the physical activity received by parents in relation to gender, adiposity and body composition, as well as in the strength of association that the types and sources of this support may represent in physical activity (^{Beets et al., 2010}; ^{Piola, Bacil, Silva, Pacifico, & Campos, 2018}; ^{Prado, Lima, Fermino, Anez, & Reis, 2014}). For example, the company of friends has a greater magnitude of effect for boys, while the family company is more strongly associated with physical activity in girls (^{Prado et al., 2014}). Since maturation influences different characteristics, it is believed that it may moderate the effect of the association between social support and physical activity (^{Bacil et al., 2015}; ^{Jackson et al., 2013}; ^{Sherar et al., 2010}; ^{Sherar et al., 2009}).

No studies were found that analyzed the moderating effect of maturation on the association among these variables in adolescents of both sexes, and this analysis may be bibrlied by the moderator effect, which could verify if sexual maturation affect the direction and/or strength of the relationship of social support and physical activity.

understanding how maturation modifies the association between social support and physical activity may contribute to the design of effective interventions to increase levels of physical activity in adolescents (^{Sherar et al., 2010}). Thus, the objective of this study was to analyze the moderating effect of maturation on the association between the social support received from parents and friends in the level of physical activity in adolescents.

MATERIAL AND METHODS

Participants

For the estimation of the sample calculation (^{Luiz & Magnanini, 2000}), an error of 5% was considered, an estimated prevalence of insufficiently active of 51% (^{Barbosa Filho, Campos, Bozza, & Lopes, 2012}), a delineation effect of 1.4 bibrlied on following equation;

Additionally an increase of 30% for possible losses and refusals and a 95% confidence interval. Thus, the minimum sample required was estimated in 675 adolescents.

The sample was selected in three stages. In the first, all the schools that offered high school in the morning period were selected (n=18). Following this, a simple random selection of one class from each high school series was performed according to the number of schoolchildren, proportional by sex, required by school. Finally, all students in each class were invited to participate in the study. Those excluded were schoolchildren under the age of 14 or older than 18 years and those with physical or cognitive limitations that limited the practice of physical activities. Adolescents who did not present the informed consent form signed by parents or guardians, those who refused to participate in the study or who missed classes on the day of collection were considered as losses. A total of 959 adolescents were evaluated; however, participants classified in the pre-pubertal maturational stage (n=68) were excluded from the study, who completed the questionnaires with unlikely, incomplete or blank responses (n=36). Thus, the analytical sample was 855 adolescents.

A posteriori analysis showed that the sample (n=855) could detect significant prevalence ratios (PR) above 1.22 as risk and below 0.80 as protection, with a power of 80% (β=20%), 95% confidence level (α = 5%) and an insufficiently active prevalence of 78.4%, calculated on software GPower 3.1.9.4.

Instruments

Social support for the practice of physical activities was evaluated with a validated questionnaire for Brazilian adolescents (^{Farias Júnior et al., 2011}). The questionnaire presents six different types of social support, coming from two different sources, parents and friends. The types of social support received from the parents were: "stimulated", "practiced together", "led", "attended", "commented" and "talked" (table 1). In relation to friends, the following types of support were evaluated: "stimulated", "practiced together", "invited", "attended", "commented" and "talked" (table 1). Response options were set out on a four-point Likert scale, and adolescents reported weekly frequency ("never", "rarely", "often", "always") with which parents and friends provided some form of assistance and / or assistance (social support) to perform physical activities during a typical week.

Table 1. Descriptive characteristics of adolescents. by maturational stage. São José dos Pinhais, Paraná, Brazil, 2014 (n=855). 

^hχ 2 for heterogenity

^t:χ 2 for linear trend

For the purposes of analysis, and in order to ensure comparability with similar studies (^{Piola et al., 2018}; ^{Prado et al., 2014}), the "rarely" and "often" response options were grouped and classified as "sometimes". Thus, the weekly social support categories were classified as "never", "sometimes" and "always".

We chose to create a variable that represents the social support score received from each source, based on the sum of the six types of support received from parents and friends. Thus, the "general social support score" was classified into three levels according to tertiles: "low" (1st tertile), "intermediate" (2nd tertile) and "high" (3rd tertile).

The practice of physical activity in leisure was evaluated by completing the "Self-Administered Physical Activity Checklist", in its version adapted and validated for Brazilian adolescents (^{Farias Júnior et al., 2012}). The adolescents reported weekly frequency and duration of participation in 22 types of physical activities, with moderate to vigorous intensities, in the last week.

The physical activity score was calculated with the sum of the product of the weekly frequency by the volume, in minutes, spent on each activity. The "walking on the move" and "walking with the dog" activities were excluded from the score because they were of mild intensity (^{Adams, Johnson, & Tudor-Locke, 2013}). A cut-off point of ≥300 min / week was used to classify adolescents as "physically active" (^{World Health Organization, 2010}).

The maturation was autoevaluated by pubic hairiness by photos, using the Tanner method (^{Hills & Byrne, 2010}; ^{Malina, Rogol, Cumming, Silva, & Figueiredo, 2015}). The stages vary from one to five, in the first the adolescent is considered as "pre-puberty" (pre-adolescence), in stages two, three and four are classified as "pubescent"; and stage five, as "post-puberty", which indicates that the maturational process is concluded (^{Lloyd, Oliver, Faigenbaum, Myer, & Croix, 2014}; ^{Malina et al., 2015}). However, for the present study, adolescents classified in stage one were excluded from the analysis (n=68) because they characterized a phase of pre-adolescence (^{Malina et al., 2015}).

In the review of the literature, some variables potentially associated with social support or level of physical activity were identified: sex, age, nutritional status and socioeconomic status (^{Bauman et al., 2012}; ^{Sallis et al., 2016}; ^{Van Der Horst et al., 2007}). These variables were selected, evaluated and tested as possible confounding factors. Gender was self-reported ("male", "female"), age classified in two age groups ("14-15 years", "16-17 years"), nutritional status (body mass index) measured body mass and height (kg / height2) and adolescents classified as "normal weight" and "overweight" (^{Conde & Monteiro, 2006}). The socioeconomic status was evaluated with the standardized questionnaire based on the number of household utensils in the residence, presence of a monthly employee, and the school attendance of the household financial officer (ABEP, 2013). For the purpose of analysis, this variable was classified into three categories: "low" (class C + D), "intermediate" (class B1 + B2) and "high" (class A1 + A2).

RESULTS

The final sample consisted of 855 adolescents (54% of girls), with a mean age of 16.1 ± 1.1 years. The highest proportion of adolescents were classified as pubertal maturational stage (68.9%), were between 16-18 years old (50.6%), had normal nutritional status (82.9%), intermediate socioeconomic status (56.0%) and performed <300 min / without moderate to vigorous leisure time physical activity (78.4%) (table 1). The male sex and the largest age group were positively associated with the post-pubertal maturational stage (p≤0.001).

The most common type of social support received from the parents was the encouragement of physical activity (22.0%), while the least reported was physical activity with adolescents (6.0%) (table 1). None of the types of social support received from parents was associated with the maturational stage. In relation to the support received from friends, the fact that they "practiced together" or "invited" to the activities was the most frequent (23.0%), while the positive reinforcement ("commenting positively") was the least reported (12.6%). The greatest stimulus, the company, the positive reinforcement, the fact that the friends attended and talked about the activities were positively associated with the post-pubertal maturational stage (p≤0.013) (table 1).

In the bivariate analysis, in relation to the support received from parents, only the social support score showed no significant positive association with physical activity, regardless of the maturational stage. Among pubertal adolescents, the fact that parents "always" attended physical activities had the greatest strength of association with the outcome (PR: 5.44; 95% CI: 3.12-9.47). Among post-pubertal adolescents, parental stimulation had the greatest positive association with physical activity (PR: 3.00; 95% CI: 1.23-7.31). Each social support types received from friends were positively associated with the physical activity of pubescent adolescents. For post-pubertal adolescents, only the fact that friends "attended activities" and the social support score were not associated with physical activity (table 2).

Table 2. Moderating effect of maturation on the bivariate association between social support, received from parents and friends, and physical activity on leisure in adolescents. São José dos Pinhais, Paraná, Brazil, 2014 (n=855). 

^*significance value for the trend test.

After adjusting for the confounding variables (table 3), only the fact that the parents participated in the activities remained positively associated with the physical activity of pubescent adolescents ("sometimes"=PR: 2.51; 95% CI:1.42) 4,44. For post-pubertal adolescents, the fact that parents "always" comment positively on activities was positively associated with physical activity (PR:2.60; 95% CI:1.01-6.10). In relation to friends, the fact that they always participated in activities (PR: 2.39; 95% CI:1.01-5.76), as well as the highest support score ("intermediate"=PR: 2.13; 95% CI:1.28-3.58 "high"=PR:1.91; 95% CI:1.11-3.28), maintained a positive association with the physical activity of pubertal adolescents.

Table 3. Moderating effect of maturation on multivariate association between social support, received from parents and friends, with physical activity in adolescents. São José dos Pinhais, Paraná, Brazil, 2014 (n=855). 

^§adjusted for gender, age, nutritional status, socioeconomic status and other social support types received from the same source.

^*significance value for the trend test.

DISCUSSION

This is the first study that analyzed the moderating effect of maturation on the association between social support received from parents and friends with leisure physical activity of adolescents of both sexes. Evidence pointed to a consistent positive association between social support and physical activity, however, they did not analyze the maturation effect (^{Beets et al., 2010}; ^{Corbí, Moreno-Murcia, & Gea-Escámez, 2018}; ^{Costa, Pureza, & Mielke, 2017}; ^{Laird et al., 2016}; G. ^{Mendonça et al., 2014}; Gerfeson Mendonça & Farias Júnior, 2015; ^{Tebar et al., 2018}; ^{Todisco et al., 2017}). The methodology used allowed the evaluation of these parameters in a representative sample of adolescents enrolled in the public-school system, which is one of the study's strengths.

The results reinforce the pre-established hypothesis by pointing out the moderating effect of maturation on the association between social support and adolescent physical activity. The presence of parents, when attending their children's activities, was positively associated with the physical activity of pubertal adolescents, while positive reinforcement was positively associated with the physical activity of postpubertal adolescents. The presence of friends, when attending activities, and the highest overall social support score were positively associated with the physical activity of pubescent adolescents.

The absence of similar studies limits, in part, the adequate comparison and discussion of the results of the present study (^{Costa et al., 2017}; Gerfeson ^{Mendonça & Farias Júnior, 2015}; ^{Tebar et al., 2018}; ^{Todisco et al., 2017}). The presence of the parents, when attending "sometimes" (PR: 2.51; 95% CI: 1.42-4.44) or "always" (PR: 4.60; 95% CI: 2.10-10.10) the activities of their children showed a significant association with the physical activity of pubescent adolescents, similar with Torre-Cruz et al study (2019). Few studies have analyzed the association between this type of social support and adolescents' physical activity according to age (^{Marques, González Valeiro, Martins, Fernández-Villarino, & Carreiro da Costa, 2017}; ^{G. Mendonça et al., 2014}). For example, Mendonça and Farias Jr (^{Gerfeson Mendonça & Farias Júnior, 2015}) did not find a moderating effect on the age group ("14-16 years" vs. "17-19 years") in the association between parental support when they see their children performing activities with the level of physical activity of adolescents from João Pessoa-Paraíba, Brazil.

Most social support types exploited in present study did not show a significant association with physical activity, indicating that perhaps physical education teachers support may also be considered as favorable to the practice of physical activities of adolescents (^{Trigueros-Ramos, Gómez, Aguilar-Parra, & León-Estrada, 2019}).

Some limitations should be considered for the proper interpretation and extrapolation of the results of this study. The sample is representative of adolescents aged 14-18, students of the public-school system, in the urban area of a medium-sized municipality in the southern region of the country. Therefore, the external validity and inference of the results are limited for adolescents of other age groups, students from private schools or from rural areas, from large capitals or regions of the country with different socioeconomic conditions. Parental social support was assessed globally, which made it impossible to analyze specific sources such as the father or mother. Similarly, social support was not assessed according to the sex of the friends (support relationship girl-girl, boy-boy, etc.). Finally, the cross-sectional design does not allow establishing the causal relationship between the variables. However, the procedures used to select the sample made it possible to reduce the selection bias, the main confounding factors were controlled and the sample size was sufficient to detect the effects in the multivariate analysis, which increases the internal validity of the study.

CONCLUSION

Maturation has a moderating effect on the association between social support and physical activity in adolescents. Regarding the social support of the parents, the fact that they attended the activities of the children was positively associated with the physical activity of pubertal adolescents, while the positive reinforcement was associated with the outcome in postpubertal adolescents. When friends attend activities and provide greater overall support, they are thus positively associated with the physical activities of pubescent adolescents.

These results are important so that health professionals can advise parents about their behavior and actions, as well as actions towards adolescents 'friends, which can stimulate adolescents' involvement, according to maturation, in activities physical and sports activities of greater intensity in leisure. Future experimental studies could analyze the effect that different types and sources of social support (father, mother, siblings, same-sex and opposite-sex friends) could provide on behavior in specific modalities / intensities of physical activity in adolescents.