The relationship between physical activity and health status has been described extensively in several studies. Active lifestyles are often associated with better health status and quality of life 12345678, however, the relationship between socio-economic status (SES) and physical activity has been investigated less 9. In regards to this subject, the definition for SES seems to be extremely variable. Giles-Corti et al., evaluates SES according to residence in low, middle or high income geographical areas 10; Lindstrom et al., considers employment exclusively 11; while Gordon-Larsen et al., distinguishes between the different socio-economic levels based on family income, for example, low (up to $26200), middle (between $26200 and $50000) and high (> $50000) 12. Anyway, most of the studies on socio-economic determinants influencing physical activity choices consider education as a discriminating factor 13141516.

In Italy there isn't an individual SES classification available, territorial data (ISTAT data) is based on personal economic consumption rather than income and education. Another possible classification of personal income comes from the Italian financial acts and the different levels of income tax rates. However, this data is not easily inferable from statistical sampling including adolescents.

The aim of the present study was to investigate the relationship between adolescents' extra-curricular physical activity and the SES of their families.

Participants were 1121 males (46.5%) and 1290 females (53.5%), aged between 11 and 17 years (median age 12 years). 33.7% of the students attended the first class, 38.5% the second and 27.8% the third one.

86.4% of them participated in school-based physical activity for two hours per week and 13.6% for three hours per week or more.

The types of school-based physical activity undertaken were as follows: 71.1% did gymnastics; 81.7% volleyball; 37.8% basketball; 10.7% football; 36.8% handball; 22.3% long jump; 21.6% high jump and 48.8% running.

71.1% of students declared to participate in extra-curricular physical activity with the following breakdown: 20.2% for one-two hours per week, 23.8% for three-four hours per week, while 56% stated they did five hours or more per week.

Concerning extra-curricular sports, 30.1% of the students played football, 19.7% did dancing/aerobics/gymnastics, 14.5% swam, 14.2% played volleyball, 5.8% played basket ball, 5.4% did fitness/body-building, 3.9% did martial arts, and 1.8% cycled.

370 students (15.3%) state to take nutritional supplementation before or after physical activity. 68.2% of students practising extra-curricular physical activity stated that they paid a monthly fee.

Concerning attitudes toward physical activity, 79.9% of participants considered constant physical activity useful in preventing obesity, 5.6% did not and 14.5% were unsure. 82.7% of students considered physical activity important in socialising, 5.9% did not and 11.6% remained unsure. 56.1% of responders considered physical activity important in character- building, 15.8% did not and 19.1% did not know.

As for unhealthy lifestyles, 37.8% stated that they were smokers or had started smoking, 43.8% drank coffee and 43.4% drank alcohol, especially beer, wine, alcoholic lemon drinks and vodka.

Table 1 shows parents' characteristics (physical activity levels, educational levels and work activities).

With regards to parental physical activity levels, fathers were significantly more active than mothers (χ² = 15.41; p = 0.0015). However, there were not significant differences between parental education (χ² = 4.113; p = 0.249).

The main parental work activities were: Office workers/Skilled workers (36.5%) and Non-skilled workers (34.3%) for fathers and Housewives (59.3%) and Office workers (28.3%) for mothers. According to our SES classification adolescents' families were categorised in the following way: very high 16.8%, high 14.5%, middle 37.2%, middle-low 30.1% and low 1.4%.

In relation to the SES of the students' family, the variables parents educational levels and work activities were considered. The educational level attained by the father seems to considerably influence the student's physical activity. Students with fathers who graduated with a Degree practice extra-curricular physical activity more frequently (χ² = 64.764; p < 0.0001) (Fig. 2) and more intensively (χ² = 21.091; p < 0.0001) (Fig. 3) than those with fathers characterised by lower educational attainment.

86.3% of students with fathers graduated with a Degree considered constant physical activity helpful in preventing obesity, compared to 85.2%, 80.7% and 69.4%, of students respectively, with fathers holding a diploma, a secondary-school leaving certificate or a primary school-leaving certificate (χ² = 35.421; p < 0.0001). 86.7% of students with fathers graduated with a Degree considered constant physical activity helpful in socialising, versus 84.5%, 88% and 75.5% respectively, of students with fathers holding a diploma, a secondary school-leaving certificate, a primary school/leaving certificate (χ² = 11.155; p = 0.084).

Mother's educational level also seemed to considerably influence the determinants of the student's physical activity (Fig. 2) and the total number of extra-curricular weekly physical activity hours (Fig. 3). Students with mothers graduated with a Degree were more likely to practice extra-curricular physical activity (χ² = 55.512; p < 0.0001) and undertook physical activity, for more than three hours per week (χ² = 19.65; p < 0.0001), than those whose mothers have a lower educational level.

86.4% of students with mothers graduated with a Degree and 86.5% of those whose mothers have a diploma considered physical activity important in socialising, compared to 84.6% and 78.1%, respectively, of students whose mothers have a secondary-school leaving or a primary school-leaving certificate (χ² = 12.817; p = 0.046).

83.5% of students with graduated mothers considered constant physical activity useful in preventing obesity, compared to 85.3%, 80.4% and 77%, respectively, of students with mothers holding a diploma, a secondary-school leaving certificate, a primary school-leaving certificate (χ² = 15.11; p = 0.019).

Furthermore, extra-curricular physical activity is related to the parents' work activity (Fig. 4). Students whose fathers are Managers/Professional or Office-workers/Skilled workers showed a significantly higher level of extra-curricular physical activity than students with fathers who were Non-Skilled workers, Unemployed or Retired from work (χ² = 39.029; p < 0.0001).

The mother's work activity also considerably influences the students' extra-curricular physical activity (χ² = 64.319; p < 0.0001). Adolescents with mothers who are Non-skilled workers/Housewives or Unemployed undertake less extra-curricular physical activity than those with mothers who were Managers/Professionals or Office-workers/Skilled workers (Fig. 4).

Weekly hours of extra-curricular physical activity show a similar trend (Fig. 5): percentages of students practising physical activity for three hours per week or more). Frequency of extra-curricular physical activity performed during the week appears to be related to the work activity level of both parents, with a statistically significant difference for fathers (χ² = 8.229; p = 0.048) and even more so for the mothers (χ² = 28.321; p < 0.0001).

Table 2 shows results of multiple logistic regression. The first model shows an association between extra-curricular physical activity and: 1) father's educational level (direct relation), 2) parents' physical activity (OR = 1.58; 95% CI: 1.24 – 2.02 for father's and OR = 1.33; 95% CI: 1.04 – 1.71 for mother's physical activity), 3) father's work activity (with a statistically significant difference between children of Managers/Professionals and Office workers/Skilled workers and children whose fathers were Unemployed), 4) attitude towards physical activity (OR = 2.25; 95% CI:1.72 – 2.94 for adolescents who considered sport important in socialising as opposed to those who did not). The likelihood of practising extra-curricular physical activity is almost 30% higher for females than it is for males.

The second model, which considers socio-economic family index, confirms the results of the first one. Moreover it shows an association between extra-curricular physical activity and SES (with ORs ranging from 1.63 and 1.96 for the indexes of High to Very high levels).

Similar results were found considering in the logistic regression models the outcome "percentages of students practising physical activity for three hours per week or more" (data not shown).

Several studies show evidence of a relationship between socio-economic level and health status. As demonstrated by Lowry et al. 6, higher family income is associated with lower alcohol and cigarette consumption and a lower level of sedentary behaviour. In Anglo-Saxon countries slums represent an obstacle to the practise of physical activity for several reasons ranging from absence of pavements, public parks and gardens to excessive traffic 10.

The likelihood of undertaking adequate physical activity is low for subjects who have a low family income 19 and for not skilled Workers or Unemployed 11.

Studies on the relationship between physical activity and parents' educational level 1416202122232526 show that higher educational levels are determinant factors for children's physical activity. In particular, Sallis and coll. showed that demographic variables are good predictors of children's physical activity 21, and ethnic and SES differences are potential correlates of physical activity among adolescents 19. In a previous study 26, in 1357 students in Central Italy, we found a strong relationship between weekly physical activity of adolescents, their BMI levels (overweight and obesity) and parents' educational levels and nutritional status, with a higher participation in physical activity shown above all in children with highly educated parents.

Results of our study confirm the relationship between family socio-economic, parents' educational levels and physical activity of adolescents.

Moreover the classification of socio-economic index used in this survey, derived by the combination of the parents' work activities (Very High, High, Middle, Middle-Low and Low), appears to be comparable to the distribution of family income levels as reported in the Italian Financial Acts (law 27/12/2002 n°289 is the latest law in chronological order) 27. In particular the Financial Act assigns a IRPEF tax rate to every income level: tax free until 7.800,00 euro (low income); 23% for 7.800,00 to 15.000,00 euro (middle-low income); 29% for 15.000,00 to 29.000,00 euro (middle income); 31% for 29.000,00 to 32.600,00 euro (middle income); 39% for 32.600,00 to 70.000,00 (middle-high income) and 45% for greater than 70.000,00 euro (very high income).

Given these results, which are in agreement with other international surveys conducted in USA 28, South Africa 29, and Canada 30, we can say that family SES plays a key role in determining extra-curricular physical activity of students: the highest educational levels and the most remunerative work activities are directly related to the sporting practices of adolescents. Children with parents characterised by higher educational levels are encouraged more than others to practice constant and demanding extra-curricular sport. Zakarian and coll. 28 demonstrated that among minority and low-SES adolescents (high school population) in USA, the prevalence rate of vigorous exercise tends to decrease with age or when this kind of popuation is not required to participate in school physical education. McVeigh and coll. 29 found that South African children belonging to the highest SES quartile were highly physically active, watched less television and had greater lean tissue than children in lower quartiles. In Canada, among adolescents aged 12–20 years, individuals from low income families had a 30% higher probability of being inactive than those from high income families 30.

Families with high SES consider physical activity useful both in preventing chronic and degenerative diseases and for socialisation, essential for the correct physical and psychological development of adolescents.

SES is therefore clearly related to parents' educational levels, and parents' low educational levels and low remunerative work activities can negatively influence the participation of children in extra-curricular physical activity.

In addition, it must be considered the inadequate knowledge about benefits of physical activity, related to low socio-economic and cultural levels 31.

The presents study has some limitations. As far as concerns internal validity, we sampled randomly the participants, and used a validated questionnaire. Misclassification bias could arise if incorrect information on both exposure (SES status) and outcome (PA). The accuracy of self-reported physical activity of adolescents, should have been avoided, given the reliability of the tool we used, which had been validated in previous studies 1718. Moreover, we collected detailed information on potential confounding factors and took it into account in the analysis of possible effects

As far as external validity a potential selection bias could have been avoided, even if we used a cross-sectional type of study design. Finally, the precision of the estimates is comparable with that had been designed for.

Neverthless, to our knowledge this is the first attempt in Italy to evaluate the relationship between SES, based either on a family income index and parents' educational status, and adolescents' extra-curricular physical activity.

In our study parents' physical activity is a good predictor of adolescents' extra-curriculum physical activity, accordingly to previous studies. In a recent review on correlates of physical activity of children and adolescents, Sallis and coll. 23 demonstrated that for children, among other factors, parental overweight status, physical activity preferences, intention to be active and program/facility access were statistically associated to physical activity. For adolescents, variables consistently associated with physical activity were perceived activity competence, intentions, but also sports conducted at community level, parent support, sibling physical activity, direct help from parents and opportunities to exercise. In that sense, social, cultural and economic deprivation have an effect on extra-curricular physical activity participation. A strong help in this task could derive from school. In fact, in Italy the vast majority of scholastic building and infrastructures belongs to municipalities and provinces, and are currently underused. A way to improve adolescents' participation in physical activity could be involving them in extra-curricular activity carried out within the schools. As an example, childhood obesity prevention might be achieved considering a variety of interventions targeting built environment, physical activity, and diet. As suggested by Dehghan and coll. (2005), some of these potential strategies for intervention in children can be implemented by targeting preschool institutions, schools or after-school care services as natural setting for influencing the diet and physical activity. In this sense, the outcomes from this study will be useful to school administrators in their bid to bridge the gap between those children who are the most and least deprived.

The author(s) declared that they have no competing interests.

All of th authors partecipated in the establishment of the research, and all read and approved the manuscript. GLT, DM and WR wrote the initial draft of the manuscript, which GLT edited.