Reference index and reduction in physical fitness tests proposed by PROESP-BR

–The purpose of this study was to create an unifying index of the PROESP-BR tests for school aged teenagers and propose a reduction in the physical fitness tests. A total of 414 adolescents between the ages of 15 and 17 years old, representative of the public school population in the city of Cuiaba were evaluated. The tests include general physical fitness, proposed by PROESP-BR. Multivariate factorial analysis was used, observing the commonality/ representativeness of each test with regard to the set, and the index was created for girls and boys. With this analysis we can choose to use the following tests: female throwing medicine ball, horizontal jump and run 20 meters; male throwing medicine ball, representing the overall performance. The created index was divided into quintiles and allowed you to view the whole performance of the six tests, as well as their distribution within the group and away from the ideal reference.


Introduction
There is growing concern about identifying the existent conditions of risk for any change in the development of children and adolescents (Zajons, Muller & Valentini, 2008).Physical capabilities are directly related to health and sporting performance, and evaluating them has helped with the diagnosis and prevention of diseases as well as in detecting sporting talents (Saraiva & Rodrigues, 2010).
Some international surveys that describe the profile of motor performance of youngsters have observed the association between low physical fitness and school performance (Koutedakis & Bouziotas, 2003).In addition, low physical fitness in adolescence has a negative reflection on adult life (Mikkelsson et al., 2006).Silva, Lima, Silva, and Prado (2009) found a prevalence of 83.1% of sedentary adolescents in the Brazilian population whereas other studies have observed variations between 39% and 93.5% of Brazilian students (Tassitano et al., 2007;Silva et al., 2005).
The lack of standardization of instruments for measuring physical activity hampers comparisons between Brazilian regions, with the Mid-Western region being prejudiced by the scarcity of studies (Silva et al., 2009;Hallal et al., 2007), as well as comparisons with other populations in this area.Therefore, it is crucial to establish equivalent normative values and reference parameters for Brazilian population establishing parameters with others studies of other regions and countries (Drews, Cardozo, Corazza, & Flôres, 2013).Thus, to get a representative number of population it is necessary to have a large scale of the samples, the goal of our study was to evaluate the largest possible number of children and adolescents.It would be interesting to have a general evaluation index for all the factors, whether it is with regard to health or to sporting performance.Therefore, the more condensed the evaluations could be, the easier the reproducibility will be.
Thus, the aim of this study was to create a general index of the tests performed by the parameters of Project Sport Brazil (PROESP-BR) and verify the possibility of reducing the number of tests established by PROESP-BR to the lowest possible number and how reliable is.

Methods
Data was analyzed from 414 adolescents representing the study population of 11.006 high school level students of the 28 schools in the state public network.All of them were within the age-range of 15 to 17 years of age for both sexes and residents within the urban perimeter of the municipality of Cuiabá-MT, situated in the Mid-Western region of Brazil.This study was performed using data Fontes (2013), which held sporting diagnosis and health of students using the parameters and classification of PROESP-BR (2011), in the following tests: abdominal resistance, throwing the medicine ball, horizontal jump, 20-meter run (displacement speed), 4x4 meter square (agility) and 9-minute run (Table 1 and 2

Statistical Analysis
Multivariate factorial analysis was used to detect the factor of correspondence of each test and afterwards create the index.This was performed with the statistical software program Minitab version 15.0 which, by means of the raw data of the variables found the matrix of correlation and determined the number of factors necessary (≥1).We used commonality for the proposed reduction of the variables with the correlation of each variable being explained by all the factors.T herefore, the higher the value of communality (closer to 1.00); the higher the power of explanation of the variable would be.For the creation of the students' rating it was necessary to create a 'gold standard' reference.Arbitrarily a "school best" was developed, which was established from the best outcome of each test for each sex.In short, no individual had the best marks in all tests, so it was necessary to use the best result obtained in each test for different schools.
The index was obtained by multiplying the standardized variables by the factors obtained, thereby we determined the amplitude between the indices obtained and the data were grouped into 05 groups, in which the lowest values corresponded to the greatest proximity to the values of the excellent adolescent students, therefore meaning the best result.

Results
To create the index of each student, it was necessary observe the score of each test on the whole (values of commonality), thus we see the possibility to reduce the amount of testing and the difference between male and female groups.Regarding the proposal for reduction in the tests, there was a variance of 65% and 58% of the importance of each variable for female and male group, respectively.The female group had two tests with high values, three with medium values, and one with a low value and in the male group there was one high, four medium and one low test value.Thus, if we consider values of correspondence above 70% to represent the set, only two tests would be necessary for the female group and one test for the male group (Tables 3 and  4).However, we observed that there was correlation between the tests themselves, and only 01 test in each group had values below 50%.

Discussion
The present study is a continuation of the work of Fontes (2013) who had identified a low performance of adolescent students in the PROESP-BR tests.Subsequently, we developed a statistical analysis that allowed a different manner of observing the behavior between the genders with regard to the power of the tests.In the female group, two tests were of high importance to represent the overall performance whereas in male group only one test seems high representative.In addition, an index that represents the overall performance of the adolescent students was developed, which provides a spatial view of the distribution of the physical capacity among these adolescent students and in comparison with the excellent adolescent students.This approach shows that these adolescent students had a low physical performance because the best classified among them were equidistant from both the excellent adolescent students and the worst classification.Moreover, it allows an approach to a specific classification for this group, considering that with regard to the original PROESP-BR classification, they are only at the margin of the excellent category.
Studies using multivariate factorial analysis have been explored to a small extent in the field of physical education up to now.In our study, the determination of an index may help future evaluations with the proposal of a smaller number of tests.Thus, we observed that the extraction of two factors According to analysis, we obtained the index of adolescent students and with the creation of the excellent female and male students within our own sample; we divided the classification into five groups, stratifying by the amplitude among the data.For the female classification, the amplitude was 19.56 resulting in the interval of 3.85 for each classification; for the male classification the amplitude was 17.49, resulting in the interval of 3.50 between the classifications.The classification of each student was determined by the distance between the school adolescent students from the excellent adolescent student, and the shorter the distance between them, the better the result (Tables 5 and  6, Figures 1 and 2).was possible, for girls and boys, these representing 65% and 58%, respectively, of the total variance in the results, showing that these tests have greater impact on the entire battery of evaluations performed.
A reduction in the number of tests may facilitate the development of the evaluations, since we observed that if some tests were not performed would not affecting the final score of the overall index.Indeed, Peterson, Alvar & Rhea (2006) stated that the horizontal jump is sufficient in the evaluation of power, speed and agility; while Guedes (2007) reported that the 20-meter run is capable of providing information about the same physical fitness level.One of the main reasons for physical follow-up is related to health; the components of force, resistance, and agility are directly related to the prevention of neuromuscular problems (Welk, Going, Morrow, & Meredith, 2011) since it has been associated with maintaince of the metabolism of some substrates and hormone regulation (Guedes,, Miranda-Neto, Germano, Lopes & Silva, 2012).Sharkey (1997), demonstrated that performance of physical tests that use greater muscle recruitment are more efficient in the diagnosis of overall physical fitness of youngsters who are not athletes.This corroborates with the finding for the male group in our study, in which the test with greatest score was the 9-minute run.On the other hand, we found that the tests with the greatest score, for the female group, were throwing the medicine ball and horizontal jump.However, in the female group this is not an absolute factor, because the horizontal jump is considered important for overall evaluation.Castro-Pinero et al. (2010), observed a strong association between the horizontal jump test and tests of muscle force of the upper and lower limbs.Furthermore, they considered that this test could be used as a general index of muscle aptness in adolescents.Further studies with different populations and biochemical analyses are necessary, in order, to better determine these inter-gender variations.
With the factors of each test, we verified that the female group had more distributed physical performance results, and only two tests were necessary to obtain the same index.Differently, in male group, which was more homogeneous, only one test could be performed providing the same information.With regard to classification, we observed that the female group had better results than the male group.This differs from the classification made by Fontes (2013) and other studies, such as Dumith et al., (2010) and Guedes, Neto & Silva (2011) which analyzed the difference between the genders in some physical fitness tests and found that boys received higher scores than girls in the majority of the tests.Silva, Nahas, Hoefelmann, Lopes and Oliveira (2008) pointed out that male adolescents are more physically active, particularly in the field of leisure and work, due to cultural influences and the strong trend towards self-determination (Viana, Andrade & Matias, 2010).
The index created may facilitate the diagnosis and follow-up of school physical education in the population diagnosed.Persisting with that which the study of Silva, Böhme, Uezu and Massa (2003) considered, the importance of creating indices that provided references for each specific population, with the justification of helping with the follow-up of lessons; and considering that in the present study the adolescent students presented low physical performance.This suggests that they are exposed to a low quality/quantity of motor activities.Similar to our results, Pelegrini, Silva, Petroski, & Glaner (2011) observed that adolescent Brazilian students presented inadequate levels of physical fitness.This situation may be due to cultural changes, as regards exposure to passive leisure, such as electronic games, TV and other entertainment (Dumith et al., 2010).According to Hallal, Knuth, Cruz, Mendes, and Malta (2010), the majority of adolescents in the Mid-Western region of Brazil are insufficiently active.
Physical Education lessons, in turn, which must contribute to improving this situation, are apparently not changing the lack of physical activity.Indeed, other studies have observed that the level of physical effort in the Physical Education lessons are not sufficient to increase physical fitness (Guedes & Guedes, 2001) both in public and private schools (Hallal et al., 2010;Silva, Lima, Silva, & Prado, 2009).Another studied has related that only 49.2% of the adolescents in major Brazilian cities responded to participating in two weekly physical education classes (Hallal, Knuth, Cruz, Mendes & Malta, 2010).Therefore, with the help of school planning, it is necessary for teachers to interact and develop socialization, challenging the processes of upbringing and lifestyles and demonstrating to the adolescent students the importance of physical activity as the opportunity to improve their quality of life (Gallahue & Donnelly, 2008).
Although we created the excellent adolescent students with the best results in our sample, this is not the ideal, because considering the general classification used by PROESP-BR, those students are not within the 98th percentile; so far, obtaining excellent results in at least two tests.The present index is representative for the sample population.Of note, we observed that the studied population lacks specific evaluations, and this instrument is useful for teachers to follow-up the development of their pupils.Furthermore, knowledge of the local reality leads us to the need for formulating public policies and specific interventions with regard to physical education lessons and the promotion of school sports.

Conclusion
With the results of factorial analyses we obtained the score of each test, allowing a reduction in the tests and showing that these tests do not have the same gradet for both sexes, therefore, it is necessary to observe the most characteristic physical qualities of each gender.From the factors analyzed, an index could be created for each pupil, helping with visualization of the overall motor performance of adolescent students, and this can be used by Physical Education teachers for following-up their pupils' performance.This index facilitated visualization of the results of the adolescent students' performance, which pointed out a low level of physical fitness in both genders and demonstrating that the intervention in physical activity may not be adequate for this population.

Figure 2 .
Figure 2. Dispersion graph according to the male group.

Figure 1 .
Figure 1.Dispersion graph according to the female group.

Table 3 .
Result of correspondence factors of each physical test stipulated by PROESP-BR of females.

Table 6 .
Index created for classification of males.
Excellent group: excellent student created N and %: quantity of students in each group.

Table 4 .
Result of correspondence factors of each physical test stipulated by PROESP-BR of males.

Table 5 .
Index created for classification of females.
Excellent group: excellent student created N and %: quantity of students in each group.