The prevalence of global physical activity among young people : a systematic review for the Report Card Brazil 2018 A prevalência de atividade física global entre jovens : uma revisão sistemática para o Report Card Brazil 2018

The aim of this study was to summarize data on the prevalence of global physical activity (PA) among children and adolescents in Brazil. This systematic review included an electronic search in 8 databases (PubMed, Scopus, Web of Science, LILACS, SPORTDiscus, BIREME, Scielo, and Google Scholar) and a manual search in the references of retrieved studies. Observational studies that assessed global PA among Brazilian children and adolescents were included. A narrative approach toward results was adopted. An initial database search reached 3276 potentially relevant titles (2534 titles after duplicate analysis); 92 (62 different studies) met all inclusion criteria. Studies were carried out between 2001 and 2015; most studies were conducted in Southern and Northeastern Brazil. Three were nationwide surveys (4.8%), but no studies included children aged 0-5 years. Eighteen different instruments for PA measurement were used; only two studies used objectively-measured PA. The most used definition was ≥ 300 minutes/week of moderate-to-vigorous PA. The overall proportion of young people who were physically active ranged from 6.5% to 92.3%. Fourteen studies (22.6%) found a prevalence of ≥ 50%. Three nationwide surveys reported the prevalence of physically active students ranging from 29.0% to 66.0%. A wide variability of PA estimates were found. Important research gaps (e.g., studies with objectively-measured PA, children aged up to 6 years-old and in Northern Brazil) should be considered to develop new research studies in Brazil.


INTRODUCTION
Physical activity is strongly associated with development, health, and quality of life among children and adolescents 1 .Despite this, only two out of ten adolescents meet physical activity guidelines (60 or more minutes per day of moderate-to-vigorous physical activity) worldwide 2 .Nationwide data show a similar trend in Brazilian adolescents 3 .Thus, policies and programs for physical activity promotion are urgent, and monitoring how and whether intervention strategies can improve physical activity practices among young populations is an important component of public health action.
In 2014, researchers, health professionals, and stakeholders built The Active Healthy Kids Global Alliance, a network focused on advancing physical activity among children and adolescents from around the world.For this, physical activity indicators (e.g., meeting physical activity guidelines, active transportation, and others) have been monitored and described in a document called a "Report Card", in order to be a useful tool for the advocacy of physical activity 4 among young people.In 2016, data from 38 countries were included, including Brazil.A grade of C-(prevalence rates from 40% to 46%) was given for global physical activity levels based on data from studies with Brazilian adolescents -this was one of the lowest grades for Brazil's 2016 Report Card 5 .
The present study aimed to systematically summarize the literature on the prevalence of global physical activity (based on international physical activity guidelines) among Brazilian children and adolescents.The summarized data were used to define the grade and recommendations for Brazil's 2018 Report Card for the global physical activity indicator.

Measured Outcome
In this systematic review, physical activity outcome was defined as any bodily movement produced by skeletal muscles that require energy expenditure 6 .Studies were considered whether they measured global physical activity (e.g., different domains combined (leisure, transportation, home and/or school combined), in order to follow the global physical activity guidelines 7 .

Study Search Strategies
A systematic search was conducted in electronic databases Medline (Pub-Med), Scopus, Web of Science (Web of Knowledge), LILACS (Literatura Latino-Americana em Ciências da Saúde), SPORTDiscus, BIREME (Biblioteca Regional de Medicina), Scielo, and Google Scholar in February 2018.The search strategy included four groups of descriptors: outcome (physical activity), physical activity evaluation methods, population (young people) and country (see Supplementary Material).The Boolean operator "OR" was used for intragroup combinations, and the "AND" operator was used for intergroup combinations.The truncation symbols ($, * or "") specific to each database were also used to increase the range of searches for the descriptor variations.Searches were conducted with the descriptors in English and Portuguese, when required.The search of the electronic databases was supplemented by a screening of the reference list of retrieved articles in order to find potentially relevant titles and the personal library.

• Selection Process
The initial analysis was performed based on the reading of the manuscripts titles and, when there was doubt regarding the inclusion of the study, a reading of the abstract was carried out.After this analysis, articles were obtained in full text version and subsequently analyzed according to established selection criteria.Subsequently, the screening of the reference list was carried out.All the process was conducted by independent peers (RC/ MK or BO/CS), and a third author (KS or VB) helped with disagreements.

• Selection Criteria
To be eligible for inclusion in this review, studies had to be: (I) articles published in peer-reviewed journals that were original research; (II) samples with Brazilian children and adolescents aged 0-18 years (or a mean age within these ranges, or separated data for individuals at this age range) (III) observational studies showing the prevalence of at least one of global physical activity (e.g., % of children and adolescents who accumulate a combined total of at least 60 minutes of daily moderate-to vigorous-intensity physical activity), regardless of whether the study dealt with this behavior as an exposure or an outcome; (IV) studies using different methods for PA assessment (e.g., self-report, structured interviews, objectively-measured PA, and steps per day); (V) a school-or population-based survey with information about the methodological procedures of representation of the target population (e.g., random sampling).

• Data Extraction and Synthesis
Information of each study was extracted by independent peers (RC/MK and BO/CS) and a third author (VB or KS) helped with disagreements.Extracted information included: authors and year of publication, local of the study, year of data collect, age range, sample type, sample size, percentage of girls, instrument description, instruments characteristics and mode of administration, and physical activity cut-off definitions.A narrative approach for the results was adopted because the heterogeneity of the study's data (no meta-analyze was performed).
Also, the proportion of young people that were considered physically active (based on the physical activity definition in each study), as well as the proportion of this outcome for boys and girls separately, when presented.
Results of the studies were presented in alphabetical and chronological order by first author name and year of publication, respectively.The results were organized in order to present methodological aspects (Table 1) and the prevalence of global physical activity (Table 2) of the included studies.
Considering the results of adolescents according to age groups, the proportion of physically active children (up to 6-10 year-olds exclusively) ranged from 10.9% 15 to 29.2% 89 .Studies with adolescents exclusively (10 years or older) presented proportions that ranged from 8.3% 17 to 92.3% 64 (Table 2).
Included studies that used the cutoff point of ≥ 300 minutes/week of moderate-to vigorous physical activity presented proportions of physically active individuals that ranged from 18.4% 86 to 92.3% 64 .Studies using the cutoff point of ≥ 60 minutes/day moderate-to vigorous physical activity presented proportions of physically active individuals that ranged from 12.6% 20 to 68.2% 65 .Studies using the cut-off score for PAQ-C/PAQ-A scores presented proportions ranging from 6.5% 12 to 23.5% 101 .Studies that used the cutoff point of the IPAQ short presented proportions varying from 66.0% 82 to 69.8% 77 .
Regarding objective measurements of PA, the population-based study that used accelerometers 61 was conducted in South Brazil with 457 adolescents who were 12-14 years old.This study found overall prevalence of physically active of 61.3% 61 .A school-based study with 1139 6-8-year-old students from Northeastern Brazil the used pedometer-measured physical activity found an overall prevalence of 35.7% physically active individuals 23 (Table 2).
Trends in the prevalence of physical activity were observed in three studies, which presented two waves of data collection.The National School Health Survey (2009 and 2012) described percentages of physically active students of 43.1% in 2009 83,84 to 29.0% in 2012 33 .However, the number of questions used to define physical activity level was different between these surveys.The Health Risk Behaviors Project (COMPAC study, a study in Santa Catarina state) showed a prevalence of physically active boys and girls of 72.1% and 71.2% in 2001 9 , and 39.3% and 23.7% in 2011, respectively 27 .A statewide survey in Pernambuco state presented the prevalence of physically active students of 34.9% in 2006 31,68 and 35.1% in 2011 96 .

DISCUSSION
The objective of this systematic review was to synthesize the studies on the prevalence of physical activity in children and adolescents in Brazil.Our findings showed that there was a wide range between studies in the prevalence of physically active children and adolescents.However, thirty-five studies found a prevalence of physically active children and adolescents lower than 50% 12,13,[23][24][25][26][27][28][29][30][31]36,14,[40][41][42][43]45,[49][50][51][52][53]16,54,[56][57][58][59]62,63,[66][67][68]17,81,[90][91][92][96][97][98]101,[18][19][20][21][22] , including two National School Health Surveys 33,84 . Thus, it seems that increasing physical activity practices among Brazilin children and adolescents is a priority. The lrge variability between studies in prevalence rates may be explained by the distinctions in the instruments/questionnaires used.Recording period, instrument structure, wording of questionnaires, physical activity domains included, and cutoff points aspects that varied between surveys (Table 1).For example, Farias Júnior et al. 73 found that estimates of physically active adolescents using simplified and detailed instruments, can have a relative difference of 100% or higher.Subjective measures of physical activity are still predominantly used in epidemiological research in Brazil 104 and other low-and middle-income countries due to the context of where the data is recorded and the low operational costs involved 105 .However, it is important that future studies consider developing and using validated and reliable instruments for physical activity measurement in epidemiological surveys in Brazil.
Inversely, this systematic review included only two studies with objectively-measured physical activity 23,61 .In Brazil, the high cost of objective measures of physical activity can still be a barrier to their use in comprehensive survey studies.Another obstacle to be overcome for optimizing the use of objective instruments, in particular accelerometers, is the problematic comparability between different models and outputs 106  surveys using objectively-measured physical activity (e.g., accelerometers) are still on the rise in Brazil, and well-designed studies should be developed in order to reduce comparability issues of physical activity data.
Most of studies in Brazil highlighted physical activity distinctions for boys and girls, which was similar to the global estimate 107 .However, when considering age groups, no studies were found that investigated the prevalence of physically active preschool children (under 6 years old).An explanation for this result may be that the measurement of physical activity with subjective instruments is unreliable and most of questionnaires were not validated for younger children.
The results of this systematic review highlighted other trends in the research on physical activity prevalence in Brazil.Since 2001, there has been an increase in research on physical activity in children and adolescents with a concomitant evolution of the robustness of the studies developed in the Brazilian scientific community.This is also supported by two versions of the National Survey of School Health (2009 and 2012), which provided information for the Surveillance System for Risk Factors for Non-communicable Chronic Diseases 34,84 .These results reflect the emphasis on epidemiological surveys in the Brazilian physical activity research agenda during this period.
The results showed that most of the studies were carried out in South and Northeast regions of Brazil.Previous systematic reviews 104,108 found a large number of studies on physical activity and health in the South and Southeast regions of Brazil.Thus, a positive finding in the current review was the relevant number of included studies in Northeastern Brazil.This may be related to the number of research groups and graduate programs in Physical Education in this region of Brazil, which can stimulate new and well-designed research practices for studies in these areas.However, the development of studies in the Northern region of Brazil is urgent, because no studies of this region were included in the current review and physical activity promotion strategies for young populations in the Northern area need to consider context-specific needs and peculiarities.Research groups focused on physical activity and health should be stimulated and integrated with other multicenter studies to promote best research practice for studies on physical activity and health in this region.
This systematic review is important because research on the physical activity level of Brazilian children and adolescents was presented with new evidence in comparison to previous reviews of this topic 109 .Also, gaps and recommendations for studying physical activity were highlighted.However, a large variability in the measurement of physical activity between studies limited the possibilities of a quantitative synthesis (meta-analysis) of the physical activity prevalence rates from all studies and of a discussion of other relevant aspects of physical activity in Brazil, such as secular trends.

CONCLUSION
The current review found that there was great variability in the measure-ment of physical activity and cut-off points among the studies with children and adolescents in Brazil.The large range of physical activity prevalence was something remarkable in this review.However, 35 studies found a prevalence of physically active children and adolescents lower than 50%.Thus, it seems that physical activity practices among Brazilian children and adolescents is still a challenge for the country.Important research gaps (e.g., studies with objectively-measured physical activity, children aged up to 6 years old and in Northern Brazil) should be considered when developing new research studies in Brazil.

Figure 1 :
Figure 1: Search process results according PRISMA flow diagram

Table 1 .
Description of the studies on the global physical activity among Brazilian children and adolescents.

Table 2 .
Description of prevalence of physical activity (%) in studies including Brazilian children and adolescents.