Sedentary behavior and motor competence in children and adolescents: a review

ABSTRACT OBJECTIVE To synthesize evidence from studies that analyzed the associations between sedentary behavior and motor competence in children and adolescents. METHODS Systematic review of original articles that analyzed possible associations between sedentary behavior and motor competence in children and adolescents (3–18 years of age), without restrictions on study design, instruments and analysis protocols. The articles were identified through searches in the PubMed, Web of Science, Academic Search Premier, Cinahl, Medline and SPORTDiscus databases, as well as in reference lists. The level of evidence was evaluated according to the amount of studies that reported statistical significance in the associations between the variables and the quality of the articles (risk of bias). RESULTS Of 2,462 initial studies, 22 composed the synthesis (two interventions, nine longitudinal and eleven cross-sectional studies). Of these, in 13, we observed negative associations between the variables, more often in the age group of seven to fourteen years. In the analysis of risk of bias, the main limitations of the studies were “convenience sampling” and “no description of sample sizing”. CONCLUSIONS The available evidence suggests that sedentary behavior is negatively associated with motor competence in elementary school children, although the evidence is uncertain in the preschool years; the synthesis of results from longitudinal studies suggests that sedentary behavior negatively affects the development of motor competence. It is important that future studies have greater control over sociocultural determinants and deepen knowledge regarding sex and age, as well as the methods and indicators used to evaluate the two variables.


INTRODUCTION
High sedentary behavior (SB) index, defined by activities with energy expenditure ≤ 1.5 metabolic equivalents (MET) while sitting or lying down during the waking period 1 , is currently a serious global public health problem 2,3 . SB during childhood and adolescence has been identified as a predictor of SB in adulthood 4,5 and development of chronic diseases such as diabetes, hypertension and overweight or obesity [6][7][8][9] .
Several factors, such as access to technologies, the decrease in the supply of public space and rates of violence, have led children and adolescents to adopt a SB during much of their day [10][11][12] , particularly in leisure options involving screen activities (tablets, smartphones, computers, video games, television) 10,13,14 . Current guidelines suggest a daily limit of 1h in screen activities for the age group of 3 to 5 years and 2h daily from 5 to 17 years 15,16 . Despite these recommendations, studies 10,17,18 have reported high screen time values (more than 3 to 4 hours per day) in children and adolescents of different nationalities, a fact probably aggravated due to the covid-19 pandemic 19 .
Along with the high rates of SB, a decline in motor competence (MC) has also been reported [20][21][22] , that is, competence in the execution of motor skills, with organization and movement control compatible with age 23 ; this factor is potentially important to decrease SB indices and increase the practice of physical activity (PA) 12,24 . A growing body of evidence has indicated that MC favors participation in physical and sports activities [25][26][27] and associates with better overall health outcomes, including adequate body weight and increased cardiorespiratory fitness 28 . Children and adolescents with low MC tend to avoid physical activity 12,17 and adopt screen activities as leisure option 17,29 , which further restricts motor development and can generate a negative behavioral cycle, increasing the likelihood of physical inactivity and excessive SB throughout life. Recent research has found a negative association between SB and MC 12,30 which suggests a reciprocal relationship between these variables 24 , with important application in health promotion policies.
However, in a meta-analysis article, the evidence of association between SB and MC was considered uncertain 31 because, in addition to identifying only three surveys with children and adolescents [32][33][34] , they only found one 33 , conducted with children from 9 to 10 years of age, with significant association. Therefore, depending on the possible impact of SB guidelines on health policies 15 and considering the importance of MC development in youth 23,24,28 , there is a need for a greater understanding of the association between MC and SB, considering a greater number of studies and the possible differences between age groups.
Thus, the objective of this systematic review was to synthesize the evidence of studies that analyzed the associations between SB and MC in children and adolescents.

Protocol and Registration
This study is a systematic review, with its protocol registered in the International Prospective Register of Systematic Reviews (PROSPERO CRD42020161554). The full text was elaborated based on the items in the list Preferred Reporting Items of Systematic Reviews and Meta-Analyses (PRISM) 35 .

Eligibility Criteria
Based on the research question, original articles published in peer-reviewed scientific journals in English, Portuguese and Spanish were sought. More specifically, other items were established from the PICO strategy 36 , considering:

Population
Heterogeneous samples of children and adolescents (without specific deficiencies or clinical cases, except for samples specifically composed of overweight or obese children) aged 3 to 18 years. For our purposes, aiming to improve the degree of comparability and presentation of evidence, the subgroups were defined as follows, considering the Brazilian education system: preschoolers, from 3 to 6 years of age; elementary school, involving children and adolescents between 6 and 14 years; and high school, which covers adolescents between 15 and 18 years.

Intervention or Exposure
Intervention studies were included that implemented strategies for the control or reduction of SB, regardless of the context (for example, at school or in the community) and characteristics (whether by theoretical, practical activities or both).
Observational studies analyzed possible associations between SB and MC, based on SB as an exposure variable and MC as an outcome variable. For the record, no restrictions were imposed on the types (e.g. screen time, sitting time), domains (e.g. leisure, school and travel) and instruments (e.g. questionnaires and motion sensors) used to measure SB or MC. Considering the current understanding of the concept, studies that addressed "sedentary" as absence of physical activity were excluded.

Comparison
In the intervention studies, no restrictions were imposed on the existence or type of activities offered to the control groups, opting, when possible (as in studies with more than one control group), for comparators that received less activity load.

Outcomes (Health Indicators)
The outcome was motor competence. To evaluate it, the gross motor skills of locomotion, object control and balance were considered.

Study Design
Cross-sectional studies, cohorts and interventions were included that showed analyses on possible associations between SB and MC, regardless of the protocol used (e.g. univariate or multivariate analyses). Case studies, descriptive studies, reviews, meta-analyses, dissertations, theses and summaries of events were excluded.

Sources of Information and Search Strategy
To recover potential studies, in March 2020, systematic searches were applied in six electronic databases: PubMed, Web of Science, Academic Search Premier, Cinahl, Medline and SPORTDiscus from the strategy applied in PubMed: An author performed the initial search and entered all the recovered articles in the Rayyan platform (https://rayyan.qcri.org), where identification and removal of inter-database duplicates was conducted. Two authors (GS and SAM) independently reviewed the articles available by titles and abstracts. The results were compared and inconsistencies were discussed until a consensus was reached. If consensus was not reached, a third author (AHNR) would define the eligibility of the study. After this phase, the same authors evaluated the full texts of the remaining articles.

Data Extraction
Data were extracted independently by two authors (GS and SAM), using an electronic spreadsheet, which was organized into two levels of information: (1) descriptive (location, design, sample and age) and (2) methodological (type and measurement of MC, type and measurement of SB, statistics and main results). In particular, the results related to the analysis between SB and MC were extracted, considering the positive, negative or null associations, according to magnitude and level of significance p < 0.05. Data were extracted independently by sex only if data from the total sample were unavailable. In the case of different results for each sex, the study was classified as uncertain association.

Risk of Bias and Evaluation of the Quality of Studies
All included studies had their risk of bias assessed by two authors (GS and SAM), independently, with the support of the senior researcher (AHNR). For this purpose, the instrument developed by Lubans et al. 27 was used, who, in turn, were based on the items of the STROBE and CONSORT guidelines. Scores of 0 (absent or inadequately described) or 1 (present and adequately described) were assigned in six questions, namely: (a) "Does the study describe the eligibility/selection criteria of participants?"; (b) "Were participants randomly selected? " ; (c) "Does the study mention sources and details of the MC assessment, and do these instruments have adequate reliability for this specific age group?"; (d) "Does the study mention sources and details of SB assessment, and do all methods have acceptable reliability?" ; (e) "Did the study report sample sizing and was it adequately sized to detect hypothetical relationships?"; (f) "Does the study mention the number of subjects who completed each of the different measurements, and did these participants complete at least 80% of the MC and SB measurements?". It was previously established that studies with scores ≤ 2 would have high risk of bias; studies in the range between 3 and 4 points, medium risk of bias; and studies with scores between 5 and 6, low risk of bias.

Summary of Results
Considering the heterogeneity between the designs and the methods adopted, since the first treatments, the construction of a descriptive synthesis of the available results was stipulated. The judgment of scientific evidence was based on Lubans et al. 27 using the percentage of studies that reported a statistically significant association, while also considering the risk of bias: (a) lack of scientific evidence, if less than 33% of the studies indicate a significant association between the variables or none of the studies considered at low risk of bias find a significant association; (b) uncertain evidence, if 34 to 59% of the studies indicate a significant association between the variables; (c) positive (or negative) evidence, if 60 to 100% of the studies indicate a significant association between variables; (d) strong evidence, if 60 to 100% of the studies indicate a significant association between variables (in the same direction), there are no studies classified as uncertain association and more than 59% of the studies are considered to be at low risk of bias (score ≥ 5).

RESULTS
The flow chart shows the selection process ( Figure 1). In summary, of the 2,462 references initially identified, we evaluated 1,336 by titles and abstracts. Of these, we referred 36 for screening by full reading of the texts and excluded 14 because they did not show SB indicators (n = 12) or did not include the age group that was the target of our study (n = 2). Finally, we included 22 studies in the systematic review 12,17,30,32,33,37-53 .
Regarding the measures, fourteen studies used product measures, such as Bruininks-Oseretsky Test of Motor Proficiency Second Edition (BOT-2), Körperkoordinationstest für Kinder (KTK) and Movement Assessment Battery for Children Second Edition (MABC-2); to evaluate the MC, seven used process measures such as Test of Gross Motor Development (TGMD) -and only one study 17 used both types. The TGMD (first, second or third version) was the most used motor test (n = 8) 17,[38][39][40][41][42]49,53 . Three studies used the PE Metrics [43][44][45] , three studies used the BOT-2 37,40,51 , two studies used the KTK 33,48 and two studies used the MABC-2 17,50 . Only one study used the Athletic Skills Track 12 test and another used the Deutsche Motorik Test 30 . Three studies used specific motor tasks (e.g. walking backwards, balancing, throwing and grasping) 46,47,52 . Regarding SB measurement, sixteen studies used accelerometer as a direct measure and six studies used questionnaires. Regarding the questionnaires, five studies used them to evaluate recreational screen time and only one study used them to evaluate sitting time ( Table 2).
As for the risk of bias, 18.2% (n = 4) of the studies obtained a low-risk score (≥ 5), 77.3% (n = 17) obtained a medium risk score, and only one study obtained a high risk score. All studies met the criteria: (a) "Does the study describe the eligibility/selection criteria of participants?" and (d) "Does the study mention sources and details of the SB assessment?". Finally, the most absent quality items were: (b) "Were participants randomly selected?" and (e) Did the study report sample sizing and was it properly sized to detect hypothetical relationships?" (Figure 2).
Of the total studies investigated (n = 22), thirteen indicated negative associations between SB and MC, six did not indicate association and three indicated uncertain associations.  In the stratum of studies evaluated with low risk of bias (n = 4), a negative association was found in three studies, highlighting as main results: SB and high MC in schoolchildren 12 ; SB and MC in preschoolers 44 ; MC in childhood and sitting time after 20 years 49 . Studies with moderate risk of bias also showed negative association results 30,33,37,40,41,43,45,46,52,53 .
Considering experimental studies (n = 2) of medium risk of bias, a study 41 found a negative association between SB and MC in the ability to move and control objects, but the other 48 found uncertain association. Considering the longitudinal studies (n = 9), two studies with low risk of bias found a negative association between SB and MC 44,49 , and five other studies with medium risk of bias 30,40,43,45,52 also found a negative association between SB and MC. In contrast, a study of high risk of bias 51 found no association and a study of moderate risk of bias 47 found uncertain association. Considering the cross-sectional studies (n = 11) a study of low risk of bias found a negative association between SB and MC 12 and four studies of medium risk of bias 33,37,46,53 found a negative association between SB and MC. In contrast, a study of low risk of bias 39 and four studies of medium risk of bias 17,32,38,50 found no association, and a study of medium risk of bias 42 found uncertain association.   Analyzing the age group of preschoolers (n = 11), two studies of low risk of bias 44,49 and three studies of medium risk of bias 40,43,53 found a negative association between SB and MC. In contrast, a study of low risk of bias 39 , three studies of medium risk of bias 17,32,50 and one study of high risk of bias 51 found no association, while one study found uncertain association 42 . Therefore, according to the established criteria, the results show uncertain evidence of association between SB and MC in this age group. Considering the age group of elementary school (n = 11), a study of low risk of bias 12 and seven studies 30,33,37,41,45,46,52 of medium risk of bias found negative association between SB and MC. In contrast, a study of medium risk of bias 38 found no association and two studies found uncertain association. Therefore, the results indicate evidence of a negative association between SB and MC in this age group (Table 3).    Does the study describe the eligibility/selection criteria of participants?
Were participants randomly selected?
Does the study mention sources and details of the MC assessment, and do these instruments have adequate reliability for this specific age group?
Did the study report sample sizing and was it adequately sized?
Did the study report sample sizing and was it adequately sized to detect hypothetical relationships?
Does the study mention the number of subjects who completed each of the different measurements, and did these participants complete at least 80% of the MC and SB measurements?

Yes N o
In short, considering the result of the associations and the risk of bias, there was scientific evidence for the negative association between SB and MC of children and adolescents in elementary school and uncertain scientific evidence for association in preschool children.

DISCUSSION
This review synthesized the results of studies that evaluated associations between SB and MC in childhood and adolescence. According to the results, there is evidence for negative association between SB and MC in elementary school years. Even though most of the studies examined used a cross-sectional design, making inferences about causality difficult, seven longitudinal studies pointed to negative associations between SB and MC, suggesting that time in SB may impair the development of MC 30,40,[43][44][45]49,52 .
A meta-analysis by Engel et al. 54 suggests that MC and PA levels may increase through interventions in childhood, corroborating the findings of a previous review 55 , in which object control skills were strongly associated with PA in boys, and locomotion skills were associated with PA in girls. A recent longitudinal study demonstrated decreased PA and increased SB between 6 and 11 years 56 . A plausible hypothesis, based on the results found in this study and the conceptual model proposed by Stodden et al. 24 , is that the adequate MC could influence the increase of PA and the decrease of SB. It is possible that children with excess SB have few opportunities for developing MC 30 , further increasing the likelihood of adoption of SB (screen activities) in leisure moments 18,29 and generating a vicious behavioral cycle with unhealthy lifestyle habits associated with the emergence of chronic-degenerative diseases 24 . Thus, it is important that public health policies consider the reciprocal and dynamic relationship between SB and MC and promote, simultaneously, the improvement of MC and the decrease of SB, particularly leisure time in screen activities during childhood and adolescence.
As for environmental factors, the theoretical model by Hulteen et al. 57 suggests that motor skills vary in importance and popularity according to the cultural and geographical context, with an important role for the maintenance of PA throughout life. Considering the role of physical activity as a form of intervention to decrease SB and contribute to the development of MC, it is possible to suggest that the higher the MC in different abilities (for example, kicking, throwing, bouncing and jumping), the greater the range of possibilities of PA practice that could replace SB. The development of MC in skills that predominate in the cultural and geographical context in question can be an effective strategy to replace SB with a PA that the child or adolescent has the competence to perform. Reinforcing this perspective, data from a Finnish study 41 point out a negative association between kicking skills and SB, being soccer one of the predominant youth sports in the country 58 .
Data from this review reinforce the importance of integrated public education and health policies focused simultaneously on improving MC and decreasing SB, particularly screen time at leisure. Interventions in children who spend excessive time in SB should include the In addition, there is a wide scope to be explored, considering environmental factors, gender and assessment methods, as well as the gap in the transition between age groups (children, adolescents, young adults and older adults). It is plausible that the causal direction between the variables is influenced by the age group. We suggest that future studies use two or more motor tests 22,61 and the use of the accelerometer in conjunction with the questionnaire for a better understanding of SB and environmental factors. We also suggest randomized selection of participants and description of the statistical power of the study. Finally, it is also important to highlight that the results found are limited to adolescents up to 14 years, providing a wide field of research among adolescents and young adults in this topic.
This study contains some limitations. Although the search was comprehensive, we could only include studies in the English language. The bias related to the selective reporting of associations in the studies may be a possibility, and adjustments in the associations between SB and MC were not considered as a function of the practice of PA. In addition, limitations in the evidence base influenced the results of this review. The practice of moderate to vigorous PA (MVPA) may partially offset the negative effects of SB 31,62 , being independent behaviors, i.e. high levels of SB do not necessarily imply low levels of MVPA and vice versa; people who meet the recommendations of MVPA (≥ 1h daily) may still have SB at many hours of the day 63 . In reality, to date, SB recommendations are restricted to screen time 1,15,64 . From a mathematical point of view, if the recommendation of at least three hours a day of PA at any intensity (including 1h of MVPA) is applied 1,15 , and considering a period of 12 hours of wakefulness, the time limit of SB would be 9 hours daily. Therefore, a greater understanding of the context in which SB manifests itself and its association with the overall development of the child or adolescent is necessary.
Another limitation is the heterogeneity in the forms used to evaluate SB 65 (questionnaires with information about total screen time or only TV/computer time or sitting time and direct accelerometry measurements), making it impossible to differentiate SB in screen/leisure activities and SB in other domains, such as sitting time in displacement or studies. The use of different motor tests to measure MC is also an important limitation.
In total, at least 10 different MC evaluations were used. In literature, measures oriented to the process or product of movement can evaluate MC, and this can influence the magnitude of the associations 22 .
Finally, the results of this systematic review demonstrate that there is evidence of a negative association between SB and MC in children and adolescents in the elementary school period, although the evidence is uncertain in the preschool years. Environmental factors, as well as the method of assessment and age group, can be determinants for a better understanding of the investigated phenomenon. To better understand the associations between SB and MC, we suggest the establishment of standardized criteria for conducting studies, highlighting the context in which SB manifests itself (for example, in leisure situations or school activities). Improving the predominant MC in the sociocultural context in which the person is inserted can contribute to the decrease of SB and promote engagement in an active lifestyle in the long term for children and adolescents.