Background

Higher levels of physical activity are associated with superior physical, social and psychological health in young people. Physical activity is inversely related to overweight and obesity in youth [1, 2] and poor body composition in childhood is associated with an increased risk of coronary heart disease in adulthood [3]. While there has been debate about whether or not physical activity levels have declined in the last 30 years [4], evidence suggests that there has been a decline in active travel to school (ATS) among children and adolescents in many countries [5–7]. For example, data from the US National Personal Transportation Survey revealed that the number of students who walked or cycled to school decreased from 41% in 1969 to only 13% in 2001 [6].

ATS includes various modes of travel such as walking, cycling, and skateboarding, and has been identified as an important source of physical activity for young people [8]. Previous reviews [9–11] have examined the relationship between ATS, body composition and physical activity and reported strong evidence for a positive association between ATS and overall physical activity level, but little evidence linking ATS to leaner body composition. However, these reviews did not stratify their results or conclusions by study quality or risk of bias [12]. While previous reviews included studies that examined the relationship between ATS and cardiorespiratory fitness, the authors were unable to determine the association due to limited evidence. Since the publication of these reviews there has been an increase in the number of studies exploring this relationship.

It has been suggested that much of the activity completed by children and adolescents is of insufficient volume and intensity to improve health-related fitness (i.e., body composition, cardiorespiratory fitness, muscular fitness and flexibility) [13]. While the exact volume and intensity of activity required to improve fitness in young people is not clear [14], it is plausible to suggest that habitual ATS has the potential to improve health-related fitness (HRF) among youth. ATS provides a regular opportunity for young people to accumulate physical activity, but evidence to support or refute the health benefits of active transportation is limited due to the focus on assessing time spent in vigorous activity [8]. Furthermore, few studies have explored the effects of ATS on HRF among young people in longitudinal studies [15]. The aim of this review, therefore, was to systematically examine the potential health benefits associated with ATS among children and adolescents. The secondary aim was to explore the quality of studies that have examined the relationship between ATS and HRF and to determine whether study quality may have confounded these relationships.

Methods

Identification of studies

The Quality of Reporting of Meta-analyses statement (QUOROM) [16] provided the structure for this review. A systematic search of seven electronic databases (EMBASE, OVID MEDLINE, PsycINFO, PubMed, Scopus, SPORTDiscus and TRIS Online) was conducted in December 2009. Studies published from 1980 up to and including the search date were considered for inclusion. Individualized search strategies for the different databases included combinations of the following key words: (child* OR adolescent OR youth OR "young people") AND (walking OR "active travel" OR "active transport*" OR cycling OR riding OR "travel mode" OR trip) AND ("health-related fitness" OR "body composition" OR "obesity" OR "weight status" OR "physical fitness" OR fitness OR "cardiovascular fitness" OR "cardiorespiratory fitness" OR "aerobic fitness"). Only articles published or accepted for publication in refereed journals were considered for the review. Conference proceedings and abstracts, therefore, were not included. In the first stage of the literature search, titles and abstracts of identified articles were checked for relevance. In the second stage, full-text articles were retrieved and considered for inclusion. In the final stage, the reference lists of retrieved full-text articles were searched and additional articles known to the authors were assessed for possible inclusion.

Criteria for inclusion/exclusion

In the first stage of the literature search, two of the four authors (DRL and PK) independently assessed the studies for inclusion. Studies were considered to be eligible for inclusion according to the following criteria: (i) participants were children or adolescents aged 5 to 18 years, (ii) study reported active school transportation in children and/or adolescents, (iii) study assessed at least one component of health-related fitness (i.e. cardiorespiratory fitness, muscular fitness, body composition or flexibility), (iv) study included quantitative analysis of the relationship between ATS and at least one component of health-related fitness, (v) study published in English.

Criteria for assessment of study quality

Two authors (DRL and PK) independently assessed the quality of the studies that met the inclusion criteria (Table 1). The criteria for assessing the quality of the studies were adapted from the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [17]. Items that were considered to be the most important in relation to study bias were included in the concise checklist. A formal quality score for each study was completed for a 6-item list by assigning a value of 0 (absent or inadequately described) or 1 (explicitly described and present) to each of the following questions listed: (i) Did the study describe the participant eligibility criteria? (ii) Were the study schools/participants randomly selected (or representative of the study population)? (iii) Did the study report the sources and details of ATS assessment and did the instruments have acceptable reliability for the specific age group? (iv) Did the study report the sources and details of HRF assessment and did all the methods have acceptable reliability? (v) Did the study report a power calculation and was the study adequately powered to detect hypothesized relationships? (vi) Did the study report the numbers of individuals who completed each of the different measures and did participants complete at least 80% of measures? Differences were resolved by a third reviewer (CEF).

Table 1 ATS study quality checklist with quality scores assigned
Full size table

Categorization of variables and level of evidence

The level of association between ATS and HRF was determined using the coding system first described by Sallis et al. [18] (Table 2). The relationship between ATS and HRF components was determined by examining the percentage of studies that reported a statistically significant relationship (i.e. between ATS and the HRF component). If only 0-33% of the included studies reported a statistically significant relationship between ATS and a given HRF component, the result was categorized as no association (0). If 34-59% of the studies reported statistically significant relationships between ATS and the HRF component, the result was categorized as indeterminate (?). If 60-100% of studies reported a positive relationship between ATS and the HRF component, the result was coded as a positive association (+). If 60-100% of studies reported a negative relationship between ATS and the HRF component, the result was coded as a negative association (-). The methods of Sallis et al. [18] were modified to address the issue of study quality, and additional coding was conducted based on studies identified as high quality. If 60-100% of high quality studies (≥ 4) found a relationship between ATS and a given HRF component, the result was coded as having strong evidence for a positive association (++) or negative association (--).

Table 2 Summary of studies examining the relationship between ATS and HRF in youth
Full size table

Results

Overview of studies

The flow of studies through the search process is reported in Figure 1. While the majority of studies (Table 3) were conducted in the US [19–25], the relationship between ATS and HRF was explored in a further 12 different countries, including the Philippines [8], England [26–29], Holland [30], France [31], Denmark [15, 32, 33], Portugal [34, 35], Australia [36, 37], China [38], Spain [39], Germany [40], Canada [41], and Brazil [42]. The study sample sizes ranged from 107 [37] to 21,345 [41]. While most of the studies involved cross-sectional study designs, Cooper and colleagues [33] examined longitudinal changes in transport to school over a 6-year study period. Similarly, Rosenberg et al [23] explored the relationship between school travel and weight status over a 2-year period. The prevalence of ATS among study participants ranged from 3 and 8% in boys and girls [22], to 72 and 79% in boys and girls [15].

Table 3 Summary of included studies
Full size table
Figure 1
figure 1

Flow diagram of study selection. Note. ATS = active travel to school, HRF = health-related fitness

Full size image

Overview of study quality

There was 90% agreement between reviewers (DRL and PK) on the study assessment criteria and full consensus was achieved with the help of a third reviewer (CEF). Eleven studies scored ≥ 4 on the study quality assessment [20, 24, 28, 31–33, 36, 40–43] (Table 1). The majority of studies used samples of schools/participants who were randomly selected and/or representative of the study population. Most of the studies described the participant eligibility criteria, although participant eligibility (e.g. were participants selected from all grades at participating schools?) was unclear in three studies [21, 29, 30]. The majority of studies reported standardized procedures for the assessment of height, weight and cardiorespiratory fitness. Two studies reported parental proxy reports for their children's height and weight and did not provide reliability/validity data [21, 37] and a further two studies used self-reported height and weight to calculate BMI [19, 30]. While the majority of studies involved large sample sizes and were adequately powered to detect hypothesized relationships, only two studies reported power calculations relevant to their study aims [42, 43]. Fifteen studies reported the number of individuals who completed each of the different measures.

Overview of findings

Twenty three studies examined the relationship between ATS and weight status/body composition (Table 2). ATS was associated with more beneficial weight status (or lower body fat) in 11 (48%) of these studies. However, when lower quality studies (< 4) were excluded from the summary, five of the nine remaining studies found ATS to be beneficially associated with weight status. Five studies examined the association between ATS and CRF [15, 25, 29, 32, 33]. Four of these studies [15, 29, 32, 33] found a positive association, while the fifth study [25] found a non-significant (p < .07) trend between ATS and CRF. More specifically, three of four studies involving cycling reported positive associations with CRF, and a 6-year longitudinal study [33] demonstrated that change in travel mode to school from non-cycling to cycling was a significant predictor of CRF at follow-up. Further evidence supporting the benefits of cycling was provided by Andersen and colleagues [15], who found that muscular fitness and flexibility were significantly better among Danish cyclists compared to walkers and those who used passive transportation.

Discussion

The aim of this review was to systematically examine the potential health benefits associated with ATS among children and adolescents. Of the 27 studies that were included in the review, 25 examined the relationship between ATS and weight status/body composition. While only 48% of these studies reported significant inverse associations between ATS and weight status/body fatness, this increased to 55% once poor quality studies were removed. Based on the findings from five studies, including one longitudinal study, it appears that ATS may be associated with superior cardiorespiratory fitness in youth. It was not possible to draw any conclusions about the associations between ATS, muscular fitness and flexibility due to the small number of studies exploring these relationships.

Despite methodological differences, these results are similar to those of previous reviews [9, 10]. Two previous ATS and health reviews reached the same conclusion about a lack of association between body weight and ATS despite not assessing the quality of their included studies. The decision to use an assessment of study quality or risk of bias was both pragmatic and justified. The application of this metric allowed for the removal the studies with the greatest risk of bias. Faulkner and colleagues' review [10] reported just over half the studies included in this review (n = 10), that had examined associations between ATS and body weight. One explanation for this difference lies in the recent increase in ATS-related publication since their review's search date (06/2008). Their review not only focused on the relationship between ATS and overall physical activity, but also adopted inclusion criteria for studies for the assessment of physical activity using an objective measure (pedometer or accelerometer). Despite these differences, the findings of both reviews were consistent, concluding that there is little evidence to suggest an association between body weight and ATS.

Previous reviews that examined the relationship between ATS and cardiorespiratory fitness were unable to draw conclusions due to the small number of studies. However, evidence from a number of recent well designed studies suggests a positive association between ATS and cardiorespiratory fitness. One partial explanation for these positive associations may be the use of cycle ergometers to measure fitness in three of the five studies, which may have favored those children who cycled to school [15, 32, 33]. However, the two other studies that used running-based fitness assessments also observed differences in fitness between ATS and non ATS groups, with one significant [29] and one borderline nonsignificant [25] result. Furthermore, a recent study, not included in the current review, found that CRF levels were higher among young Swedish and Estonians adolescents who cycled to school compared to those who walked or used passive transportation methods [44]. There appears, therefore, to be some evidence supporting the observation that fitness is associated with ATS, but future studies must account for loss to follow-up and validated assessments of ATS to improve confidence in these observations.

Similar limitations to the overall quality of studies were observed, particularly in the definition and measurement of ATS. There were a range of definitions for classifying a participant as an active traveler. These definitions used different categories for frequency, duration and type of activity that counted as ATS. For example, studies often used questions about "usual mode of travel to and from school", and only 11 of the 27 studies reported an acceptable reliability of their methodology for ATS assessment in the sample population. Thus, standardized definition and measurement should be addressed in future studies, as active travel is a collection of behaviors that vary by purpose and duration [45], e.g. journey to and from school. The potential for misclassification of ATS was also reflected in the actual measurement of ATS mode, with no study reporting a validated measure of ATS by travel mode. With the advent of both portable global positioning system devices and digital image capture systems, the opportunity to improve this potential source of bias must be addressed. These new systems, which allow the confirmation of journey mode, duration and distance, are currently being evaluated in a number of international studies [46, 47]. Such systems would also address the issue of possible confounding or mediation of ATS behavior and health outcome by the built environment characteristics of study areas [48].

The majority of studies did not examine associations by population sub- groups, e.g. gender or age. Indeed, study populations were treated as homogeneous, being usually dichotomized into ATS v non-ATS groups. Such an approach is limited as it ignores the broader social, environmental and personal level correlates of behavior. Indeed, Panter and colleagues [48] highlighted the importance of facilities to assist active travel and urban design in the neighborhood, as well as shorter distances and road safety in-route in relation to active travel. A recent review of qualitative studies of children's experiences of active travel not only supported these observations, but also highlighted the importance of the constraining influence of parents' restrictions on independent movement, as well as children's own fears of traffic [49]. All these factors clearly have a major impact on children's ability to undertake ATS and may provide an explanation why some studies fail to report consistent associations [9, 10]. Only two studies reported power calculations in their methods sections. While the majority of the studies included large samples and were likely to have adequate power to detect hypothesized relationships, the reporting of power calculations should be considered by others in the future.

Body composition/weight status was measured in a variety of ways, with most studies employing BMI z-scores or percentiles to classify youth as healthy, overweight or obese. Two studies [21, 37] used parental proxy reports of their children's height and weight and both studies found no association between BMI and method of transportation to school. Alternatively, two studies [19, 30] used students' self-reported BMI and found significant associations between ATS and weight status. More consistency in using a more objective assessment of this exposure variable in future studies would help to establish the association in relation to ATS.

This is the first systematic review to report the relationship between ATS and HRF in youth. In addition, the assessment of study eligibility and quality was conducted independently by two authors and consensus was reached for every decision. However, the limitations of this review should be noted. First, the measurement and classification of ATS were different across studies, which may explain some of the inconsistencies found. Second, the quality assessment was not extensive and additional criteria may provide additional insights. Finally, the search strategy was limited to published studies identified through the selected search engines. As more studies continue to be published, it will be important to reconsider and refine these findings. Despite these shortcomings, it seems that the majority of higher quality studies reported consistent associations between ATS and the important health indexes of weight status/body composition and cardiorespiratory fitness. With such health benefits for children, the promotion of ATS provides a real opportunity for public health interventions to tackle the double challenge of obesity and poor fitness.

Conclusions

There is some evidence to suggest that active travel to school is associated with a healthier body composition and level of cardiorespiratory fitness among youth. However, as this review included only two longitudinal studies, more interventions and longitudinal studies exploring the relationship between changes in health-related fitness and active travel are needed to investigate the causal nature of such relationships. While it is important that researchers continue to monitor children and adolescents school transportation behaviours, there is a need to broaden the research on young people's active transport to include journeys to destinations other than school, e.g. walking/cycling trips from home to a variety of neighbourhood destinations. Distance to school has been identified as a barrier to active transportation [50] and public health strategies promoting active transportation to more proximal destinations are clearly warranted.