In this population-based longitudinal study of 9–10 year old children, boys and girls who changed to an active mode of travel increased their total daily MVPA by an average of 9 min and 6 min respectively. This represents 12% of boys’ and 13% of girls’ total daily time spent in MVPA at follow-up. The total daily MVPA of children who did not change their travel mode, or who changed to a passive mode, decreased over the same period. It should be noted however that boys and girls who changed to an active mode spent less time in MVPA (6 min/day and 22 min/day, respectively) at baseline than children in the other exposure categories and may therefore have had greater potential to increase their physical activity levels. However, models were adjusted for students’ baseline time spent in MVPA, to control for potential regression to the mean. The results presented here, although using a more proximal outcome, support previous work by Cooper et al. who found that boys and girls who took up cycling to school during a 6-year period had better cardiorespiratory fitness than those who did not cycle to school at either baseline or follow-up
For boys who changed to an active mode of travel, no significant association was observed with time spent in MVPA on weekdays. However, a significant association was observed with daily time spent in MVPA over the entire week. These findings suggest that the observed association between active travel and overall physical activity may reflect the combined effect of the journeys themselves, spontaneous play during the journeys and an encouragement of active behaviour in other areas of children’s lives
. It is also possible that increases in active travel may reflect a change in travel mode preference resulting from a more general increase in physical activity; in other words, children who become more active in general may then choose to travel to school by active means. Further analyses using multiple time points are needed to investigate these potential causal pathways.
Children who consistently used either an active or a passive mode of travel recorded decreases in MVPA over the one-year study period, as did children who changed from an active to a passive mode. This suggests that while taking up an active mode of travel may negate the downward trend in MVPA that is typically seen in this age group
[14, 17], continuing active travel is not sufficient to do so. The decline in MVPA seen among children who remained active travellers highlights a need for interventions that prevent a decline in physical activity in other areas of children’s lives, for example during and after school, as well as interventions to promote active travel.
In this study only 9.2% of pupils changed to an active mode of travel, while 40.6% consistently used passive modes. In 2005, 43% of UK children were driven to school
. Interventions that promote active modes of travel to school therefore have the potential to increase the activity levels of large numbers of pupils, which may have important public health implications
[1, 18]. However, a recent systematic review found that to date, interventions promoting active travel to school have produced only small effects, suggesting a need for further research as to how best to achieve such behaviour change in practice
Two-level regression models showed that parental education — an indicator of socioeconomic status (SES) — predicted changes in girls’ time spent in MVPA but not boys’. Since boys participate in more active play in their free time than girls
, it is possible that parents encourage their daughters, or girls choose, to accumulate MVPA in other ways such as organised sports. The costs associated with participating in organised sports (such as those of equipment and membership fees) might contribute to explaining why parental SES appears to be an important influence on girls’ time spent in MVPA and not associated with boys’.
Strengths and limitations
Strengths of this study include its prospective design, the use of objective measures of physical activity and that it was conducted in a large population-based sample of UK primary school children. The one year interval separating baseline and follow-up data collection ensured seasonally matched data, thereby reducing the risk of artefactual changes in either usual travel behaviour or measured physical activity attributable to different weather conditions at the two time points. It also provided sufficient time to result in changes to children’s usual travel behaviour. However, despite the prospective design, it remains unclear whether children become more physically active and then changed to an active mode of travel, or vice versa. A longitudinal dataset including measurement at three or more time points or a controlled intervention study would be needed to answer this question with confidence.
Although the analysis sample comprised approximately one-third of the pupils recruited at baseline, we found no significant differences between the characteristics (sex, BMI, urban rural/status, parental education or baseline MVPA) of pupils included in and excluded from the analysis. However, we have previously shown that in comparison to the Norfolk population a smaller proportion of obese children took part in the SPEEDY study, and that the demographic profile of the county of Norfolk (with only 3.8% non-white children) is not representative of the whole of the UK
The measure of travel mode was relatively crude and did not allow multi-modal trips or the daily breakdown of travel behaviour to be ascertained. Moreover, to our knowledge there is no published validation of questions to ascertain ‘usual travel mode’ in children. However, similar measures have been used in previous studies examining the association between travel mode and physical activity in children
[8, 21]. Furthermore, an insufficient proportion (6%) of pupils used public transport at baseline to allow public transport to be treated as a separate category. Consequently, if we had disaggregated travel mode further into car, public transport, walking and cycling our statistical models would had been underpowered to detect associations. As in previous studies, public transport was therefore combined with passive travel
. Although travelling by public transport usually includes some walking, school bus services in the UK often pick up pupils close to their homes and it is therefore reasonable to assume that these journeys might have included relatively little active travel. In support of this assumption, van Sluijs et al. found that UK children who walked to school had significantly higher average weekly accelerometer counts per minute and total minutes of MVPA than those classified as using the car or public transport to travel to school
. Moreover, removing pupils who used public transport from the current analyses made no substantive differences to the findings (results not shown). Distance from home to school has consistently been found to be associated with mode of travel to and from school
[24, 25]. Although network distance was included as a covariate in analysis, it was not possible to stratify the analysis by this measure owing to small cell sizes.
Accelerometers as used in this study are calibrated to record ambulatory activity (hip movement) and therefore underestimate the intensity of physical activity undertaken whilst cycling. In this study only 9% and 7% of pupils reported cycling to school at baseline and follow-up respectively, so the influence of this limitation of measurement is likely to be small. Criteria for classifying accelerometer non-wear time vary within the literature
[26, 27]. For the present analysis 10 minutes of continuous zero counts were used to define non-wear. It appears reasonable to assume that 9–10 year old children are unlikely to remain sufficiently still to register no movement on an accelerometer for more than 10 minutes when awake, and this criterion has been applied in previous studies