Physical activity in young people is associated with multiple mental, social, and physical health benefits, both in the short and long term,1 but many young people worldwide do not meet the World Health Organization guidelines for moderate to vigorous physical activity (MVPA).2,3 Moreover, they spend large amounts of time being sedentary (SED), posing further health risks.4,5 Ecological models suggest that physical activity is influenced by factors at the individual, interpersonal, environmental, regional/national policy, and global levels.6 In recognition of the need to promote physical activity, physical activity programs have been evaluated that target factors at the individual, interpersonal, and organizational levels, generally with limited effects.7–9 This has led to calls for higher level action, such as regional, state, or national policies affecting young people’s activity behavior.10,11 These affect a greater number of individuals, potentially enabling a shift in the population-level distribution of physical activity and its associated health outcomes.
Global trends in city planning and urban design encourage car use by separating people’s residences from the destinations they need to access regularly, such as shops and schools. Although the environmental and health impacts of environments conducive to car use are well documented,12 in the past 25 years, United Kingdom (UK) car ownership increased by 40% to 32.9 million in 2023.13 A similar increase has been observed in Australia.14 UK and United States (US) households now have access to an average of 1.2 and 2.0 cars, respectively.15,16 From a political perspective, it may not be desirable or feasible to target reductions in car ownership per se. Aiming to make multiple car ownership less attractive may be a more feasible policy option. Indeed, most change in car ownership occurs between single and 2-car ownership.17 Multiple car ownership is particularly common in families: In the UK and Australia, households with 5- to 11-year-olds are most likely to own ≥ 2 cars,14,18 and becoming a parent is associated with car acquisition.18
Household car ownership may impact children’s activity levels. Panter et al’s19 framework for environmental determinants of active travel in youth proposes household car access as a determinant of travel mode choice and physical activity and that associations may be moderated by age, sex, and socioeconomic position (SEP). Evidence shows that children living in households with access to a car are less likely to walk or cycle to school,20 and associations between household car ownership and accelerometer-measured time spent SED and in MVPA have been shown in adults21 and UK22–24 and Ugandan25 primary school children but not in South African primary school children.26 Previous research exploring associations with children’s activity has mostly been cross-sectional, relied on self-reported activity,27,28 and focused on car ownership per se (ie, no vs any car),20,22,26 lacking the heterogeneity and sample size to assess differences between single and multiple car ownership and the longitudinal data to assess associations with changes in exposure.
The primary research questions addressed here, therefore, are: (1) Is number of cars in the household (none, 1, or ≥ 2 cars) associated with children’s accelerometer-assessed time spent in SED and MVPA, and (2) Is change in number of cars in the household associated with change in children’s time spent in SED and MVPA? Secondary research questions address whether the association differs by demographic characteristics and day of week, to what extent travel mode to school explains this association, and associations with secondary outcomes (intensity categories and adiposity, as a closely related health outcome of physical activity behavior1).
Methods
The International Children’s Accelerometry Database (ICAD) is a large, multicountry data pooling project. ICAD pooled data from 23 pediatric studies from 10 countries across Europe, North and South America, and Australia, all of which used the waist-worn ActiGraph accelerometer (Pensacola, FL). The original set-up of ICAD and its update have been described in detail elsewhere.29,30 This analysis used data from the second release of the database (ICAD2).
Research Ethics Approval
All studies contributing data to ICAD received ethical approval for their data collection and sharing of data and obtained written parental consent and child assent prior to data collection (further details available in study-specific publications).
Sample
Studies were included if they provided data on household car ownership on at least one measurement occasion, enabling inclusion of 5 studies from 3 countries: Australia (CLAN, HEAPS), the UK (PEACH, SPEEDY), and Brazil (Pelotas; see Supplementary Table S1 [available online] for study overview). All studies provided data for cross-sectional analyses, with only Pelotas not contributing to the longitudinal analyses. All data were collected between 2001 and 2013 and pooled in 2016–17. Participants were eligible for inclusion in the analyses described here if they provided at least one wave of valid data for household car ownership and accelerometry.
Main Exposure: Household Car Ownership
Parent-reported household car ownership was available in all studies; additional child-reported data were used in the SPEEDY study. Household car ownership was categorized into no, 1, or ≥2 cars (see Supplementary Table S2 [available online] for data harmonization procedures) based on data availability and distribution. We defined change in car ownership as “no change,” “increase,” and “decrease.”
Outcome Assessment
Raw accelerometer files were reprocessed using specifically developed and commercially available software (KineSoft version 3.30) to create standardized outcomes. The processing protocol is freely available online (http://www.mrc-epid.cam.ac.uk/research/studies/icad/). Raw Actigraph data files were reintegrated to files with 60-second epochs to facilitate comparison across studies, and periods of at least 60 minutes of consecutive zeros (with a tolerance of 2 min of nonzero interruptions) were identified and excluded,31 as was any time between 11 PM and 6 AM. Accelerometer-registered wear time was defined as average daily valid wear time across all valid days. Intensity thresholds corresponding to SED and MVPA were derived using previously applied cut points (≤100 counts per minute [cpm] and ≥ 2000 cpm, respectively),32 with the MVPA threshold corresponding to a walking speed of ∼3 to 4 km/h in children.33 Average minutes per day spent SED and in MVPA were calculated by summing all minutes wherein the activity count met the criterion for that intensity, divided by the number of valid days. For secondary analyses, intensity-specific cut points for moderate physical activity (MPA, 2000–3999 cpm) and vigorous physical activity (VPA, ≥4000 cpm) were applied.32 Participants with ≥8 hours wear time on ≥4 days (including 1 weekend day) were included in the analyses. Analyses were stratified for weekdays (including ≥3 weekdays) and weekends (including ≥1 weekend day).
Covariates
Participant sex was self- or parent reported, and age in years was preferentially calculated from date of birth and first date of accelerometer measurement. Where this was unavailable, we used age at assessment as provided by the original investigators. Measured height (in meters) and weight (in kilograms) were used to ascertain age- and sex-specific BMI z-score (2006 World Health Organization child growth standards, implemented as “zscore06” in Stata). Parent-reported maternal education was used as a proxy for family SEP (up to and including completion of compulsory education [low/medium SEP] vs any postcompulsory education or vocational training [high SEP]). Travel mode to school was predominantly child reported, except for the younger cohorts of CLAN and HEAPS in which it was parent reported. A dichotomous variable indicated use of an active (walk, cycle, scooter) or inactive (car, bus, etc) travel mode to school. Full harmonization details for all covariates are available on the ICAD website: (http://www.mrc-epid.cam.ac.uk/research/studies/icad/data-harmonisation/).
Statistical Analyses
Analyses followed a prespecified protocol and were performed using Stata (version 14.2). Descriptive statistics are presented as means (SD) or proportions for continuous and categorical/dichotomous variables, respectively. Drop-out analyses compared participants (1) included in the cross-sectional analyses with those excluded and (2) included in the longitudinal analyses versus those only included in the cross-sectional analyses. For the cross-sectional analyses, all waves of data were included, and a 3-level structure was applied (wave, child, study). Clustering at school level was not included due to the lack of information available for one study (Pelotas) and for consistency with other ICAD analyses. Complete-case mixed-effects linear regression models with random intercepts at child and study level were used to estimate the association between car ownership (categorical, using “no car” as reference) and SED/MVPA, adjusting for accelerometer wear time (model 1), sex, SEP, age, and BMI z-score (model 2).
The longitudinal association between change in car ownership and change in each outcome was assessed using complete-case mixed-effects linear regression models with 2 levels (child, study) and random intercepts at the study level. For each participant, data from the longest available follow-up were used. Models of accelerometer-based outcomes were adjusted for sex, SEP, baseline age, BMI z-score, and car ownership, and change in age and accelerometer wear time.
Secondary analyses using the same modeling approach estimated the association for SED and MVPA separately for week and weekend days and for MPA, VPA, and BMI z-score. For the latter, model 1 assessed unadjusted associations, whereas model 2 adjusted for sex, SEP, and age. The model with change in BMI z-score as the outcome was adjusted for sex, SEP, baseline age and car ownership, and change in age. A sensitivity analysis to investigate the impact of choice of accelerometer cut point was conducted for the main MVPA analyses (using ≥ 3000 cpm).
In the main cross-sectional models, effect modification by country, sex, SEP, and age (categorized as ≤ 7, 8–11, and ≥ 12 y) was investigated by including an interaction between the modifier of interest and the main exposure variable and performing a Wald test. Longitudinally, only effect modification by country was investigated. A post hoc analysis explored whether the association differed by study by testing the interaction between study and car ownership. In cross-sectional models, the extent to which any association was explained by travel mode to school (active vs inactive) was assessed by calculating the percentage change in the beta-coefficient comparing model 2 with and without travel model included as a covariate (both models fit to the same sample of individuals). This was not explored longitudinally due to limited longitudinal data on travel mode.
Results
Of 6396 participants across the 5 studies, 4193 provided sufficient data for inclusion in the cross-sectional analyses (contributing 5971 observations across the waves) and 1333 for longitudinal analyses (from 4 studies). Figure 1 presents a flowchart of studies and participants; Table 1 provides a summary of the baseline characteristics for the cross-sectional sample. Those included in the cross-sectional analyses were slightly younger and engaged in more MVPA than those excluded (Supplementary Table S3 [available online]). Those included in the longitudinal analyses were younger, had lower BMI z-scores, and were more likely to be from higher SEP households than those excluded.
—Flowchart of participants. ICAD = International Children’s Accelerometry Database.
Citation: Journal of Physical Activity and Health 21, 12; 10.1123/jpah.2024-0044
Descriptive Characteristics of Participants in Cross-Sectional Analyses (N = 4193)
| Variable | Mean (SD) or N (%) |
|---|---|
| Age in years | 10.3 (2.0) |
| BMI z score | 0.5 (1.1) |
| Sex (% female) | 2239 (53.4%) |
| Maternal education (low/medium SEP) | 2084 (49.7%) |
| Car ownership | |
| No car | 377 (9.0%) |
| 1 car | 1246 (29.7%) |
| 2 or more cars | 2570 (61.3%) |
| Total daily average, min: | |
| SED | 349.1 (104.2) |
| MVPA | 67.3 (28.3) |
| Monitor wear time | 766.5 (94.6) |
Abbreviations: BMI z score, standardized body mass index; MVPA, moderate to vigorous physical activity; SED, sedentary; SEP, socioeconomic position.
Cross-Sectional Associations
Table 2 shows that in adjusted cross-sectional models, compared with no household car access, children in households with access to one or more cars spent more time in SED and less time in MVPA, with no dose–response relationship. These findings were similar across weekdays and weekends but more pronounced for MPA than VPA, with the latter not statistically significant. There was no evidence that car ownership was associated with BMI z-score. Sensitivity analysis using a higher cut point for MVPA attenuated the effect estimate but did not alter the conclusions (data not shown).
Cross-Sectional Association Between Household Car Ownership and Children’s Accelerometer-Measured Activity Intensity (in Minutes Per Day) and BMI z Score (SD Units)
| Model 1a | Model 2b | |||
|---|---|---|---|---|
| 1 car | ≥2 cars | 1 car | ≥2 cars | |
| Total daily average | ||||
| SED | 17.0 (8.2 to 25.7) | 18.0 (9.2 to 26.8) | 14.1 (6.7 to 21.5) | 12.8 (5.3 to 20.4) |
| MVPA | −8.6 (−12.1 to −5.2) | −8.8 (−12.3 to −5.2) | −8.8 (−11.9 to −5.7) | −8.8 (−12.0 to −5.7) |
| Weekday average | ||||
| SED | 15.3 (6.4 to 24.2) | 17.2 (8.1 to 26.2) | 13.5 (5.9 to 21.2) | 13.0 (5.2 to 20.8) |
| MVPA | −8.2 (−11.8 to −4.6) | −9.5 (−13.1 to −5.9) | −8.5 (−11.8 to −5.3) | −9.8 (−13.1 to −6.5) |
| Weekend day average | ||||
| SED | 17.7 (6.6 to 28.7) | 17.1 (5.9 to 28.3) | 13.5 (3.6 to 23.4) | 10.5 (0.5 to 20.5) |
| MVPA | −9.0 (−13.9 to −4.1) | −6.6 (−11.5 to −1.6) | −8.6 (−13.3 to −4.0) | −6.0 (−10.6 to −1.3) |
| Total daily average | ||||
| MPA | −7.7 (−10.2 to −5.2) | −8.4 (−11.0 to −5.9) | −7.7 (−9.9 to −5.5) | −8.3 (−10.6 to −6.1) |
| VPA | −0.9 (−2.3 to 0.6) | −0.3 (−1.8 to 1.2) | −1.0 (−2.4 to 0.4) | −0.5 (−1.9 to 0.9) |
| BMI z-score | 0.01 (−0.1 to 0.1)c | −0.04 (−0.2 to 0.1)c | 0.03 (−0.1 to 0.1)d | −0.01 (−0.1 to 0.1)d |
Abbreviations: BMI z-score, standardized body mass index; SED, sedentary; MVPA, moderate to vigorous physical activity; MPA, moderate physical activity; SEP, socioeconomic position; VPA, vigorous physical activity.
N observations = 5971. Numbers in Cells Represent Beta Coefficient and 95% Confidence Interval, Using “No Car” as Reference Category
aMultilevel linear regression model with random intercept adjusted for accelerometer wear time. bAdditionally adjusted for age, sex, BMI z-score (except where BMI z-score is outcome), and maternal education (SEP). cFor BMI z-score, model 1 is an unadjusted multilevel linear regression model with random intercept. dFor BMI z-score, model 2 is a multilevel linear regression model with random intercept adjusted for sex, SEP, and age.
Effect modification analyses in the adjusted models (Figure 2) demonstrated significant interactions with sex for both SED and MVPA (see Supplementary Table S4 [available online] for results of all interactions tests). For both outcomes, the association was stronger in boys than girls; the association with SED in girls did not reach statistical significance. In addition, there was a suggestion of a country by SED interaction, with stronger associations and a clear dose–response relationship in Brazilian children but no association in UK/Australian children. No differences were observed between the UK and Australia. A significant interaction with SEP indicated a positive association with SED in children from low/medium SEP households but no association in those from higher SEP households.
—Results of interaction analyses subgroup estimates for average daily MVPA (a) and SED (b). MVPA indicates moderate to vigorous physical activity; SED, sedentary.
Citation: Journal of Physical Activity and Health 21, 12; 10.1123/jpah.2024-0044
Inclusion of travel mode to school led to an attenuation of the associations with daily SED and MVPA (Supplementary Table S5 [available online]). Specifically, in 1-car households, travel mode to school attenuated the magnitude of the associations with both SED and MVPA by 14.4%. Attenuation of the association was larger for greater car ownership: 24.0% (SED) and 21.1% (MVPA). In all analyses, the main associations between car ownership and activity intensity remained statistically significant after inclusion of travel mode to school.
Longitudinal Associations
Participants included in longitudinal analyses (N = 1333) had a baseline age of 9.8 (2.0) years, 54.5% were girls, and 58.2% came from a high SES household (Supplementary Table S3 [available online]). Mean follow-up was 3.3 (1.4) years, during which time spent in SED increased by 94.3 (83.8) minutes per day and MVPA decreased by 11.5 (27.9) minutes per day. The majority of households (73.1%) had access to ≥ 2 cars at baseline; 8.1% of households increased the numbers of cars (89.8% 1-car households at baseline), 7.7% decreased (93.1% ≥2-car households at baseline), and the remaining 84.2% reported no change. Supplementary Table S6 (available online) presents the descriptive data on change in car ownership and change in SED and MVPA by change in car ownership.
There was no evidence of associations between increased car ownership and change in SED, MVPA, or BMI z-score (Table 3). In contrast, a decrease in car ownership was associated with a greater decrease in MVPA but not with change in SED or BMI z-score. The association with change in MVPA was most pronounced at weekends and of similar magnitude for MPA and VPA. There was no difference in association by country (Wald test: SED: P = .957; MVPA: P = .752) or study (SED: P = .593; MVPA: P = .791).
Longitudinal Association Between Change in Household Car Ownership (Reference Category: No Change) and Children’s Accelerometer-Measured Activity Intensity (in Minutes Per Day) and BMI z-Score (SD Units). Numbers in Cells Represent Beta Coefficient and 95% Confidence Interval
| More cars at follow-up | Fewer cars at follow-up | |
|---|---|---|
| Total daily average | ||
| SED | 2.5 (−10.6 to 15.7) | 0.4 (−11.5 to 12.3) |
| MVPA | −3.9 (−10.0 to 2.1) | −8.4 (−13.9 to −3.0) |
| Weekday average | ||
| SED | 0.2 (−13.8 to 14.3) | 1.9 (−10.8 to 14.6) |
| MVPA | −3.4 (−9.8 to 2.9) | −6.8 (−12.5 to −1.0) |
| Weekend day average | ||
| SED | 2.8 (−18.4 to 24.0) | −0.9 (−20.1 to 18.3) |
| MVPA | −3.3 (−13.9 to 7.3) | −12.3 (−21.9 to −2.8) |
| Total daily average | ||
| MPA | −3.9 (−8.2 to 0.3) | −4.0 (−7.8 to −0.1) |
| VPA | −0.1 (−3.2 to 3.0) | −4.4 (−7.2 to −1.6) |
| BMI z-score | 0.05 (−0.1 to 0.2) | −0.03 (−0.1 to 0.1) |
Abbreviations: BMI z-score, standardized body mass index; MPA, moderate physical activity; MVPA, moderate to vigorous physical activity; SED, sedentary; SEP, socioeconomic position; VPA, vigorous physical activity.
aMultilevel linear regression model with random intercept adjusted for baseline data for age, sex, BMI z-score (except when BMI z-score is outcome), maternal education (SEP), and car ownership, and change in age and accelerometer wear time (except when BMI z-score is outcome; N observations = 1333).
Sensitivity analysis using a higher cut point for MVPA attenuated the size of the estimate but did not alter the conclusion (data not shown). Post hoc analyses stratified by baseline car ownership showed similar lack of associations for gaining a second car and similar, but stronger, associations for those losing access to a second car (data not shown). Too few participants lost or gained access to a car to run stratified models (N = 7 and N = 11, respectively).
Discussion
Summary of Main Findings
In this large, heterogeneous international sample of school-aged children, we confirmed that household car ownership was cross-sectionally associated with children spending more time in SED and less time in MVPA, but there was no evidence of a dose-dependent association. These associations were similar across weekdays and weekends but more pronounced for MPA than VPA and could be explained, in part, by children’s travel mode to school. Associations with SED were stronger for boys, in those from lower SEP households, and in Brazilian children relative to their respective reference groups. There was no association between car ownership and BMI z-score. Longitudinally, against a backdrop of decreasing MVPA and increasing SED, those children living in households losing access to a second car recorded larger decreases in MVPA compared with children living in households where car access remained the same. The magnitude of this association was larger at weekends. No associations were observed between reduced car ownership and change in SED or BMI z-score or for increased car ownership with any outcome.
Comparison With Previous Literature
Previous systematic reviews also identified that, cross-sectionally, car access was negatively associated with physical activity in children and adolescents,20,27,28 although this evidence predominantly relied on associations with reported activity and active travel. More recent publications have supported these findings using accelerometer-measured activity intensity22,23 but, in contrast to our findings, also reported that MVPA decreased incrementally with increasing car access.23 This study adds to this limited evidence by using a large heterogeneous, international sample with accelerometer-measured activity, allowing for exploration of difference in effect for subgroups and investigating associations with a range of outcomes. The observation that the positive association with time spent SED is strongest for boys and those from lower SEP backgrounds is novel and speaks to the different ways in which subgroups accumulate their physical activity, such as using active travel when access to a car is limited. This issue warrants consideration in future interventions and policy action.
Our results also corroborate the findings from a recent multicountry cross-sectional study in adults, which reported a negative association between car ownership and device-measured MVPA, although they found that the strength of the association was context dependent with larger differences in the US compared with other countries.21 In our analysis, few country-specific associations were observed. Cross-sectionally, we observed a stronger, dose–response relationship among Brazilian participants for SED. It is important to note that the distribution of car ownership differed by study, with Brazilian participants making up 57.7% of the “no car” sample. Unfortunately, we were unable to include the Brazilian study in the longitudinal analyses due to lack of follow-up data. Car ownership is known to be context dependent and influenced by economic (eg, income), environmental (eg, home location), and social (eg, family composition) factors.34 In the context of a high-density, lower resource setting, greater car ownership, therefore, appears to be associated with more time spent in SED independent of family-level SEP. The lack of a similar association for MVPA suggests that greater car ownership may be associated with a less incidental activity of lower intensity (eg, light physical activity) without impacting higher intensity activity (eg, MVPA). Previous research indicated that adults who reported higher levels of recreational walking and overall physical activity were more likely to travel further by car,35 suggesting that car access can facilitate higher intensity leisure activity.
Prospectively, we found that losing access to a second car was associated with greater decreases in MVPA, particularly at the weekend, whereas acquiring a second car was not associated with changes in activity. Change in family car ownership may be due to a variety of factors, including changes in family composition, income, or home/work location, all of which may also influence children’s access to places to be active. Unmeasured confounding may exist; and therefore, it is not possible to know whether the observed changes in MVPA are directly attributable to reduced car access. The findings do, however, highlight that some climate change mitigation strategies, such as discouraging multiple car ownership, may inadvertently be negatively associated with children’s physical activity, particularly during leisure time. Complementary actions that facilitate access to leisure activities via public transport or active travel may be needed. This work also highlights the complex interplay between social, economic, and environmental factors impacting car ownership, car use, and children’s activity levels. Further quantitative analyses into changes in ownership and use of car in the context of families as well as qualitative investigations of the (implicit) value of family car ownership and use may help shed light on this.
Emerging literature highlights not only the potential of increases in physical activity to contribute positively to climate change36 but also how the changing climate may impact participation in physical activity37 and how physical activity participation and promotion may increase carbon emissions (eg, through motorized transport to activity venues, heating or cooling of sports venues) and have other detrimental effects on planetary health.37–39 The findings presented here suggest that children may be particularly dependent on motorized transport to access places offering opportunities for higher intensity physical activity during leisure time, which has also been observed in South African children.26 This may be because of decentralized locations of youth sport opportunities, observed globally,40 which leads to greater dependency on (safe) motorized travel for participation. For example, Bunds et al40 showed that US families with a child in a regionally competitive youth swim team would drive, on average, 2222 miles (3500 km) annually to attend training and competitions. In addition, a study in Norwegian adolescent soccer players highlighted that despite most players living within cycling distance (<3 km), the majority reported being driven to practice.41 This phenomenon is referred to as the “insularization of children’s individual life spaces”: the ferrying of children from home to school to leisure activities.42 This has consequences not only for planetary health but also for inequalities in physical activity participation as low-income families may not have the resources required to facilitate access. Future research should further develop theories to guide research into the complex relationship between household car access and physical activity, including consideration of car use patterns, access to alternative modes of transport, and urban design features. Research should also consider the role of organized sports for both young people’s physical activity and planetary health and identifying ways to improve all children’s access to leisure-time activities.
Strengths and Limitations of this Study
This study is the largest analysis to date of the association between car ownership and children’s device-measured activity intensity and anthropometry outcomes as well as the first to assess this association prospectively. The pooling and harmonization of data in ICAD ensure comparability of data across different studies and contexts. We observed substantial heterogeneity in car ownership, and change in car ownership, which often is lacking in single-study analyses. We adjusted the models for relevant confounders and explored interactions by key characteristics. As processing decisions may affect accelerometer-based variables, we used preestablished and frequently applied processing methods and a frequently used accelerometer cut point to define MVPA. Sensitivity analyses using a different cut point demonstrated limited impact on the main conclusions drawn.
We were unable to account for car use and car dependency or (changes in) family or economic circumstances that may impact car ownership and use. In addition, data were collected between 2001 and 2013. Although this may impact on the prevalence of exposures and outcomes, it is unclear whether this would impact the associations between them. We acknowledge that travel and activity habits were significantly affected by the policy response to the COVID-19 pandemic and that contemporary research is required to confirm the generalizability of these findings to the postpandemic context. The samples included are not nationally representative, although they are largely representative of the population from which they were drawn. Although there was sufficient heterogeneity in exposures and outcomes to run the longitudinal analyses, the sample size was substantially reduced compared with the cross-sectional sample due to nonresponse and missing data. Some differences were observed between those included and excluded from the longitudinal analyses, affecting the generalizability of these findings. Future research should study these associations in other large heterogeneous samples.
Implications
Reducing car access and use are important policy targets to achieve improvements in air pollution and congestion and related population health indicators. Targeting multiple car ownership may be politically more attractive than targeting car ownership per se, but our findings suggest that such policies may have mixed impacts on children’s activity. The cross-sectional results from these analyses suggest that children living in households with access to either one or multiple cars have similar levels of activity and adiposity and that potentially making multiple car ownership less attractive would have limited impact. However, the longitudinal results suggest that losing access to multiple cars may have negative implications for children’s activity levels, particularly at weekends, which implicates car ownership as an enabler of activity outside of school for some children. It is unclear how these changes are related to concurrent changes in economic or family circumstances or reduced access to activity opportunities. It does, however, suggest that the effect of interventions on out-of-school activity9 might depend on car access and that facilitating activity as part of daily life is critical to countering this impact. Our finding that travel mode to school partly explained the associations observed, combined with previous evidence suggesting that switching to more active modes does not necessarily lead to compensatory decreases in activity,43 indicates that active travel could serve this purpose. Our results suggest that the development of urban and transport planning strategies that promote public transport and active travel and reduce car dependency, such as carpooling and future-proofing new developments, should be prioritized and include consideration of locations of young people’s activity. Future policy decisions on reductions in car ownership and use should consider ways of negating the potential negative impact on children’s activity opportunities.
Acknowledgments
We would like to thank all participants and funders of the original studies that contributed data to ICAD. We gratefully acknowledge the past contributions of Prof Chris Riddoch, Prof Ken Judge, Prof Ashley Cooper, and Dr Pippa Griew to the development of ICAD. The ICAD collaborators include: Dr DW Esliger, School of Sports, Exercise and Health Sciences, Loughborough University, the United Kingdom; Prof A Page, Centre for Exercise, Nutrition and Health Sciences, University of Bristol, the United Kingdom (Personal and Environmental Associations with Children’s Health [PEACH]); and Prof A Timperio, Institute for Physical Activity and Nutrition (IPAN), School of Exercise and Nutrition Sciences, Deakin University, Geelong, Australia (Healthy Eating and Play Study [HEAPS]). The ICAD was made possible thanks to the sharing of data from the following contributors (study name): Prof LB Andersen, Faculty of Teacher Education and Sport, Western Norway University of Applied Sciences, Sogndal, Norway (Copenhagen School Child Intervention Study [CoSCIS]); Prof S Anderssen, Norwegian School for Sport Science, Oslo, Norway (European Youth Heart Study [EYHS], Norway); Prof G Cardon, Department of Movement and Sports Sciences, Ghent University, Belgium (Belgium Preschool Study); Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS), Hyattsville, MD, the United States (National Health and Nutrition Examination Survey [NHANES]); Dr R Davey, Centre for Research and Action in Public Health, University of Canberra, Australia (Children’s Health and Activity Monitoring for Schools [CHAMPS]); Dr P Hallal, Postgraduate Program in Epidemiology, Federal University of Pelotas, Brazil (1993 Pelotas Birth Cohort); Prof KF Janz, Department of Health and Human Physiology, Department of Epidemiology, University of Iowa, Iowa City, the United States (Iowa Bone Development Study); Prof S Kriemler, Epidemiology, Biostatistics and Prevention Institute, University of Zürich, Switzerland (Kinder-Sportstudie [KISS]); Dr N Møller, University of Southern Denmark, Odense, Denmark (European Youth Heart Study [EYHS], Denmark); Dr K Northstone, School of Social and Community Medicine, University of Bristol, the United Kingdom (Avon Longitudinal Study of Parents and Children [ALSPAC]); Dr A Page, Centre for Exercise, Nutrition and Health Sciences, University of Bristol, the United Kingdom (Personal and Environmental Associations with Children’s Health [PEACH]); Prof R Pate, Department of Exercise Science, University of South Carolina, Columbia, the United States (Physical Activity in Preschool Children [CHAMPS-US] and Project Trial of Activity for Adolescent Girls [Project TAAG]); Dr JJ Puder, Service of Endocrinology, Diabetes and Metabolism, Centre Hospitalier Universitaire Vaudois, University of Lausanne, Switzerland (Ballabeina Study); Prof J Reilly, Physical Activity for Health Group, School of Psychological Sciences and Health, University of Strathclyde, Glasgow, the United Kingdom (Movement and Activity Glasgow Intervention in Children [MAGIC]); Prof J Salmon, Institute for Physical Activity and Nutrition (IPAN), School of Exercise and Nutrition Sciences, Deakin University, Geelong, Australia (Children Living in Active Neigbourhoods [CLAN] and Healthy Eating and Play Study [HEAPS]); Prof LB Sardinha, Exercise and Health Laboratory, Faculty of Human Movement, Universidade de Lisboa, Lisbon, Portugal (European Youth Heart Study [EYHS], Portugal); Dr EMF van Sluijs, MRC Epidemiology Unit, University of Cambridge, the United Kingdom (Sport, Physical Activity and Eating Behavior: Environmental Determinants in Young People [SPEEDY]). This work was funded by UKRI grants [MC_UU_12015/3; MC_UU_12015/6; MC_UU_12015/7; MC_UU_00006/5]. For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) license to any Author Accepted Manuscript version arising. Funding: The ICAD data pooling project and SPEEDY study were funded through grants from the National Prevention Research Initiative (Grant Numbers: G0501294; G0701877; http://www.mrc.ac.uk/research/initiatives/national-prevention-research-initiative-npri/). The funding partners relevant to this award are: British Heart Foundation; Cancer Research UK; Department of Health; Diabetes UK; Economic and Social Research Council; Medical Research Council; Research and Development Office for the Northern Ireland Health and Social Services; Chief Scientist Office; Scottish Executive Health Department; the Stroke Association; Welsh Assembly Government; and World Cancer Research Fund. The CLAN study was supported by grants from the Financial Markets Foundation for Children and the National Health and Medical Research Council (NHMRC, ID: 274309 and 374241). HEAPS was funded by the Victorian Health Promotion Foundation (baseline) and Australian Research Council (follow-ups; DP 0664206). This work was additionally supported by the Medical Research Council [MC_UU_12015/3; MC_UU_12015/6; MC_UU_12015/7; MC_UU_00006/5], the Research Council of Norway (249932/F20), Bristol University, Loughborough University, and Norwegian School of Sport Sciences. Data are available upon reasonable request: The original data request and prespecified analysis plan are available upon reasonable request (please contact lead author). The data cannot be made openly available because of ethical and legal considerations. Nonidentifiable data can be made available to bona fide researchers on submission of a reasonable request, following standard ICAD data request procedures (see for information: http://www.mrc-epid.cam.ac.uk/research/studies/icad/). The principles and processes for accessing and sharing data are outlined in the MRC Epidemiology Unit Data Access and Data Sharing Policy.
References
- 1.↑
Poitras VJ, Gray CE, Borghese MM, et al. Systematic review of the relationships between objectively measured physical activity and health indicators in school-aged children and youth. Appl Physiol Nutr Metab. 2016;41(6 Suppl. 3):S197–S239. PubMed ID: 27306431 doi:
- 2.↑
Hallal PC, Andersen LB, Bull FC, et al. Global physical activity levels: surveillance progress, pitfalls, and prospects. Lancet. 2012;380(9838):247–257. PubMed ID: 22818937 doi:
- 3.↑
Roman-Vinas B, Chaput JP, Katzmarzyk PT, et al. Proportion of children meeting recommendations for 24-hour movement guidelines and associations with adiposity in a 12-country study. Int J Behav Nutr Phys Act. 2016;13(1):123.
- 4.↑
van Ekris E, Altenburg TM, Singh AS, Proper KI, Heymans MW, Chinapaw MJM. An evidence-update on the prospective relationship between childhood sedentary behaviour and biomedical health indicators: a systematic review and meta-analysis. Obes Rev. 2017;18(6):712–714. PubMed ID: 28474511 doi:
- 5.↑
Hoare E, Milton K, Foster C, Allender S. The associations between sedentary behaviour and mental health among adolescents: a systematic review. Int J Behav Nutr Phys Act. 2016;13(1):108. doi:
- 6.↑
Bauman AE, Reis RS, Sallis JF, et al. Correlates of physical activity: why are some people physically active and others not? Lancet. 2012;380(9838):258–271. PubMed ID: 22818938 doi:
- 7.↑
van Sluijs EMF, McMinn AM, Griffin SJ. Effectiveness of interventions to promote physical activity in children and adolescents: systematic review of controlled trials. BMJ. 2007;335(7622):703. doi:
- 8.
Metcalf B, Henley W, Wilkin T. Effectiveness of intervention on physical activity of children: systematic review and meta-analysis of controlled trials with objectively measured outcomes (EarlyBird 54). BMJ. 2012;345:e5888.
- 9.↑
Love R, Adams J, van Sluijs EM. Are effects of school-based physical activity interventions effective and equitable? A meta-analysis of cluster randomized controlled trials with accelerometer-assessed activity. Obes Rev. 2019;20(6):859–870. PubMed ID: 30628172 doi:
- 10.↑
Reis RS, Salvo D, Ogilvie D, et al. Scaling up physical activity interventions worldwide: stepping up to larger and smarter approaches to get people moving. Lancet. 2016;388(10051):1337–1348. PubMed ID: 27475273 doi:
- 11.↑
van Sluijs EMF, Ekelund U, Crochemore-Silva I, et al. Physical activity behaviours in adolescence: current evidence and opportunities for intervention. Lancet. 2021;398(10298):429–442. PubMed ID: 34302767 doi:
- 12.↑
Giles-Corti B, Vernez-Moudon A, Reis R, et al. City planning and population health: a global challenge. Lancet. 2016;388(10062):2912–2924. PubMed ID: 27671668 doi:
- 13.↑
Yurday E. Number of cars in the UK 2024. 2023. Accessed October 31, 2023. https://www.nimblefins.co.uk/cheap-car-insurance/number-cars-great-britain#nogo.
- 14.↑
Loader C. Update on Australian transport trends (December 2019). 2019. Accessed December 8, 2023. https://chartingtransport.com/2019/12/30/update-on-australian-transport-trends-december-2019/
- 15.↑
Office for National Statistics. Census: Key Statistics for England and Wales, March 2011. The defining characteristics of the population: who we are, how we live and what we do. 2012. Accessed June 6, 2018. https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/bulletins/2011censuskeystatisticsforenglandandwales/2012-12-11#car-or-van-availability.
- 16.↑
Sivak M. Has Motorization in the U.S. Peaked? Part 9: Vehicle Ownership and Distance Driven, 1984 to 2015. University of Michigan; 2017.
- 17.↑
Dargay J, Hanly M. Volatility of car ownership, commuting mode and time in the UK. Transp Res Part A Policy Pract. 2007;41(10):934–948.
- 18.↑
Clark B, Chatterjee K, Melia S. Changes in level of household car ownership: the role of life events and spatial context. Transportation. 2016;43(4):565–599. doi:
- 19.↑
Panter JR, Jones AP, van Sluijs EM. Environmental determinants of active travel in youth: a review and framework for future research. Int J Behav Nutr Phys Act. 2008;5(1):34. doi:
- 20.↑
Ikeda E, Hinckson E, Witten K, Smith M. Associations of children’s active school travel with perceptions of the physical environment and characteristics of the social environment: a systematic review. Health & place. 2018;54:118–131. PubMed ID: 30261352 doi:
- 21.↑
Shoham DA, Dugas LR, Bovet P, et al. Association of car ownership and physical activity across the spectrum of human development: Modeling the Epidemiologic Transition Study (METS). BMC Public Health. 2015;15(1):173. doi:
- 22.↑
McMinn AM, van Sluijs EMF, Nightingale CM, et al. Family and home correlates of children’s physical activity in a multi-ethnic population: the cross-sectional child heart and health study in England (CHASE). Int J Behav Nutr Phys Act. 2011;8(1):11. doi:
- 23.↑
Pouliou T, Sera F, Griffiths L, et al. Environmental influences on children’s physical activity. J Epidemiol Community Health. 2015;69(1):77–85. PubMed ID: 25359920 doi:
- 24.↑
Pulsford RM, Griew P, Page AS, Cooper AR, Hillsdon MM. Socioeconomic position and childhood sedentary time: evidence from the Peach project. Int J Behav Nutr Phys Act. 2013;10:105.
- 25.↑
Nakabazzi B, Wachira LM, Oyeyemi AL, Ssenyonga R, Onywera VO. Prevalence and socio-demographic correlates of accelerometer measured physical activity levels of school-going children in Kampala city, Uganda. PLoS One. 2020;15(7):e0235211. doi:
- 26.↑
Gerber M, Lang C, Beckmann J, et al. Associations between household socioeconomic status, car ownership, physical activity, and cardiorespiratory fitness in South African primary schoolchildren living in marginalized communities. J Phys Act Health. 2021;18(8):883–894. PubMed ID: 34172589 doi:
- 27.↑
Ferreira I, van der Horst K, Wendel-Vos W, Kremers S, van Lenthe FJ, Brug J. Environmental correlates of physical activity in youth – a review and update. Obes Rev. 2007;8(2):129–154. PubMed ID: 17300279 doi:
- 28.↑
Carlin A, Perchoux C, Puggina A, et al. A life course examination of the physical environmental determinants of physical activity behaviour: a “determinants of diet and physical activity” (DEDIPAC) umbrella systematic literature review. PLoS One. 2017;12(8):e0182083. doi:
- 29.↑
Atkin AJ, Biddle SJH, Broyles ST, et al. Harmonising data on the correlates of physical activity and sedentary behaviour in young people: methods and lessons learnt from the international children’s accelerometry database (ICAD). Int J Behav Nutr Phys Act. 2017;14(1):174. doi:
- 30.↑
Sherar LB, Griew P, Esliger DW, et al. International children’s accelerometry database (ICAD): design and methods. BMC Public Health. 2011;11(1):485. doi:
- 31.↑
Salmon J, Telford A, Crawford D. The Children’s Leisure Activities Study (CLASS). Summary Report. VIC: Centre for Physical Activity and Nutrition Research; 2004.
- 32.↑
Steele RM, van Sluijs EMF, Cassidy A, Griffin SJ, Ekelund U. Targeting sedentary time or moderate- and vigorous-intensity activity: independent relations with adiposity in a population-based sample of 10-y-old British children. Am J Clin Nutr. 2009;90(5):1185–1192. PubMed ID: 19776141 doi:
- 33.↑
Trost SG, Ward DS, Moorehead SM, Watson PD, Riner W, Burke JR. Validity of the computer science and applications (CSA) activity monitor in children. Med Sci Sports Exerc. 1998;30(4):629–633. PubMed ID: 9565947 doi:
- 34.↑
Brand C, Goodman A, Rutter H, Song Y, Ogilvie D. Associations of individual, household and environmental characteristics with carbon dioxide emissions from motorised passenger travel. Appl Energy. 2013;104(100):158–169. PubMed ID: 24882922 doi:
- 35.↑
Goodman A, Brand C, Ogilvie D, Iconnect C. Associations of health, physical activity and weight status with motorised travel and transport carbon dioxide emissions: a cross-sectional, observational study. Environ Health. 2012;11:52.
- 36.↑
Salvo D, Garcia L, Reis RS, et al. Physical activity promotion and the united nations sustainable development goals: building synergies to maximize impact. J Phys Act Health. 2021;18(10):1163–1180. PubMed ID: 34257157 doi:
- 37.↑
Bernard P, Chevance G, Kingsbury C, et al. Climate change, physical activity and sport: a systematic review. Sports Med. 2021;51(5):1041–1059. PubMed ID: 33689139 doi:
- 38.
Gelius P, Till M, Messing S, Tcymbal A, Abu-Omar K. Must we tell people to be less active? The dilemma of physical activity recommendations from a holistic health perspective. Ger J Exerc Sport Res. 2023;54:21–28.
- 39.↑
Siefken K, Abu-Omar K. Striking a balance: physical activity and planetary health. J Phys Act Health. 2023;20(12):1081–1083. PubMed ID: 37890840 doi:
- 40.↑
Bunds KS, Kanters MA, Venditti RA, Rajagopalan N, Casper JN, Carlton TA. Organized youth sports and commuting behavior: the environmental impact of decentralized community sport facilities. Transp Res D Transp Environ. 2018;65:387–395. doi:
- 41.↑
Laxdal A, Johansen BT, Bere E, Solstad BE. Towards more sustainable sports: analyzing the travel behavior of adolescent soccer players in southern norway. Int J Environ Res Public Health. 2022;19(15):9373. doi:
- 42.↑
Zeiher H. Children’s islands in space and time: The impact of spatial differentiation on children’s ways of shaping social life. In: Du Bois-Reymond M, Sunker H, Kruger HH, eds. Childhood in Europe. Approaches—Trends—Findings. Peter Lang; 2001:139–159.
- 43.↑
Smith L, Sahlqvist S, Ogilvie D, et al. Is a change in mode of travel to school associated with a change in overall physical activity levels in children? Longitudinal results from the Speedy study. Int J Behav Nutr Phys Act. 2012;9(1):134. doi:
