Levels of sedentary time and physical activity
This is the first study to show that young South Asian and white British children from a deprived urban location spend half of their daily time sedentary, just over 40% in light physical activity, and remaining time in MVPA. Wide variability exists in the levels of objectively measured movement behaviours in young children (27). In part these differences are due to methodological inconsistencies, including different accelerometer types and placement, data reduction and processing, all of which prevent a direct (like-for-like) comparison of results across study populations. Even so, our results appear to be similar to studies that have incorporated the same accelerometer, device placement and cutpoints in children of comparable ages from the Netherlands (28), Sweden (29), Switzerland (30) and America (31); though the children in the current study participated in slightly more sedentary time and less light physical activity. Another study conducted in England, by Hesketh and colleagues, used a different measurement procedure and was performed in a predominantly white population, from a smaller and more affluent city in the South of the country (32). That study found that children performed a similar amount of MVPA, but nearly 60% of daily time was spent in light physical activity and only one-third was sedentary time (32). The estimate provided by Hesketh et al. differs considerably from a recent meta-analysis, which calculated that children aged 2–6y spend 51.4% (95% CI: 48.4 to 54.4%) of daily time being sedentary (33). This concurs with our novel results in young South Asian and white British children, who also spent approximately half of daily time in sedentary behaviour.
Patterns of sedentary time and physical activity
This is the first study to compare the levels and patterns of movement behaviours in young South Asian and white British children. Similar studies have been performed in school-aged children (most of them using data more than a decade old) and they have consistently found that South Asian children are less physically active compared to white British children (8,34). This was not observed in the current sample of young children. Ultimately, we found no significant ethnic differences in levels of sedentary time or physical activity, even though there was some indication that South Asian boys were less sedentary and more active than white British boys. This could indicate that meaningful ethnic differences only manifest when children are older, which would imply that early prevention is key to preventing ethnic inequalities in physical activity. Alternatively, second generation UK South Asians are more likely to be educated beyond school (35), and they purportedly have different attitudes to physical activity and are more active compared to the first (migrant) generation (8). Plausibly, as knowledge, attitudes and behaviours in the youngest generation of UK South Asians and their UK born parents have continued to change, this may have translated to parity in children’s physical activity levels compared to the white British. Adding support to this hypothesis, we recently reported that there were no differences in the levels of light physical activity or MVPA performed by present-day South Asian and white British primary school-aged children aged 6 to 8y (36). New studies of contemporary datasets are warranted, to revisit the presumption that South Asian children are less active than white British children.
Sedentary time did not significantly vary by ethnicity in this study of young children, but it is unfortunate that we were not able to scrutinise different types of sedentary behaviour. We have previously shown that compared to white British children, Pakistani infants watch more television (37), which is associated with myriad adverse health outcomes including obesity (38). South Asian primary school children have consistently been shown to be more sedentary compared to white British children (36,39). This is speculated to be because of Madrassa, an after-school religious commitment that starts around age 5y for South Asian children (40,41). Most children in this study were measured when they were younger than 5y and as such we may have missed this transition; we found that sedentary time decreased with advancing age in both ethnic groups, and levels of MVPA and vector magnitude CPM increased. This spike in MVPA (which precedes an age-related decline throughout the life-course (42)) has been reported previously and is often attributed to young children mastering movement skills (43,44). We observed that older white British children performed more light physical activity than younger white British children, but there were no such differences in South Asian children. This could be meaningful for future health inequalities if the trend was to continue, because light physical activity is associated with lower adiposity and higher aerobic fitness in mid-childhood (45).
A recent review and meta-analysis reported that young boys are more physically active (46) and less sedentary (33) than young girls, but there are several large and methodologically sound studies that report no differences by sex (e.g. (47–49)). We found no differences in movement behaviours between white British boys and girls. This is in contrast to Hesketh et al., who found that white British boys accumulated more MVPA than girls (32). The difference in results could be explained by incomparable measurement procedures combined with the already described differences between study populations. Although we found no gender differences in white British children, South Asian boys performed more MVPA and registered more vector magnitude CPM than South Asian girls. Islamic religious settings like mosques or madrassa, with support and training from local authorities to organise and encourage physical activity, could provide important opportunities to reduce the observed gender differences in South Asian children (50).
It is uncertain whether physical activity levels differ between week and weekend days in young children (46). We found that MVPA and vector magnitude CPM were higher on weekends in both ethnic groups. In white British children of comparable age, Hesketh et al. found that overall levels of light physical activity and MVPA did not vary between weekdays and weekends, but there were differences between particular segments of days (e.g. children performed more MVPA on weekend afternoons and evenings compared to weekdays) (32). With regard to sedentary time, our results for white British children concur with meta-analysed results from four studies of young children, which together indicated there were no differences in sedentary time by type of day (33). We did observe, however, that sedentary time was markedly higher on weekdays than weekends in South Asian children. As mentioned, most children in this study were measured when they were younger than 5y, hence before most of them begin to attend Madrassa every weekday after school. It is still possible, nonetheless, that Madrassa may inadvertently increase sedentary time in young children by intruding upon opportunities for physical activity. For example, South Asian parents are often preoccupied on weekday afternoons as they chaperon older siblings between school, home and the mosque, leaving less time for after-school physical activities (40).
There is inconsistent evidence for seasonal differences in the physical activity of young children (46). We found that white British children were less sedentary and more physically active in spring compared to winter, and to some extent were less sedentary and more light physically active in summer than winter. South Asian children were markedly less sedentary and more physically active in all seasons relative to winter. This highlights that the movement behaviours of young South Asian children are markedly hampered in winter, more so than their white British counterparts. This could be due to a combination of cultural and environmental differences. It has been suggested that South Asian parents prefer to keep children indoors given inclement weather or winter colds (17,41). Additionally, South Asian families taking part in this study lived in the most deprived areas. They may have had less access to quality indoor play areas and more concerns about neighbourhood and traffic safety on dark winter evenings (51).
A recent review found that there was no evidence in young children for associations of parental education or other markers of socio-economic status with physical activity (46). Aside from parental education, geography based measures of neighbourhood deprivation have most commonly been used to signify socio-economic status. In line with earlier studies, we found no evidence that movement behaviours varied by area deprivation. It is possible, however, that an area level marker of socio-economic status may have introduced misclassification errors and biased associations to the null (52). Supporting this assertion, we found that household income varied markedly within each group of area deprivation (e.g., in the most deprived 10% of areas the distribution of household income was <£6,999 (30.7%), £7,000 to 16,999 (50.5%) and ≥£17,000 (18.8%)). With this in mind, it is valuable that we were also able to investigate household income as an individual level marker of socio-economic status. Consistent with the results for area deprivation, no associations were apparent in white British children, but South Asian children in the highest compared to the lowest income households were more sedentary and performed less light physical activity. We have previously reported that TV viewing is prevalent in young South Asian children who watch more TV overall (37) and in the evening than white British children (38). It is plausible that South Asian children in higher than lower income households may have more access to TVs and other screen-based electronic media (including laptops, tablets, games consoles, and smartphones, equipment which can be prohibitively expensive), thereby encouraging more sedentary time and displacing light physical activity. Unlike physical activity participation in older children, which can require sports equipment or fees for club membership, physical activity performed by young children is mostly informal and play-based, and rarely incurs financial cost (53). This might explain why levels of MVPA in our sample of young children did not vary by household income.
In young children there is inconsistent evidence for associations between physical activity with adiposity (1) and associations for sedentary behaviours are predominantly null (2). It is a limitation, however, that the majority of studies have used BMI as a proxy measure of adiposity. This may also explain why we failed to find evidence of variation in movement behaviours by weight status. We have previously reported (in a study that consisted largely of the same sample of children) dose-dependent associations between physical activity intensity with lower sum of skinfolds, a more direct marker of adiposity (16). It is unfortunate that the diminutive number of underweight children in the current analysis precludes any meaningful inference for this particular category. Very low levels of physical activity have previously been observed in young Ethiopian children with severe acute malnutrition (54). Additional studies should investigate the movement behaviours of underweight children in western society, who are at elevated risk of adverse health and development delays and tend to be from disadvantaged social backgrounds (55).
Previous studies have reported null associations between parental age with child physical activity (46) and sedentary behaviour (56). One exception was a study of Canadian children, in whom the prevalence of performing ≥60 minutes of MVPA per day was lower in 5y old children whose mothers were older at delivery (49). In both ethnic groups, we observed that children with older mothers tended to be more sedentary and less physically active. Beyond childhood there is an age-related decline in physical activity (42) and in young children maternal-child activity levels are closely related (28,57). For these reasons, it is conceivable that lower physical activity and higher sedentary time performed by older mothers may have translated to more of the same behaviours in their children. There might also be biological consequences of having an older mother that predispose to suboptimal movement profiles (58). For instance, children born to women aged ≥30y are at elevated risk of prematurity, which appears to predict decreased lung function (59) and delayed motor development (60). Birth weight is also lower in children born to older mothers, and lower birth weight is associated with childhood morbidity (61), and diminished aerobic and neuromuscular fitness (62). This may explain our observation that South Asian children of high birth weight performed more MVPA and registered more vector magnitude CPM than their normal birth weight peers. We observed no comparable associations in white British children. This apparent ethnic difference could be because South Asian children are typically smaller at birth (63) and hence a threshold of >3500g constituted a more extreme birth weight for South Asian than white British children. A recent best evidence synthesis concluded that there is limited evidence that birth weight predicts movement behaviours, with the exception of extreme birth weights predicting later physical activity (64).
Prevalence and factors associated with physical inactivity
Nearly all valid days were characterised by ≥180 minutes of total physical activity at any intensity, which until recently was the recommendation for all children younger than 5y. However, on two-thirds of days children were deemed to be inactive, as they did not satisfy the new international guideline daily amount of physical activity for 3 to 4y olds (≥180 minutes of total physical activity, including ≥60 minutes MVPA (3–7). There were no significant ethnic differences in the likelihood of meeting guidelines, although there was some indication that white British boys were more likely to be inactive compared to South Asian boys. In both ethnic groups, every additional year of age was associated with lower odds of physical inactivity. South Asian children were less likely to be physically inactive at weekends compared to weekdays, and relative to winter were half as likely to be physically inactive in all other seasons. White British children were less likely to be physically inactive in spring than winter. By far the strongest association was found for South Asian girls, who were 3.5 times more likely to be physically inactive compared to South Asian boys. Our results highlight that certain periods may be better targets for reducing the prevalence of childhood inactivity (e.g. focussing on South Asian children at weekdays and in winter) and that targeting children from a young age may pay dividends. South Asian girls appear to have the most to gain from a physical activity intervention.
Strengths & weaknesses
Data were harmonised from three studies to permit investigation of a relatively large sample of South Asian and white British children from a materially deprived urban location. This is an important population to study as it is high risk for childhood obesity and subsequent adverse health, and can aid understanding of ethnic inequalities (65). It is unfortunate that, due to small numbers, all children of South Asian origin were considered as one group. This precluded a more detailed description of movement behaviours by specific countries of origin. Triaxial accelerometry and a short sampling interval were advantageously used to investigate the whole range of movement behaviours, including light physical activity, for which the evidence of health benefits continue to emerge (1). Combining the accelerometer data with contextual information to explain patterns of movement behaviours, and to identify the types of sedentary behaviour performed by children, would have been beneficial (66). Nonetheless, this is the first study to quantify levels of movement behaviours in a young group of bi-ethnic children from a deprived location, and to investigate patterns of movement by numerous socio-demographic, temporal and perinatal characteristics. The prevalence of healthy weight, overweight and obesity in this study sample closely matched that of all children aged 4 to 5y in Bradford who were measured in 2017/18 as part of the National Child Measurement Programme (67). This provides further reassurance that the study sample represented the source population. Our results are likely generalisable to other young South Asian and white British children who are living in materially deprived UK cities.