Covid-19 lockdown: Ethnic differences in children’s self-reported physical activity and the importance of leaving the home environment; a longitudinal and cross-sectional study from the Born in Bradford birth cohort study

Background In England, the onset of COVID-19 and a rapidly increasing infection rate resulted in a lockdown (March-June 2020) which placed strict restrictions on movement of the public, including children. Using data collected from children living in a multi-ethnic city with high levels of deprivation, this study aimed to: (1) report children’s self-reported physical activity (PA) during the first COVID-19 UK lockdown and identify associated factors; (2) examine changes of children’s self-reported PA prior to and during the first UK lockdown. Methods This study is part of the Born in Bradford (BiB) COVID-19 Research Study. PA (amended Youth Activity Profile), sleep, sedentary behaviours, daily frequency/time/destination/activity when leaving the home, were self-reported by 949 children (9–13 years). A sub-sample (n = 634) also self-reported PA (Physical Activity Questionnaire for Children) pre-pandemic (2017-February 2020). Univariate analysis assessed differences in PA between sex and ethnicity groups; multivariable logistic regression identified factors associated with children’s PA. Differences in children's levels of being sufficiently active prior to and during the lockdown were examined using the McNemar test; and multivariable logistic regression was used to identify factors explaining change. Results During the pandemic, White British (WB) children were more sufficiently active (34.1%) compared to Pakistani Heritage children (PH) (22.8%) or ‘Other’ ethnicity children (O) (22.8%). WB children reported leaving the home more frequently and for longer periods than PH and O children. Modifiable variables related to being sufficiently active were frequency, duration, type of activity, and destination away from the home environment. There was a large reduction in children being sufficiently active during the first COVID-19 lockdown (28.9%) compared to pre-pandemic (69.4%). Conclusions Promoting safe extended periods of PA everyday outdoors is important for all children, in particular for children from ethnic minority groups. Children’s PA during the first COVID-19 UK lockdown has drastically reduced from before. Policy and decision makers, and practitioners should consider the findings in order to begin to understand the impact and consequences that COVID-19 has had upon children’s PA which is a key and vital behaviour for health and development. Supplementary Information The online version contains supplementary material available at 10.1186/s12966-021-01183-y.


Background
In England, the immediate response to the first wave of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)-COVID19-pandemic was a stringent lockdown implemented on 23 rd March 2020 [1]. The government placed extreme restrictions on movement of the public stating that "during the emergency period, no person may leave the place where they are living without reasonable excuse", which included shopping for food and medical supplies [1]. Furthermore, guidance stipulated that members of the public could also leave the home for a short bout (60 min) of local daily exercise. All playgrounds and indoor and outdoor play facilities (e.g. skate parks, soft play centres) were closed, in addition to leisure facilities and gyms. Schools were closed for most children with the exception of vulnerable children and children of key workers (those working across health, social and public sectors). The lockdown measures were eased in England on 4 th July 2020 [2]. However, at the time of writing, two further national lockdowns have occurred in England, in November 2020 [3] and January 2021 [3,4].
Before the COVID-19 pandemic, national and international epidemiological data (whether device or selfreported measured) report that up to 80% of children and young people in high-income countries are not sufficiently physically active for health and well-being (e.g. achieve 60 min of MVPA per day) [5,6]. Within England, recent survey data suggests that the 53.2% of children, aged 5-16 years, were not achieving physical activity (PA) guidelines [7]. Of specific concern, levels of inactivity were higher in children from ethnic minority groups, especially those with South Asian heritage [8,9]. Such low levels of PA place children at risk of poor physical and mental wellbeing in addition to having a negative impact on school performance [10][11][12][13][14]. Within South Asian communities, such risks are high as children present with increased rates of Obesity and type-II diabetes [15]. Further, early evidence suggest such populations are more likely to suffer the most during and after the pandemic [4,[16][17][18][19][20]. Research conducted during COVID-19 has already reported low levels and significant reductions of children's PA [21,22]. It is essential to understand the impact of the pandemic on PA levels and behaviours for different ethnic groups for two reasons, first to prevent inactive behaviours becoming entrenched and second, to tailor support for different populations by addressing the root causes of PA inequality within different populations [23].
The Born in Bradford (BiB) research programme [24] provides a premium opportunity to study the impact of COVID-19 lockdown on school-aged children living in a deprived and ethnically diverse city. To date, BiB has tracked/monitored the health, wellbeing, and determinants of health of over 30,000 Bradford residents (parents and children) since 2007 [24]. The latest round of data collection occurred pre COVID (2017-March 2020, n = 7500, aged between 6-11 years) [25], establishing a pre-COVID baseline; providing a unique opportunity to understand the impact of the COVID lockdown on physical activity behaviour in an ethnically diverse sample of school-aged children and young people. Further, the BiB cohort study will follow participants throughout the duration of the pandemic and in the following years, to understand the impact of the crisis on health and wellbeing trajectories [26].
The current study is part of the wider Born in Bradford COVID-19 Research Study [26] and aims to: 1) report children's self-reported physical activity (PA) during the first COVID-19 UK lockdown and identify associated factors; 2) examine changes of children's self-reported PA prior to and during the first UK lockdown. overcrowded [28]. Almost a quarter of Bradford children live in poverty and 25% are living with obesity at age 11, and the rates of childhood obesity are 6% higher among the same age group of South Asian children compared to White British children [28]. Such socio-economic and structural characteristics of Bradford make the community particularly vulnerable to COVID-19.

Participants and procedure
Participants were children aged 9-13 who were invited to take part in the BiB COVID-19 research study [26] following a protocol approved by the Health Research Authority and Bradford/Leeds research ethics committee (reference: 16/YH/0320).
The parents/carers of 5,298 children aged 9-13 years who are participants in the existing BiB birth cohort study and who had engaged in a recent follow-up data collection wave pre-COVID-19 (2017-early March 2020) [25], were contacted by trained researchers via telephone to invite their child to take part in a survey. Following verbal consent from parents/carers, children received a survey via post to be completed and returned to the research team using pre-paid envelopes [26]. Completion of the survey was deemed as participation assent from the child. Overall, 970 children returned surveys during the period of May 21st to July 31st 2020 and 949 children (97.8%) provided enough data to be included for the analysis of the first two aims of this research.

Demographic measures
Children's age, sex, ethnicity, and home postcode-derived Index of Multiple Deprivation (IMD) [29] were extracted from the BiB cohort dataset. Three categories of ethnicity were used for the analysis, White British (WB), Pakistani Heritage (PH) (the two largest groups in the sample) and 'Other' (O) (any other ethnic group). Index of Multiple Deprivation (IMD) deciles [30] were categorised into either the 'most deprived nationally' (most deprived 10% areas in England), '2 nd most deprived nationally' (10-20% most deprived areas), '3 rd most deprived nationally' (30-40% most deprived areas), and '4 th or more most deprived nationally' (40%-100% most deprived areas). Child's school attendance during the April-June 2020 lockdown period was included in the survey.

During COVID-19 Lockdown
During the first COVID-19 lockdown PA, sedentary behaviours, screen-time, sleep, activity (frequency, duration, type, and place) away from home environment were all measured by child self-report (Table 1). Self-reported PA was measured using a modified version of the validated seven day recall questionnaire, the Youth Activity Profile-English Youth Version (YAP) [31,32]. The YAP requires children to report the frequency and/or duration of physical activities through different segments of a usual day (i.e. before school, break time at school, lunch at school, after school). During the first lockdown most children were not attending school, so this format was not appropriate, and for the same reason, neither was the PA questionnaire for children (PAQ-C) [33] which was the questionnaire completed by children in data collection pre-COVID [25]. The choice to amend the YAP, which was originally based upon the PAQ-C, and to not use the PAQ-C was due to the YAP specifically including an item asking directly for an estimation of time in overall MVPA across weekend days. Following consultation with the lead author of the English child version of the YAP (also a co-author of this study-SF) a decision was made to ask the YAP-weekend item along with an overall weekday item, using the same wording ( Table 1). The YAP was also used to estimate sedentary behaviours whilst watching television, playing video games, using a mobile phone, a computer/tablet during COVID- 19 restrictions. An additional question of 'doing school work' was also included to capture the amount of time children spent doing sedentary school work during COVID-19 restrictions (Table 1). A binary variable of meeting screen time (ST) guidelines (< 2 h a day) [34,35] was calculated by the values of each answer for sedentary screen behaviours (Table 1). Children's average sleep time was estimated by children reporting the time they normally go to sleep and the time they normally wake up. Sleep time was categorised into meeting sleep guidelines [34] (9-to-11 h a day) or more or less.
Because of the uniqueness of the COVID-19 lockdown and subsequent reduced opportunities for children to being physically active as they normally would be, children were asked to answer questions on the frequency they normally left the home, the duration they would normally leave for, the type of activities they usually did when leaving the home, and where they usually went (Table 1).

Before COVID-19
For the sub-sample of children with available data from before COVID-19, PA levels were measured by children completing the PAQ-C with the support of trained researchers during school time (2017-2019). The PAQ-C is a validated PA seven day recall questionnaire, that measures general levels of MVPA of children aged 8-14 years by assessing participation in different physical activities as well as activity during physical education, lunch break, recess (play time), before school, after school, evenings and weekends [33,36]. The scoring of the PAQ-C is based upon an average of all questions  asked, with a score between 1 (low activity rating) to 5 (high activity rating) [33,36]. Cut-off values indicating whether children were sufficiently active (relating to cardio-respiratory fitness [37]) were applied (2.7 aggregate score [out of 5] for girls, 2.9 aggregate score [out of 5] for boys).

Statistical analysis
Descriptive statistics for all variables were generated. Continuous variables were described using mean and standard deviation; categorical variables using counts and proportions. between pre-post COVID, ethnicity, IMD) were investigated using logistic regression through simultaneous entry of independent variables. The outcome variable was coded 0 for the 'absence of negative change' for being sufficiently active and 1 for the 'presence of negative change' being sufficiently active. An alpha value ≤ 0.05 was considered statistically significant.

Descriptive Statistics
Out of a total n = 5,298 eligible children, n = 970 (18.3%) children agreed to take part, completed and returned a survey in spring 2020. A total of n = 949 (17.9%) had completed PA data and were included in the analysis for aim 1. A total of n = 634 (66.8%, based upon 949 children) children had matched PA data prior to COVID-19.
The characteristics for both pre and during COVID-19 samples are reported in Table 2.
During COVID-19 lockdown: Levels of self-reported physical activity and activity away from the home   once a day. However, 30% reported that they had normally stayed at home. Of the children who reported leaving the home at least once a day, the majority of children reported leaving between 31-60 min (54%). The most frequently reported type of activities outside of the home was walking (77%) and riding a bike/scooter (41.9%), and the most frequent reported places for children to go was the street (33.7%) and park (34.5%).

During COVID-19 lockdown: Sex and ethnicity behaviour differences during the COVID-19 lockdown
Univariate sex and ethnicity differences are reported in   jogging, riding a bike/scooter, and playing sports and games.
For the multivariable analysis, summaries of logistic regression models (1, 2, 3, 4) are reported in Table 5 (a full results table is found in Additional file 1: Appendix 1 -supplementary material). In model 1, variables that decreased the odds of being sufficiently active were age (years) (Odds ratio, OR = 0.82, 95%CI 0.72-0.94), and ethnicity (reference: WB); PH children (OR 0.64, 95%CI 0.44-0.92), Other (OR = 0.57, 95%CI 0.35-0.90). In model 2 (which included sedentary behaviours), being a girl (OR = 0.63, 95%CI 0.45-0.88) and playing on video games ≥ 3 h a day (OR = 0.43, 95%CI 0.28-0.67) significantly decreased the odds of being sufficiently active, in addition to age and ethnicity. In model 3 (which included daily frequency of leaving the home), age, being a girl, being from another ethnic group and playing video games (≥ 3 h a day) still decreased the odds of being sufficiently active; however, being of PH no longer did. Leaving the home at least once a day significantly increased the odds (OR = 1.57 95%CI(1.04-2.36)), with the odds increasing further for children who reported leaving the home more than once a day (OR = 2.73, 95%CI 1.66-4.48). In model 4, (which included duration, place and type of activity), age and playing videos for ≥ 3 h/day significantly decreased the odds, but leaving the home for 31-60 min significantly increased the odds (OR = 2.21, 95%CI 1.01-4.8), and the odds increased further for children reporting leaving the home for ≥ 60 min (OR = 7.9, 95%CI 3.5-18.0). Children reporting that the place they usually went too was the shop which reduced the odds of children being sufficiently active (OR = 0.36, 95%CI 0.13-0.98). Odds were increased for children reporting that usually took part in running/jogging (OR = 2.13, 95%CI 1.30-3.47), riding a bike/scooter (OR = 1.52, 95%CI 1.01-2.31), and playing sports and games (OR = 2.13, 95%CI 3.4-2.70). * corrected p-value = .006, non-significant ** corrected p-value = .008, non-significant *** corrected p-value = .008, non-significant

Changes in children being sufficiently physically active before and during the COVID-19 lockdown
The sub-sample's pre-COVID-19 PA (sub-sample, n = 643) was PAQ-C score 3.2 (SD = 0.77), with 69.4% (n = 440) found to be sufficiently active. During COVID-19 the proportion of children being sufficiently active reduced to 28.7% (n = 183). The magnitude of change was statistically significant (see Table 6), with 47.5% of children changing from being sufficiently active before COVID-19 to not being sufficiently active during COVID-19. A small number of children who were not sufficiently active pre-COVID did report being sufficiently active during COVID-19 (7.0%, n = 44), leading to a 40.5% reduction. A logistic regression model ( Table 7) predicted that the age difference between the two measurement periods, ethnicity and sex did not significantly increase or decrease the odds of children negatively changing from being sufficiently active from before COVID-19 to during COVID-19.

Discussion
The purpose of this study was to investigate the levels, factors associated and change of children's self-reported PA during the first COVID-19 in England. Results show levels of children reporting being sufficiently active has drastically reduced from before COVID-19. Only a quarter of children reported being sufficiently active enough to benefit their health during the first COVID-19 lockdown, and this reduced greatly from before COVID-19, independently of increased age. These findings are similar to other studies [19,21,38] and are unsurprising when considering the sharp change in the systems in which children's PA would usually occur (i.e. school, sport clubs, parks, playgrounds, active travel). Daily PA outside of the home environment was allowed and has been consistently allowed by the UK-government during the first lockdown and throughout the pandemic, but not actively promoted [38,39]. This is unsurprising due to the priority being to reduce mixing of individual households. As the BiB COVID-19 study [26] progresses further studies will be able to report on changes in PA during the pandemic and during differing restriction circumstances. It is likely, given the ongoing restrictions, that PA levels will remain lower than pre-pandemic. The short-and long-term health implications for reduced PA for a sustained period of time during childhood is unknown and this is something which the cohort study aims to investigate. Early life is particularly important for habit formation and has been shown that PA tracks from across the life course of young people [40][41][42] so there is a possibility of long term health implications associated with reduced PA across the lifespan, triggered by reduced PA during the COVID-19 pandemic. This is something that requires careful monitoring and preventative interventions to reduce the likelihood of ongoing low PA levels.
Previous non-pandemic research has shown an association between children's PA levels and time spent outdoors, [43,44] with the current study's findings highlight how important time away from the home environment was for being active. Worryingly, 29.7% of children reported that they didn't leave the home on a usual day during lockdown and this was strongly associated with not being sufficiently active (OR = 1.6 once a day, OR = 2.7 ≥ once a day). For those who did leave the home, just under half of children (46%) did so for longer than 60 min and leaving the home environment for this amount of time was found to be important for children being sufficiently active (i.e. MVPA-60 min guidelines, OR = 7.9). The government guidance during the first and all subsequent lockdowns (November 2020, Januarycurrently 2021) has been to minimise the time spent outside of the home, and there has been a common misconception that exercising away from the home should be for no longer than one hour [45]. This study illustrates the importance of extending the amount of time away from the home for children to be physically active, and if this can be done safely, should be promoted.
The places children most frequently reported going to were, the streets, and parks, and the most frequent activities reported were walking and riding a bike/scooter. The results showed that children who reported going to the shops had reduced odds of being sufficiently active, therefore illustrating that getting out of the home environment to places which are conducive to being active (e.g. streets, parks, greenspaces) should be encouraged, whilst adhering to current COVID-19 guidelines and taking always necessary precautions (e.g. staying 2 m apart). Furthermore, as would be expected, children who reported engaging in more vigorous PA such as riding a bike/scooter and playing sports and games were more likely to be sufficiently active than those who reported just walking; suggesting that campaigns should focus on the promotion of these more vigorous types of activities, but also acknowledge that any PA is worthwhile and should be promoted.
There were large differences in whether children reported leaving the home and for how long between ethnic groups. PH and O children left the home significantly less often than their WB peers and for shorter periods. When frequency of leaving the home was controlled for, ethnic PA differences no longer existed between WB and PH children, therefore highlighting an inequality in a key factor for why children were sufficiently active during COVID-19 lockdown (i.e. more PH children stayed at home than WB, therefore were less active). Because of the importance of leaving the home environment to be sufficiently active during COVID-19 lockdown, it is important for policy, strategy and practice to consider why some children were leaving the home environment and why others were not, particularly between different ethnic groups. The current study did not directly ask children to report why they had not left their home so this could not be examined. It may have been that those who did not leave the home environment were living in areas less conducive for PA. The following environmental determinants of children and adolescents PA have previously been identified: walkability, availability/access/ proximity to recreational facilities, environment aesthetics, negative street characteristics [46]. All such determinants were not explored in the current study and should be considered in future research to possibly explore the ethnic differences found.
A further influence upon whether children were leaving the home in the current study may have been worries and stress experienced by families during lockdown. Mothers of the children from the sample of this study, who mostly live in areas of high deprivation reported numerous difficulties during the spring 2020 lockdown with many insecurities (financial, employment, housing, clinical symptoms of anxiety and depression) and high levels of anxiety about becoming ill or dying from COVID-19 [47] furthermore, COVID-19 has disproportionately impacted ethnic minority groups such as Pakistani, South Asian, and Black ethnicities more than White British, with greater ill health and death reported [47]. Anxiety and fear of ill health and death could be greater within PH and O groups leading to them not wanting to leave the home environment. Negative mainstream media reporting on ethnic minorities violating lockdown protocols and government guidelines, and fear of getting labelled when outside of home could have been another reason for ethnic minority children not leaving home during lockdown. More research using qualitative and anthropological methodologies are required to begin to understand this complex phenomenon, because there is a risk of the exacerbation of PA inequalities between ethnic groups, which were well established pre-COVID-19 [4,16,23,48].
The guidance to stay at home during lockdown periods and anxieties surrounding leaving the home, and the association between leaving the house and physical activity, has created a demand for home-based PA interventions for children, with numerous options being made available [49,50]. Previous research on the determinants of children's home PA are unclear with inconsistent findings [51]. These programmes which have been developed rapidly may not be evidence based or grounded in behaviour change theory. Moreover, there is a dearth of literature regarding the feasibility, acceptability, efficacy and effectiveness of such home-based PA programmes/ interventions [52]. Home-based PA will likely remain in demand for the foreseeable future as part of the gradual   (7), p = 0.0014 reopening of society, and the changing of the home environment from one promoting mainly sedentary time activities to more physically active activities [51] is a topic of priority to further understand how best is it for children to be active within their home environments. A concern of the COVID-19 lockdown(s) has been a possible increase of sedentary (particularly screen behaviours) and disturbances of sleep [18,[53][54][55]. Findings from this study showed that the majority of children reported meeting sleep guidelines (9-11 h a day), engaged in more than one hour a day of TV viewing, playing video games on a console, and doing school work; but also only a small proportion of children (10.1%) reported in meeting ST-guidelines recommendations. Data from Canadian young people also found similar low levels of ST compliance during the first COVID-19 lockdown (11.3%) [19]. But the current study has most likely underestimated the amount of ST, due to the use of a screen for school work (school work item queried any school work, whether using a screen or not) not being factored in the estimate. The high non-compliance of ST recommendations is higher compared to a UK sample of young people pre-COVID (23.1%) [56], which is unsurprising for children restricted to the home environment for much of their time. Of all of the sedentary ST behaviours, playing a video game on a console for ≥ 3 h decreased the odds (OR = 0.43-0.52) of children being sufficiently active. However, it must be noted that significantly more boys reported playing video games for more than 3 h a day than girls, and this may explain why there was not a found difference between the sexes, which is usually commonly found in physical activity data. These findings suggests that alongside promotion of leaving the home to support PA during and following the pandemic, reducing the use of sedentary ST, in particular video game usage (especially for boys) also needs addressing by public health campaigns. The ethnic and sex differences of sleep and sedentary behaviours (video games, mobile usage, school work) found in this study should be further explored with different outcomes such as educational, emotional and mental health which all have been associated previously with sleep and sedentary behaviours [57].
The limitations of this study include use of two different child self-reported PA questionnaires (for pre and during COVID) with one questionnaire (amended-YAP) not being formally validated. The BiB study had previously used the validated PAQ-C questionnaire due to availability of children's questionnaires for previous cohort data collection [25], before the YAP had been published. In the current study, the PAQ-C was decided by authors not to be suitable for use during lockdown with the majority of children not attending school and being restricted to their homes. Irrespective of the questionnaires used, an obvious limitation of this study is the use of children's self-reported behaviour and recall. A further limitation is that causality of the variables associated with PA-guidelines cannot be implied, and neither can the direction of association, this is due to the crosssectional nature of data presented. However, the circumstances of COVID-19 and the ability to rapidly survey and receive data from 979 children and continue to follow and collect further data in the future, is a strength of this city-wide cohort study. The ongoing study is providing insights into the lives of children and families during an ongoing pandemic and provide scientific insight for policy makers to make evidence informed decisions and guidance [26,58].

Conclusion
The findings of this study are important for practitioners, policy and decision makers to consider in order to begin to understand the impact and consequences that the drastic but required COVID-19 measures (i.e. lockdown) has had upon children's PA which is a key and vital behaviour for health and development. Key associations have been identified between self-reported PA and the frequency and length of time children went outside of the home. COVID-19 guidelines should factor that many children will not be sufficiently active just in the home environment. Leaving the home for physical activity/ exercise for a minimum of 60 min, and preferably longer each day safely (staying in household bubbles, social distancing, wearing face coverings where necessary) should be actively prioritised and promoted through campaigns and initiatives. Findings should be considered now during the ongoing COVID-19 crisis to support children's PA and short-term health and wellbeing; and, once COVID-19 is under control. Policies and interventions to facilitate 'recovery' after COVID-19 will be required to prevent potential long-term health problems associated with low levels of PA during the pandemic.