Scottish children's physical activity levels: study analysis

Chapter 4 Discussion and Conclusions

The purpose of this report was to investigate the physical activity levels of Scottish children; using objective and self-reported PA data collected from a cohort of children drawn from the Growing Up in Scotland study. A key objective of this report was to explore the gender and socio-economic patterning of activity and assess whether these patterns differed by the physical activity measurement assessment used.

The project successfully collected objectively measured and self-reported physical activity levels in a large, nationally representative, sample of Scottish 10-11 year old children employing a study protocol that was based on a postal method: contact with participants was either by phone, or traditional paper-based materials delivered by post. The overall response rate was 40%, with 36% of the possible sample providing sufficient data to be included in this report. The data presented are likely to be broadly representative of the children in Scotland who will be making the transition into secondary schooling and the availability of this data will prove to be a valuable longitudinal resource to understand health and health behaviours as this cohort of children age.

4.1 Overall physical activity levels by accelerometry and self-report

Overall physical activity levels were measured using the standardised unit of counts per minute ( CPM). For this age group, this value was 648 CPM. This is consistent with accelerometry data from over 27,500 children using the International Children's accelerometry database ( ICAD) which holds data from 20 studies in 10 countries (Cooper et al., 2015). Individual studies from the ICAD database can be seen in Table 13, alongside the setting, age group, and mean CPM. These studies include UK examples from different parts of England, such as East Anglia ( SPEEDY; Corder, van Sluijs, Ekelund, Jones, & Griffin, 2010); Bristol ( PEACH; Page, Cooper, Griew, Davis, & Hillsdon, 2009) ( ALSPAC; Riddoch et al., 2007); European studies such as the European Youth Heart Study ( EYHS; Riddoch et al., 2004); and studies from the USA ( NHANES; Troiano et al., 2008).

Table 13 - Descriptive information from large UK and international physical activity studies

SPACES: Studying Physical Activity in Children's Environments across Scotland
SPEEDY: Sport, Physical activity and Eating behaviour: Environmental Determinants in Young people
PEACH: Personal and Environmental Associations with Children's Health
NHANES: National Health and Nutrition Examination Survey
ALSPAC: Avon Longitudinal Study of Parents and Children EYHS: European Youth Heart Study
MCS: Millennium Cohort Study

The consistency across studies with regard to the main physical outcome of CPM suggests that the use of accelerometry in Scottish 10-11 year old children is a valid method of measuring physical activity.

Although total physical activity takes into consideration all activity, all levels of intensity, including sedentary time, MVPA is very specific to those periods where children are being active at a level that is optimal for health and well-being. On average, children spent 73 mins per day in MVPA (76 mins per day weekday and 64 mins per day weekend. The comparison of the results within this report with other studies is difficult due to the inconsistent approaches used to identify intensity classifications. However, the MCS study (Griffiths, et al., 2013), which used similar methods but with younger children (7 year olds), found that Scottish children spent 61.8 mins in MVPA, which was slightly more than Wales (61.6 mins), and England (60.6 mins) and around 9% higher than children in Northern Ireland (56.6 mins). Cooper and colleagues (2015) presented MVPA data as a relative measure ( i.e. percentage of wear time that is spent in MVPA) across 7 countries (9-10 year olds) and found a range of between 5-10%. With results from this report suggesting that Scottish 10-11 year old children spend 8% of their wear-time in MVPA, the results are internationally comparable for a similar age group.

With children only spending 8% of their waking day in MVPA, they must spend the rest of their time in either light activity or sedentary activity. Indeed, we found that Scottish children are sedentary for 7.5 hours, higher than that observed in the Scottish sample of the MCS where 7-year old children spent 6.4 hours sedentary (Griffiths, et al., 2013). Cooper and colleagues (2015) have identified a significant positive age-related trend where sedentary time increases by 20-25% from children aged 5-6 through to 17-18 years old. The higher sedentary time presented in this report could be a reflection of this age related increase. Evidence from other, similarly aged children, provides support for our findings. For example, the ALSPAC study reported a mean sedentary time of just over 7 hours in 9 year old boys and girls, and the SPEEDY study reported a mean time of approximately 7.6 hours in 10 year old children. These figures, therefore, compare equally across the UK, provide support for the measurement method in Scottish children, and provide useful data on current levels of sedentary behaviour among Scottish 10-11 year old children.

In addition to providing data on overall physical activity levels, the objectively measured approach provided an opportunity to investigate the proportion of children who met the Chief Medical Officers physical activity guidelines. As outlined, the method used had a large impact on the outcome. When using the threshold approach, where all valid days per person were required to be over 60 minutes of MVPA on valid days before being classified as meeting the guidelines, only 11% of children were classified as meeting the guidelines. Another way of calculating the proportional outcome is by removing the necessity to accumulate 60 minutes of MVPA every day. Instead, if children have an average of 60 minutes or more across their valid days then this would constitute adherence to the guidelines. In situations where participants may accumulate 55 or 58 minutes on any given day, this approach seems to provide a fairer representation. When taking this approach, the proportion of children who met the guidelines increased from 11% to 60%. These figures are however lower than those presented in the SHeS for children aged 8-10 (81% vs. 83%, boys and girls respectively). Objective measures have, in recent years, demonstrated that the proportions of children who meet the physical activity guidelines are lower than those reported in national surveys such as the SHeS (Cooper, et al., 2015). At this point it should be noted that both approaches are completely different, and therefore cannot be compared like for like; the accelerometer used in this report measures body acceleration, whereas

SHeS measures physical activity behaviours as recalled by the parent of child. However, with regards to the SHeS, a number of factors can provide possible explanations for the apparent discrepancy: i) parents act as a proxy and report on their child's behalf - this may introduce an element of reporting error due to parents recalling the physical activity behaviours of their children; ii) potentially most importantly, the SHeS automatically considers any activity as at least MVPA - this will introduce misclassification error due to light activity being incorrectly classified as MVPA. Results from this report suggest that children spend 4.2 hours in light activity so it could be argued that more children would be categorised as meeting the guidelines using an objective method if light activity were to be included in the classification. Similarly, if light activity were to be removed from the calculations in the SHeS, there is a considerable possibility that the proportion of children who meet the physical activity guidelines would decrease substantially. The Health Behaviour in School-Aged Children ( HBSC) survey is a national source of self-reported physical activity data in Scotland that aims to remove light activity from its calculations. Instead of including all activity as MVPA, it specifically asks about activity that is at least of moderate intensity. Results are still higher (30% of boys and 21% of girls; 11 years old) than the accelerometry data presented in this report but lower than those reported by the SHeS. There is a pressing need to investigate both approaches and the findings of this report suggest that further work to reduce the discrepancy between measures is warranted.

Although self-reported approaches have their drawbacks, we introduced a self-report method of measurement within this study for very good reasons. Conducting large scale, objective measures of physical activity across a whole country are costly and involve a number of logistical issues in administrating such studies and can involve an increased burden on participants compared to wholly self-reported methods. If an appropriate self-report questionnaire is chosen, particularly one which takes cognitive ability and respondents' burden into account, self-reported methods should be able to return valuable data, i.e. to differentiate physical activity patterns in gender or across socioeconomic groups. Initially, we assessed the ability of the PAQ-C questionnaire to capture physical activity levels in children of this age group. Overall, it performed moderately well. Validity coefficients were acceptable and this type of self-reported questionnaire was able to accurately record general levels of physical activity. This was further confirmed by the questionnaire's ability to discriminate those who met and those who did not meet the CMO physical activity guidelines - the mean PAQ-C scores were significantly lower in those who did not meet the guidelines. Recent data from England (Voss, et al., 2013) has found mean PAQ-C scores of approximately 3.10 in 10 year olds, and 2.90 in 11 year olds. Another study in England found a mean score of 3.49 in children aged 9-11 years. Internationally, Spanish data in 10-11 year old children have demonstrated a mean score of 3.24 (Benitez-Porres et al., 2016). All scores reported are similar to that found in this study (mean PAQ-C score of 3.10) which suggest comparability nationally and internationally. Currently, there are no uniform and meaningful thresholds that can be applied to differentiate youth based on their activity levels, although, internationally, PAQ-C scores of less than 2 have been considered low activity, 2-3 as moderate activity, and more than 3 as high activity in children between 8 and 12 years old (Chen, Lee, Chiu, & Jeng, 2008). The next stage for the PAQ-C questionnaire could be to try to develop activity thresholds to categorise youth in Scotland.

4.1.1 Gender differences in physical activity by accelerometry and self-report

Results from this report suggest that boys are more active than girls. This pattern was present across most of the outcome variables and measurement approaches. Boys were more active than girls on weekdays using CPM as an outcome (674 CPM vs 628 CPM), MVPA on both weekdays (81.1 vs. 70.2 mins) and weekend days (67.6 vs. 60.0 mins), and the PAQ-C also found significant differences between boys and girls (3.19 vs. 3.05, boys and girls respectively). Regarding the proportion of children meeting the CMO physical activity guidelines, boys were also more likely to meet these guidelines (69%) compared to girls (52%) when using the averaging approach.

Interestingly, gender differences were not observed for overall physical activity at the weekend (634 CPM vs. 639 CPM, boys and girls respectively), and further work is required to explore potential reasons for differences in weekday and weekend activities.

The Scottish Health Survey reports little evidence of gender differences in physical activity guideline adherence for either the 8-10 years or 11-12 years age groups, with differences not emerging until the 13-15 year old age group. Our data suggests this difference exists earlier, within the 10-11 year old age group, and this finding has also been noted in international data using objectively measured PA (Cooper, et al., 2015). Findings that gender differences emerge earlier than previously suggested in studies of Scottish children suggest that there is a need to develop policies to reduce the gender gap at an earlier age.

4.1.2 Social patterning of physical activity using accelerometry

Both methods of physical activity measurement found little significant observable pattern/ differences across the quintiles of SIMD in either physical activity levels or adherence to the CMO guidelines. This is similar to mothers' reports of the children's physical activity at age 6 (of the same children) which found no clear social gradient in the patterning of children's physical activity, and that reported by the Millennium Cohort Study (Griffiths, et al., 2013). We did however find that MVPA was slightly higher among children from the most deprived quintile compared to other quintiles. Further work is required to investigate the reasons for this pattern but one explanation may be that of active travel to school. It has been suggested that active travel and deprivation follow a 'U shaped' distribution, where active travel is high in the most deprived and the least deprived, and lower in between ( GCPH, 2011). Transport Scotland produced a report in 2014 that suggested 59% of children in the lowest income category and 62% of the most deprived quintile of SIMD reported their usual mode of transport to school to be an active one (walk/bicycle). Both figures were higher than any other income category or deprivation quintile [20] .

There are a small number of methodological or analytical limitations in our study, largely based around known limitations of accelerometers or the analysis of the data extracted. Waist mounted devices are typically poor at recording the acceleration associated with cycling or upper body dominant activities ( e.g. throwing, lifting). In addition, the devices were removed for water-based activities and contact sports so there is the possibility that we have underestimated these activities and their associated activity level. Furthermore, the threshold cut points used for this study, although shown to be accurate and reliable in this age group (Trost, 2007), can misclassify motionless standing as sedentary due to the way 'counts' are used to classify intensity; motionless standing will record low 'counts' due to limited movement. However, this behaviour is not sedentary (defined as "any waking behaviour characterised by an energy expenditure ≤ 1.5 metabolic equivalents ( METs) while in a sitting or reclining posture" (Owen, Healy, Matthews, & Dunstan, 2010)). Finally, as previously noted, using a different threshold classification will influence the resulting physical activity outcomes. As such, MVPA may have been over or underestimated depending on the threshold used. Importantly, the threshold used in all analyses is one that has demonstrated accuracy across all levels of intensity and is being used internationally with similar aged children to those in this report (Cooper, et al., 2015; Trost, 2007).

4.2 Conclusions

Valid and reliable measurement methods are crucial for the ongoing accurate assessment of physical activity; the outcomes of which are of significant importance to the development and evaluation of various initiatives, strategies, and policies in Scotland. Accelerometry is viewed as one of the most valid methods we have at our disposal for measuring actual levels of physical activity, however, the self-report method used within this report was able to identify similar patterns and differences, or lack thereof, across gender and area based categories of deprivation. This suggests that the PAQ-C could be used in future surveys of the GUS children, particularly where identifying potential differences by gender or area deprivation is important.

The physical activity levels of Scottish 10-11 year olds tell a positive story if framed in average time spent in MVPA per day. However, depending on the definition used in investigating the proportion of children who meet PA guidelines, either 11% or 60% of Scottish 10-11 year olds meet the current UK recommendations. These figures highlight the continuing issue surrounding accurate physical activity measurement, and the need for careful consideration when using this type of data.