Scottish children's physical activity levels: study analysis

Chapter 3 Findings

3.1 Objective Measurement of physical activity

3.1.1 Overall Activity Levels - weekday and weekend activity by gender

When assessing total physical activity, counts per minute ( CPM) were used as a means of exploring gender differences in overall activity levels. During weekdays, total physical activity was significantly higher in boys (674.2 CPM) than girls (627.6 CPM). Although boys had higher total activity across all valid days, this was not statistically significant (two tailed, p=0.09). Moreover, no differences existed on weekends (Table 4).

3.1.2 Time spent Sedentary and in MVPA - weekday and weekend activity by gender

Time spent sedentary and in moderate to vigorous physical activity ( MVPA) was analysed using the mean score derived from all valid days for each individual ( i.e. the mean was calculated from the valid days of an individual and then averaged across the sample). On average, children spent 7.5 hours sedentary each day (Table 4). Although no large difference was evident between boys and girls, girls tended to be more sedentary and the difference seen during weekday sedentary time approached statistical significance (p=0.06). On average, children spent 73 mins per day in MVPA, with boys exhibiting significantly higher levels than girls on both weekdays (81.1 vs. 70.2 mins) and weekend days (67.6 vs. 60.0 mins).

Table 4 - Time spent sedentary and in specific intensities of physical activity by gender and overall

	Girls	Boys	Overall
Counts per Minute ( CPM)	634.0 (607.0 - 661.0)	662.7 (639.0 - 687.0)	648.5 (627.0 - 669.9)
Weekdays **	627.6 (601.9 - 653.3)	674.2 (652.0 - 696.4)	650.6 (630.7 - 670.4)
Weekends	639.0 (586.3 - 691.6)	634.2 (591.1 - 677.3)	637.4 (595.8 - 678.9)
Sedentary (in hours)	7.6 (7.5 - 7.7)	7.4 (7.3 - 7.6)	7.5 (7.4 - 7.6)
Weekdays	7.7 (7.6 - 7.9)	7.6 (7.4 - 7.7)	7.7 (7.5 - 7.8)
Weekends	7.1 (7.0 - 7.3)	7.1 (6.9 - 7.3)	7.1 (7.0 - 7.3)
MVPA (in minutes) ***	67.7 (64.5 - 71.0)	77.5 (74.1 - 81.0)	72.6 (70.0 - 75.3)
Weekdays ***	70.2 (66.7 - 73.7)	81.1 (77.7 - 84.5)	75.7 (72.9 - 78.4)
Weekends *	60.0 (55.4 - 64.7)	67.6 (62.3 - 72.9)	63.8 (60.0 - 67.7)
Bases
Weighted	414	359	773
Unweighted	417	357	774

Statistically significant difference between boys and girls: *p <0.05, **p<0.01, ***p<0.001
Significance testing adjusted for season of measurement, mean wear time, and number of valid days
Figures are Estimated Marginal Means (95% CI)
Figures rounded to 1 decimal place

3.1.2.1 Sedentary time - Area deprivation and Gender

Mean time spent sedentary per quintile of the Scottish Index of Multiple Deprivation ( QSIMD) is summarised in Table 5. Data is further analysed as mean time spent sedentary per individual during weekdays and weekend days per QSIMD. There were no statistically significant differences between quintiles of SIMD, either for all participants or for boys and girls separately. Participants in all quintiles (both boys and girls) spent, on average, greater than 7 hours sedentary per day. Figure 1 shows a slight tendency for those in more deprived areas to spend less time sedentary per day but there were no statistically significant differences found. The pattern was similar for weekday and weekend day analyses, with lower levels across all quintiles during the weekend compared to weekdays.

Table 5 - Sedentary time by QSIMD

Sedentary time in minutes	Girls	Boys	Overall
All days
Most deprived	7.6 (7.1 - 8.0)	7.1 (6.7 - 7.5)	7.3 (7.0 - 7.7)
2	7.7 (7.3 - 8.1)	7.6 (7.3 - 7.9)	7.7 (7.4 - 7.9)
3	7.4 (7.3 - 7.6)	7.5 (7.2 - 7.9)	7.5 (7.3 - 7.7)
4	7.5 (7.4 - 7.7)	7.5 (7.3 - 7.7)	7.5 (7.4 - 7.6)
Least deprived	7.6 (7.5 - 7.7)	7.7 (7.5 - 7.8)	7.6 (7.5 - 7.7)
Weekdays
Most deprived	7.7 (7.3 - 8.2)	7.2 (6.7 - 7.6)	7.4 (7.1 - 7.8)
2	7.8 (7.4 - 8.2)	7.6 (7.3 - 8.0)	7.7 (7.5 - 8.0)
3	7.6 (7.4 - 7.8)	7.6 (7.3 - 8.0)	7.6 (7.4 - 7.8)
4	7.7 (7.5 - 7.9)	7.6 (7.4 - 7.8)	7.7 (7.4 - 7.7)
Least deprived	7.8 (7.7 - 7.9)	7.8 (7.6 - 8.0)	7.8 (7.7 - 7.8)
Weekends
Most deprived	7.1 (6.6 - 7.7)	7.0 (6.5 - 7.5)	7.1 (6.7 - 7.4)
2	7.4 (6.9 - 7.8)	7.3 (7.0 - 7.6)	7.3 (7.0 - 7.6)
3	7.0 (6.7 - 7.3)	7.2 (6.7 - 7.6)	7.1 (6.8 - 7.3)
4	7.1 (6.9 - 7.3)	6.9 (6.6 - 7.2)	7.0 (6.8 - 7.2)
Least deprived	7.1 (6.9 - 7.2)	7.3 (7.1 - 7.5)	7.2 (7.0 - 7.3)
Bases
Weighted	84	81	164
	81	59	139
	84	59	143
	86	80	166
	80	81	161
Unweighted	35	29	64
	55	44	99
	96	73	169
	110	97	207
	121	114	235

Statistically significant difference between quintile levels in 'Overall' 'Boys' and 'Girls': *p <0.05, **p<0.01, ***p<0.001
Bonferroni correction applied
Significance testing adjusted for season of measurement, mean wear time, and number of valid days
Figures are Estimated Marginal Means (95% CI)
Figures rounded to 1 decimal place

Figure 1 - Mean time sedentary per Quintile of Deprivation ( SIMD) - All Days

Error Bars: 95% Cl

3.1.2.2 MVPA - Area deprivation and Gender

Mean time spent in MVPA per quintile of the Scottish Index of Multiple Deprivation ( QSIMD) is summarised in Table 6. Data is further analysed as mean time spent per individual in MVPA during weekdays and weekend days per SIMD quintile. There were no statistically significant differences between quintiles of SIMD, either for all participants or for boys and girls separately. Participants in all quintiles (both boys and girls) engaged in, on average, more than 60 minutes of MVPA per day. Figure 2 shows a tendency for those in more deprived areas to engage in higher level of MVPA per day but there were no significant differences found. The pattern was similar for weekday and weekend day analyses.

Table 6 - MVPA by QSIMD

MVPA in minutes	Girls	Boys	Overall
All days
Most deprived	69.7 (59.6 - 79.7)	80.8 (70.2 - 91.4)	75.2 (67.7 - 82.7)
2	65.8 (57.0 - 74.6)	77.8 (71.6 - 84.0)	71.8 (66.1 - 77.5)
3	71.2 (64.8 - 77.6)	75.2 (68.7 - 81.7)	73.2 (68.7 - 77.8)
4	65.3 (59.7 - 70.9)	77.5 (71.5 - 83.5)	71.4 (67.2 - 75.6)
Least deprived	67.0 (63.3 - 70.7)	76.2 (71.5 - 80.8)	71.6 (68.3 - 74.9)
Weekdays
Most deprived	73.1 (61.4 - 84.8)	83.2 (73.2 - 93.3)	78.2 (70.1 - 86.3)
2	69.1 (60.1 - 77.3)	84.5 (76.7 - 92.2)	76.8 (71.1 - 82.5)
3	73.3 (66.9 - 80.1)	78.7 (72.1 - 85.3)	76.0 (71.4 - 80.6)
4	66.7 (61.9 - 71.7)	79.8 (74.0 - 85.6)	73.3 (69.2 - 77.3)
Least deprived	68.9 (65.2 - 72.5)	80.1 (75.1 - 84.8)	74.4 (71.1 - 77.7)
Weekends
Most deprived	60.1 (46.4 - 73.9)	71.0 (54.2 - 87.9)	65.6 (54.4 - 76.8)
2	56.3 (41.4 - 71.3)	62.6 (56.1 - 69.1)	59.5 (51.4 - 67.5)
3	64.0 (55.3 - 72.8)	65.1 (55.0 - 75.3)	64.6 (58.0 - 71.2)
4	59.6 (50.0 - 69.2)	72.0 (62.6 - 81.2)	65.8 (58.6 - 72.9)
Least deprived	60.7 (55.5 - 65.8)	65.8 (60.1 - 71.6)	63.2 (59.1 - 67.3)
Bases
Weighted	84	81	164
	81	59	139
	84	59	143
	86	80	166
	80	81	161
Unweighted	35	29	64
	55	44	99
	96	73	169
	110	97	207
	121	114	235

No Statistically significant difference between quintile levels in 'Overall' 'Boys' and 'Girls'
Bonferroni correction applied
Significance testing adjusted for season of measurement, mean wear time, and number of valid days
Figures are Estimated Marginal Means (95% CI)
Figures rounded to 1 decimal place

Figure 2 - Mean time in MVPA per Quintile of deprivation rank ( SIMD) - All Days

Error Bars: 95% Cl

3.1.3 Meeting the PA guidelines using 'Threshold' approach

Using a threshold approach to physical activity guidelines adherence, as described in section 3.3.2, only 12% of boys and 11% of girls meet the levels of activity recommended for their age group (that is at least 60 minutes in moderate to vigorous physical activity per day, across valid days (see Table 7 & Figure 3). No statistically significant differences existed between boys and girls.

Table 7 - Meeting the physical activity guidelines: Threshold approach

60 min MVPA per day	Girls	Boys	Total	P value
No	371	317	688
	(89.5%)	(88.4%)	(89.0%)
Yes	44	42	85
	(10.5%)	(11.6%)	(11.0%)
Total	415	359	773	0.769
	(100%)	(100%)	(100%)
Bases
Weighted	415	359	773
Unweighted	417	357	774

Figure 3 - Meeting the guidelines according to threshold approach

3.1.3.1 Guideline adherence by area deprivation - Threshold approach

There was no significant pattern by area deprivation for a higher likelihood of meeting the guidelines when using the 'threshold' approach to guideline adherence, as can be seen in Table 8 and Figure 4. A higher proportion of those in the most deprived areas met the guidelines, but the difference between quintiles was not statistically significant using a Pearson chi ^[2] test.

Table 8 - Meeting the physical activity guidelines: Threshold approach by QSIMD

	QSIMD					P value
Meets guidelines	1 (Most deprived)	2	3	4	5 (Least deprived)
No	135	124	130	150	150
	(82.0%)	(89.2%)	(90.8%)	(90.1%)	(93.1%)
Yes	30	15	13	16	11	0.30
	(18.0%)	(10.8%)	(9.2%)	(9.9%)	(6.9%)
Bases
Weighted	164	139	143	166	161
Unweighted	64	99	169	207	235

Figure 4 - Meeting the guidelines by Quintile of area deprivation ( SIMD) - Threshold approach

3.1.4 Meeting the PA guidelines using 'Averaging' approach

Another approach to look at adherence to guidelines is to obtain an average across all valid days with more than 1 min MVPA. That is, minutes spent in MVPA were averaged across valid days and participants with at least 60 mins per day were marked as meeting the guidelines. As shown in Table 9 and Figure 5, this approach substantially increases the proportion of children meeting the guidelines. Using this method, 60% of the children overall meet the recommended daily guidelines, and boys are significantly more likely to meet the guidelines than girls (69.4% vs. 52.4%, p <0.001).

Table 9 - Meeting the physical activity guidelines: Average approach

Meets guidelines	Girls	Boys	Overall
No	197	110	307
	(47.6%)	(30.6%)	(39.7%)
Yes *	217	249	466
	(52.4%)	(69.4%)	(60.3%)
Total	414	359	773
	(100%)	(100%)	(100%)
Bases
Weighted	415	359	773
Unweighted	417	357	774

* Statistically significant difference between girls and boys p<0.001

Figure 5 - Proportion meeting UK Chief Medical Officers PA guidelines using

Averaging method

3.1.4.1 Guideline adherence by area deprivation - Averaging approach

There was no clear pattern by area deprivation in meeting the guidelines in one quintile or another when using the 'averaging' approach to guideline adherence, as can be seen in Table 10 and Figure 6. The difference between quintiles was not statistically significant using a Pearson chi ^[2] test.

Table 10 - Meeting the physical activity guidelines: Average approach by QSIMD

	QSIMD					P value
Meets guidelines	1 (Most deprived)	2	3	4	5 (Least deprived)
No	64	48	60	68	68
	(38.7%)	(34.3%)	(42.1%)	(40.7%)	(42.3%)
Yes	100	91	83	98	93	0.811
	(61.3%)	(65.7%)	(57.9%)	(59.3%)	(57.7%)
Bases
Weighted	164	139	143	166	161
Unweighted	64	99	169	207	235

Figure 6 - Meeting the guidelines by Quintile of area deprivation ( SIMD) - Averaging approach

3.2 Self-reported physical activity

The first part of this section describes the ability of the PAQ-C to accurately measure physical activity levels in this age group using the accelerometer data as the reference (or criterion from which we can compare). This is followed by the gender and socio-economic analyses by PAQ-C score.

3.2.1 Validation of PAQ-C

As previously noted, the PAQ-C did not provide information that permitted direct comparison with the accelerometer output - the questionnaire does not provide time spent in MVPA or counts per minute. This meant that we could not measure the direct agreement between both measures ( i.e. do both measures produce identical outcomes). However, we were able to assess the ability of the PAQ-C to measure physical activity by assessing the monotonic/ linear relationship between the accelerometer (which we termed the criterion) and the questionnaire.

We tested the relationship between the two measures by way of the Pearson correlation coefficient (r), and the Spearman's rank order correlation coefficients (rho). The first analysis tested for a linear relationship between both methods ( i.e. the change in one variable being associated with a proportional change in the other variable); whereas the second analysis tested for a monotonic relationship ( i.e. both measures increase/decrease concurrently but not necessarily at the same rate) using ranked scores rather than raw data. The scatter plots that visually demonstrate these analyses can be seen in Appendix C; the scores presented in the following sections describe the strength of the relationship (0 = no relationship, +/-1 = perfect positive or negative monotonic/linear relationship).

3.2.1.1 Accelerometer and PAQ-C - MVPA

All graphs demonstrated a positive relationship between PAQ-C scores and time spent in MVPA (r = 0.39, rho=0.39). That is, as one of the measures increases, so does the other. This remains the case for both boys and girls when analysed separately, (girls: r=0.38, rho=0.34; boys: r=0.37, rho=0.39). The questionnaire was designed to measure activities within the MVPA range, and therefore should demonstrate a reasonable relationship with accelerometry derived MVPA; our results supported this. When we removed MVPA from the analysis and substituted three other physical activity outcomes (light, moderate, and vigorous activity), the correlations were weaker: light (r=0.14), moderate (r=0.30) and vigorous (r=0.37) and all were all lower than the r=0.39 reported for MVPA, indicating a stronger relationship between PAQ-C scores and accelerometry derived MVPA.

3.2.1.2 Accelerometer and PAQ-C - counts per minute ( CPM)

When analysing the relationship between PAQ-C scores and CPM (considered to represent 'overall' physical activity), a similar relationship, as with MVPA, was evident (r=0.36, rho=0.38) overall, and for boys (r=0.35, rho=0.36) and girls (r=0.37, rho=0.39). There was a strong positive correlation between MVPA and CPM (r=0.92) so the similarity between MVPA and CPM is expected. The 'counts per minute construct' integrates all activity data that has been recorded and includes physical activity of all intensities. It is therefore a useful measure to demonstrate 'total' PA of an individual. The PAQ-C questionnaire, in addition to MVPA, will inevitably capture time spent in light activity, and as such, the CPM construct was included to assess its relationship with PAQ-C scores.

3.2.1.3 PAQ-C scores between those who do and do not meet the CMO guidelines - Averaging method

We investigated whether PAQ-C detected differences in scores dependent on whether participants met, or did not meet the PA guidelines (using the averaging approach). As can be seen in Table 11 and Figure 7, PAQ-C scores were significantly higher in those participants who met the guidelines compared to those who didn't, further demonstrating the questionnaire's ability to discriminate between MVPA activity levels.

Table 11 PAQ-C score by those who meet and do not meet the physical activity guidelines: Average approach

Average MVPA per day	Girls	Boys	Overall
Meets Guidelines	3.23 (3.14-3.32)	3.30 (3.19-3.41)	3.26 (3.18-3.33)
Does not Meet Guidelines	2.82 (2.71-2.94)	2.90 (2.74-3.05)	2.87 (2.78-2.96)
Bases
Weighted	361	317	678
Unweighted	375	321	696
P value	<0.001	<0.001	<0.001

Statistically significant difference between those who meet and do not meet the CMO guidelines - overall, boys, and girls Significance testing adjusted for season of measurement. Figures are Estimated Marginal Means (95% CI). Figures are rounded to two decimal places

Figure 7 - PAQ-C scores as a function of guideline adherence grouping

Error Bars: 95% Cl

3.2.2 PAQ-C scores - Overall, by Gender, and by SIMD

The previous section provides support on the use of the PAQ-C as an acceptable method to capture physical activity in children of this age group. We then used the questionnaire to investigate its ability to capture potential differences between boys and girls, and across quintiles of deprivation.

Boys scored higher on the PAQ-C than girls (3.19 vs. 3.05); this was statistically significant. There were no statistically significant differences for the mean PAQ-C scores between quintiles of SIMD (Table 12 and Figure 8). Both results were similar to those demonstrated by the accelerometers.

Table 12 PAQ-C score by QSIMD and Gender

	Girls	Boys	Overall
PAQ-C scores *	3.05 (2.99 - 3.12)	3.19 (3.10 - 3.23)	3.12 (3.06 - 3.19)
QSIMD
Most deprived	3.07 (2.83 - 3.32)	3.16 (2.83 - 3.49)	3.12 (2.88 - 3.35)
2	3.04 (2.89 - 3.18)	3.27 (2.97 - 3.55)	3.15 (2.98 - 3.33)
3	3.12 (2.94 - 3.30)	3.08 (2.92 - 3.22)	3.10 (2.99 - 3.21)
4	3.00 (2.84 - 3.10)	3.23 (3.09 - 3.36)	3.10 (3.00 - 3.20)
Least deprived	3.07 (3.00 - 3.20)	3.24 (3.13 - 3.34)	3.16 (3.08 - 3.23)
Bases (Gender)
Weighted	361	317	678
Unweighted	375	321	696
Bases (Gender* SIMD)
Weighted	64	74	138
	77	48	124
	72	52	123
	77	72	149
	72	71	143
Unweighted	29	23	52
	52	39	91
	86	67	153
	99	90	189
	110	100	210

Statistically significant difference between boys and girls: *p <0.05, Bonferroni correction applied,
Significance testing adjusted for season of measurement. Figures are Estimated Marginal Means (95% CI). Figures are rounded to nearest two decimal places

Figure 8 - PAQ-C scores by quintile of deprivation ( SIMD)

Error Bars: 95% Cl