Growing up in Scotland: overweight and obesity at age 10

9. Appendix

9.1. Summary of the multivariable analysis

A key point of interest in this report is to examine what features of children's lives are associated with them being overweight or obese at age 10 and which are associated with movement into and out of overweight/obesity between the ages of 6 and 10. When examining such associations it is important to consider the influence of underlying factors and the inter-relationships between such factors and other things we are interested in.

For example, as well as being associated with BMI status, the number of hours per week children spend on screens is likely to be related to household/parental rules about such activities as well as household income level. Simple analysis may identify an association between duration of screen time and children's BMI status – for example, that children who spend greater amounts of time using screens have higher levels overweight/obesity than children who have lower screen time. However, this association may be occurring simply because of an underlying relationship between screen time and household income. Thus, rather than screen time being associated with children's BMI in its own right, the relationship found in the analysis may be due to the influence of other factors, particularly social background factors such as household income.

To 'control' for the influence of other factors (e.g. household income) multivariable regression analysis was used. This form of analysis allows the examination of the relationships between an outcome variable (e.g. overweight/obesity) and multiple explanatory variables (e.g. weekly screen time, household income) whilst controlling for the inter-relationships between each of the explanatory variables. This means it is possible to identify whether there is an association between any single explanatory variable and the outcome variable also when other relevant variables have been controlled for. For example, to look at whether there is a relationship between screen time and overweight/obesity that does not simply occur because screen time and household income are related.

To examine whether the various demographic, social background and other factors selected were associated with children's overweight/obesity at age 10, for each factor, bivariate analysis was first undertaken to look at the relationship between the measures and each of the following outcomes: overweight/obesity at age 10; moving into overweight between ages 6 and 10; moving out of overweight between ages 6 and 10.

Where the bivariate analysis showed a statistically significant association between the factor and the outcome in question, that factor was included in a multivariable regression model.

A different approach to the regression modelling was used for examining factors associated with overweight/obesity at age 10 and for examining factors associated with moving into or out of overweight between ages 6 and 10.

To examine factors associated with overweight/obesity at age 10, the regression analysis was conducted in a series of steps:

• First, a regression model was run which included only demographic (sex, ethnicity) and socio-economic variables (household income, parental educational qualifications, social class and area deprivation)
• Second, weight status at age 6 was added to a model containing only significant variables from amongst the demographic and socio-economic variables
• Third, a separate model was run including those other factors found to be statistically significantly associated with weight status in the cross-sectional analysis conducted earlier in section 5 (birth weight, maternal BMI, frequency at which breakfast eaten, screen time, TV in bedroom, sleep).
• Finally, a combined model was then run including the statistically significant variables from the first three steps.

In all models, the outcome variable was whether or not the child was overweight or obese at age 10.

For analysis of factors associated with moving into or out of overweight between ages 6 and 10, only a small number of variables showed a statistically significant relationship with the outcome variable in the initial bivariate analysis. As such, all statistically significant variables from this initial step were entered into a single model. In one model the outcome variable indicated whether the child had moved into overweight/obesity between age 6 and age 10. In the other model the outcome variable indicated whether the child had moved out of overweight/obesity between age 6 and age 10. In both cases, the reference category was all other children.

9.2. Interpreting the tables

The results for the binary logistic regression analysis are presented as odds ratios, all of which have a significance value and 95% confidence intervals attached. Logistic regression compares the odds of a reference category (shown in the tables) with that of the other categories. An odds ratio of greater than one indicates that the group in question is more likely to demonstrate this characteristic than is the chosen reference category. An odds ratio of less than one means they are less likely. For example, in the 'odds ratio' column of Table A:1 which contains the results of the first step of the regression analysis seeking to identify factors related to a child being overweight or obese at age 10, the category of 'No qualification' returns an odds ratio of 1.82. This indicates that the odds of children whose parents have no qualifications being overweight or obese at age 10 are 1.82 times greater than they are for children whose parents are degree educated (the reference category) whilst holding all other variables in the model constant.

Note that an odds ratio cannot be interpreted in the same way as a co-efficient. An odds ratio of 2 does not mean 'two times as likely' but instead means 'the odds are two times higher'. To understand an odds ratio we first need to describe the meaning of odds. The definition of odds is similar but significantly different to that of probability. This is best explained in the form of an example. If 200 individuals out of a population of 1000 experienced persistent poverty, the probability (p) of experiencing persistent poverty is 200/1000, thus p=0.2. The probability of not experiencing persistent poverty is therefore 1-p = 0.8. The odds of experiencing persistent poverty are calculated as the quotient of these two mutually exclusive events. So, the odds in favour of experiencing persistent poverty to not experiencing persistent poverty, is therefore 0.2/0.8=0.25. Suppose that 150 out of 300 people living in social rented housing experience persistent poverty compared to 50 out of 150 who live in owner occupied housing. The odds of a person living in social rented housing of experiencing persistent poverty are 0.5/0.5=1.0. The odds of a person living in owner occupied housing of experiencing persistent poverty is 0.33/0.66=0.5. The odds ratio of experiencing persistent poverty is the ratio of these odds, 1.0/0.5=2.0. Thus the odds of experiencing persistent poverty are twice as high among people who live in social rented housing (compared to people who live in owner occupied housing – the 'reference category'). This is not the same as being 'twice as likely' to experience the outcome.

Categories which have a p-value greater than 0.05 are not considered to be statistically significant. However, cases where the p-value is only a little beyond this (e.g. 0.06) have been reported. These can be indicative of other variables which are having some impact on the model and if a lower significance threshold was set (e.g. of 0.10/90% rather than 0.05/95%) would be considered statistically significant.

9.3. Results

9.3.1. Factors associated with overweight or obesity at age 10

Appendix table A:1 Demographic and socio-economic variables only

	n (weighted)	p-value	Odds ratio	95% confidence interval
				Lower	Upper
Child's sex (ref: Female)	1419	.951
Male	1331		1.01	0.83	1.22
Child's ethnicity (ref: White)	2635	.239
Other	115		0.75	0.47	1.21
Household equivalised income quintile (ref: Top Quintile (>=£39,216))	409	.613
Missing	182		0.78	0.50	1.20
1 Bottom Quintile (<£13,450)	699		0.91	0.64	1.31
2nd Quintile (>=£13,451 <£22,827)	549		0.78	0.58	1.06
3rd Quintile (>=£22,827< £29,375)	502		0.98	0.76	1.27
4th Quintile (>=£29,375< £39,216)	409		0.88	0.67	1.17
Highest Education level of Respondent and Partner (ref: Degree or equivalent)	1056	.000
No qualifications	155		1.82	1.13	2.94
Lower level Standard Grades and Vocational qualifications / other qualifications	128		2.24	1.27	3.94
Upper level Standard Grades and Intermediate Vocational qualifications	499		1.40	1.02	1.91
Higher grades and Upper level vocational qualifications	913		1.56	1.27	1.92
Household NS-SEC (ref: Managerial and professional)	1342	.589
Intermediate occupations	389		1.03	0.76	1.38
Small employers and own account workers	225		1.35	0.93	1.96
Lower supervisory and technical occupations	193		1.08	0.72	1.60
5.00 occupations	602		1.04	0.73	1.49
Area deprivation – SIMD 2012 quintiles (ref: Least deprived .936 - 7.936)	570	.015
2 7.938 - 13.674	582		1.17	0.89	1.53
3 13.674 - 20.970	536		1.41	1.05	1.90
4 20.976 - 33.807	498		1.57	1.18	2.09
5 Most deprived 33.817 - 89.894	564		1.51	1.08	2.12

Appendix table A:2 Age 6 BMI status and socio-economic variables only

	n (weighted)	p-value	Odds ratio	95% confidence interval
				Lower	Upper
Child's BMI at age 6 (ref: underweight/healthy weight (< 85th percentile))	2010	<.001
Missing	102		3.04	1.86	4.99
Overweight (85th to < 95th percentile)	354		8.31	6.37	10.84
Obese (95th percentile or above)	285		44.69	28.93	69.04
Highest Education level of Respondent and Partner (ref: Degree or equivalent)	1056	.038
No qualifications	155		1.632	0.909	2.930
Lower level Standard Grades and Vocational qualifications / other qualifications	128		2.071	1.059	4.050
Upper level Standard Grades and Intermediate Vocational qualifications	499		1.352	0.946	1.931
Higher grades and Upper level vocational qualifications	913		1.314	1.036	1.667
Area deprivation – SIMD 2012 quintiles (ref: Least deprived .936 - 7.936)	570	.035
2 7.938 - 13.674	582		1.128	0.813	1.565
3 13.674 - 20.970	536		1.362	0.990	1.874
4 20.976 - 33.807	498		1.461	1.024	2.084
5 Most deprived 33.817 - 89.894	564		1.585	1.125	2.232

Appendix table A:3 Other factors associated with overweight and obesity

	n (weighted)	p-value	Odds ratio	95% confidence interval
				Lower	Upper
Child's birth weight in kilograms (ref: 3 - < 3.5 kg)	990	.019
< 2.5 kg (low birth weight)	180		0.63	0.40	0.99
2.5 - <3kg	408		1.00	0.72	1.37
3.5 - <4 kg	820		1.14	0.91	1.45
4 kg or above	352		1.48	1.04	2.12
Mother's BMI (ref: less than 25)	1015	<.001
Missing	420		2.11	1.51	2.95
25 to less than 30	717		1.38	1.06	1.78
30 and over	598		2.77	2.12	3.63
How often does child eat breakfast (ref: always)	2223	.043
Never	66		0.89	0.42	1.86
Occasionally	242		1.48	1.09	2.00
Quite often	219		1.38	0.93	2.03
Average weekly screen time (ref: less than 14 hours)	788	<.001
Missing	47		3.38	1.56	7.29
14 - <21 hours	971		1.34	1.09	1.64
21 - <28 hours	578		1.75	1.35	2.28
28 hours or more	367		1.67	1.22	2.28
Child has TV in bedroom (ref: No)	934	<.001
Yes	1816		1.40	1.18	1.66
Child gets recommended amount of sleep (ref: Yes for at least two age points)	1862	.447
Not for at least two age points	888		1.08	.88	1.33

Appendix table A:4 Combined model with demographic, socio-economic and other factors

	n (weighted)	p-value	Odds ratio	95% confidence interval
				Lower	Upper
Child's BMI at age 6 (ref: underweight/healthy weight (< 85th percentile))	2010	< .001
Missing	102		2.609	1.570	4.336
Overweight (85th to < 95th percentile)	354		8.099	6.071	10.804
Obese (95th percentile or above)	285		43.231	28.288	66.068
Highest Education level of Respondent and Partner (ref: Degree or equivalent)	1056	.636
No qualifications	155		1.209	0.678	2.156
Lower level Standard Grades and Vocational qualifications / other qualifications	128		1.691	0.802	3.566
Upper level Standard Grades and Intermediate Vocational qualifications	499		1.097	0.757	1.589
Higher grades and Upper level vocational qualifications	913		1.104	0.863	1.413
SIMD 2012 quintiles (ref: 1 Least deprived .936 - 7.936)	570	.868
2 7.938 - 13.674	582		1.016	0.729	1.418
3 13.674 - 20.970	536		1.110	0.807	1.525
4 20.976 - 33.807	498		1.189	0.817	1.730
5 Most deprived 33.817 - 89.894	564		1.112	0.757	1.632
Child's birth weight in kilograms (ref: 3 - < 3.5 kg)	990	.600
< 2.5 kg (low birth weight)	180		0.845	.499	1.431
2.5 - <3kg	408		1.152	.790	1.681
3.5 - <4 kg	820		1.173	.899	1.529
4 kg or above	352		1.260	.862	1.840
Mother's BMI (ref: less than 25)	1015	<.001
Missing	420		1.647	1.108	2.449
25 to less than 30 (Overweight)	717		1.240	0.947	1.625
30 and over (Obese)	598		2.297	1.650	3.196
How often does child eat breakfast (ref: always)	2223	.046
Never	66		0.956	.428	2.138
Occasionally	242		1.533	1.069	2.198
Quite often	219		1.552	.976	2.469
Average weekly screen time (ref: less than 14 hours)	788	<.001
Missing	47		4.088	1.779	9.394
14 - <21 hours	971		1.329	1.030	1.715
21 - <28 hours	578		1.638	1.258	2.133
28 hours or more	367		1.661	1.182	2.335
Child has TV in bedroom (ref: No)	934	.017
Yes	1816		1.303	1.043	1.629

9.3.2. Findings: Factors associated with a move into or out of overweight or obesity between ages 6 and 10

Appendix table A:5 Factors predicting a move into overweight or obesity between ages 6 and 10

	n (weighted)	p-value	Odds ratio	95% confidence interval
				Lower	Upper
SIMD 2012 quintiles (ref: Least deprived .936 - 7.936)	407	.473
2 7.938 - 13.674	432		1.37	0.95	1.97
3 13.674 - 20.970	365		1.08	0.74	1.56
4 20.976 - 33.807	362		1.30	0.85	1.99
5 Most deprived 33.817 - 89.894	445		1.12	0.75	1.69
Mother's BMI (ref: less than 25)	826	<.001
Missing	235		1.94	1.17	3.21
25 to less than 30	554		1.34	0.98	1.82
30 and over	395		2.51	1.75	3.61
How often does child eat breakfast (ref: always)	1627	.063
Never	50		0.87	0.35	2.18
Occasionally	172		1.58	1.03	2.43
Quite often	160		1.68	1.02	2.75
Average weekly screen time (ref: less than 14 hours)	607	.001
Missing	36		4.74	1.97	11.40
14 - <21 hours	712		1.54	1.10	2.14
21 - <28 hours	400		1.80	1.27	2.56
28 hours or more	254		1.77	1.12	2.81
Child has TV in bedroom (ref: No)	720	.057
Yes	1289		1.28	0.99	1.66
Child gets recommended amount of sleep (ref: Yes for at least two age points)	1370	.314
Not for at least two age points	639		1.14	0.88	1.47

Appendix table A:6 Factors predicting a move out of overweight or obesity between ages 6 and 10

	n (weighted)	p-value	Odds ratio	95% confidence interval
				Lower	Upper
Household equivalied income quintile (ref: Top Quintile (>=£39,216))	83	.040
Missing	40		1.42	0.53	3.80
Bottom Quintile (<£13,450)	186		1.06	0.46	2.47
2nd Quintile (>=£13,451 <£22,827)	126		2.42	1.27	4.63
3rd Quintile (>=£22,827< £29,375)	120		1.38	0.71	2.67
4th Quintile (>=£29,375< £39,216)	81		1.26	0.61	2.60
Highest Education level of Respondent and Partner (ref: Degree or equivalent)	181	.355
No qualifications	42		3.27	0.95	11.22
Lower level Standard Grades and Vocational qualifications / other qualifications	42		0.54	0.11	2.68
Upper level Standard Grades and Intermediate Vocational qualifications	132		0.96	0.53	1.74
Higher grades and Upper level vocational qualifications	239		1.13	0.66	1.92
Household NS-SEC (ref: Managerial and professional)	287	.249
Intermediate occupations	89		0.87	0.52	1.46
Small employers and own account workers	64		0.36	0.13	0.98
Lower supervisory and technical occupations	41		0.63	0.24	1.69
Semi-routine and routine occupations	154		0.52	0.19	1.40
SIMD 2012 quintiles (ref: Least deprived .936 - 7.936)	113	.086
2 7.938 - 13.674	123		0.55	0.30	1.01
3 13.674 - 20.970	117		0.66	0.34	1.31
4 20.976 - 33.807	146		0.66	0.29	1.50
5 Most deprived 33.817 - 89.894	137		0.30	0.12	0.70
Child's birth weight in kilograms (ref: 3 - < 3.5 kg)	218	.685
< 2.5 kg (low birth weight)	21		1.34	0.41	4.36
2.5 - <3kg	69		0.82	0.37	1.83
3.5 - <4 kg	207		1.16	0.75	1.80
4 kg or above	122		0.78	0.41	1.45
Child frequency of eating unhealthy snacks (ref: Up to once a day)	141	.005
More than once a day, up to twice a day	227		0.43	0.24	0.75
More than twice a day, up to three times a day	130		0.39	0.22	0.69
More than three times a day	139		0.49	0.25	0.96
Child has TV in bedroom (ref: No)	184	.038
Yes	452		0.60	0.37	0.97