Growing up in Scotland: overweight obesity and activity

CHAPTER 3 OVERWEIGHT AND OBESITY

3.1 Key findings

• 22% of children were overweight (including obese), 9% were obese.
• Risk factors for children's overweight and obesity include child characteristics (higher birthweight), maternal characteristics (high BMI and poor physical health), neighbourhood characteristics (low child friendliness), dietary practices and low parental supervision
• Children who were overweight and obesity at age 6 were very likely to have been classified as overweight or obese at age 4
• Physical activity and sedentary behaviour were not clearly associated with children's overweight and/or obesity

3.2 Prevalence and tracking of children's overweight and obesity

3.2.1 Prevalence of overweight and obesity at age 6

Reliable height and weight measurements were obtained from 3035 children (96% of the eligible sample at sweep 6). Following standard procedures developed by ISD Scotland, children with BMI values more than three standard deviations from the mean population value were excluded ( n=43, 1%), leaving 2992 cases.

Of these, three-quarters (76%) had BMI in the healthy range, 22% were overweight including obese, consisting of 13% who were overweight but not obese, and 9% who were obese or morbidly obese (Figure 3.1). It is not possible to make a precise comparison of these figures collected in 2010-11 for GUS children approaching their sixth birthday with the prevalence of children's overweight and obesity measured in the 2010 Scottish Health Survey, SHeS (Gray and Leyland 2011), as the SHeS does not provide a detalied breakdown by age. SHeS data found that 26% of children aged 2-6 years were overweight (including 11% obese), while 31% of children aged 7-11 years were overweight (including 18% obese).

Figure 3.1 Children's BMI classification at sweep 6

Note: Numbers above each histogram show estimated percentage in each BMI category, with error bars indicating 95% confidence intervals. N values for number of cases in each BMI category are unweighted values. n=2992

The prevalence of overweight (including obesity) at age 6 (22%) was slightly reduced from that at age 4 (26%). However, the prevalence of obesity was similar at both ages (see Figure 3.2).

Figure 3.2 Prevalence of children's overweight and obesity at ages 4 and 6

Note: Numbers above each histogram show estimated percentage in each BMI category, with error bars indicating 95% confidence intervals.

3.2.2 Association between overweight/obesity and children's height

Figure 3.3 shows that there was a strong association between height and overweight or obesity, with taller children more likely to be classified as obese or overweight. Other research has recognised this association, and considered whether the BMI measure might be misclassifying taller children (Freedman et al. 2003). However, this work found that taller children had greater skinfold thickness and raised insulin levels, and that childhood BMI was more strongly linked with adult adiposity than alternative indices relating height and weight measurements. The researchers concluded that BMI was more useful than alternative ways of relating weight and height measurements, and that the association between children's height and BMI does indeed reflect greater adiposity in taller children.

Figure 3.3 Association between children's height and overweight or obesity

3.2.3 Tracking of overweight and/or obesity

This section examines associations between children's overweight/obesity at age 4 and obesity/overweight two years later, referred to as "tracking" of overweight and/or obesity.

The vast majority (90%) of children classified as healthy weight at age 4 remained in this category two years later (Table 3.1), with only 8% becoming overweight or obese. There was considerable tracking of overweight and/or obesity, in that children who were overweight or obese at age 4 were much more likely than healthy weight children to have weight problems two years later. However, there was considerable movement out of the overweight and obese categories over the two-year period.

Overall, 60% of children classified as either overweight, obese or morbidly obese at age 4 (combining three columns of Table 3.1) were also overweight or obese at age 6, but 40% became of healthy weight. Almost half (48%) of children classified as either obese or morbidly obese at age 4 (combining two columns of Table 3.1) were similarly classified at age 6, 37% were reclassified as overweight and 15% became a healthy weight.

Table 3.1 Association between children's BMI at age 4 and age 6

	Age 4 BMI classification
	Underweight	Healthy weight	Overweight	Obese	Morbidly obese	Bases
Age 6 BMI classification	%	%	%	%	%	Wtd	Uwtd
Underweight	25	2	-	-	-	47	47
Healthy weight	75	90	54	21	8	2069	2136
Overweight	-	6	31	45	25	351	369
Obese	-	1	6	14	21	103	109
Morbidly obese	-	1	9	20	47	137	120
Total	100	100	100	100	100	2707	2781

3.3 Exploration of child and family factors associated with children's overweight and obesity

Important note: From this point forward, the term "overweight" is used to include all children at or above the 85th BMI percentile ( i.e. including those who were obese or morbidly obese). The term "obese" includes all children at or above the 95th BMI percentile ( i.e. including those who were morbidly obese).

Associations between individual child and family characteristics and children's overweight and obesity were explored (Table 3.2). Note that these associations do not take account of other factors that simultaneously may play a role.

Greater birthweight (standardised for gender, gestation and parity) was strongly associated with being overweight or obese at age 6. Neither gender, ethnicity, birth order or number of siblings was clearly associated with being overweight or obese, although there was a trend for children from larger families to be at lower risk. The GUS sample contains only a small percentage of children from ethnic minority groups, so is not as well placed to explore ethnic differences in overweight and obesity as the Millennium Cohort Study, which has a boosted ethnic sample.

Table 3.2 Child and family characteristics: associated with with children's overweight and obesity at age 6, bivariate analyses

		% overweight or obese	p	% obese	p	Bases
						Wtd	Uwtd
Gender	Male	23	NS	8	NS	1502	1521
	Female	22		10		1433	1471
Ethnicity	White	22	NS	9	NS	2814	2892
	Minority group	20		11		120	99
Birthweight (standardised)	1 - lowest decile	12	***	7	?	325	300
	2	14		6		311	298
	3	15		7		300	293
	4	23		12		294	303
	5	23		8		289	295
	6	23		9		292	301
	7	25		11		288	302
	8	26		9		283	299
	9	32		14		269	296
	10 - highest decile	32		11		279	298
Birth order	First born	23	NS	9	NS	1472	1445
	Not first born	21		9		1463	1547
Number of children in household	One	25	?	11	?	585	552
	Two	22		9		1538	1643
	Three	22		9		621	635
	Four or more	14		5		192	162

Notes: p=probability of difference between groups based on adjusted F where NS=not significant, † p<0.1, * p<0.05, ** p<0.01, *** p<0.001.

3.4 Exploration of parental factors associated with children's overweight and obesity

There were a number of statistically significant associations ( p<0.05) between individual parental factors and children's overweight and/or obesity, shown in Table 3.3. Note that these associations do not take account of other possible confounding influences. Significant associations with children's overweight and/or obesity included:

• Mother's overweight or obesity
• Nutritional practices related to
  • Infant nutrition including duration of breastfeeding and the timing of solids
  • Frequency of consumption of sweets and crisps as a toddler (age 2)
  • Eating the main meal in a room without a dining area (ages 2 and 5)
  • Skipping breakfast (age 5)
  • Use of a local take-away (age 6)
• Sedentary behaviour related to
  • TV in the child's bedroom
  • Screen time from sweep 4 onwards
• Low levels of parental supervision

Parental factors that were not significantly associated with children's overweight and/or obesity at age 6 are not shown in Table 3.3. These included parental sedentary behaviour and exercise, children's pattern of eating at age 2 (snacking between meals, composition of snacks), children's fruit and vegetable consumption, children's soft drink consumption, children's frequent eating of sweets and crisps at age 5, frequent use of a take-away, restaurant or convenience food for main meal at ages 3 and 5, children's physical activity, children's sleeping duration, patterns and problems; as well as most aspects of general patterns of parenting including warmth, joint parent-child activities, rules, household chaos and conflict.

Table 3.3 Parental factors associated with children's overweight and obesity at age 6, results of bivariate analyses

		% overweight	p	% obese	p	Bases
						Wtd	Uwtd
Parental example
Mother's BMI (sweep 6)	Underweight or healthy	16	***	5	***	1147	1200
	Overweight	22		10		808	845
	Obese	29		14		619	606
Health-related practices - nutrition
Breastfeeding duration (age 1)	None	24	NS	11	***	1147	973
	<4 months	21		8		1210	1329
	4 months plus	20		6		578	689
Introduction of solids (age 1)	<4 months	28	**	12	?	434	384
	4 months or later	21		9		2498	2603
Crisps - frequency (age 2)	Once a week or less	17	*	5	*	538	585
	More than once a week but not every day	23		9		1079	1139
	Daily	24		11		975	966
	More than once daily	24		10		342	301
Sweets - frequency (age 2)	Once a week or less	17	*	4	**	628	685
	More than once a week but not every day	24		10		1070	1136
	Daily	22		11		888	870
	More than once daily	25		10		347	299
Where child eats main meal (age 2)	Kitchen/dining room/combined living and dining room	20	*	8	*	1784	1994
	Other room (living room, bedroom, other)	25		11		1031	880
Eats breakfast (age 5)	Yes	22	?	9	***	2791	2860
	No	29		19		144	131
Where child eats main meal (age 5)	Kitchen/dining room/combined living and dining room	20	***	8	***	1901	2109
	Other room (living room, bedroom, other)	27		13		928	793
Use of local take-away/fast food outlet (age 6)	I don't use it	16	*	7	NS	314	345
	I use it sometimes	23		10		2046	2073
	I use it often	22		8		364	330
	There isn't one in this area	22		7		211	243
TV in bedroom (age 4)	Yes	24	*	10	*	1332	1180
	No	21		8		1574	1783
TV in bedroom (age 5)	Yes	25	***	11	***	1512	1368
	No	19		7		1414	1613
Weekday TV time (age 4)	None	21	*	9	*	152	165
	Up to 30 minutes	16		6		78	84
	30 minutes to 1 hour	21		9		155	170
	1 to 2 hours	21		8		1120	1189
	2 to 3 hours	25		11		843	852
	3 to 4 hours	21		7		319	297
	4 to 5 hours	21		9		130	115
	5 or more hours	34		17		104	86
Weekday screen time (age 5)	Under 1 hour	20	*	8	NS	279	314
	1 to under 2 hours	19		8		876	956
	2 to under 3 hours	22		9		904	918
	3 to under 5 hours	26		11		691	656
	5 hrs plus	23		13		157	125
Weekday screen time (age 6)	Under 1 hour	19	*	8	?	190	212
	1 hr to under 2 hours	19		7		734	803
	2 hrs to under 3 hours	24		9		744	757
	3 hrs plus	25		11		763	698
General patterns of parenting
Supervision (age 4)	Low	27	*	15	***	568	534
	Medium - low	21		8		923	908
	Medium - high	21		7		1015	1065
	High	21		9		430	485

Notes: p=probability of difference between groups based on adjusted F where NS=not significant, † p<0.1, * p<0.05, ** p<0.01, *** p<0.001.

3.5 Exploration of family and neighbourhood constraints associated with children's overweight and obesity

Distal (background) family and neighbourhood constraints on children's healthy weight identified as having statistically significant individual associations with children's overweight and/or obesity at sweep 6 (see Table 3.4 and Table 3.5) included:

• Lower maternal education (below HNC/ HND)
• Biological father not present in household
• Bottom 40% of average household income
• Poor maternal physical health over sweeps 1, 3 and 5
• Maternal smoking (pregnancy, sweep 1)
• Area deprivation (two most deprived quintiles)

These findings suggest social inequalities in the patterning of children's overweight and obesity. As in the previous section, note that these associations do not take account of possible confounding influences.

Other possible family and neighbourhood constraints examined (mother's working hours, mother's mental health and urban-rural area) were not clearly associated with children's overweight and obesity (and are not included in Table 3.4 and Table 3.5).

Proximal family and neighbourhood constraints significantly associated with children's overweight and obesity (see Table 3.4 and Table 3.5) included:

• Cost affecting mother's preparation of food (age 5, but not age 2)
• Mother's feelings of personal safety when out alone by day and night (age 4 but not age 6)
• Child-friendliness of neighbourhood (age 4, but not age 6)
• Antisocial behaviour problems in neighbourhood (age 6)

Proximal family factors that did not show clear protective effects in helping to prevent children's overweight or obesity included mothers' views on desirable levels for children's physical activity and concerns for the child's weight.

Proximal neighbourhood factors not clearly asociated with children's overweight and/or obesity (and not shown in Table 3.4 and Table 3.5) included access to local facilities such as a garden, playpark, local swimming pool or green space and the quality or safety of playparks.

Table 3.4 Family constraints associated with children's overweight and obesity at age 6, results of bivariate analyses

		% overweight or obese	p	% obese	p	Bases
						Wtd	Uwtd
Distal family factors
Mother's education	Up to Standard Grades or equivalent	23	*	11	**	1104	936
	Higher Grades or equivalent	25		11		649	669
	Upper level VQs ( e.g. HNC, HND)	20		7		370	400
	Degree or equivalent	19		6		805	980
Biological father in household (ages 1-5)	Not present throughout	25	?	12	**	850	640
	Present throughout	21		8		2086	2352
Household income (equivalised, average quintile, ages 1-5)	Top 60%	21	NS	7	**	1499	1784
	Bottom 40%	23		11		1421	1194
Mother's physical health (ages 1,3 and 5)	Poor	26	**	11	**	995	980
	Medium	22		9		1003	1022
	Good	19		7		929	984
Smoked in pregnancy	No	22	NS	8	**	2222	2396
	Yes	24		12		703	586
Proximal family factors
Mother affected by food costs (age 5)	A lot	29	NS	19	*	94	74
	A fair amount	24		11		251	242
	A little	23		9		881	899
	Not at all	21		8		1710	1777

Notes: p=probability of difference between groups based on adjusted F where NS=not significant, † p<0.1, * p<0.05, ** p<0.01, *** p<0.001.

Table 3.5 Neighbourhood constraints associated with children's overweight and obesity at age 6, results of bivariate analyses

		% overweight or obese	p	% obese	p	Bases
						Wtd	Uwtd
Distal neighbourhood factors
Area deprivation ( SIMD quintile)	1 least deprived	20	NS	7	*	576	697
	2	20		7		615	706
	3	22		8		566	618
	4	26		11		552	511
	5 - most deprived	23		12		627	459
Proximal neighbourhood factors
Neighbourhood safety (age 4)	High	22	NS	8	**	1886	2021
	Low	23		11		1049	970
Child friendly neighbourhood (age 4)	Good	20	*	8	**	1102	1238
	Average	23		7		760	795
	Poor	24		12		1073	958
Antisocial behaviour in area last year (age 6)	None	22	NS	8	*	1503	1638
	One	21		8		625	640
	Two	22		9		295	296
	Three	22		8		197	183
	Four or more	25		15		312	233

Notes: p=probability based on adjusted F where NS=not significant, † p<0.1, * p<0.05, ** p<0.01, *** p<0.001.

3.6 Multivariate modelling of factors associated with children's overweight and obesity at age 6

More sophisticated analyses examined associations between parenting, family and neighbourhood factors and children's overweight and/or obesity, controlling for various factors simultaneously. This section reports the main findings of multivariate analyses using logistic regression models of children's overweight (including obese) and obese at age 6. Detailed results are provided in the Technical Appendix. The reference group for the model of overweight was all children with healthy weight plus the small number of underweight children. In the obesity model, this reference group in addition included overweight children.

Several of the potential risk factors were highly correlated with one another. These included TV in bedroom with high screen time, consumption of crisps with consumption of sweets, child-friendliness of neighbourhood with low perceived safety and antisocial behaviour problems, and eating in a room not primarily designed for food consumption with use of a local takeaway. Exploratory models to assess the strength of different effects was used to guide further selection of variables for the multivariate models. Bedroom TV, child-friendliness of the neighbourhood and eating in a room without a dining area were selected ¹¹ . Snacking on sweets and crisps were combined to create a composite measure of snacking on both foods.

Children's overweight and obesity at age 4 was not included in the models, although as section 3.2.3 showed, children's BMI classification at age 4 was a very strong predictor of overweight and obesity two years later. ¹²

All modelling controlled for the child's gender and their age in months at the sweep 6 interview, ethnic group, birthweight, and number of children in the family. In respect of these standard controls for child characteristics and family size, all models found a significant effect of:

• higher birthweight on an increased likelihood of overweight or obesity
• being in a family with four or more children on a reduced likelihood of being overweight or obese

There were no statistically significant differences in overweight or obesity according to the child's gender, age at sweep 6 interview or ethnic group (white/minority).

Various stages of modelling looked at the separate influences of parenting and of family and neighbourhood constraints (see the Technical Appendix for details). A combined model was then used to examine all factors simultaneously. In this combined model the following factors emerged as significantly associated with children's overweight and obesity (summarised in Figure 3.4):

• Mother's overweight or obesity
• Frequency of sweet and crisp consumption as a toddler
• Mother's poor physical health

Additional factors associated with children's overweight (including obesity) were:

• Eating main meal in a room without a dining area

Additional factors associated with children's obesity were:

• Low parental supervision
• Skipping breakfast
• Low child-friendliness of neighbourhood

Figure 3.4 Summary of risk factors for children's overweight and/or obesity at sweep 6: results of final multivariate modelling

Note: Shaded areas indicate increased or decreased risk of either overweight and obesity, or obesity. Model adjusted for all factors together, see Tables 1.1 and 1.2 in the Technical Appendix for further details