GROWING UP IN SCOTLAND: THE CIRCUMSTANCES AND EXPERIENCES OF 3 YEAR OLD CHILDREN LIVING IN SCOTLAND IN 2007/08 AND 2013

Chapter 3 Child Health

This section compares the health of children in Scotland who were aged 3 in 2013 and 2007/08. This section presents data on children's general health, prevalence of longstanding illnesses and disabilities and the number of accidents children have had.

We know from previous analyses of GUS data that there is a correlation between the parent's assessment of their child's health and household income, area deprivation and parental education - children living in more advantaged circumstances have better general health (see Bromley and Cunningham-Burley, 2010). This chapter explores whether this relationship is also apparent for BC2 children at age 3, and whether there have been any changes in the nature of this relationship since BC1 children were the same age.

3.1 Child's general health

Parents were asked to rate their child's health in general, with response options ranging from 'very good' to 'very bad'. As Table 4.1 shows, for both cohorts the majority (70% in BC1, 69% in BC2) of parents rated their child's general health as very good. Conversely, only 1% or less, rated their child's health as bad or very bad. There were no statistically significant differences between the cohorts on this measure.

Table 3.1 Child's general health by cohort

	BC1	BC2
	%	%
Very good	70	69
Good	24	26
Fair	5	5
Bad or very bad	0	1
Unweighted bases	4193	5019

Differences by cohort are not significant.

Table 3.2 shows parent assessed child general health by household income for the two cohorts. For both cohorts child general health varied by income and this relationship was statistically significant. Children in the highest income households were more likely to be assessed as having very good health (77% in BC1, 74% in BC2), compared with children in the lowest income households (62% in BC1, 65% in BC2).

The larger differences in assessed general health between those in the highest and lowest income quintiles in BC1, compared with BC2, might suggest that the relationship between household income and child general health was stronger in BC1 compared with BC2. However, this relationship was further investigated^[8] and regression analysis showed that there were no statistically significant changes in the nature of the relationship between assessed general health and income between the cohorts. Looking at the table, whilst the difference between the proportion of children reported as having very good health in the lowest and highest income groups decreased between the cohorts, the likelihood of having good health nevertheless remains closely linked to level of household income. Furthermore, the apparent reduction in the proportion of children with very good health in the highest two income groups is unexpected. As will be shown below, the pattern of improving health amongst the most disadvantaged group and declining health amongst the most advantaged is not repeated consistently across the other measures of socio-economic circumstances nor in other health measures. As such, it does not appear to suggest a trend towards less health inequality between children in the most and least advantaged groups.

Table 3.2 Child's general health, by equivalised household income (quintiles) and cohort

	Lowest quintile	2nd quintile	3rd quintile	4th quintile	Highest quintile
	%	%	%	%	%
BC1
Very good	62	68	74	77	77
Good	30	25	22	20	19
Fair	7	7	4	3	3
Bad or very bad	1	0	1	0	0
BC2
Very good	65	69	70	71	74
Good	29	26	23	24	22
Fair	5	4	6	5	4
Bad or very bad	1	1	0	0	1
Unweighted bases - BC1	783	803	761	858	721
Unweighted bases - BC2	981	778	814	776	1021

Tested on very good health: differences by income - p < .001; differences by cohort - NS; cohort*income - p < .05.

Table 3.3 shows a statistically significant association between parent reported child general health and parental level of education. Among children with the most educated parents around three quarters (76% in BC1 and 72% in BC2) were assessed as having very good health. By comparison, among those whose parents had no qualifications, less than two thirds (66% in BC1 and 61% in BC2) were assessed as such. The relationship between parental level of education and child general health did not differ between the two cohorts.

Table 3.3 Child's general health, by parental level of education and cohort

	No qualifications	Lower Standard Grades or VQs or Other	Upper level SGs or Intermediate VQs	Higher grades and upper level VQs	Degree level academic and vocational qualifications
	%	%	%	%	%
BC1
Very good	66	63	64	71	76
Good	26	33	28	24	20
Fair	7	5	7	5	4
Bad or very bad	1	-	1	0	0
BC2
Very good	61	64	65	68	72
Good	31	30	29	26	23
Fair	7	4	5	5	5
Bad or very bad	1	3	1	1	1
Unweighted bases - BC1	200	206	804	1379	1597
Unweighted bases - BC2	164	217	714	1428	2318

Tested on very good health: differences by parental education - p < .001; differences by cohort - p < .05; cohort*parental education - NS.

For both cohorts, child general health was also associated with maternal age, and this relationship was statistically significant. Table 3.4 shows that BC1 children born to mothers over 40 (74%) were more likely to be assessed as having very good health than those born to the youngest mothers (63%). In BC2, the relationship between maternal age and child general health showed a different pattern, with children whose mothers were aged 20 to 29 being less likely to be reported as having very good health than children from other mothers (66% compared with 71% of mothers in all the other age groups). Overall, however, changes between the two cohorts were not statistically significant, and the relationship between maternal age and parent assessed child general health was similar in both cohorts. This may be related, in part, to the small numbers of parents in the youngest and oldest age groups.

When specifically comparing children born to teenage mothers with children born to mothers aged 20 and over, there is a statistically significant difference in the relationship between maternal age and the proportion of parents assessing their child as having very good health across the cohorts. In BC1, children born to teenage mothers were less likely than children born to mothers aged 20 and over to be assessed as having very good health. For BC2 children the relationship is slightly different: the proportion of children born to teenage mothers with very good health was 71% which was the same proportion amongst children with older mothers (aged 30 years and older). It should also be noted, however, that when looking simply at good health (i.e. 'good' and 'very good' health combined), there was no difference in the relationship between maternal age and child health between the cohorts.

Table 3.4 Child's general health, by maternal age at child's birth and cohort

	Under 20 years old	20 to 29 years old	30 to 39 years old	40 or older
	%	%	%	%
BC1
Very good	63	67	74	74
Good	31	25	22	24
Fair	6	7	4	3
Bad or very bad	0	1	0	0
BC2
Very good	71	66	71	71
Good	25	27	24	24
Fair	3	6	5	5
Bad or very bad	1	1	1	0
Unweighted bases - BC1	221	1565	2229	155
Unweighted bases - BC2	217	1979	2573	227

Tested on very good health: differences by maternal age - p < .001; differences by cohort - NS; cohort*maternal age - NS.

Table 3.5 shows the variation in parental assessed child general health by area deprivation for both cohorts. In both BC1 and BC2 children living in the least deprived areas were more likely to be assessed as having very good health (76% in BC1, 73% in BC2) than those living in the most deprived areas (64% in BC1, 63% in BC2). This relationship was statistically significant for both cohorts. There were no statistically significant changes in the nature of this relationship between the cohorts.

Table 3.5 Child's general health, by area deprivation (quintiles) and cohort

	1 Most deprived	2	3	4	5 Least deprived
	%	%	%	%	%
BC1
Very good	64	68	71	74	76
Good	28	26	24	21	20
Fair	7	6	5	5	3
Bad or very bad	1	0	1	0	-
BC2
Very good	63	65	70	73	73
Good	31	26	24	22	23
Fair	5	7	5	4	3
Bad or very bad	1	1	1	0	1
Unweighted bases - BC1	833	698	873	884	905
Unweighted bases - BC2	943	936	1031	1064	1011

Tested on very good health: differences by area deprivation - p < 0.001; differences by cohort - NS; cohort*area deprivation - NS.

3.2 Longstanding illnesses and disabilities

At the age 3 interview parents were asked whether their child had any long-term conditions that affected their health. A long-term condition was described as any illness or disability that troubled the child long-term and was expected to last for more than a year. No examples of conditions were provided, but if the carer answered 'yes', then further details were collected.

Table 3.6 shows the proportion of 3-year-old children who had a longstanding illness or disability for both cohorts. The table shows that a slightly higher proportion of children in BC2 (17%) had a long-term health condition compared with children in BC1 (14%). This difference was statistically significant. It is also worth noting that this difference between the cohorts was not apparent when comparisons were made at age 10 months.^[9] The majority of children who had a longstanding illness or disability had only one condition - just 2% of all children in BC1 and 3% of children in BC2 had two or more longstanding illnesses or disabilities (Table 3.7).

Table 3.6 Child's longstanding illnesses or disabilities by cohort

	BC1	BC2
	%	%
No	86	83
Yes	14	17
Unweighted bases	4193	5019

Tested on 'Yes': difference between cohorts - p < .01.

Table 3.7 Child's number of longstanding illnesses or disabilities by cohort

	BC1	BC2
	%	%
None*	86	83
One*	12	14
Two or more*	2	3
Unweighted bases	4193	5019

Differences by cohort on items marked * are statistically significant at p < .05 or less.

The proportion of children who had any longstanding illnesses or disabilities by household income is shown in Table 3.8. There was a statistically significant relationship between household income and the proportion of children who had any long-term health conditions in both cohorts: children living in high-income households were slightly less likely to have any longstanding illnesses or disabilities than those living in households with lower incomes. For example, 14% of BC2 children living in the highest income households had a long-term health condition while this was the case for 19% of those living in the lowest income households. The increase in longstanding health conditions in BC2 was evident across all income groups. However, there was no statistically significant difference in the relationship between household income and prevalence of longstanding illnesses or disabilities between the two cohorts.

Table 3.8 Child's longstanding illnesses or disabilities, by equivalised household income (quintiles) and cohort

	Lowest quintile	2nd quintile	3rd quintile	4th quintile	Highest quintile
	%	%	%	%	%
BC1
No	83	86	85	87	87
Yes	17	14	15	13	13
BC2
No	81	82	82	85	86
Yes	19	18	18	15	14
Unweighted bases - BC1	783	803	761	858	721
Unweighted bases - BC2	981	778	814	776	1021

Tested on 'yes': differences by income - p < .01; differences by cohort - p < .01; cohort*income - NS.

Table 3.9 shows the proportion of children who have long-term health conditions by parental level of education. The table suggests that, for BC1, children whose parents had no qualifications were more likely to have a long-term condition than those whose parents had degree level qualifications. However, this apparent relationship between having a long-term condition and parental level of education was not statistically significant. For BC2, there was no clear pattern of association between the prevalence of long-term health conditions and parental level of education. This is in line with previous analysis of GUS data collected when the BC2 children were aged 10 months, which found that prevalence of longstanding illnesses or disabilities did not vary significantly by parental level of education (Bradshaw et al, 2013). There were no statistically significant changes in the nature of the relationship between long-term conditions and parental level of education between BC1 and BC2.

Table 3.9 Child's longstanding illnesses or disabilities by parental level of education and cohort

	No qualifications	Lower Standard Grades or VQs or Other	Upper level SGs or Intermediate VQs	Higher grades and upper level VQs	Degree level academic and vocational qualifications
	%	%	%	%	%
BC1
No	81	84	84	86	87
Yes	19	16	16	14	13
BC2
No	83	83	81	84	83
Yes	17	17	19	16	17
Unweighted bases - BC1	200	206	804	1379	1597
Unweighted bases - BC2	164	217	714	1428	2318

Tested on 'yes': differences by parental education - NS; differences by cohort - p < .001; cohort*parental education - NS.

There was no statistically significant correlation between maternal age and prevalence of longstanding illnesses or disabilities for either of the cohorts (Table 3.10).^[10] With the exception of children whose mothers were aged 20 to 29, children in all groups showed an increase in longstanding illness between cohorts.

Table 3.10 Child's longstanding illnesses or disabilities, by maternal age at child's birth and cohort

	Under 20 years old	20 to 29 years old	30 to 39 years old	40 or older
	%	%	%	%
BC1
No	87	83	87	90
Yes	13	17	13	10
BC2
No	84	83	83	82
Yes	16	17	17	18
Unweighted bases - BC1	221	1565	2229	155
Unweighted bases - BC2	217	1979	2573	227

Tested on 'yes': differences by maternal age - NS; differences by cohort - p < .01; cohort*maternal age - NS.

Table 3.11 shows the proportion of children having at least one long-term health condition by area deprivation. Analysis of GUS data collected when BC2 were aged 10 months did not find any relationship between area deprivation and prevalence of longstanding illnesses or disabilities (Bradshaw et al, 2013). By age 3, however, there was a relationship: children living in the most deprived areas were more likely to have a longstanding illness or disability than those living in the least deprived areas. This relationship was statistically significant for both cohorts. In BC1, for example, 18% of children living in areas in the most deprived quintile had a longstanding illness compared with 13% children living in areas in the least deprived quintile.

Comparing BC2 with BC1 at age 3, we see that there was a slight increase in the proportion of children in almost all SIMD quintiles (except the most deprived) who were reported as having a longstanding illness. There was no statistically significant change in the nature of the relationship between long-standing illness and area deprivation across the two cohorts.

Table 3.11 Child's longstanding illnesses or disabilities by area deprivation (quintiles) and cohort

	1 Most deprived	2	3	4	5 Least deprived
	%	%	%	%	%
BC1
No	82	85	87	87	87
Yes	18	15	13	13	13
BC2
No	83	81	84	84	84
Yes	17	19	16	16	16
Unweighted bases - BC1	833	698	873	884	905
Unweighted bases - BC2	943	936	1031	1064	1011

Tested on 'yes': differences by area deprivation - p < .01; differences by cohort - p < .01; cohort*area deprivation - NS.

3.3 Accidents

Parents were asked about the number of accidents the cohort child had had since the time of the last interview. It is important to note that, due to differences in the frequency of sweeps of data collection, the reference period for the two cohorts is different: BC1 parents were asked about the number of accidents in the past year whilst BC2 parents were asked about the number of accidents in the previous two years. This means that the figures are not directly comparable.^[11] It is possible, however, to look at and compare trends in the relationship between number of accidents and the various measures of social disadvantage for each cohort.

Table 3.12 gives an overview of the total number of accidents that children were reported to have had. As may be expected given the longer reference period, children in BC2 were more likely to have had an accident.

Table 3.12 Number of accidents in last year (BC1) or last two years (BC2)

	BC1	BC2
	%	%
None	81	68
One	17	25
Two or more	2	7
Unweighted bases	4193	5019

Differences by cohort are not comparable due to different reference periods.

For both cohorts, the number of accidents a child had within the reference period (whether one or two years) was correlated with the level of household income: children living in the highest income quintiles were more likely than children living in the lower income quintiles to have had no accidents (Table 3.13). For example, 71% of BC2 children in the highest income quintile had not had any accidents in the past two years compared with 67% for children in the lowest income quintile. There was no statistically significant difference in the relationship between income and number of accidents between the two cohorts.

Table 3.13 Number of accidents in last year (BC1) or two years (BC2), by equivalised household income (quintiles) and cohort

	Lowest quintile	2nd quintile	3rd quintile	4th quintile	Highest quintile
	%	%	%	%	%
BC1
None	79	81	80	84	83
One	18	17	17	14	15
Two or more	3	2	3	2	2
BC2
None	67	68	67	68	71
One	25	25	25	25	24
Two or more	8	7	8	7	5
Unweighted bases - BC1	783	803	761	858	721
Unweighted bases - BC2	981	778	814	776	1021

Tested on 'none': differences by income - p < .05; cohort*income - NS. Note that due to differences in the measure used for BC1 and BC2 a direct comparison by cohort of values for each sub-group is not valid. As such the p-value has not been reported.

Table 3.14 shows the number of accidents by parental level of education for each cohort. The nature of the relationship between parental level of education and number of accidents differed between the cohorts. For BC1 the figures suggest that children with highly educated parents were more likely to have had no accidents in the reference period than children whose parents had lower level or no qualifications. However, further analysis of the BC1 data showed that this relationship was not statistically significant.^[12] For BC2, on the other hand, the figures suggest that children whose parents were educated to at least degree level were more likely to have had accidents than children whose parents had lower level or no qualifications, and further analysis shows that this was a statistically significant relationship.

Table 3.14 Number of accidents in last year (BC1) or two years (BC2), by parental level of education and cohort

	No qualifications	Lower Standard Grades or VQs or Other	Upper level SGs or Intermediate VQs	Higher grades and upper level VQs	Degree level academic and vocational qualifications
	%	%	%	%	%
BC1
None	78	79	81	79	83
One	20	18	17	18	15
Two or more	2	3	3	3	2
BC2
None	74	77	66	65	69
One	20	15	25	26	25
Two or more	6	8	9	9	6
Unweighted bases - BC1	200	206	804	1379	1597
Unweighted bases - BC2	164	217	714	1428	2318

Tested on 'none': differences by parental education - p < .01; cohort*parental education - p < .05. Note that due to differences in the measure used for BC1 and BC2 a direct comparison by cohort of values for each sub-group is not valid. As such the p-value has not been reported.

The number of accidents was also correlated with maternal age at the child's birth (Table 3.15). Children born to younger mothers were more likely to have had two or more accidents in the reference period than those born to older mothers. For example, 10% of BC2 children born to teenage mothers had had two or more accidents while this was the case for only 5% of those born to mothers over 40. There was no statistically significant difference in the relationship between maternal age and number of accidents between the two cohorts.

Table 3.15 Number of accidents in last year (BC1) or two years (BC2), by maternal age at child's birth and cohort

	Under 20 years old	20 to 29 years old	30 to 39 years old	40 or older
	%	%	%	%
BC1
None	79	79	83	79
One	17	19	15	18
Two or more	4	2	2	3
BC2
None	66	67	69	71
One	24	25	25	24
Two or more	10	8	6	5
Unweighted bases - BC1	221	1565	2229	155
Unweighted bases - BC2	217	1979	2573	227

Tested on 'none': differences by maternal age - p < .05; cohort*maternal age - NS. Note that due to differences in the measure used for BC1 and BC2 a direct comparison by cohort of values for each sub-group is not valid. As such the p-value has not been reported.

Area deprivation was not correlated with number of accidents for either of the cohorts (Table 3.16).

Table 3.16 Number of accidents in last year (BC1) or two years (BC2), by area deprivation (quintiles) and cohort

	1 Most deprived	2	3	4	5 Least deprived
	%	%	%	%	%
BC1
None	78	81	82	81	84
One	19	17	15	18	15
Two or more	3	2	3	2	2
BC2
None	68	67	69	68	68
One	24	25	25	25	25
Two or more	8	8	6	7	7
Unweighted bases - BC1	833	698	873	884	905
Unweighted bases - BC2	943	936	1031	1064	1011

Tested on 'none': differences by area deprivation - NS; cohort*area deprivation - NS. Note that due to differences in the measure used for BC1 and BC2 a direct comparison by cohort of values for each sub-group is not valid. As such the p-value has not been reported.