Growing Up In Scotland: Maternal mental health and its impact on child behaviour and development

Chapter 3 FINDINGS

3.1 Summary of main findings

• Almost a third (31%) of the GUS mothers overall experienced poor mental health at some point in the four survey years, although at any one time, the incidence of poor mental health was much lower, affecting 12-16% of mothers
• Almost 1 in 6 (17%) had poor mental well-being recorded on one occasion only. These women were classified as having "brief" mental health problems
• Almost 1 in 7 (14%) had poor mental well-being recorded on at least two occasions. These women were classified as having "repeated" mental health problems
• Mental health difficulties were associated with a mother's social characteristics (e.g. lone parenting, low income and living in an area of deprivation) and with reported relationship difficulties and poor social support
• Children's social, behavioural, cognitive and emotional development were also associated with social factors
• Children whose mothers were emotionally well throughout the survey period had better social, behavioural and emotional development than those whose mothers had brief mental health problems, and they in turn, had better development than those whose mothers had repeated mental health problems
• The relationships between maternal mental health and children's social, emotional and behavioural development at 46 months was still statistically significant after we took account of social factors
• Cognitive development at 34 months was not statistically associated with the mother's mental well-being

3.2 Incidence of maternal mental health problems

Maternal mental health was assessed at each sweep of the survey. As can be seen in Table 3.1, at any one sweep, 12-16% of all mothers interviewed were found to be experiencing poorer mental health (see Appendix A for a detailed description of how mental health was assessed and defined).

Table 3.1 Incidence of poorer mental health at each sweep

	%
Poor mental health when child aged 10 months	14
Poor mental health when child aged 22 months	12
Poor mental health when child aged 34 months	16
Poor mental health when child aged 46 months	12
Bases
Weighted	3844
Unweighted	3884

Base: All natural mothers who responded at sweeps 1-4

Whereas about two thirds (69%) of GUS mothers had good or average health at all sweeps, almost a third (31%) of the GUS mothers had scores which indicated that they were experiencing mental health problems at some point in the four years after the birth of the cohort baby (see Figure 3-A).

Figure 3 A Instances of poorer maternal mental health (%)

In all, 1 in 7 (14%) of all the women was defined as having "repeated" poor mental health - with scores indicative of poorer mental health at two or more sweeps of the survey.

As we were measuring at four discrete points in time, it cannot be assumed that repeated negative scores indicate continuous mental health problems. Nor can it be assumed that those we have classified as having "brief" emotional health problems were experiencing transitory or fleeting difficulties. It may simply have been that they had problems at other times as well, but not at the point when a survey interview was carried out.

In the following sections, we distinguish between three distinct groups comprising those who were classified as having:

• good mental health at every sweep (that is, had scores in the good/average range)
• a "brief" mental health problem (poor mental health at one sweep only - 1 or 2 or 3 or 4)
• a "repeated" mental health problem (poor mental health at any two (or more) sweeps) ⁶

It was relatively unusual for someone to be included in this "repeated" category if she had been defined as having average or good mental health at Sweep 1: only 1 in 5 (20%) of these women (that is, those with good or average mental health at sweep 1) had scores indicating poor mental health at a further point. On the other hand, of those who had poor mental health at Sweep 1, two-thirds (67%) went on to have poor mental health scores recorded at a subsequent sweep or sweeps.

Table 3.2 Mother's mental health at Sweep 1 (10 months after child's birth) by mother's mental health group

Maternal mental health over time	Mental health assessment at Sweep 1
	Good/average (%)	Poor (%)
Good/average mental health at all sweeps	80	-
Brief mental health problems (any one sweep)	15	33
Repeated mental health problems (at two+ sweeps)	5	67
Weighted bases	3219	555
Unweighted bases	3267	512

3.3 The social patterning of mental health problems

The mental health of the GUS mothers was patterned by social and economic factors. The following figures provide some basic descriptive information about the characteristics of women who experience brief, repeated or no mental health problems in the child's early years.

Lone parents were found to be both more likely to experience a mental health problem while their child was young, and were also more likely to have a repeated mental health problem (Figure 3 B).

Figure 3 B Maternal mental health group by family type (%)

The mother's age was less obviously associated with poor mental health (Figure 3 C), although it is worth noting that although teenage mothers were significantly more likely than older mothers to have a brief period of poor mental health, the difference was much smaller for repeated mental health problems.

Figure 3 C Maternal mental health group by age of mother at birth of the child (%)

There were also stark differences by income (Figure 3-D) with 82% of those in the highest income group having good/average mental health throughout the four sweeps, compared with just 52% of those in the lowest equivalised income group. In contrast, four times as many women in the lowest income group had repeated mental health problems as those in the highest income group (24% compared with 6%).

Figure 3 D Maternal mental health group by Household equivalised income (%) ⁷

Household employment status also showed a relationship with maternal mental health (Figure 3-E) In particular, in households where there was no-one in paid employment, rates of poor maternal mental health were significantly higher. It appeared to make little difference whether the employment was full or part-time.

Figure 3 E Maternal mental health group by household employment status (%)

The factors that could affect mental health may, of course, be inter-related. For example, low income is likely to be connected to unemployment. As such, we wanted to explore the extent to which different factors were important in their own right. We therefore used multivariate regression analysis to look at the independent associations between maternal mental health and the various indicators of socio-economic status. Regression analysis allows the association between an explanatory variable (such as low income) and an outcome variable (such as repeated mental health problems) to be explored while controlling for other variables (such as lone parenthood) which may affect the outcome variable - mental health. As well as considering various socio-economic characteristics, including those described above, a range of additional psycho-social factors were also considered which have been shown to be associated with maternal mental health in previous research. These include access to social support and the quality of the couple relationship (Brown and Harris, 1978; Martin et al., 1989). ⁸

Living in an area of deprivation was the only characteristic associated with both brief and repeated mental health problems. Repeated mental health problems were additionally associated with experiencing brief or persistent poverty, low social support and a poor relationship with a partner as well as living in a large urban locality. Having no access to a garden and having a low household income were associated with brief mental health problems as was reported difficulty in coping shortly after the birth. Table 3.3 shows the factors that had an independent statistically significant association with maternal mental health after we took account of the inter-relationships between variables.

Although socio-economic factors were significantly and independently associated with brief mental health problems, the relationship was more pronounced in relation to repeated mental health difficulties.

Table 3.3 Independent significant associations with brief and repeated mental health problems

	Brief mental health problems	Repeated mental health problems
Living in low income household	v	-
Living in brief or persistent poverty	-	v
Area of deprivation	v	v
No access to a garden	v	-
Urban locality	-	v
Did not cope well as a couple after the birth	v	-
Low relationship strength	-	v
Low social support	-	v
In receipt of incapacity benefit	-	v

3.4 The impact of maternal mental health on early child outcomes

Figure 3-F shows the percentage of children for each maternal mental health category who had poorer child outcome scores. At a basic level, maternal mental health was significantly associated with children's emotional well-being, cognitive development, behaviour and their social relationships with peers at 46 months. For example, almost a quarter (27%) of children whose mothers had good/average mental health at all sweeps had poor reported relations with their peers at 3 years of age. However, this figure almost doubles to 48% for those whose mothers had repeated mental health problems (see Figure 3-F).

There was a clear effect of the duration of exposure to poor mental health for all of the child outcomes: those children exposed to repeated poor maternal mental health had poorer outcomes than those exposed "briefly" who, in turn, had poorer outcomes than those whose mothers were defined as having good or average mental health throughout. In other words, the greater or more prolonged the exposure to maternal mental health problems, the greater the likelihood that the child would have poorer outcomes. These differences were more marked for some outcomes than others but were statistically significant for all.

Figure 3 F Poor child outcomes in relation to maternal mental health status (%)

It is perhaps not surprising that many of the factors associated with poorer maternal mental health were also associated with child outcomes. The following analysis therefore assessed whether maternal mental health status continued to be significantly associated with child outcomes once other socio-economic factors were controlled for: that is, whether it was simply that the same socio-economic factors were associated with maternal mental health and child outcomes or whether maternal mental health affected child outcomes, regardless of other socio-economic factors. The following sections consider the relationship between maternal mental health and each of the selected child outcomes (see Figure 3-F).

Relations with peers

Children's social skills were assessed in terms of the quality of his/her relations with peers at age 46 months via a sub-scale of Goodman's "Strengths and Difficulties Questionnaire" ( SDQ). Appendix A provides further detail about the measures used. Looking at just the relationship between the child's social skills and maternal mental health without taking into account any other factors, we found that children exposed to a mother with repeated mental health problems were almost twice as likely to have poorer relations with peers than those whose mothers remained mentally well throughout the four year period or who had only brief episodes of poor mental health in that time. The more exposed a child was to poor maternal mental health, the greater the likelihood that they would be experiencing difficulties in their peer relationships at age 46 months.

When we added in other factors, we found that that maternal mental health continued to exert a significant independent impact on a child's social skills. However, other socio-economic, household and family characteristics were also independently associated with peer problems. In particular, children in low income households, children with no siblings, and those whose parents (or step-parents) had a poor relationship were more likely to have poor pre-school relationships with their peers. In other words, children who do not experience positive relationships within the home appear to have more negative social relationships outside the home. Boys were also found to have poorer peer relationships than girls at this age.

Emotional well-being

A very similar pattern was observed in relation to children's emotional well-being at 46 months with, once again, an almost two-fold difference between those with an emotionally well mother and those children whose mothers had repeated problems. Again too, those children whose mothers had brief episodes of poorer mental health were more likely than those with repeated exposure to fare well emotionally, but had poorer outcomes than children of mentally well mothers.

The child's emotional well-being was also clearly and independently associated with his or her mother's mental state, even when taking into account other socio-demographic factors, including the mother's social class status and household income, both of which were also independently associated with children's emotional outcomes. Having a poor couple relationship was also weakly associated with greater emotional difficulties.

Behaviour

Children's behaviour at sweep 4, when they were almost 4 years old, was assessed using the conduct and hyperactivity sub-scales of the SDQ. This included a list of statements such as "child often has temper tantrums or hot tempers" and "child often lies or cheats", to which the respondent had to code whether this was "Not true" of their child, "Somewhat true", or "Certainly true".

Maternal mental health remained significantly and independently associated with behavioural outcomes at 46 months: children whose mothers had either brief or repeated mental health problems were more likely than those whose mothers were emotionally well to have behavioural difficulties at 46 months. Demographic and socio-economic factors were again also significantly independently associated with behavioural outcomes: children with a younger mother, those with lower social support, who had lived in persistent poverty and in an area of higher deprivation had poorer behavioural outcomes than other children.

Cognitive development

Cognitive outcomes had a rather different relationship with maternal mental health than the social, emotional and behavioural outcomes. While cognitive outcomes (both naming vocabulary and picture similarities) at 34 months were significantly associated with maternal mental health, there was a less marked gradient compared with the other outcomes we examined. More importantly, the association was no longer significant once other social and economic factors were taken into account.

Factors which were found to be significantly related to lower cognitive scores included maternal characteristics (low maternal educational attainment and younger age) and socio-economic factors (living in an area of deprivation, an urban area of residence, larger family size and living in persistent poverty during the early years).

Table 3.4 summarises the relationships between maternal mental health and the selected child development outcomes that continued to show a significant independent relationship after controlling for variables with were related to each other. See also table B.2 in Appendix B.

Table 3.4 Summary of significant independent association between maternal mental health and poorer child outcomes

	Child outcomes
	Behaviour	Emotional	Peer	Cognitive: Picture task	Cognitive: Naming task
Maternal characteristics
Poor mental health (brief and repeated)	v	v	v	-	-
Age (younger)	v	-	-	-	v
Unemployment or low status occupation	v	v	-	-	-
Low educational attainment	-	-	-	v	v
Family/child characteristics
Male	-	-	v	-	-
Not an only child	-	-	v	v	v
Household characteristics
Area of deprivation	v	-	-	v	v
Urban locality	-	-	-	v	v
Low household income	-	v	v	-	-
Persistent poverty	v	-	-	v	v
Psycho-social characteristics
Poor couple relationship	-	v	v	-	-
Lower social support	v	-	-	-	-

The cognitive assessments were conducted at an earlier stage of development than the other measures which may explain at least some of the difference with this outcome. It may also be that very early maternal mental well-being may be particularly important in terms of shaping children's cognitive abilities, possibly mediated by poor attachment. We were not able to explore this as we were unable validly to identify mothers who may have been depressed in the early postpartum period where there may be a more pronounced potential impact. However, there is also strong evidence that children's cognitive development, particularly in relation to language acquisition, is closely tied to parental social position and education.