The Scottish Government

« Previous | Contents | Next »

Listen

CHAPTER THREE Differences in maternal behaviour by measures of social disadvantage

Introduction

3.1 Chapter 2 mapped out the patterns of advantage and disadvantage in terms of socio-economic and demographic characteristics by maternal age and family type. This exploration highlighted the inter-relationship between many of these characteristics and demonstrated that age and family type themselves are important factors contributing to a mother's relative social position with concentrated disadvantage evident in particular amongst mothers aged under 25, and lone parents who do not live with other adults.

3.2 Data collected on various maternal behaviours and service use, namely breastfeeding, smoking and attendance at ante-natal classes allow an initial consideration of the relationship between social disadvantage and these behaviours and outcomes.

Breastfeeding

3.3 Younger mothers, lone mothers, those with fewer educational qualifications, on low incomes, and those living in areas of relative deprivation were less likely to breastfeed than older mothers, those in couple families, those with more educational qualifications, living in high income households and those living in less deprived areas respectively (Table 9). Differences in likelihood of breastfeeding by maternal age and level of educational qualifications are particularly notable. Whereas 70% of mothers with Higher grades or above had breastfed the sample child only 29% of mothers with no qualifications had done so. A similarly sized gap exists by maternal age - around one-third (33%) of mothers in the youngest age group had breastfed the sample child compared with over two-thirds (68%) of mothers aged 30 to 34, and 72% of mothers in the oldest age group. Again, the distinction between the under 25s and over 25s is prominent in these analyses.

Table 9 Whether child was ever breastfed by selected measures of advantage/disadvantage

Base: Natural mothers in the birth cohort

3.4 Logistic regression was undertaken to identify more precisely the key measures of advantage or disadvantage, amongst those being considered, which appeared to influence whether or not a mother had decided to breastfeed. Whilst most measures remained significant in the regression analysis, mother's level of education was proven to be the strongest predictor of breastfeeding. The odds of mothers who were educated to higher grade breastfeeding were 2.6 times higher than for those who had no qualifications. Clearly the data cannot tell us what is it about education that produces such results or whether education itself is a proxy for other relevant social processes. The Hosmer-Lemeshow test revealed a good fit for the model, although the Nagelkerke R ² effect size demonstrated fairly weak explanatory power suggesting there are other important factors at work here.

Table 2 Logitistic regression model detailing factors related to child having been breastfed: birth cohort

Attendance at ante-natal classes

3.5 Earlier analysis of GUS data (Anderson et al, 2007) illustrated that parity was the strongest factor affecting attendance at ante-natal classes with first-time mothers significantly more likely to attend than those who already had children. As such, to examine differences in attendance by measures of disadvantage, cross-sectional analysis was restricted to first-time mothers only. The data in Table 11 show that amongst first-time mothers, non-attendance at ante-natal classes was associated with younger age, lower income, socio-economic classification and lower educational attainment.

Table 11 Whether mother attended ante-natal classes by selected measures of advantage/disadvantage

Base: Primaparous natural mothers in the birth cohort

3.6 Again, logistic regression analysis was run to pinpoint the key measures influencing attendance at ante-natal class. As with breastfeeding, several measures remained significant in the regression analysis however, after parity (which was by far the strongest variable), maternal age at the child's birth emerged as the next strongest predictor of attendance at ante-natal classes. The odds of mothers in each of the three age groups above 25 years attending ante-natal classes were at least three times higher than for mothers aged under 20, and around twice as high as for mothers aged between 20 and 24. Whilst the Nagelkerke R ² effect suggests this model has better predictive efficacy than the breastfeeding model (although it is still moderate), the Hosmer and Lemeshow test indicates that the model is of poor fit.

Table 12 Logitistic regression model detailing factors related to attendance at ante-natal classes: birth cohort

Smoking

3.7 Younger mothers, those with lower educational attainment and those living
in more deprived areas were more likely to say they smoked. Differences by area of deprivation are particularly stark. Mothers living in an area in the most deprived quintile were around 4 times more likely to say they smoked than mothers living in the least deprived quintile (44% compared with 10%). Likelihood of smoking decreased with age - around two-fifths (41%) of mothers aged 20-24 said they smoked compared with one-fifth (19%) of mothers aged 30 to 34.

Table 13 Whether mother smoked by selected measures of advantage/disadvantage

Base: Primiparous natural mothers in the birth cohort

3.8 Logistic regression analysis was run again to identify those variables which were most strongly predictive of smoking. Living in social housing, having a home in a deprived area and lack of educational qualifications were all similarly strongly predictive of smoking. Renting from the local authority was the strongest predictor: the odds of mothers who lived in a home rented from the local authority smoking were around 2.4 times higher than for mothers who owned their home (or were buying it with a mortgage). Maternal age did not feature in this model. Whilst the parameters of the model suggest good fit, Nagelkerke's R ² indicates only weak explanatory power. Clearly, renting from the local authority, while being the strongest predictor in this model, is likely to be related to a range of factors that influence smoking rather than itself being a direct influence on that behaviour.

Table 14 Logitistic regression model detailing factors related to mother smoking

Summary

3.9 This chapter has demonstrated clearly that adverse maternal behaviour appears to be closely related to maternal age and socio-economic circumstances. Younger mothers (particularly those aged under 25), lone mothers, those with lower or fewer educational qualifications, on low incomes, and those living in areas of relative deprivation were less likely to breastfeed, while low socio-economic status, poor educational attainment, area deprivation and renting from the local authority were also associated with maternal smoking. Similarly, controlling for parity, non-attendance at ante-natal classes was associated with younger age, lower income, socio-economic position and educational attainment.

3.10 Considerable attention has rightly focused on the association between adversity and negative social and health behaviours and outcomes for mothers and children. Rather less attention has been paid to understanding, within a group with generally poor outcomes, whether and how "healthy" or more positive outcomes may be produced. The next chapter takes an approach to the exploration of the impact of social disadvantage on the potential health and well-being of mothers and their infants that aims to identify resilience; in other words identifying those who report more positive health related behaviours although in disadvantaged circumstances.

« Previous | Contents | Next »