نابرابری وضعیت تاهل در سیگار کشیدن مادر در دوران بارداری، شیردهی و افسردگی مادر

Abstract One of the dramatic recent changes in family life in Western nations has been the rise in non-marital childbearing. Much of this increase is attributable to the growth in cohabitation. But in some countries, notably the UK (and the USA) this is much less the case with significant proportions of children being born to parents who are not living together. This study uses data from the Millennium Cohort Study, a British birth cohort established in 2001, to examine whether the closeness of the tie between parents, as assessed by their partnership status at birth, is related to smoking during pregnancy, breastfeeding and maternal depression. Four sets of parents are distinguished representing a hierarchy of bonding or connectedness: married and cohabiting parents, and two groups of solo mothers, those closely involved with the father at the time of the birth and those not in a relationship. Smoking in pregnancy, breastfeeding and maternal depression tests for trend, adjusted for socio-demographic factors, showed that there was a statistically increased risk of adverse health and health behaviours by degree of parental connectedness. There were also consistent and statistically significant differences between married and non-married mothers. Particularly noteworthy was the finding that cohabiting mothers have greater risk of adverse outcomes than married women. Among the non-married set, there were also differences in risk of adverse outcomes. For smoking in pregnancy, the key difference for continuing to smoke throughout the pregnancy lay between mothers involved with partners and those lacking an intimate relationship. For breastfeeding, stronger parental bonds were associated with initiation of breastfeeding, with a clear difference between cohabiting mothers compared to solo mothers. There was also an increased risk of maternal depression with looser parental bonding, and among non-married groups this increased risk was most noticeable among cohabiting mothers when compared with solo mothers.

Introduction There have been major changes in the demography of family life in recent decades with one of the most dramatic being the rise in non-marital childbearing. In UK, for example, the proportion of births occurring outside of marriage was 12 per cent in 1981, 30 per cent in 1991 and in 2004 stood at 42 per cent (ONS, 2005). Similar developments have occurred across many Western nations, with most of the rise in non-marital childbearing being attributable to the growth in cohabitation (Andersson, 2002, Kiernan, 1999). However, this is much less the case in Britain and the USA, where there have been notable increases in children being born to parents who are not living together at the time of the birth (Kiernan, 2004a). Little is known about these developments in family life and particularly whether new parents who are more loosely bonded differ in their behaviours and experiences during pregnancy and post-birth. In UK it is estimated from the Millennium Cohort Study that children born in 2000–2001, 60 per cent of children were born to married couples, 25 per cent to cohabiting couples and 15 per cent to solo mothers (Kiernan & Smith, 2003). The absence of the legal bond of marriage among cohabiting couples may represent less economic or emotional security, which may lie behind the higher dissolution rates invariably found among cohabiting parents compared with married parents (Bumpass & Lu, 2000; Kiernan, 1999). Solo mothers are the most socio-economically disadvantaged of parents and are less likely to have the support of a partner (Kiernan, 2002; Marsh & Perry, 2003). From pregnancy, through birth, and into infancy the health-related behaviour of the child's mother, as well as broader aspects of the family environment, matter for the long- and short-term healthy development of the child. In this study we focus on three aspects of maternal health and health-related behaviour that have important implications for child development: (1) maternal smoking during pregnancy, (2) breastfeeding and (3) maternal depression. Foetal exposure to cigarette smoke during pregnancy is associated with multiple deleterious short-term outcomes, including medical complications of pregnancy and birth (Castles, Adams, Melvin, Kelsch, & Boulton, 1999), intrauterine foetal growth retardation (Horta, Victora, Menezes, Halpern, & Barros, 1997), preterm delivery (Shah & Bracken, 2000), low birth weight, (Walsh, 1994), infant mortality (DiFranza & Lew, 1995) and negative temperament in early childhood (Brook, Brook, & Whiteman, 2000). While some of these short-term outcomes themselves exert a long-term impact on child health and development, foetal exposure to maternal smoking during pregnancy appears to also have independent long-term effects on cognition (Ernst, Moolchan, & Robinson, 2001; Najman et al., 2004; Olds, 1997) and behaviour (Rodriguez & Bohlin, 2005; Wakschlag, Pickett, Cook, Benowitz, & Leventhal, 2002). The same pattern is recognized in relation to breastfeeding. In the short-term, breastfeeding is beneficial for the physical health of the infant (less diarrhoea, respiratory infections, otitis media and fewer clinic, emergency and hospital visits) (American Academy of Pediatrics Work Group on Breastfeeding, 1997; Heinig & Dewey, 1996). These short-term effects have a long-term impact on child health and development, while breastfeeding also has independent, beneficial, long-term effects on health (lower rates of diabetes, Crohn's disease, lymphoma, atopic disease and obesity), and intellectual development (Horwood & Fergusson, 1998). It is well known that maternal depression increases the risk of emotional and behavioural problems among offspring (Brennan, Hammen, Anderson, & Bor, 2000; Cummings & Davies, 1994; Downey & Coyne, 1990; Kim-Cohen, Moffitt, Taylor, Pawlby, & Caspi, 2005). These problems can persist far beyond childhood and there is increasing evidence for inter-generational transmission of psychopathology and its risk factors (Serbin & Karp, 2003; Warner, Weissman, Mufson, & Wickramaratne, 1999). Maternal depression has also been shown to have adverse effects on cognitive and language development among offspring (Cox, Puckering, Pound, & Mills, 1987; Hay et al., 2001; Whiffen & Gotlib, 1989). These three facets of maternal health and health-related behaviour are interrelated. Depression and cigarette smoking are highly correlated and depressed women are less likely to quit smoking (Borrelli, Bock, King, Pinto, & Marcus, 1996; Pritchard, 1994); there is a substantial, albeit inconclusive, research literature that attempts to determine whether smoking causes depression, or vice versa (Goodman & Capitman, 2000; Hanna, Faden, & Dufour, 1994; Kendler et al., 1993). Women who smoke are less likely to breastfeed than non-smokers and, if they do initiate breastfeeding, they do not breastfeed for as long (Hill & Aldag, 1996; Minchin, 1991). While breastfeeding seems neither to increase or decrease the risk of maternal depression (Cox, Connor, & Kendall, 1982), depressed mothers report more difficulties with breastfeeding, are less responsive to infant feeding cues, and more likely to view breastfeeding difficulties in psychological terms—as a rejection of them and their milk (Nordstrom et al., 1988; Tamminen & Salmelin, 1991). It therefore makes sense to seek common risk factors and/or risk markers for maternal smoking in pregnancy, breastfeeding and maternal depression. Single motherhood per se has long been recognized as a socio-demographic risk factor associated with smoking during pregnancy (Ebrahim, Floyd, Merritt, Decoufle, & Holtzman, 2000; LeClere & Wilson, 1997), breastfeeding (Wright, et al., 1988), and maternal depression (Brown & Moran, 1997), although some evidence suggests that this association may not be consistent for breastfeeding (Scott & Binns, 1999). However, it is impossible within the current literature to discern whether or not it is marriage itself, cohabitation, lack of an intimate partner, or a broader lack of social support that makes the most difference to, or acts as the best marker of, maternal health and health-related behaviours. Most studies have been unable to distinguish cohabiting mothers from married mothers, as, for example, in two studies of smoking during pregnancy from Sweden, where married and cohabiting mothers are grouped together (Cnattingius, 1989; Thue, Schei, & Jacobsen, 1995). Moreover, few studies have been able to distinguish between solo mothers who are in a relationship with the father even though they do not live together and those who are without a partner relationship. Nevertheless, there is evidence that lack of paternal support, independently of marital status, affects whether or not pregnant women quit smoking (Dejin-Karlsson et al., 1996), as does the quality of marital relationship (Wakschlag et al., 2003). It has also been shown that fathers’ preferences are a major factor in mothers’ decisions about breastfeeding (Arora, McJunkin, Wehrer, & Kuhn, 2000) and that the amount of support that a woman receives from significant people in her life affects her risk of postpartum depression (O’Hara, 1986). Beyond the bare facts of marital status, cohabitation and partner support there may be underlying issues, such as socio-economic status and mothers’ own experiences of early childhood and adolescence that may link marital status and family structure to maternal health and health-related behaviours. In this study, to begin to examine whether the closeness of the tie between parents, as assessed by their partnership status, is related to smoking during pregnancy, breastfeeding and maternal depression, we use data from the Millennium Cohort Study, a British birth cohort established in 2001. We distinguish four sets of parents, married and cohabiting parents and two groups of solo parents. For solo mothers, as a gauge of the strength of the relationship between the mother and father at the time of the birth we use the mother's self-report of whether or not she was in a close relationship with the father of the baby. Thus the four sets of parents can be viewed as representing a hierarchy of bonding, with the married having the strongest bonds and the solo mothers where the mother does not report a close relationship having the loosest ones. We hypothesized that there would be significant differences in maternal health and health-related behaviours between these groups, with trends towards a decreased likelihood of smoking during pregnancy and maternal depression, and an increased likelihood of breastfeeding, associated with increasingly strong partnership bonds.

Results Table 3 shows the relative risk ratios for unmarried mothers in different partnership settings relative to married mothers for whether they gave up or continued to smoke compared with those who were non-smokers. Model One is the simple model with no controls, while Model Two includes as covariates maternal age, parity, educational attainment and ethnicity. Among those mothers who continued to smoke there is a gradient according to degree of partner bonding in the extent to which this occurred: from married mothers down through cohabiting (relative risk of 4.7 compared with married mothers); to solo mothers who were involved with the father at the time of the birth (relative risk of 7.6); and on to those who were not in a close relationship—who had the highest risk (relative risk of 9.2) of continuing to smoke. This pattern persisted after adjusting for socio-demographic factors, although the risk ratios were attenuated. An adjusted Wald test for Model Two showed that the differences in the estimates between the cohabiting group and the two solo mother groups were significantly different (cohabiting vs. closely involved, F-test 14.55, p-value <0.01; cohabiting vs. not in a relationship, F-test 29.3, p-value <0.01); the difference in the estimates between the two groups of solo mothers was smaller (F-test 3.09, p-value=0.08). The test for trend of increasing risk of continuing to smoke with decreasing parental bonding was significant (p-value <0.01). The association between giving up smoking during pregnancy and partnership status was less clear-cut ( Table 3). Compared with the married mothers, the unmarried groups of cohabiting and solo mothers were less likely to give up smoking i.e. they had higher risks of not quitting smoking, but within the unmarried set the differences in quitting were only weakly associated with the level of bonding between the parents. The difference between cohabiting mothers and those who were not living with the father but were closely involved had similar risks of quitting (Model Two, adjusted Wald test, F-test=0.01, p-value=0.92). However, the difference between the cohabitants and the solo mothers who were not in a relationship was statistically significant (Model Two, adjusted Wald test, F -test=0.59, p=0.015p=0.015), as was the test for trend relating quitting smoking to the hierarchy of bonding (p-value <0.01). Table 3. Smoking behaviour according to partnership status at birth: relative risk ratios and confidence intervals from multi-nomial regression analysis Non-smoker vs. continued Non-smoker vs. not quit Continued smoking vs. quit Model 1 Partnership status at birth Married 1.00 1.00 1.00 Cohabiting 4.71⁎⁎⁎ (4.2–5.3) 3.51⁎⁎⁎ (3.1–4.0) 1.34⁎⁎⁎ (1.2–1.6) Solo—closely involved with the father 7.60⁎⁎⁎ (6.4–9.0) 3.92⁎⁎⁎ (3.2–4.8) 1.94⁎⁎⁎ (1.6–2.4) Solo—not involved with the father 9.19⁎⁎⁎ (7.9–10.7) 4.77⁎⁎⁎ (4.0–5.7) 1.93⁎⁎⁎(1.6–2.3) Model 2 Partnership status at birth Married 1.00 1.00 1.00 Cohabiting 2.89⁎⁎⁎ (2.5–3.3) 2.35⁎⁎⁎ (2.0–2.7) 1.22⁎⁎ (1.1–1.4) Solo—closely involved with the father 3.90⁎⁎⁎ (3.3–4.6) 2.38⁎⁎⁎ (1.9–3.0) 1.64⁎⁎⁎(1.3–2.1) Solo—not involved with the father 4.57⁎⁎⁎ (3.8–5.4) 2.99⁎⁎⁎ (2.4–3.6) 1.52⁎⁎⁎ (1.3–1.9) ⁎⁎ p<0.01p<0.01. ⁎⁎⁎ p<0.001p<0.001. Table options Turning to breastfeeding, Table 4 shows the relative risks for the groups of unmarried mothers in different partnership settings relative to married mothers for whether they never breastfed, or breastfed for less than 6 months, compared with those who breastfed for 6 months or longer. Relative to married mothers, there is a gradient between the degree of bonding between parents and the extent to which babies are ever breastfed. In the unadjusted model, the relative risk of a cohabiting mother not breastfeeding are 3.1 times greater than that of a married mother, and the risks for solo mothers are 5.4 times and 6.7 times, respectively for those who were in, and not in, a relationship with the father. An adjusted Wald test showed that the differences in the estimates between the cohabiting group and the two groups of solo mothers were significantly different (cohabiting vs. closely involved, F-test=18.67, p-value <0.01; cohabiting vs. not in a relationship, F -test=39.25, p<0.01p<0.01) and the difference between two groups of solo mothers was smaller and not statistically significantly different (F-test=2.66, p-value=0.10). After adjustment for maternal socio-demographic factors, the risks are reduced, suggesting that these factors account for some of the observed differences between the groups of women, although differences in ever breastfeeding between some groups of women continue to be statistically significant and the test for trend by hierarchy of bonding was also significant (p-value <0.01). The analogous adjusted Wald tests are: cohabiting vs. closely involved, F-test=10.9, p-value <0.01; cohabiting vs. not in a relationship, F-test=13.93, p-value <0.001; and there was no difference between the two groups of solo mothers, F -test 0.82, p=0.36p=0.36. In Model One with no adjustments, and in Model Two with adjustments for maternal socio-demographic factors, relative to married mothers, all of the unmarried groups have an increased risk of breastfeeding for less than the recommended 6 months compared to breastfeeding for 6 months or longer. None of the pair-wise group comparisons in the adjusted model are statistically significant (adjusted Wald tests: cohabiting vs. closely involved, F-test=0.35, p-value=0.55; cohabiting vs. not in a relationship, F-test=0.03, p-value=0.85; closely involved vs. not in a relationship, F -test 0.11, p=0.74p=0.74), although the test for trend of hierarchy of bonding with duration of breastfeeding is significant (p<0.01p<0.01). Table 4. Breast-feeding behaviour according to partnership status at birth: Relative risk ratios and 95% confidence intervals from multi-nomial regression analysis Model 1 Never vs. breast fed for 6 months or more Breast fed for 6 months or less vs. breast fed for 6 months or more Never breast fed vs. breast fed for less than 6 months Model 1 Partnership status at birth Married 1.00 1.00 1.00 Cohabiting 3.10⁎⁎⁎ (2.7–3.5) 1.84⁎⁎⁎ (1.6–2.1) 1.68⁎⁎⁎ (1.5–1.9) Solo—closely involved with the father 5.36⁎⁎⁎ (4.2–6.9) 2.08⁎⁎⁎ (1.7–2.6) 2.56⁎⁎⁎(2.2–3.0) Solo—not involved with the father 6.74⁎⁎⁎ (5.3–8.6) 2.13⁎⁎⁎ (1.7–2.6) 3.17⁎⁎⁎(2.7–3.7) Model 2 Partnership status at birth Married 1.00 1.00 1.00 Cohabiting 1.59⁎⁎⁎ (1.4–1.8) 1.27⁎⁎⁎ (1.1–1.5) 1.24⁎⁎⁎ (1.1–1.4) Solo—closely involved with the father 2.36⁎⁎⁎ (1.8–3.0) 1.37⁎⁎⁎ (1.1–1.7) 1.72⁎⁎⁎(1.4–2.1) Solo—not involved with the father 2.73⁎⁎⁎ (2.1–3.6) 1.30⁎ (1.0–1.7) 2.10⁎⁎⁎ (1.7–2.5) ⁎ p<0.05p<0.05. ⁎⁎⁎ p<0.001p<0.001. Table options Models estimating the impact of parental bonding on maternal post-natal depression are shown in Table 5 and paint a broadly similar picture to that seen for smoking during pregnancy and breastfeeding. As the degree of connection between the parents weakens, so the risk of a mother reporting having been low or sad increases. After taking into account the socio-demographic characteristics, compared with married mothers; cohabiting mothers were 1.3 times more likely to report that they had felt depressed, the solo mothers who were closely involved with father were 1.5 times more likely to do so, and the solo mothers who were not in a relationship were 1.7 times more likely to report being depressed. However, while the differences between the two groups of solo mothers and reported depression was not statistically significant, the test for trend of increasing depression with lower bonding was significant (p-value <0.01). Table 5. Mothers report of depression post birth, according to partnership status at birth: odds ratios and 95% confidence intervals from logistic regression analysis Mother depressed or not Mother depressed or not Partnership status at birth Model 1 Model 2 Married 1.00 1.00 Cohabiting 1.40⁎⁎⁎ (1.3–1.5) 1.29⁎⁎⁎ (1.2–1.4) Solo—closely involved with the father 1.74⁎⁎⁎ (1.5–2.0) 1.47⁎⁎⁎ (1.3–1.7) Solo—not involved with the father 2.01⁎⁎⁎ (1.7–2.3) 1.67⁎⁎⁎ (1.4–1.9) ⁎⁎⁎ p<0.001p<0.001. Table options It was reported above in Table 2 that compared with White mothers Asian and Black mothers were more likely to breastfeed and less likely to smoke. Given these differences in behaviour we investigated whether the closeness of the relationship between parents mattered in the same ways for these two sets of ethnic minority mothers. The analyses showed that with respect to breastfeeding that there was little evidence of a trend in the propensity to breastfeed and the closeness of the relationship between the parents amongst Asian and Black mothers, but that there was a trend with respect to giving up smoking. The findings in relation to breastfeeding suggest that cultural differences in breastfeeding habits may be more salient than a mother's family circumstances.