Final Report Summary - INNODEVELOPMENT (The Impact of Breastfeeding on Children’s Cognitive and Non-cognitive Development: One Step Closer to Addressing the Notion of Causality) Background and Main Objectives: Childhood cognitive and behavioural problems carry high social burden and economic cost to societies given the increased risks of school absenteeism/drop-out, lower educational achievement and occupational attainment, increased delinquency/criminality and substance abuse. Research suggests that prenatal and early prevention yields the greatest returns on investments given the positive sustaining impacts on children’s development. The literature on child cognitive and behavioural development would suggest that breastfeeding may be a potential target for investment given the seemingly protective effects observed. The World Health Organization (WHO) has put forth recommendations for exclusive breastfeeding during the first six months following birth, and partial breastfeeding for up to two years of age to promote the healthy growth and development of infants. However, large debate remains ongoing as to whether or not breastfeeding is casually related to children’s cognitive development in particular, given the reliance on observational studies, and little is known about the ‘effects’ of breastfeeding on children’s language and behavioural problems. This EU-funded project, InnoDevelopment (The Impact of Breastfeeding on Children’s Cognitive and Non-cognitive Development: One Step Closer to Addressing the Notion of Causality), aimed to capitalize on the use of a quasi-experimental approach using a large national cohort study in an effort to better understand potential ‘causality’ associated with breastfeeding and children’s cognitive and non-cognitive outcomes. The project also sought to better understand the micro-level determinants associated with mother’s selection into breastfeeding engagement and to compare these determinants across cohorts in Europe (France) and North America (Canada). Our aim here was to contribute to findings within a more intercontinental context, which may help to promote international growth and programming for increased breastfeeding engagement. Finally, Innodevelopment also examined the influence of maternal parenting practices towards children’s cognitive and non-cognitive outcomes. Cohort(s): To address the main objective, the Growing Up in Ireland (GUI) cohorts were used. The GUI is the most significant longitudinal study of its kind in Ireland related to child development that commenced in 2007 with two nationally representative cohorts of children. The first cohort (the Child Cohort) is comprised of 8, 500 children and their families, assessed when children were nine and 13 years of age (the first phase) and the second cohort (the Infant Cohort) is comprised of 11, 134, children and their families, assessed when children were nine months, three and five years of age. To address the second objective, the Quebec Longitudinal Study of Child Development (QLSCD; Quebec) and the Étude des Déterminants pre- et postnatals précoces du développement et de la santé de l’ENfant (EDEN; France) were used. The QLSCD included ~2,000 mothers selected from the Quebec Birth Registry, who gave birth to singletons between 1997 and 1998. For EDEN, ~2,000 pregnant women who were 18 years or older and less than 24 weeks of gestation, were recruited from two university hospitals between 2003 and 2006.To address the final objective, we used data from 173 children and their families enrolled the Preparing for Life (PFL) programme in Ireland. The PFL programme is a prenatally commencing early intervention and prevention programme, concentrated in several disadvantaged communities with high rates of unemployment, welfare dependency and low levels of education in Northern Dublin, to help promote children’s school readiness. The trial was registered with controlled-trials.com (trial number: ISRCTN04631728) and was conducted and reported in conformity with CONSORT guidelines.Main Results and Conclusions.1) The impact of breastfeeding for children and adolescents’ cognitive and non-cognitive development:We focused on two areas of cognitive development in early childhood - namely non-verbal reasoning and expressive vocabulary. We also focused on two areas of externalizing behaviours in early childhood, late childhood, and early adolescence – hyperactivity and conduct problems. These two GUI studies used a multi-informant repeated measures approach. We made use of propensity score matching (PSM), which matches children who were breastfed to those who were not, but who have a similar probability of being breastfed based on their measured characteristics, consequently diminishing group differences and mimicking random assignment. The results from the Infant Cohort, published in Pediatrics (2017), revealed benefits in favour of children who were breastfed for any duration, either exclusively or partially, on essentially all cognitive and non-cognitive outcomes at ages three and five, prior to matching. Following matching and adjustment for multiple testing, none of the initial observed ‘effects’ of breastfeeding remained statistically significant. This is with one exception, whereby children who were breastfed exclusively for a minimum of six months had lower maternal-rated hyperactivity at age three as compared to children who were not breastfed. This result was not found at age five. Similar results were found in the Child Cohort (research in submission) whereby benefits of any duration of breastfeeding on reduced hyperactivity and conduct problems (both maternal and teacher ratings) were found prior to matching. Post matching results revealed effects of breastfeeding on reduced maternal rated hyperactivity and conduct problems at age 9, but not 13, if the child had been breastfed a minimum of six months. Effects of sex analysis revealed that it was the boys in the sample driving the effect on hyperactivity. These results were not however replicated when using teacher reports.Taken together, our results contribute to the current state of knowledge in important ways. First, the difference in ‘effects’ pre and post matching highlight the importance of using quasi-experimental statistical approaches that help to reduce selection bias between groups, which may have been driving the associations that are typically found in observational studies. Second, the importance of a multi-informant approach is also suggested given that the results were not replicated across informants. Third, while we were unable to find support for a direct causal effect of breastfeeding on cognitive development (i.e. non-verbal reasoning and vocabulary) in early childhood using the GUI Infant cohort, our findings would suggest that longer durations of breastfeeding may in fact help to reduce externalizing behaviours (i.e. hyperactivity and conduct problems) in the short term. These benefits however, do not appear to be sustained in the long term (i.e. into adolescence when other factors may exert a more influential role in emergence and/or continuation of externalizing problems). Fourth, the benefits found for externalizing problems were in line with the WHO recommendations for exclusive breastfeeding up to six months and partial feeding thereafter, given that no effects were observed for children who were breastfed less than six months. Finally, it is important to note that our findings in no way contradict the many medical benefits of breastfeeding found for both infant and mother’s health.2) Micro-level determinants of selection into breastfeeding engagement - A multi-cohort perspective:Examination of breastfeeding initiation rates across regions would reveal large variation. While cultural and/or geographical regions may influence particular micro-level determinants of breastfeeding initiation, the benefits of breastfeeding to all societies make gaining a better understanding of these factors, an important goal for research. In 2003, the WHO suggested that both clinical and population-based research ought to focus on findings pertaining to breastfeeding references for international growth, which are critical to the long-term global goals of increasing breastfeeding engagement. We focused on micro-level determinants of breastfeeding initiation (i.e. maternal, familial and medical factors) using two French-speaking cohorts from developed countries to better understand differences in determinants of maternal selection into breastfeeding. Pearson’s chi-square tests, logistic regression, and linear contrasts were estimated. Heterogeneity between cohorts was examined through interaction terms. The results, published in PloS ONE (2016), revelaed that initiation rates across cohorts were similar (i.e. 72% and 73% in Quebec and France respectively). For mothers in both cohorts, having higher levels of education, not smoking during pregnancy, and being born outside of their respective countries increased the odds of initiating breastfeeding with their infant. Notably, effects of maternal education differed by cohort at the university level, whereby having ‘some university’ was not statistically significant for mothers in France. Differences in breastfeeding determinants by cohort were also found. For example, being a younger mother, delivering by caesarean section and having had previous children all reduced the odds of breastfeeding initiation for mothers in Canada. These results were not found for mothers in France. We conclude that while we observed a few similar determinants between mothers in Canada and France, programming efforts to increase breastfeeding initiation ought to be tailored to the characteristics of specific geographical regions, which may be heavily impacted by the social, cultural and political climate of the region, in addition to individual and family level factors.3) Impacts of maternal warmth on toddler’s development - support for transactional models in disadvantaged communities:Cognitive and non-cognitive development in childhood are argued to be largely intertwined, with bidirectional impacts of one on the other. Moreover, research continues to demonstrate the impacts of parenting behaviours on children’s cognitive and non-cognitive development, particularly in early childhood, prior to entrance into formal schooling. However, there remains a dearth of research examining the mother-child relationship within a reciprocal interdependence framework (i.e. the cognitions, emotions or behaviours of one will affect those of the other and vice versa). Studies focusing on these transactional processes within families from impoverished communities during the toddler years are limited. Further, a majority of studies that do examine transactional models between mother and child, focus on harsh or hostile parenting rather than the impacts of warm or positive parenting practices. In this study we focused on children’s emerging communication abilities and social competence starting in the early toddler years. The study used a repeated measures approach with assessments of communication, social competence, and maternal warmth at 6, 12, 18, 24 and 36 months. We implemented cross-lagged path analysis, which is advantageous in its ability to test multiple directional paths of association simultaneously, to first examine the bi-directional impacts of emerging communication ability on social competence and vice versa. We also examined indirect and direct paths of association in testing for mediating and transactional effects of maternal warmth.Our findings, published in European Child & Adolescent Psychiatry (2016), suggested that between 12 and 24 months, the paths of association are bi-directional. That is, toddlers with good communication abilities at 12 months had higher levels of social competence at the following assessment period (i.e. 18 months) and in turn they also had better communication ability by 24 months. These paths of association were also found in the other direction (i.e. higher levels of social competence at 12 months to better language at 18 months). Surprisingly, no support for a mediating role of maternal warmth was found. However, we found support for a transactional model of reciprocal interdependence with maternal warmth and toddler’s social competence, whereby social competence at 18 months was significantly associated with increased maternal warmth at 24 months and in turn, maternal warmth at 24 months was significantly associated with a higher level of social competence for the toddler at 36 months. These results suggest that more socially competent toddlers may be better able to engage during dyadic interactions, resulting in better quality and more enjoyable interactions for the mother. This may help to promote more maternal satisfaction and increase a mother’s desire to engage in more nurturing behaviours during activities with her toddler. Noteworthy, our findings are of interest in the context of the sample characteristics, (i.e. almost 2/3 of the mothers were single during pregnancy, over 1/2 of them were not employed during pregnancy and only 3.5% of mothers had a higher-level education). Thus, a toddler’s level of social competence may be particularly important to the response styles of low SES mothers whose resources (personal or financial) are already limited and who may be experiencing higher levels of drain as a result. Our findings would suggest that not only do parents exert influential impacts upon their toddler’s behaviours but also that toddlers’ behaviour can exert positive influences on a mother’s expressed levels of warmth towards her child. This ought to be considered during prevention and/or intervention programming efforts.
