Trade‐offs underlying maternal breastfeeding decisions: a conceptual model

Abstract

This paper presents a new conceptual model that generates predictions about breastfeeding decisions and identifies interactions that affect outcomes. We offer a contextual approach to infant feeding that models multi‐directional influences by expanding on the evolutionary parent–offspring conflict and situation‐specific breastfeeding theories. The main hypothesis generated from our framework suggests that simultaneously addressing breastfeeding costs and benefits, in relation to how they are interpreted by mothers, will be most effective. Our approach focuses on contributors to the attitudes and commitment underlying breastfeeding outcomes, beginning in the prenatal period. We conclude that some maternal–offspring conflict is inherent with the dynamic infant feeding relationship. Guidance that anticipates and addresses family trade‐offs over time can be incorporated into breastfeeding support for families.

Introduction

Infant feeding has required extensive maternal time and energy throughout mammalian, primate and hominin existence. Breastfeeding is a dynamic process comprising regulation between the mother's and infant's interconnected physiological, psychological and behavioural systems (Winberg 2005). Lactation can be understood as the final stage of labour (Labbok 2001) and as the physiological completion of the woman's current reproductive cycle (Lawrence & Lawrence 2005).

Despite the importance for child health, exclusive breastfeeding is globally rare (World Health Organization 2010) and this carries a huge disease burden (Labbok et al. 2004; Horta et al. 2007; Black et al. 2008; Hauck et al. 2011). Current costs of suboptimal breastfeeding related to paediatric disease in the United States are estimated to be billions of dollars and hundreds of lives per year (Bartick & Reinhold 2010). Although there are continual improvements in the composition of formula, human milk is biologically superior and clinically optimal in the vast majority of circumstances (World Health Organization 2009). The American Academy of Pediatrics confirms that all infant feeding substitutes differ ‘markedly’ from the species‐specific human milk (2005, p. 496).

In an evolutionary framework, breastfeeding benefits mothers because it promotes the health of their offspring and themselves. In addition to the parental gratification that can be conferred through the breastfeeding relationship (Dykes & Flacking 2010; Rempel & Rempel 2010), lactation has short‐ and long‐term physical effects on the mother (Blackburn 2007; Stuebe & Schwarz 2010), including delayed resumption of fecundity (Bellagio Consensus Statement 1988; Valeggia & Ellison 2009). Infant feeding plays a vital role in maternal and child health, yet there is low adherence to medical recommendations and personal goals are often unrealized. Many women who intend to breastfeed supplement with formula or terminate breastfeeding in the early post‐partum period (Lavender et al. 2005; Grummer‐Strawn et al. 2008; Declercq et al. 2009; McQueen et al. 2011). Recognition of possible asymmetries in the costs and benefits between the dyads may be a key for enabling better initiation rates and facilitating maintenance of the breastfeeding relationship.

Infant feeding is among the most intensive aspects of parenting, so maternal strategies for breastfeeding will be adopted in ways consistent with expectations of pay‐offs (Tracer 2009). Strategies are embedded within cultural expectations, affected by environmental constraints and influence many aspects of families’ lives (Lavender et al. 2006). Not surprisingly, parental attitudes are increasingly identified in the literature as central to breastfeeding outcomes (e.g. Bai et al. 2010; Dyson et al. 2010; Wojcicki et al. 2010), as are hospital practices that affect maternal access to newborns, and women's views towards infant feeding (e.g. Merten et al. 2005; Bartington et al. 2006; Abrahams & Labbok 2009; Cramton et al. 2009; Declercq et al. 2009). Many women report having both positive and negative feelings about breastfeeding (Forster & McLachlan 2010; Andrew & Harvey 2011), which suggests that mothers continually perform ‘balancing acts’ with infant feeding.

Frequent breastfeeding is biologically appropriate for infants because of their small stomachs and the low solute composition of human milk (Blackburn 2007). However, this is associated with fragmented maternal sleep (Tikotzky et al. 2010), which is a major concern of many parents (Sadeh et al. 2011). Feeding occupies a large proportion of the infant's waking time and much of early mother–infant interaction. Therefore, the feeding experience has consequences for the dyad's overall relationship (Pearson et al. 2011). Similarly, mother–infant interactions have consequences for the feeding experience, such as assessment of and response to infant cues regarding hunger and satiety. Thus, infant feeding trade‐offs may be expected to change over time as the dyadic relationship changes, with consequences for breastfeeding outcomes.

Maternal intent plays a central, yet inadequately understood, role in breastfeeding (Nommsen‐Rivers et al. 2010b). That families’ prenatal ideas and reasoning affect infant feeding outcomes is well known (Alexander et al. 2010; MacGregor & Hughes 2010; Nommsen‐Rivers et al. 2010a), yet breastfeeding plans are not static. Only a few studies have begun to address the ways in which parents rework their attitudes and subsequent behaviour in response to infant cues (e.g. Mentro et al. 2002; Mizuno et al. 2004; Howard et al. 2006; Hodges et al. 2008) and the changing maternal landscape (e.g., Bai et al. 2010; Burns et al. 2010; Sheehan et al. 2010). Attention to the infant feeding issues most salient to mothers has led to the recurrent Western theme of breastfeeding requiring perseverance (Burns et al. 2010) because the process can involve overcoming or circumventing psychological, practical and social obstacles (Stewart‐Knox et al. 2003). Overall, decision criteria for the initiation of breastfeeding may often be very different than reasons for its maintenance (Rothman 2000).

Evolutionary life history theory predicts that – whether consciously or not – organisms prioritise resources based on predicted costs and benefits over their lifespan. This is ultimately because growth, maintenance and reproductive effort are sometimes conflicting pursuits (Bentley 2007). Reiches et al. (2009) summarize this ‘energy budget’:

It is adaptive to commit to these expenditures only when prospects for success are reasonable and only to a degree that optimizes lifetime reproductive success (p. 442).

The prediction is that animals, including humans, will preferentially invest in close relatives, individuals with perceived high potential for future reproduction and those who incur relatively low costs (Strassman & Mace 2008).

Parents and offspring inherently confront some conflicts of interest because offspring only share a portion of their parent's genes. The infant strives to be as healthy as possible without draining the caregiver to a degree that she/he can no longer invest; the parent strives to raise healthy offspring that survive to reproduce at a minimal cost (Darwin 1871; Trivers 1974; Haig 1993, 2008; Vitzthum 2008). These parent and infant strategies are largely subconscious – they are not scheming to take advantage of one another (Tracer 2009). Rather, these cognitive and behavioural patterns were selected over the course of evolutionary history to optimize inclusive fitness. This facultative or contingent response has been analysed in relation to infant feeding by various anthropologists, including Scheper‐Hughes (1992), McDade & Worthman (1998), Ball & Panter‐Brick (2001), Worthman & Kuzara (2005) and Sellen (2007). Mothers have the option of expending a portion of their finite time and effort towards breastfeeding, or they can employ alternate pathways for infant feeding such as human milk substitutes, donor human milk, wet nurses and/or lack of engagement.

The model we propose below is a tool to conceptualise inherent breastfeeding trade‐offs and to illustrate how this balance can be altered by exogenous and endogenous factors. Our approach is consistent with the situation‐specific theory of breastfeeding (STB) (Nelson 2006) in that balancing is modelled as occurring within the mother–infant dyad and between the dyad and their broader network of relationships:

Simultaneous consideration of the parts and the whole, that is, the individual mother/infant dyad and the broader breastfeeding context, is necessary, as is attention to our approach to breastfeeding interventions and examination of our perceptions of the professional role (p. 23).

We integrate parent–offspring conflict theory and the STB to assist in the construction of questions and methods to better understand the multi‐directional influences that contribute to women's strength of breastfeeding intent and the continual feedback affecting their perseverance.

Key messages

• • Some maternal–offspring conflict is inherent within the dynamic infant feeding relationship.
• • A new trade‐off model is offered that generates predictions about breastfeeding decisions and identifies interactions that affect infant feeding outcomes.
• • Simultaneous reduction in the costs and increase in the benefits of breastfeeding (and maternal perceptions of these) will be most effective in facilitating breastfeeding.
• • Explicit acknowledgement of individual families’ trade‐offs with infant feeding over time may aid in the development of improved promotion and support strategies.

Model

Figure 1 expands the parent–offspring conflict model put forth by Trivers (1974) to illustrate breastfeeding costs and benefits among individual dyads over a specific period of time. This model illustrates that trade‐offs underlie infant feeding decisions (investment) and this figure enables predictions based on marginal returns (the degree to which breastfeeding is ‘worth it’ given the context).

Figure 1.

Theoretical mother–infant breastfeeding trade‐offs over a specific period of time (expanded from Trivers 1974).

The degree of investment, comprising both time and effort, that a mother could devote to breastfeeding a particular infant is portrayed on the X‐axis. The benefits to the infant and the costs to the mother of the various degrees of breastfeeding input are depicted by the Y‐axis. The variables in this model are defined by the mother's perceptions of the benefits and costs in addition to the physical effects. The way in which units are measured would depend on the defined time period (investment over hours, weeks, years, etc.) and on the scope of benefits and costs measured (physiological regulation, health outcomes, satisfaction, etc.). Although Trivers conceptualised parental investment as encompassing the feeding of young, in his model cost was only gauged in relation to inclusive fitness (ability to raise offspring who, in turn, reproduce). This new model therefore applies the concept to the specific investment of breastfeeding.

Figure 1 illustrates the marginal cost and benefit of various levels of breastfeeding investment over a defined time. The main point is that over a certain period, ceteris paribus, the optimum investment is at I₁ for the mother but it is at the greater level of I₂ for the infant. This model holds for circumstances in which no medical breastfeeding contraindications exist (e.g. American Academy of Pediatrics 2005; World Health Organization 2009).

The absence of maternal investment results in zero benefit to the infant because (for the sake of simplicity) the model assumes that maternal time and energy invested is ‘measured’ in terms of human milk ingested, which is advantageous for the infant. Maternal cost does not intercept the Y‐axis at 0 because of the physical benefits that lactation provides for women. Maternal benefit is built into the model, in that a change in maternal cost represents an equal but opposite change in maternal benefit. The biopsychosocial context in which breastfeeding occurs interacts to create different slopes, and therefore different optima, for individual dyads. Yet, for all applications there is theoretically a peak in the benefit to the infant, shown in Fig. 1 at B₂. Past this point he/she would not breastfeed any more beyond the particular time period if given the opportunity. Although not shown in Fig. 1, the benefit to the infant would eventually curve back down if maternal costs reached a level that resulted in maternal depletion (see Jasienska 2009), which would eventually detrimentally affect the infant's condition. For all women, there is theoretically a maximum ‘profit’ where the difference between benefit to the infant and the cost to herself is greatest. This point (labelled I₁ on Fig. 1) is the level of maternal investment at which she is able to provide the greatest benefit to the infant at the lowest cost to herself.

Trivers (1974) contended that parents do not invest indiscriminately; both his and our models predict different optima for parent and child. With reference to Fig. 1, the mother will subconsciously resist investment beyond I₁ because the additional time and effort incurs a greater cost to herself (C₂–C₁) with only modest additional benefit to her infant (B₂–B₁). This tendency arises because of subjective utility maximization in the face of uncertain outcomes in return for the investment (Salehnejad 2007; Sloman & Wride 2009).

The hypotheses generated from this model are: H1: Reduction in maternal cost (or perception of cost) promotes breastfeeding, while holding infant benefit constant. H2: Increase in infant benefit (or perception of benefit) promotes breastfeeding, while holding maternal cost constant. H3: Reduction in maternal cost and increase in infant benefit (or perceptions thereof) will be more effective than H1 or H2 in promoting breastfeeding.

Known breastfeeding influences can be modelled as shifting the model's maternal cost and infant benefit lines. For example, maternal knowledge of infant health as being improved by breastfeeding (shifts the infant benefit line upwards and right); maternal perception of infants as uninterested in feeding or that breastfeeding does not satisfy the infant (shifts the infant benefit line downward and left); maternal tiredness, latching difficulty, pain, embarrassment, perception of formula feeding as being ‘easier’ than breastfeeding, or advice from people important to the mother to supplement or not breastfeed (shifts the maternal cost line left); and maternal perception of lactation as providing health benefits to herself, of breastfeeding as a positive experience, or of frequent infant breastfeeding as expected and ‘normal’ (shifts the maternal cost line right). This list is not comprehensive; the purpose here is to suggest how conceptualizing maternal feeding experience in the form of a cost‐benefit model might consolidate existing knowledge and offer testable hypotheses for the development of breastfeeding interventions.

Discussion

The conceptual model focuses on inherent trade‐offs in the breastfeeding landscape and asserts that mothers repeatedly renegotiate the balance between self and child care. Certain decisions will be conscious but many are likely to be mediated by our evolved psychology to maximize marginal returns on investment and therefore occur semi‐ or unconsciously, and may be rationalised in a variety of ways.

Our model avoids needless complexity (see Foster 2010) and the simplicity of the model is advantageous because the need for balance between maternal investment and returns is clear. Defining the units of measurement when testing the theoretical predictions will require careful consideration of the questions posed. The figure complements existing Venn diagrams, flow charts and other multi‐level representations of infant feeding influences (see Martens & Young 1997; Tiedje et al. 2002; Hector et al. 2005; Nelson 2006; Labbok 2008; Sheehan et al. 2010). It builds on the situation‐specific theory of breastfeeding by offering straightforward predictions about breastfeeding decisions and outcomes under various conditions. The components we propose as shifting breastfeeding cost and benefits map directly on Bandura's (2004) core determinants of social cognitive theory for health promotion: knowledge of risks/benefits, perceived self‐efficacy, outcome expectations and perceived impediments/facilitators.

This paper suggests public health messaging may be most effective by explicitly addressing family dynamics in infant and young child feeding. Specifically, instead of focusing on how ‘babies are born to be breastfed’ or separate ads on maternal health consequences, all promotions could test including at least the dyad. For example, ‘exclusive breastfeeding is the healthiest start for us both’ and ‘human milk and skin‐to‐skin contact help us all get to know each other’. Communication of empowering messages may raise maternally perceived benefits or reduce maternally perceived costs such as ‘frequent breastfeeding is normal, including through the night’. Physiologically based information on the demand and supply nature of lactation and evidence‐based suggestions for ameliorating sleep disruptions can be offered, instead of sidestepping these types of prevalent concerns. Tipping the balancing scales towards biologically optimal feeding necessitates practical, cultural and emotional support, not stigmatizing discourse. Lupton (2000) found that although breastfeeding is often an integral part of mothering ideals, women reported that it is difficult to achieve. Infant feeding is more than a health issue and mothers commonly explain deviation from breastfeeding plans as a pragmatic solution to make things easier (Lee 2007).

The choices contemporary mothers have for infant feeding methods (and constraints imposed on dyads) often lead to biologically suboptimal infant and young child feeding. Anything that decreases breastfeeding burdens, increases satisfaction and/or promotes awareness of the benefits should influence outcomes, but program effect sizes will likely be small in the absence of comprehensive support. Interventions can address internal processes such as promoting healthful infant or maternal functioning as well as external factors such as more extended and paid maternity leave. Although adherence to the 10 Steps (World Health Organization 1989) is helpful to minimise iatrogenic disruption of breastfeeding relationships, health care practitioners could take breastfeeding support to the next level by routine provision of information on culturally appropriate and affordable community resources. Bandura (2004) suggests interactive computer‐assisted feedback for individualized communication, for instance. This means that support could be tested within individual women over the course of their pregnancy, perinatal and post‐partum experiences.

The human and financial costs of suboptimal infant feeding are enormous. Yet in the United States, institutionalised lactation support such as a national guideline for hospital lactation consultant staffing does not exist (see Mannel 2010) and breast pumps were first made tax deductible only in 2011 (IRS, 2011). More substantial financial and time investment during the prenatal and perinatal stages should prevent even greater financial and human costs during the latter post‐natal period and across the lifespan. Future investigations could focus on how we better promote interactions conducive to breastfeeding, including skin‐to‐skin contact, rooming‐in arrangements and consistent messaging within and across health care providers.

It is not surprising that conveying ‘breast is best’ is inadequate for the reality of exclusive breastfeeding for at least 6 months. Initiation rates are relatively high, but the mechanisms governing breastfeeding outcomes may be improved by person‐centred, repeated measures studies. In this way, subgroup trajectories – especially the periods of greatest vulnerability – can be identified and anticipated. Attention to the interaction of both endogenous and exogenous factors on infant feeding over time, such as prenatal expectations, childbirth events, infant cues, maternal conditions, social support and the physical environment, is essential. Our approach is consistent with the systems perspective of the developmental science framework (Magnusson 1988; Magnusson & Cairns 1996; Lerner et al. 2005) and the concept of equifinality – that the same end state (e.g. breastfeeding outcomes) can occur through ‘a variety of different initial conditions and through different processes’ (Cicchetti & Rogosch 1996, p. 597). Various factors are known to impact initial infant feeding decisions, but reworking of infant care motives and goals and the weight of particular factors at different time points is less well explored. This paper is a contribution to this exploration.

Conclusion

A more holistic understanding of infant feeding decisions and the dynamic nature of these cost‐benefit influences over time is vital. Explicit acknowledgement of maternal, family and broader trade‐offs with breastfeeding may guide translational research, lead to more realistic prenatal breastfeeding discussions and promote more effective post‐partum support of desired infant feeding trajectories.

Our framework suggests that promoting only infant benefits, such as with the ‘breast is best’ public health message, without comprehensive maternal, family, institutional and other support is insufficient.

Source of funding

Kristin P. Tully is supported by the National Institutes of Child Health and Human Developmental Training Grant to the Center for Developmental Science at the University of North Carolina at Chapel Hill, 5T32HD007376‐22.

Conflicts of interest

The authors declare that they have no conflicts of interest.

Contributions

KP Tully conceived the model while conducting doctoral research with the supervision of HL Ball. Both authors helped to formulate ideas, KP Tully led the writing, and both authors reviewed drafts of the manuscript.