Cost-Effectiveness and Equity Trade-Off for Breastfeeding Interventions

Abstract

Many factors influence mothers’ decisions to begin and continue breastfeeding (BF). These include individual, societal and policy factors. In this chapter, we address these factors including the social differences in BF practice among Irish women and discuss important policy implications (efficiency and equity). It is well-documented that BF practice is different for mothers with different social backgrounds. Traditionally, evaluations of BF support interventions have focused on either the effectiveness or the equity issues, but rarely analysed both in a joint framework. The aim of this chapter is to discuss the cost-effectiveness and equity trade-off for BF interventions. We identify different BF support interventions and focus on social differences and their influence for maintaining BF practices. We illustrate how the “Distributional cost-effectiveness (DCEA) framework” can be applied to these interventions and how some interventions may be more effective in changing behaviour and outcomes for mothers with different social-economic status (SES), which may change the inequality in effectiveness and reduce the health equity.

1. Introduction

Ireland has the lowest breastfeeding (BF) rate among Western countries at about 61% compared to rates of 94% in Sweden, 95% in Australia, 81% in the UK, and 79% in the USA [1, 2, 3, 4, 5]. The Irish proportion of babies being exclusively BF on discharge from hospital is even lower at 44% [3]. Figure 1 shows a slowly growing proportion of babies are breastfed on discharge from hospital from 2011 to 2020, although around two out of five babies are still being fed with artificial feed. Ireland has introduced many BF strategies and actions such as; the European Perinatal Health reports, reports documenting BF actions from collaborations with the WHO Baby Friendly Hospital Initiative and the BF in the Healthy Ireland Health Service Executive Action Plan 2016–2021 [6]. However, Irish BF rates still continue to lag behind Western counterparts.

Figure 1.

Types of infant feeding at discharge from hospital, Ireland 2011–2020.

It is widely recognised that BF provides health benefits to both the baby and the BF mother [7]. It is reasonable to assume that there will be some form of exposure-response relationship, although it may be difficult to quantify based on observational data. Indeed, systematic reviews have identified that studies use different methodologies to analyse this relationship, and the empirical results show great variation, dependent on the intensity (exclusive or partial) and duration of the BF practice [8]. Systematic reviews have reported evidence that suggests that babies who are exclusively breastfed over 6 months have lower risks of developing infections and other complications in comparison with babies who have been exclusively fed with artificial feeding [9]. Studies have identified that BF reduces the risk of a range of infectious diseases and BF has a preventive effect against developing obesity [10].

The types of infectious diseases being present from mothers who don’t BF versus those who do include otitis media, gastroenteritis, childhood pneumonia, as well as childhood obesity, type 1 and type 2 diabetes and cancers [8, 11]. The duration of BF is essential in terms of the child health outcomes. BF for greater than 6 months has protecting influences from childhood overweight/obesity [10]. Child intellectual activity such as reading and remembering has been shown to improve with higher rates and longer BF duration [12]. A recent study in 2022 demonstrated that longer BF duration was associated with improved cognitive score of children from age 5 up to age 14 [13]. Breastmilk contains antibodies and protective factors that transfers cytokines and growth factors to the infant which can stimulate the infant’s immune system and reduce the risk of infection. Lactoferrin has anti-inflammatory, antioxidant, antiviral and antimicrobial effects and lactoferrin supplementation has been shown recently to reduce the cytokine IL6 and associated systemic inflammation [14]. BF over a long period reduces the risk of breast cancer in comparison with mothers who have not been BF their babies [15].

Beyond the health benefits of BF, a higher BF rate can offer substantial financial savings in the long term by reducing infant infections such as gastrointestinal and respiratory infections as well as otitis media, necrotising enterocolitis and breast cancer, which are all significantly reduced in women who are BF the infant [16]. In addition, in a review article from 2022 we identified BF as a mediator for youth adiposity in Ireland and the UK [17]. In the UK for example the cost to the economy from infant infections alone was £89 million and approx. £959 million for breast cancer treatments [18, 19]. Not many papers have studied the cost-effectiveness of promoting BF in Ireland, but one reported that BF for low-birth weight babies was associated with lower costs and greater health benefits [20].

2. Factors relating to Irish women’s decision to BF

Irish women’s decision to BF are influenced by individual, societal and political factors as illustrated in Figure 2.

Figure 2.

Factors influencing BF decisions.

Individual factors relating to BF decisions include women’s perceptions of healthcare professionals (HCP), level of education and awareness among others. Building trust with HCPs is important for women to feel more supported in their BF choice and more in control during their pregnancy. Individual factors are influenced by women’s SES. SES reflects an individual’s position in society and is determined by an individual’s social and economic status such as income, education level, occupation and area of residence. BF rates vary between urban and rural settings. This variation is partly explained by the level of maternal education and household income [21]. In Ireland, women living in socioeconomically disadvantaged areas are less likely to BF compared to those living in a more socioeconomically advantaged area, and they are also less likely to continue BF [22]. The proportion of babies that are artificially fed are lowest for women in higher professional work, indicative of educational attainment (Figure 3).

Figure 3.

Proportion of babies fed with artificial feeding by mothers’ socioeconomic status, Ireland, 2020.

Societal factors include advice offered to women from HCPs during pre and post pregnancy care. Immigrant women living in Ireland felt that HCPs “did not sufficiently stress the importance of BF in Irish hospital setting” [23]. They made a number of suggestions for governments and HCPs to support women to exclusively breastfeed in Ireland including BF counselling, more advice from HCPs, and that governments should develop better BF facilities in public places such as public parks or shopping malls. The health service executive (HSE) published the National Standards for Antenatal Education which identifies elements that should be included in all antenatal education programmes [24]. One of the themes of this was having an educated workforce to provide BF skills to women. More of these supports are needed for BF women in Ireland. In a study in three countries: Australia, Ireland and Sweden, exploring women’s perceptions of what best encouraged them to BF, the main factors were support and social networks which included face-to-face or online informal support, societal support from HCPs, and the supportive BF culture of the particular country [25].

Societal factors include the culture of a country which has an important role as to whether BF practices take place and for how long. Ireland’s BF culture reported women “feeling embarrassed” about BF in public [26]. In a comparison study with Sweden and Australia, Ireland ranked highest for having a “BF environment not suitable” [27]. The cultural views on BF practices differ in other EU states. Swedish culture for example is reported to be supportive towards public BF practices, where Swedish women found engaging in BF practices in public to be “easy and enjoyable” [28]. In order to create a BF-supportive culture it is essential to instil positive attitudes towards BF women early on and encourage continued self-efficacy, beliefs in their ability to breastfeed. Supportive environments that promote increased BF include, more social, family and partner supports. Women should feel empowered, supported and look favourability on BF practices in public.

Policy factors influencing BF decisions include: National policies i.e. maternal and parental leave policies, Irish marketing practices including the marketing of infant formula, BF promotion policies and policies related to the level of resources and supports provided to women returning to the work environment after leave [29]. Other policies are hospital/HCP guidelines i.e. the baby friendly hospital initiative (BFHI), educational policies i.e. healthcare staff educational requirements and individual policies such as those providing one-to-one support to women.

Paid maternity and parental leave policies have a role in BF decisions. Evidence has demonstrated that the longer the period of paid maternity or parental leave, the higher the rates of BF and the higher the odds of initiating BF in the first instance [30, 31]. BF has been associated with multiple positive benefits for both mother and child and it is recommended by the World Health Organisation (WHO) and HSE to exclusively breastfeed for a minimum of 6 months [32]. The duration of maternity leave and sharing parental leave, has also been associated with longer BF rates [33]. Ireland has the fourth shortest period of paid parental and maternity leave compared to other European countries [34]. In Ireland women can take 26 weeks (182 days) paid maternity leave with up to an additional 16 weeks of unpaid leave [35]. Maternity leave is paid by the Irish government or by employers, but employers in Ireland are not under obligation to pay maternity leave [35]. The duration of leave means that many Irish women feel compiled to begin work sooner than they wish leading to discontinuation of BF practices [36]. A study in three Organisation for Economic Co-operation and Development (OECD) countries ranked Sweden as the country providing the longest paid maternity and parental leave from all three countries and having the highest proportion of women returning to full time work after leave [37]. In contrast to Ireland, Swedish parental leave is paid up to 68 weeks (480 days) and each parent is entitled to 120 days each which they can either choose to divide between them or take solely by one parent [38]. The OECD report indicated that 77.8% of women return to full-time employment after parental leave in Sweden versus 61.5% in the US and just 44.4% in Ireland [37].

The marketing and advertising of infant formula can discourage women from BF which may contribute to the low BF rates in Ireland. Ireland is one of the biggest producers of infant formula in the world which results in higher advertising and marketing of infant formula contributing to lower BF practices [39]. The Code of Marketing of Breast-milk Substitutes (BMS) from the WHO states that marketing of breast milk substitutes should not be allowed up to 36 months, and EU food law states that companies are not allowed to make health claims about infant formula or compare breast milk and infant formula or imply it is superior to BF [40]. The Food Safety Authority of Ireland (FSAI) audits the labelling of infant formula in Ireland. However, breaches of milk formula marketing rules do occur. Milk formula manufacturers also sponsor certain helplines and other agencies and offer free gifts for health care workers (HCWs). The result of this is negligible on patients attending care during pregnancy. Governments and other policy-makers are urged to do more. A proposed amendment to the Online Safety bill was recently made to restrict marketing of infant formula and follow on formula [41]. This would legally restrict the marketing of follow-on infant formula as a proposed alternative to BF practice.

3. Interventions to support BF

Interventions to support BF include Individual, Societal, and Policy-based interventions.

For Societal (environment, education, counselling and support) interventions, different approaches have been studied. Some previous reviews indicated that methods such as one-to-one education, counselling and support provided to women over a long time period is an effective method of promoting BF [42, 43]. This may be even more effective for women of lower SES. Internet support may be a useful adjunct to face-to-face care. The settings where the interventions were conducted varied, as did the training which was provided to those performing the interventions. For peer support, there are considerable differences which include the study populations, the definition of peers, and the definition of counsellors, peer counsellor training protocols, peer visit schedules, and outcome ascertainment methods between trials. In the majority of studies, the interventions were compared to ‘routine care’, the definition of which seems to vary considerably between countries.

The acute maternity care settings are central for implementation of structured programmes to support BF. The content of such programmes could replace an existing programme, such as the BFHI, in full or in part, or be specifically developed to reflect local needs. A review of 21 studies (mostly observational) and five systematic reviews observed the introduction of a structured BF programme which offered some improvement in the duration of BF but these structured programmes were not always statistically significant [44]. A Cochrane review of 52 randomised control trials (RCTs) indicated that providing support to women during BF is effective in increasing the rates of exclusive BF [42, 45]. Having the appropriate supports in place would be effective in health care for increasing BF initiation, however support mechanisms should be offered at all heath care settings. Support may be offered by professional health care staff, through peer to peer support groups either in person or online. The media also plays a role in relaying evidence based supportive advice to women on BF. Support that is offered face-to-face is more likely to be the most effective.

Providing professional support to women has been shown to promote BF for a longer period of time [46]. Support should be tailored to the setting and the needs of the population group. In the same Cochrane review, a subgroup analysis of two studies (162 women) evaluating the effect of repeated informal BF education programmes personalised to each woman’s needs showed a statistically significant increase in the number of women starting to breastfeed as a result of the intervention. The reviews that addressed interventions among adolescent mothers showed mixed results, but it is clear that peer support and educational interventions improve BF rates, especially when these are targeted at individuals.

Exclusive BF during hospital stay also significantly increases BF duration outside the hospital irrespective of SES [47, 48, 49]. In addition, regardless of SES, BF mothers generally need practical knowledge and experienced support in order to attain BF success, however disadvantaged mothers may require extra support in order to overcome BF problems [50, 51, 52]. Education for women should be aimed at increasing their knowledge of BF practice and to acknowledge the possible lack of exposure to BF they may be engaging in along with trying to allay fears that their baby will not be satisfied by breastmilk alone. All new mothers need support from professionals and others to reassure them that exclusively BF will provide sufficient nutrition for their infants. Education about the benefits of BF should also be extended to their partners and the community [53]. Developing healthcare professionals’ capabilities to educate disadvantaged groups, their social networks and the public about BF is crucial [54].

Educational resources and BF supports for women should include adequate lactation consultant visits. In New Zealand for example at 3,650 annual births, there were 3 full-time lactation consultants in the Women’s Health Service and 0.8 full-time in the Neonatal Intensive Care Unit at the hospital setting in 2017, which if applied to Irish figures of 62,039 births would equate to approximately 50 full time lactation consultants to meet demands in the Women’s Health Service [55]. In Ireland there are significantly less numbers of lactation consultants and there is a significant need for additional lactation consultants in hospitals, with some having no specialist lactation posts in lactation [55]. In order to increase the availability and affordability of lactation consultant services for BF women, more public funding should be made available. In addition, private demand could be increased by health insurance plans providing coverage for lactation consultant services. There are a number of volunteer support groups across Ireland that support women and offer guidance and advice on BF practices. Some of these support groups include: La Leche League of Ireland, Friends of Breastfeeding, and Cuidiú support groups, among others. Many of these groups were set up by women wanting a better understanding of BF and more supports and better accessibility to these supports for Irish women. While there are many groups setup, some areas of Ireland are still limited in terms of accessibility and location. More advice from HCPs, more resources and encouragement to attend these groups early in an infant’s stage of growth are needed.

Policy interventions can further support BF practices. A barrier is a lack of resources and supports to BF women returning to work. The Work Life Balance and Miscellaneous Provisions Bill (2022) was recently proposed in Ireland to provide parents with rights to better support a work life balance including, extending BF breaks by allowing women to take paid time from work each day for BF [56]. This will assist women and support BF practices but much more needs to be done [25]. BF duration and pressure to return to work due to poor pay is forcing women to stop BF early or to not begin at all [36].

For policy-based interventions under hospital and HCP guidelines, the BFHI guidelines indicate that interventions including structured programmes to promote BF are required as well as promoting early skin-to-skin contact (SSC), the practice of rooming-in, where the baby’s crib is kept beside the mother for mother-infant contact, and avoiding supplementary infant feeding are important [57].

4. Other factors to consider in relation to BF interventions

It is also important to consider other influencing factors such as the timing of the intervention with respect to the pregnancy, who delivers the intervention, the target group, and the use of e-health technology as an add on to in-person supports. With technology and telephone support forming an ever increasing integral part of healthcare delivery, the development of web-based, text messaging, electronic prompts and interactive computer interventions promote and support BF [58]. Technology for assisting women in public to find facilities and services for BF are very useful. Further investment in technology is also needed to support women to find facilities that are available in their area. A study conducted demonstrated the advantages of the “FeedFinder App” whereby women shared views on various facilities and level of privacy [59]. Women should feel empowered to breastfeed in public and technology also can have a positive role. Business owners may also become more motivated to have positively reviewed facilities as it may enhance their customer numbers and instil more awareness on the benefits of BF facilities in public.

5. Demonstrating health economic value of BF interventions

Whether BF interventions are offering better opportunities, better knowledge or developing a supportive environment, the end result of these programmes is determined by the women’s willingness to engage and continue BF. Evaluations of these interventions must therefore consider to which extent they are able to influence women’s BF inclinations. This may be observed in the proportion of women who begin BF and continue to exclusively breastfeed their babies which should be for at least 6 months as recommended in various guidelines and recommendations e.g. by the WHO. Women’s BF practices are difficult to monitor and systematically observe. Most data collection about BF will rely on accurate self-reporting and holds a risk of misclassification. This would imply that there are various errors in reports of the duration of exclusive and partial BF periods.

In an Irish context, there is a National Perinatal Reporting System (NPRS), which captures all births and this system obtains data on the infants’ feeding at the time of discharge from the birth hospital or midwifery service [3]. Three categories of feeding are used: Artificial feed, Breast, or Combined artificial and breastfeed (see Figure 1). The accurate reporting of this variable requires that the researchers have access to information about the real BF practice. Such information may be obtained through direct observation or conversation with the woman, through qualitative studies or through analysing and reporting records in the medical charts. It is clear that the time of discharge from hospital will be between one-three days after the birth [3]. After discharge from hospital, BF practice may depend on the women’s level of knowledge and support provided.

“Growing Up in Ireland” (GUI) is a national longitudinal study of children in Ireland. This represents a different way of collecting data about BF. The first data collection point is when the child is 9 month old. Questions about BF are included in the personal interviews of predominantly the mothers. An analysis of BF practices sampled 11,134 infants out of 41,185 born during December 2007 and 2008, and 65% of the sampled families agreed to take part [60]. The BF initiation rate was 56% and reduced to 48% when the baby was brought home from hospital. In all, 44% of the babies were reported to be ever exclusively breastfed and 8% were reported to be breastfed 9 months after the birth [60]. The study reported substantially lower BF initiation rates for mothers born in Ireland (48.8%) in comparison with mothers born elsewhere (82.3%). The study reported higher BF rates for women aged over 30 years (~60%) and women of higher SES (>75%) [60].

Although the full picture of BF practice in Ireland is not conveyed through these data sources, it is clear, that the WHO recommendations of exclusively BF during the first 6 month of the infant’s life, is not followed for all babies. It would appear that there continues to be scope for further public programmes, to support BF practices in Ireland. With limited resources allocated to healthcare and with competing demands for these resources, it is relevant to analyse and document the value from investments to better support BF practice.

To inform decisions about what might be appropriate investments, the target groups, configuration, content and organisation of BF support programmes should be clearly specified to enable a careful analysis of their potential costs, effects, cost-effectiveness and potential impact on social equality.

6. Models for BF support

It is worth carefully considering different models for how health interventions can be implemented especially across different social groups. One model could be providing BF support to all women who have given birth. It is then crucial to consider whether the acceptance of the BF service will be similar for all social groups. This is unlikely to be the case especially if the mothers have to make an active effort themselves to get contacts with the BF support. There may be mothers who are well experienced with approaching HCPs and following their advice, and there may be mothers who find it much more difficult. This difference may relate to the mothers’ social and cultural background including demographic and social factors and health literacy.

It is likely that mothers with lower health literacy may be less likely to get in contact with the BF support, and if that is the case, there may be an unequal effect of the programme across different social groups. A universally available intervention is thus likely to be more favorable for mothers with higher health literacy and less favourable for mothers with lower health literacy. This will be expressed as variation in the uptake of BF support and by implication, the proportion of mothers who successfully initiate BF after the birth. When having established contacts between the BF support and mothers, there may also be variation among social groups in their acceptance, and willingness and opportunities to follow the advice and recommendations to breastfeed. Mothers from some social groups may find it easier and more appropriate than mothers from other social groups. This will in particular be expressed in social variation in the duration of BF (e.g. the proportion who is exclusively BF 6 months after the birth).

The health benefit of BF for both the baby and the mother may also be different for different social groups. Very little empirical research has explored this. It is generally challenging to demonstrate the positive health benefits of BF, in different social groups although several meta-analyses have identified health benefits. The causal inference is generally challenging. The challenges relate partly to whether the effect found in studies from different populations and jurisdictions can be assumed to apply in other contexts. Furthermore, there may be different valuations of the health benefits from different social groups. If there is social variation in the health benefits from BF by social groups, then that is also something that should be considered when evaluating BF interventions.

7. Models for outcomes

An American group of clinicians developed a model framework to estimate the impact of changes in BF practice on population health and costs [61, 62]. They analysed the health impact of higher BF rates on a range of maternal diseases including number of cases, deaths and costs related to pre-menopausal ovarian cancer, breast cancer, hypertension, diabetes, and myocardial infarction. For the baby they considered cases, deaths and costs related to acute lymphoblastic leukaemia, crohn’s disease, ulcerative colitis, sudden infant death syndrome, otitis media, gastrointestinal illness, obesity, lower respiratory tract infection and necrotizing enterocolitis. They modelled the outcomes for 21 different BF scenarios specified by monthly rates of exclusive or any BF until 12 months after the birth. They employed BF rates for different US states and reported results separately for the non-Hispanic white, non-Hispanic blacks and Hispanic sub-populations.

Behind their analysis was a mathematical model, which probably represents the most detailed approach to simulating the health outcomes from BF. The results should be interpreted with due consideration of the underlying assumptions and their validity in relation to the analysed population. Similar, although less comprehensive models of BF outcomes have been reported for Mexico and the UK [63, 64, 65].

8. Costing BF support programmes

8.1 Gross-costing approach

Costs of BF support programmes can be determined using either a gross-costing or a micro-costing approach. The gross-costing approach determines the cost of the programme by identifying the aggregated amount of resources required to implement and run the programme. The cost analysis of the programme will need to consider which resources are required and their associated cost. The costs may relate to the salary, consumables, transport and overheads. The gross-cost of a programme could be determined based on assumptions of the number of staff and related resources as illustrated with fictitious numbers in Table 1.

Resource	Cost per month (€]
Salary (10.0 full time lactation consultant)	60,000
Secretarial support (2 FT)	8000
Consumables	2000
Transport	5000
Consultation and office space	10,000
Overhead 40% of above	30,000
Total cost	115,000

Table 1.

Illustrative gross-costing of a BF support programme.

These total costs can then be apportioned to participants or activities conducted as part of the programme in order to assess the average cost per:

• Peer-support groups established.

• Mothers offered BF support.

• Individual and group contacts.

8.2 Micro-costing approach

The alternative micro-costing approach is based on explicit assumptions of resource use for different activities in the programme. Four different activities are outlined in Table 2 with number of activities included in each.

	Peer-support (groups of 8)	6-month BF support of mothers initiating LC services	Single, individual contacts	Group contacts
Initiation	2 h	1 h	0.5 h	2 h
Contacts	1 h	5 contacts @1 h	5 contacts @1 h
Number of participants	8	1 over 6 months	1	6 per group
Facilities	Private homes/public space	Hospital and private home	Hospital	Meeting space
Transport cost	€20	€40	€0	€0
Total cost per woman	€100	€300	€250	€50

Table 2.

Illustrative micro-costing of a BF support programme.

Each activity is specified by nominal (ideal) numbers of contacts in a month or year, or number of contacts for mothers who enter the programme (e.g. 3–5 contacts during the first 12 months after the birth). The contacts may be counted as individual sessions or as group sessions.

9. Cost-effectiveness of BF support programmes

When there is information about the cost of different configurations of BF support programmes and the associated outcomes can be modelled based on descriptions of the target group, uptake and adherence to the programme, the cost-effectiveness and “value for money” can be assessed.

In cost-effectiveness analyses, different programmes are compared against a base case, which often is the current situation. The required cost of the programmes is related to the health outcomes and the potential downstream healthcare costs in so-called incremental cost-effectiveness ratios (ICER). The incremental cost is the additional cost of the BF programme and the cost of future healthcare (which might be savings, if the BF programme is successful in reducing the disease incidence for the babies and their mothers). These incremental costs should be related to a valid measure of effect, which could be gains in the proportion of exclusively 6-month breastfed babies, time without disease for the baby or mother, life expectancy, health-related quality of life, or quality-adjusted life years (QALYs).

By comparing the ICERs for different configurations of the BF programmes, it is possible to identify those programmes which are more cost-effective (lower cost per unit of effect) and should be preferred over the less cost-effective programmes (higher cost per unit of effect) [20]. If decision makers want to assess to what extent a BF programme represents “good value for money” in comparison with investments in other healthcare interventions, there is a need for an effect measure that can be used to compare health outcomes across different populations and interventions. The QALY measure is a composite indicator that combines health-related quality of life with life expectancy. This outcome measure can also be used in the modelling of health outcomes from BF programmes. If the incremental cost per QALY is less than €35,000 the programme is traditionally considered to represent “good value for money” and would be recommended to be implemented [66]. The QALY outcome measure was used in a UK analysis and was enumerated separately for the baby and the mother, and combined into an ICER at €52,000 per QALY, which is not considered to represent “good value for money” [64].

A number of additional cost-effectiveness analyses have been reported. A Mexican model-based cost effectiveness study compared two optimal BF scenarios with 95% of all babies being breastfed for the first 6-months exclusively and 95% of all babies being breastfed between 12 and 36 months of their life in comparison with the suboptimal practice reported in 2012 [63]. A microsimulation model was developed to follow a synthetic cohort of 100,000 15-year-old women in terms of their fertility, BF practice, and incidence of breast cancer, treatment costs, mortality rates, employment and income status over their remaining life time. The two scenarios were simulated and the difference in outcomes from the base case were interpreted as the potential impact of more frequent and longer BF practice. The base case model was calibrated with best local data from available sources. The impact of BF on breast cancer risk was expressed in terms of relative risk rates obtained from previous meta-analyses. An increase of 6-month exclusive BF rate from 14 to 95% was modelled to avoid 573 breast cancer cases and 126 premature deaths per 100,000 women during their lifetime. This resulted in a gain in 2629 DALYs and a total cost saving at US$14 Mio. If the BF during 12 and 36 months increased from 33 to 95%, the analysis showed a lower, but still positive effect on health outcomes and costs.

A different UK-based model assessed the costs-effectiveness of BF for preterm infants [65]. They used a cohort of 51,700 preterm infants (2013, in England and Wales) and compared health outcomes and healthcare costs in a scenario where all premature infants were fed with human milk during their first 6-month lifetime in comparison with a scenario where all premature infants were fed formula milk. From literature reviews, they identified odds ratios and costs of the infants experiencing any of the following outcomes: sepsis, necrotising enterocolitis, sudden infant death syndrome, acute otitis media, childhood leukaemia, childhood obesity and the associated impact of developing type-2 diabetes and coronary heart disease, neurodevelopment impairment and disability. They found that the BF scenario provided better health outcomes and lower healthcare costs with a potential saving per infant at £583 and a mean QALY gain at 0.088.

A more recent UK-based model was developed to inform the national guidance on postnatal care in England [64]. A decision-analytic cost-effectiveness model was developed to evaluate a range of interventions added to standard care. These interventions aimed at increasing BF included different models of education, advice and support provided by professionals or peers before or immediately after, or within the first 8 weeks after the birth. The primary outcome was “any BF, 16-26 weeks after birth”. The derived outcomes for the infant related to gastrointestinal infection, respiratory tract infection, acute otitis media, and mortality due to infections disease and due to sudden infant death syndrome. The derived outcome for the mother was breast cancer. The assumed impact from BF were obtained from comprehensive literature reviews conducted on behalf of the Lancet BF Support Group [67]. In the base-case analysis, the BF support model reduced the number of disease episodes in both infants and mothers. The QALY gain per 1000 babies and women were modelled at 0.16 for infants and 1.09 for mothers. However, the modelled intervention cost at £84 per baby exceeded the modelled cost-savings £10, resulting in an ICER of £52,000 per QALY which is normally not considered to be cost-effective. By changing the assumptions related to e.g. intervention cost and effects, the model indicated situations where the ICER might suggest cost-effectiveness and dominance.

In their discussion of the results, the authors point out that the quality of data regarding the associations between BF and avoidance of certain diseases are relatively low, due to studies with risk of bias related to the randomisation process, selective reporting and missing outcomes. They also comment on the variations related to the models of interventions. BF interventions may vary in model of delivery (face-to-face, telephone, individually vs. groups, or combinations), the number of contacts, the duration of the intervention, the place of delivery (home, hospital, community setting or combinations), the person delivering the intervention (peer supporter, lactation consultant, midwife, health visitor or combinations). Furthermore, there might be variations in the target group, and their recruitment and motivation to breastfeed. Finally, they remark that their definition of “any BF” as an outcome and the definition of standard care generally are poorly described.

10. Distributional cost-effectiveness

In traditional cost-effectiveness analysis, the focus is on the efficiency of the whole programme delivered to a specified target group. This assumes that each individual who receives the services have similar rates of uptake, adherence and effect. This is a strong assumption for BF programmes. In practice, the base case BF rate, the uptake, the adherence and effect may vary for different social groups. Younger mothers might have lower BF rates than older mothers, and they might be less likely to take part in BF programmes, and may be less likely to adhere to the recommendations. Therefore, both babies and the younger mothers may experience poorer health outcomes from BF than older mothers. A cost-effective BF programme may have potential for increasing the inequality in the health outcomes, and thus reduce the health equity in the population. Economists describe this situation as the trade-off between efficiency and equity.

This efficiency-equity trade-off is particularly relevant to consider for BF programmes as the reviewed material clearly have demonstrated, there are variations in BF practice by sub-populations defined by age, geography, rural/urban, nationality, culturally, educational and labour market position. These factors: which can be denoted as SES, increase the complexities of the economic analysis. This is particularly related to the validity of data that not only need to be related to the whole population but also should consider differences between social groups. However, recent developments in methods for health economic evaluations have developed frameworks that enables these trade-off analyses [68]. With the application of model-based analysis, it is possible to conduct systematic analyses of the efficiency-equity trade-off and provide additional insight, which may support the development of appropriate programmes for BF interventions. The model-based analysis may also be used to highlight the need for data analyses.

The definition of relevant social groups represents a major challenge. The previous section has identified studies that have used different characteristics to define social groups. This is a specific research area and is beyond the focus of this chapter. Education or labour market position is often used as an indicator for social groups. It may also work well for BF interventions and has the advantage that national data are available on birth and BF practice for different educational groups.

The principles of analysing the efficiency-equity trade-off for a BF-intervention is first to assess the difference in cost-effectiveness of the intervention for the different SES groups and then analyse the distribution of health outcomes among these groups. The idea is to represent interventions in a so-called health equity impact plane, where the X-axis represents incremental change in equity and the Y-axis represents incremental change in efficiency (Figure 4). There are thus four situations identified by the four quadrants. The NE quadrant (denoted A) represents programmes that both increase efficiency and improve equity. The SW quadrant (B) represents programmes that are both inefficient and harms equity. Programmes in the NW and SE quadrant (C and D) are either efficient or improve equity at the expense of the other [69].

Figure 4.

Equity-efficiency impact.

There is clear evidence that mothers with different social backgrounds have different propensity to start BF and continue when they have started. If such variations are to be considered when designing and implementing BF intervention programmes, there is a risk of increasing the inequality in efficiency and inequity in health. This could for example arise if a peer-to-peer programme was introduced as a general offer. Women with certain backgrounds would enjoy and adhere to such programmes while others will identify barriers for taking part and have difficulties in overcoming these barriers. The result might be that they refuse to take part and thus experience no or little support to begin and maintain BF practices.

It is well documented that higher educational attainments are associated with increased BF [70]. Higher educational attainment leads to better job mobility and wealth, which may facilitate BF through increased financial support. Mothers with higher income can afford private lactation consultations, breast pumps and paid support of childcare and other household duties that may facilitate BF. Therefore, active BF practice is more common in groups of mothers with higher education.

As studies have shown, BF is associated with positive outcomes for both the baby and mother. When the practice of BF is different in different groups of mothers, the outcomes from BF will also be different. The implication is that the difference in BF practice adds to an inequitable health distribution. If a BF intervention (e.g. community lactation consultants) is made available as a general offer to all mothers, there may be variation in how groups of mothers perceive the desirability of the offer. Mothers who are motivated and well-acquainted with using the services offered by the ante-natal hospital service, may be more likely to seek the advice from the lactation consultants. The implication could be that none of the BF support models may fit all, resulting in variation in behavioural change, costs and outcomes. If such variation exists, then it becomes relevant to assess which BF support models are a better fit to certain groups, and assess their implications in terms of efficiency and equity.

11. Analysing the distributional cost-effectiveness of BF

Figure 5 below illustrates different elements of the analysis of distributional cost-effectiveness. For simplicity, there are three social groups, which could be mothers outside the labour market and mothers with different positions in the labour market.

Figure 5.

Illustration of the modelling of potential health impact from BF support programmes.

To analyse the potential benefit in terms of efficiency and equity of a new programme, there needs to be a base case situation. The illustration would be the proportion of mothers in each social group who are exclusively or partially BF their babies at particular times after the birth. Both the proportion who initiate BF and the duration may be different for different social groups and thus the average time with BF will vary.

For each group, the BF practice will result in health outcomes for the babies and mothers. It is likely, there is a relationship between the duration of BF and the occurrence of adverse events such as infections for the baby and positive effects such as reduced risk of obesity for the baby and reduced risk of breast and ovarian cancer for the mother. These relationships can be used to model the health outcomes for the babies and mothers. Unless there is a clear indication that the health effects may differ for different social groups, it is reasonable to assume that the health effects occur independently of social characteristics.

It is clear, that the assumptions about the relationship between BF practice and health effects are crucial for any analysis and pose a major challenge for any analysis. However, as has been shown above, there are a number of examples where researchers have developed such models to predict health outcomes from BF.

Based on the modelling of health effects for different social groups, the distribution of health outcomes can be assumed to be representing the base case situation. The social variation in health outcomes could be described in terms of absolute or relative gaps between the group with the best outcomes and other groups. This could express the mean difference in the composite QALY measure between the groups – e.g. the most disadvantaged group experience on average 5 fewer QALYs than the best group. This could also be expressed in relative terms, e.g. the most disadvantaged group experience 10% less QALYs than the best group. There also are a number of equality indices based e.g. on Lorenz-curves, where the index can be interpreted as a measure of deviation from a fully equal distribution.

The next part in the analysis is to specify the design and effect of the BF programme. In some analyses, it could be expected that several different designs could be envisaged and specified in terms of their target group, the organisation and provision of the intervention. Each design or model of care would be associated with different use of resources and costs and will have different consequences in terms of the proportion of mothers who initiate and continue BF their babies.

The consequences of these alternative models of care can then be analysed using the same modelling framework as was used in the baseline situation. This implies that the change in health benefits is determined by the change in BF practice as specified as part of the model of care. The health benefits with the intervention and without the intervention is compared across the social groups. The distribution of health gains (e.g. in terms of QALYs) for each social group can be calculated and the total cost-effectiveness can be compared for the whole population in terms of the ICER. The distribution of outcomes across social groups may be expressed as the health gain in each social group or by using one of the proposed equity indices. The net gain in health benefit (QALYs) may be expressed as the net gain, where the cost of the intervention programme is expressed in forgone health gains as the opportunity cost of the intervention programme.

When several models of BF support programmes are compared in terms of their efficiency-equity trade-off, they may be represented in the equity-efficiency impact plane as shown in Figure 4. Models in quadrant A have positive impact on both equity and efficiency. Models in quadrant B have negative impacts on both equity and efficiency, while models in quadrant C and D have positive impact on either efficiency or equity but not both. Interventions in quadrant A are clearly preferable. When several models are in quadrant A, decision makers can make an assessment on which model best satisfy policy objectives.

12. Conclusion

Individual, societal and policy factors influence mothers’ decisions to begin and continue BF and interventions targeted towards these factors are essential, in particular the cost-effectiveness and equity trade-off components associated with them. BF may be encouraged through social interventions where the perception, understanding and culture around BF is influenced, e.g. by offering information and advice on proper BF technique and maintenance of BF for a period of time after the birth. In addition, specifically targeted programmes may be devised to support and encourage women to breastfeed. These can be offered as part of maternity services that are offered by public, private and voluntary organisations at different phases during the ante and post-natal care. There may be programmes that are targeted to individual women during the immediate post-natal phase or to groups of women in a peer-support programme. Interventions may include high-level policies such as labour market regulations which makes continuous BF easier for mothers e.g. by extending periods of maternity leave and offering good facilities for BF. BF legislation and BF practices should be revisited in Ireland and further BF supporting policy implementations should be considered with e.g. a law stating that mothers have a right to breastfeed or a law requiring employers to prove BF facilities at work [71].

Some interventions may be more effective in changing the BF behaviour for mothers with different SES. Universally provided interventions are likely to be more effective for some SES groups because their willingness to initiate and continue BF will be impacted differently. Investments in BF support therefore have the risk of providing an unequal health benefit and may increase the inequity in health. To address both the efficiency and equity issues, it is relevant to evaluate future and current BF interventions using the “Distributional cost-effectiveness framework”. The framework will enable evaluation of the impact on efficiency and equity of more targeted BF interventions. This will provide decision-makers with stronger arguments for offering differentiated BF interventions to different SES groups.

Conflict of interest

The authors declare no conflict of interest.

Appendices and nomenclature

BF	Breastfeeding
BMS	Breast-milk Substitutes
HCP	Healthcare professionals
DCEA	Distributional cost-effectiveness analysis
BFHI	The Baby Friendly Hospital Initiative
SES	Socioeconomic status
SSC	Skin-to-Skin Contact
FSAI	Food Safety Authority of Ireland
HSE	Health Service Executive
HIQA	Health Information Quality Authority
RCTs	Randomised control trials
GUI	Growing Up in Ireland
QALYs	Quality-adjusted life years
ICER	Incremental cost-effectiveness ratios
WHO	World Health Organisation
OECD	Organisation for Economic Co-operation and Development
HCWs	Health Care Workers
NPRS	National Perinatal Reporting System