Systematic review of implementation research of effective breastfeeding support interventions for healthy women with healthy term babies

Anna Gavine; Albert Farre; Fiona Lynn; Shona Shinwell; Phyllis Buchanan; Joyce Marshall; Sara Cumming; Louise Wallace; Angie Wade; Elayne Ahern; Laura Hay; Marianne Cranwell; Alison McFadden

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Gavine A, Farre A, Lynn F, et al. Lessons for the UK on implementation and evaluation of breastfeeding support: evidence syntheses and stakeholder engagement. Southampton (UK): National Institute for Health and Care Research; 2024 Jul. (Health and Social Care Delivery Research, No. 12.20.)

Lessons for the UK on implementation and evaluation of breastfeeding support: evidence syntheses and stakeholder engagement.

Show details

Contents

< Prev Next >

Chapter 4Systematic review of implementation research of effective breastfeeding support interventions for healthy women with healthy term babies

Introduction

The Cochrane review update undertaken in our review 1 confirmed that there is ample evidence to know that breastfeeding women need support to be available and to be provided, and that such support is likely to make a difference. Such an evidence base also suggests that one key research question for the future is to identify how such support can best be provided consistently across countries and settings.

Therefore, there is now a need to improve the evidence base around scaling up issues for breastfeeding support interventions, which will require a greater emphasis on implementation and quality improvement approaches rather than effectiveness studies. To enable further advances in this area, it will be fundamental to identify and synthesise available qualitative and process evaluation data on existing interventions. The overall aim of this review was to conduct a theoretically informed mixed-methods synthesis of process evaluations of breastfeeding support interventions identified as effective in review 1.

Objectives

To identify qualitative and quantitative data from process evaluation studies linked to breastfeeding support interventions identified as effective in review 1.
To synthesise the views and experiences of those involved in receiving or delivering breastfeeding support interventions identified as effective in review 1.
To identify the contextual factors (barriers/facilitators) affecting the implementation of breastfeeding support interventions identified as effective in review 1.

Methods

The protocol for this systematic review is registered on PROSPERO (CRD42021229769).

Search strategy

We systematically searched six electronic databases (MEDLINE, CINAHL Plus, PsycInfo, ASSIA, Scopus and Web of Science). Searches were conducted in March 2022 using combinations of index terms and free-text words relating to ‘breastfeeding support’ AND ‘implementation research’ (a sample search strategy for MEDLINE is provided in Appendix 1). No restrictions were applied on publication date and publication language. Reference lists of all included studies and relevant systematic reviews were scanned for eligible studies. Supplementary searches were conducted based on the name of interventions identified in Gavine et al.,⁸³ included articles’ authors, and forwards and backwards citation checking.

Eligibility criteria

Inclusion criteria

Studies were included if they reported findings of primary research exploring the views and experiences of any participants involved in either delivering or receiving any of the breastfeeding support interventions identified as effective in Gavine et al.,⁸³ including breastfeeding women and babies and their families, service providers, managers, commissioners and policy-makers.

Qualitative and quantitative studies, either standalone or in mixed-methods designs, were included. Studies reported any type of process evaluation outcome relating to the selected interventions, including any subjective participant-reported outcomes and constructs such as attitudes, views, beliefs, perceptions, understandings or experiences.

There were no restrictions on publication date or language of publication.

Exclusion criteria

Articles only reporting on impact evaluation results of breastfeeding support interventions (i.e. effectiveness of interventions) were excluded.

Studies that focused specifically on women or infants with additional care needs were excluded. For mothers this could mean coexisting medical problems (e.g. diabetes, HIV) or pregnancy-related complications (e.g. pre-eclampsia). For infants this could include preterm birth, low birthweight or additional care in a neonatal unit.

Studies relating to interventions taking place in the antenatal period alone were excluded from this review, as were interventions described as solely educational or promotional.

Selection process

Two reviewers independently screened titles, abstracts and relevant full texts against the predetermined eligibility criteria. Any discrepancies were resolved through discussion and consultation with a third reviewer.

Data extraction and quality appraisal

Data extraction was undertaken independently by two reviewers using a piloted data extraction form. Any discrepancies were resolved through discussion and consultation with a third reviewer. The table of characteristics is presented in Appendix 2, Table 13.

Quality appraisal of included studies was conducted by two reviewers, using a self-developed tool derived from a set of criteria previously used in other National Institute for Health and Care Research (NIHR)-funded work to assess the quality of process evaluations.⁸⁹ Studies were not excluded based on the quality/adequacy of the reporting. Instead, the quality of studies was taken into consideration during data synthesis by exploring whether any particular finding or group of findings were dependent, either exclusively or disproportionately, on one or more studies classed as ‘low quality’ or ‘inadequately reported’. Any discrepancies were resolved by discussion and the involvement of a third reviewer where necessary. See Appendix 2, Table 14.

Data synthesis

We adopted a mixed-methods synthesis approach. We first undertook two preliminary syntheses of quantitative (synthesis 1) and qualitative (synthesis 2) process evaluation studies, and then integrated qualitative and quantitative process evaluation data into a theoretically informed cross-study synthesis (synthesis 3).

For synthesis 1 we used narrative methods⁹⁰ to synthesise quantitative findings from included process evaluations. Two reviewers independently assessed the tabulated characteristics of the included quantitative studies and agreed the criteria to organise the included studies. For synthesis 2 we used a data-driven approach to thematic synthesis⁹¹ to synthesise qualitative findings from included process evaluations. This involved three overlapping and interrelated stages: (1) line-by-line coding of findings from primary studies, (2) categorisation of codes into descriptive themes and (3) development of analytical themes to describe or explain previous descriptive themes. To ensure the robustness of the synthesis, various techniques to enhance trustworthiness were undertaken, including audit trail, multiple coding, reviewer triangulation and team discussions. Finally, for synthesis 3, we adopted a theory-driven approach to thematic synthesis⁹¹ to synthesise and bring together quantitative and qualitative findings from included primary studies. This synthesis was informed by the Consolidated Framework for Implementation Research (CFIR),⁹² a comprehensive framework that characterises the contextual determinants of implementation and can be used to inform implementation theory development and verification of what works where and why across multiple contexts.

Results

The searches identified 2894 records, which were assessed against the inclusion criteria. Title and abstract screening resulted in 243 records considered eligible or inconclusive. Full-text articles were then retrieved and assessed for eligibility. Two records could not be retrieved. Of the 241 records screened at full text, 225 were excluded. The main reason for exclusion was studies not being linked to an intervention identified as effective in review 1 (n = 84), followed by standalone studies that were not linked to any intervention (n = 51) and studies not involving implementation research and/or process evaluation data (e.g. pre-implementation or intervention development studies) from eligible interventions (n = 50). Other reasons for exclusion were studies linked to either interventions (n = 26) or populations (n = 6) not eligible for inclusion in review 1, and systematic reviews (n = 4) and other publication types not reporting primary research findings (n = 4). The remaining 16 studies were included in the final synthesis (Figure 3). The 16 studies are linked to 10 RCTs of effective interventions from review 1.

FIGURE 3

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.

Summary of included studies

A summary of key characteristics of the included studies is presented in Appendix 2, Table 13.

Twelve studies contributed qualitative data to the synthesis, comprising eight qualitative⁹³^–¹⁰⁰ and four mixed-methods¹⁰¹^–¹⁰⁴ process evaluation studies; and eight studies contributed quantitative data to the final synthesis, comprising four quantitative¹⁰⁵^–¹⁰⁸ and four mixed-methods studies.¹⁰¹^–¹⁰⁴

Studies reported data from ten countries: nine from HICs (five in the USA, two in Australia and one each in Canada and the UK); and seven from LMICs (four in Uganda, two in South Africa and one in Pakistan). All of the studies from Uganda and South Africa were evaluations of aspects of the PROMISE-EBF RCTs.¹⁰⁹

Study settings included rural and urban areas and hospital and community facilities. In eight of the studies in HICs, the target populations were low-income or disadvantaged populations, or those living in areas with low breastfeeding rates.

Study samples ranged from 26 to 130 mothers, 12 to 254 peer counsellors, 13 to 28 healthcare staff and 2 to 409 other stakeholders, including supervisors, programme managers and co-ordinators, and unspecified key informants. Other forms of data included observations, diaries and daily activity logs.

Process evaluations included in this review were linked to effective interventions identified in review 1 (for details, see Appendix 2, Table 13).

The descriptions of linked interventions were coded against a taxonomy of behaviour change techniques. The most commonly identified behaviour change techniques related to social support, goals and planning, and feedback and monitoring. A summary of the behaviour change techniques identified across all the linked interventions is provided in Appendix 2, Table 15.

Quality appraisal

The quality of the 16 process evaluations was mixed (see Appendix 2, Table 14). Seven studies were judged to have made a fairly thorough attempt to increase rigour and minimise bias in sampling, data collection and analysis.⁹³^,⁹⁶^,⁹⁷^,¹⁰⁰^,¹⁰¹^,¹⁰³^,¹⁰⁷ A further six studies were assessed to have taken at least a few steps to increase rigour of sampling, data collection and analysis.⁹⁴^,⁹⁸^,⁹⁹^,¹⁰²^,¹⁰⁴^,¹⁰⁵ For the remaining three studies, judgements for at least one element of sampling, data collection or data analysis were hindered by poor reporting.⁹⁵^,¹⁰⁶^,¹⁰⁸ All studies’ findings were judged to be at least fairly well supported by the data. The findings of three studies were judged to have limited breadth and/or depth.⁹³^,¹⁰⁶^,¹⁰⁸ In Andaya et al.,⁹³ the evaluation was based on exit interviews lasting 8–12 minutes. Chapman et al.¹⁰⁶ report only coverage of the intervention. Ridgeway et al.¹⁰⁸ do not report responses to open-ended questions in their survey. Seven studies were judged not to have privileged the perspectives of breastfeeding women.⁹⁴^,⁹⁷^,⁹⁸^,¹⁰²^,¹⁰⁴^,¹⁰⁶^,¹⁰⁸ Two studies were judged to have low reliability of findings⁹⁴^,⁹⁵ and one study was judged to have low usefulness.⁹⁴^,⁹⁵

Stakeholders’ perceptions and experiences

Stages 1 and 2 of our mixed-methods synthesis resulted in the categorisation of primary quantitative and qualitative data from included studies into 86 descriptive themes. Building on these findings, further analytical work and team discussion was undertaken, and the initial descriptive themes were grouped around a resulting set of 18 factors affecting the implementation of effective interventions, which in turn informed our preliminary, data-driven synthesis conclusions. These revolved around the following three analytical themes:

that qualitative/quantitative monitoring data and feedback are provided for women and/or professionals to reflect on and evaluate the progress, quality and experience of implementing the new breastfeeding support intervention
that breastfeeding support needs of women/families served by the implementing organisation (including any barriers to/facilitators of meeting those needs) are known
that individuals involved in the new breastfeeding support intervention are appropriately trained, have confidence in their capabilities, and are able to execute the courses of action required to achieve the desired implementation/intervention goals.

For the final stage of our thematic synthesis, we mapped our descriptive and analytical themes against the domains of the CFIR. Our three analytical themes and subthemes aligned across five subdomains of the implementation process domain (assessing needs, assessing context, tailoring strategies, engaging, and reflecting and evaluating) of the CFIR framework.

Our final three overarching, theoretically informed analytical themes are described below. Table 4 shows the distribution of primary studies underpinning each analytical theme and their mapping against the relevant CFIR subdomains.

TABLE 4

Included studies mapped against relevant subdomains of the CFIR

Assessing the needs of those delivering and receiving breastfeeding support interventions

Some reported having to deal with feelings of frustration when running breastfeeding support services with low attendance rates.¹⁰² This was a particular challenge for those running services located in small or rural areas. For those juggling a breastfeeding support role with healthcare provider roles, the pressure of the competing demands in the context of low attendance rates could make them feel like their time might have been better spent on other activities.¹⁰²

One key strategy reported to both identify and address the needs of breastfeeding support providers was through training.⁹⁴^–⁹⁶^,⁹⁸^,¹⁰³^,¹⁰⁵^,¹⁰⁷ Studies largely reported that intervention deliverers felt training prepared them well, in terms of both counselling skills and technical competence (e.g. being able to show how to breastfeed correctly). All of this was perceived as key to ensuring consistency in intervention delivery.

Other issues that could be addressed through training were to do with the practical expectations of undertaking the breastfeeding supporter role. Uncertainties about safety, transport and reimbursement while delivering support were among the most reported needs for those delivering community-based interventions,⁹⁴^,⁹⁵ as well as around more complex issues, such as managing difficult scenarios or the interplay of cultural beliefs and breastfeeding practice. The last was particularly relevant to lay breastfeeding supporters delivering interventions at community level. They noted the importance of acknowledging that trainees themselves belong to a range of communities that might be systematically exposed to certain issues/inequities more than others (e.g. rural isolation, HIV prevalence in the community) and/or might hold cultural beliefs about breastfeeding or breastfeeding-related practices that could act as barriers. These should be identified and addressed in a culturally sensitive manner and without antagonising the communities, enabling lay providers to appropriately and inclusively support breastfeeding women from a range of communities.⁹⁴^,⁹⁵^,⁹⁷

Those in implementation leadership roles also emphasised the importance of effective management and supervision. This was reported as a key facilitator of some interventions,⁹⁴^,⁹⁷ particularly to ensure that certain needs of intervention deliverers continue to be addressed beyond the provision of formal training. For example, for those engaged in interventions relying on peer, lay and/or volunteer supporters, there was an important need to provide them with ongoing emotional support, including mentoring and motivation.

Overall, the breastfeeding supporters felt that their role was important, satisfying and rewarding,¹⁵ with implications that were perceived to go beyond the specific breastfeeding support encounters to act as triggers of the wider support network of the breastfeeding women.⁹⁵^,⁹⁶

The needs, preferences and priorities of recipients of breastfeeding support interventions were echoed in five studies.

Breastfeeding women perceived the provision of support as positive, important and needed.⁹⁹^,¹⁰⁷ Key to this was being offered the opportunity to ask questions and being allowed to spend enough time to address any issues.¹⁰³^,¹⁰⁴ Also important was accessing support flexibly as needed, rather than having to fit support around fixed working hours or at times that might not be convenient (particularly if receiving support visits at home or after starting paid work after maternity leave).¹⁰¹^,¹⁰³^,¹⁰⁴

Assessing the context and optimising delivery of and engagement with breastfeeding support interventions

Some studies reported a range of contextual factors affecting the implementation and delivery of breastfeeding support interventions. These included identification of appropriate settings and accessible, available spaces to deliver breastfeeding support;⁹⁵^,¹⁰² consideration of environmental factors that are considered breastfeeding promoting (and avoidance of those that are not) in the intervention delivery settings (e.g. use of breastfeeding promotion leaflets, posters and videos);¹⁰⁵ and availability of and alignment with local policies and procedures, as well as with existing practices, in maternity care.⁹⁸^,¹⁰⁵ Studies also reported examples of tailoring implementation strategies to address barriers, leverage facilitators and optimise how breastfeeding support interventions fit the context. These included strategies to promote and encourage engagement, such as ensuring embeddedness within the community,⁹⁵^,⁹⁶ addressing challenges to recruit breastfeeding supporters,¹⁰² favouring lay language;¹⁰³ teamwork and positive interactions with other breastfeeding supporters and healthcare professionals;⁹⁶^,¹⁰⁵ responsiveness of support content and language to address known barriers and common issues;¹⁰⁰^,¹⁰³^,¹⁰⁶^,¹⁰⁸ and continuity/accessibility of interventions across the continuum of care.⁹³^,¹⁰³

Reflecting and evaluating the success of implementing and providing breastfeeding support

Included studies reported a broad range of reflective and evaluative accounts about the success of implementation processes and about how impactful breastfeeding support interventions were perceived by women.

Reports about the success of implementation focused on issues relating to key implementation outcomes such as satisfaction,¹⁰³^,¹⁰⁴^,¹⁰⁷ fidelity,¹⁰³ convenience¹⁰¹^,¹⁰³^,¹⁰⁴ or usefulness.¹⁰¹^,¹⁰⁴^,¹⁰⁷ Other studies reported on the key drivers that enabled successful engagement between mothers and breastfeeding supporters,⁹⁷^,¹⁰⁴^,¹⁰⁷ including elements of responsiveness/tailoring and content areas addressed in support encounters.⁹⁵^,⁹⁷^,¹⁰⁴^,¹⁰⁶^,¹⁰⁸ Some studies reported data on the views and experiences of enacting the role of breastfeeding supporter⁹⁵^,⁹⁶^,⁹⁸^,¹⁰⁵^,¹⁰⁷ and breastfeeding supporter’s supervisor/lead,⁹⁷^,¹⁰⁷ all of which documented positive perceptions by those undertaking and/or interacting with those roles. Other studies looked at factors affecting the scale-up of breastfeeding support interventions, including key barriers (e.g. stigma around exclusive breastfeeding, economic barriers and limited resources, health facilities, lack of supportive policies, low male involvement, negative sociocultural beliefs) and facilitators (e.g. promotion at health system level, engagement of professional associations and active collaborations with existing groups, the media and appropriate role models).⁹⁸^,⁹⁹

Some studies included reports of perceived meaningfulness and impact of breastfeeding support interventions from women’s perspectives, which can be considered reflective accounts that add to the existing body of evidence about the success of breastfeeding support interventions. Women perceived breastfeeding support interventions as beneficial to women, babies and the wider community;¹⁰² and helpful for improving breastfeeding knowledge,⁹³ ensuring the early establishment of breastfeeding⁹³ and enabling women to recognise feeding patterns and problems.¹⁰¹ Breastfeeding supporters were perceived by women as allies who bolstered their conﬁdence in their decision to breastfeed, particularly for those who were faced with a lack of encouragement from family or hospital staff.⁹³

The provision of practical information about breastfeeding mechanics and hands-on support were perceived as useful and enabled women to feel reassured and encouraged to continue breastfeeding.⁹³ The element of responsiveness in terms of support content areas afforded by breastfeeding support interventions helped make interventions meaningful for women in the context of their specific breastfeeding support encounters.⁹⁵^,⁹⁷^,¹⁰⁴^,¹⁰⁶^,¹⁰⁸ The most commonly reported issues addressed were reassurance, general breastfeeding information, supply and demand, breastfeeding positioning and attachment, feed frequency, normal infant behaviour, expressing and breast pump use, nipple pain/damage issues and not having enough milk. More interactive intervention components (e.g. monitoring systems, telephone-based support) were appreciated and seen as useful but perceived as a ‘mixed fit’ for breastfeeding support. Women saw these modes of support as an addition to rather than a replacement for face-to-face support.¹⁰¹^,¹⁰³

Chapter summary

This review comprised 16 studies linked to 10 interventions identified as effective in review 1, which reported the views and experiences of those delivering or receiving breastfeeding support. The quality of the included studies was mixed, but all study findings were judged to be at least fairly well supported by the data.

The synthesis resulted in three overarching themes, theoretically informed by the CFIR: (1) assessing the needs of those delivering and receiving breastfeeding support interventions; (2) assessing the context and optimising delivery and engagement with breastfeeding support interventions; and (3) reflecting and evaluating the success of implementing and providing breastfeeding support.

Included studies identified several implementation challenges relating to the needs, preferences and priorities of those delivering and receiving breastfeeding support interventions. Breastfeeding supporter training was a commonly reported implementation strategy, which also enabled implementation teams to identify and address breastfeeding supporters’ needs. Included studies reported a range of contextual factors (e.g. alignment with local policies) affecting the implementation and delivery of breastfeeding support interventions as well as a range of tailoring strategies (e.g. community involvement, use of lay language, responsive support content/information) to address contextual factors. Reports about implementation success focused on issues relating to key implementation outcomes such as satisfaction, fidelity and usefulness.

This work was produced by Gavine et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

Bookshelf ID: NBK605422

Contents