7

Conclusions and Future Directions

This chapter summarizes the committee’s conclusions about the consistency of existing feeding recommendations for infants and children from birth to 24 months of age, and the types of evidence used. In highlighting areas of inconsistency and differences in evidence types, the committee comments on possible explanations for these apparent discrepancies. Considering the level of consistency seen across many recommendations, the committee provides comments on the ability to harmonize the process of developing feeding recommendations. The committee also offers a recommendation to support communication and dissemination of feeding guidance based on its review of compiled materials and its collective expertise. Finally, the committee identifies gaps in the evidence related to feeding recommendations and communication and dissemination of feeding guidance.

CONSISTENCY AMONG EXISTING FEEDING RECOMMENDATIONS AND TYPES OF EVIDENCE USED

Across the 26 topic areas related to what to feed and how to feed, the committee sorted the recommendations in the 43 guideline documents into a variety of different themes. For many of those themes, the recommendations from the various sources were consistent or generally consistent (i.e., some nuances in wording) (see Tables 4-3 and 5-2 in Chapters 4 and 5, respectively). When there was inconsistency, it often related to the specific ages or age ranges stated in the recommendation or the specific target group for the recommendation (see Box 7-1).

For some themes, the committee could not determine consistency because there was only one recommendation on that theme or the recommendations were too heterogeneous in the type of advice offered (though not necessarily inconsistent). Examples of such themes include

• Avoiding use of nonnutritive sweeteners,
• Limiting the percentage of energy from sugar,
• Using vitamin C–rich foods to enhance iron absorption,
• Consuming iodine from foods,
• Recommended types of fat from foods, and
• Recommended bottle-feeding behaviors or types of bottles and cups to promote infant self-regulation of intake.

Lack of evidence for certain topics may account for why there was only one recommendation on some themes. It was sometimes challenging to judge the consistency of recommendations because of differences in wording and/or lack of clarity in the guidance documents. Similarly, it was sometimes difficult to determine which statements were intended to be formal feeding recommendations as opposed to simply providing feeding information. There were also instances where it was unclear if the guideline document identified was the most current guidance, or when guidance was considered out-of-date owing to the emergence of new evidence.

The type of evidence cited to support the recommendations was variable. Variations included whether a scientific review of the evidence formed the basis for each recommendation, and if so, the nature of the scientific review (e.g., systematic review or narrative review) and whether the evidence was graded (in the case of systematic reviews). In some cases, there was no explicit link between recommendations and supporting evidence and it was difficult to map recommendations to the evidence presented. Some organizations or consortia released multiple documents for different audiences, making it difficult to identify the original evidence base used to formulate the recommendations. Despite these challenges, the committee attempted to categorize the type of evidence presented in each document, or in some cases in supporting documents, for each theme.

Overall, most of the recommendations mapped to narrative reviews. For most of the themes, at least one of the recommendations mapped to a systematic review, but formal grading of the evidence and systematic indication of additional considerations related to the recommendations, such as potential benefits versus harms, side effects, or risks, was rare. For some themes, the guideline document indicated that there was little or no evidence to support a recommendation (e.g., for the use of nonnutritive sweeteners).

Some of the observed inconsistency across recommendations may stem from differences among organizations in the process used for review of the available evidence, the criteria for drawing conclusions from that evidence, and the process for developing the wording of the recommendations. Although some organizations conducted systematic reviews of the evidence for select topics, there is a relative paucity of published randomized controlled trials related to infant feeding, which limits the strength of the conclusions that can be drawn. Because of the challenges of conducting a formal systematic review, some organizations based their recommendations on “expert-informed” consensus, but this approach can result in different conclusions depending on which body of experts is involved in the process (Kredo et al., 2016).

Inconsistency across the recommendations reviewed by the committee could also be related to the dates of publication; for some topics, the

evidence has evolved substantially in recent years and older guideline documents may now be outdated, as was the case for food allergies. The cycle for updating guidelines may differ across organizations and may be driven by a set schedule (e.g., every 5 years), or by the availability of a substantial body of new evidence. In addition, there could be variability in recommendations stemming from differing objectives of the organization creating the guidance. For example, some organizations (such as dental health organizations) emphasize oral health, while others may emphasize short- or long-term health and developmental outcomes. Organizations also may have different target populations, such as a certain age range (e.g., infants versus toddlers) or geographic focus (e.g., recommendations for a specific country or population versus global recommendations). Finally, recommendations from organizations may also vary based on their interpretation of the risks versus the benefits of making certain recommendations.

Although there is often considerable variability in the wording of recommendations for feeding infants and children from birth to 24 months, the committee noted that many organizations had recommendations on the same topics, so there is potential for collaboration on feeding guidelines. In many cases, the differences in wording were subtle. Collaboration on feeding guidelines, starting at the development phase and continuing through the dissemination phase, is likely to facilitate the harmonization of guidance, as described below.

HARMONIZING THE DEVELOPMENT OF FUTURE FEEDING GUIDELINES

The guideline documents reviewed reflected a mix of approaches, indicating substantial opportunity to improve and ideally harmonize the planning, development, communication, and dissemination of future feeding guidelines across organizations. Established criteria for high-quality guidelines have been published (Brouwers et al., 2010a,b,c; Grimmer et al., 2014; IOM, 2011; Kredo et al., 2016; Qaseem et al., 2012; Schünemann et al., 2014; WHO, 2012). The sections below describe several key elements to consider during development of future guidelines and recommendations for feeding infants and young children, with respect to three aspects of this process: (1) planning to develop the guidelines, (2) developing the guidelines, and (3) planning for dissemination and implementation of the guidelines.

Planning to Develop the Guidelines

The theory of collaborative advantage suggests that involvement of multiple stakeholder organizations (or a consortium of organizations) starting from the planning stage of the feeding guideline process could be benefi-

cial (Vangen and Huxham, 2013). Collaboration allows for joint decisions related to the scope of the guidelines, the relevant stakeholders that should be included on the guidelines panel, and the process for establishing consensus and approval of the guidelines (Qaseem et al., 2012). If a formal systematic review of the evidence is to be conducted, which is considered best practice (Handu et al., 2018; IOM, 2011; Kredo et al., 2016; Qaseem et al., 2012; Schünemann et al., 2014; WHO, 2012), this process can be centralized to enhance quality and efficiency and reduce the cost and time burden for each participating organization.

The specialized skills and resources required to conduct a high-quality systematic review are often beyond the reach of an individual organization, which may result in reliance on narrative reviews (Allen and Olkin, 1999; McGowan and Sampson, 2005). Robust and frequent collaboration between organizations can also facilitate continued evolution of uniform standards for quality nutrition guidelines, avoid duplication of effort, and establish consistent guidance across organizations, with proper accommodations for specific audiences and contexts considered when finalizing the wording and presentation of recommendations and during the dissemination process. Harmonization of the process for developing feeding guidelines does not necessarily imply that the specific feeding recommendations need to be exactly the same across countries, communities, and professional organizations. In fact, the need to customize guidance for particular target audiences and contexts means that the wording and presentation of recommendations may vary, especially in the communication and dissemination products. However, general consistency is particularly important for reinforcing harmonized guidance to target audiences, such as health care providers, parents, and guardians.

Including key stakeholders, communication experts, and representatives of the target audience at the beginning stages can help to ensure that the final product is appropriately designed and disseminated (Gagliardi et al., 2015). These types of collaborators can help design formative and implementation research to identify barriers and facilitators to the uptake of recommendations by the target audiences (Gagliardi et al., 2015). This allows for adaptation for specific populations groups and setting up a feedback loop for continued fine-tuning of the recommendations in the future.

There are several existing examples of coordination among organizations to develop feeding guidelines for infants and young children, as noted in Chapter 3. Most of these were collaborations among organizations within a country (e.g., in Canada, among the Breastfeeding Committee for Canada, Canadian Paediatric Society, Dietitians of Canada, and Health Canada [Health Canada et al., 2014, 2015]), but there is also potential for cross-country collaboration. One example is the guideline recommendations produced by the European Society for Paediatric Gastroenterol-

ogy, Hepatology and Nutrition (ESPGHAN), a professional society whose membership spans most European countries (ESPGHAN, 2020). More broadly, a recent report has offered recommendations to unify the approach to developing nutrient reference values that would be acceptable globally (NASEM, 2018). A similar approach could be taken for infant and young child feeding guidelines.

Although involvement of multiple stakeholder organizations could be beneficial, there may be substantial challenges to this approach, particularly related to establishing collaborative goals and trust and to differences in cultural and leadership approaches (Vangen and Huxham, 2013). There needs to be active acknowledgment and management of inevitable differences and tensions that arise, to maximize collaborative advantage and avoid collaborative inertia, or the tendency for collaborative activities to progress slowly (Vangen and Huxham, 2013). There are existing frameworks, such as the theory of collaborative advantage and collective impact (Kania and Kramer, 2011; Vangen and Huxham, 2013), that explicitly outline considerations for effectively collaborating across organizations. The barriers described above can also negatively influence post hoc approval of guidelines from other organizations; it may be more difficult to address and effectively manage them after guideline development than when the other organizations are involved from the beginning (Kania and Kramer, 2011).

Developing the Guidelines

Established criteria for high-quality guidelines exist (IOM, 2011; Kredo et al., 2016; Qaseem et al., 2012; Schünemann et al., 2014; WHO, 2011) and have been operationalized into tools that can be used to assess guideline quality (AGREE Next Steps Consortium, 2017; Brouwers et al., 2010a,b,c; Grimmer et al., 2014). These criteria and tools recommend that the guidelines explicitly describe:

• The scope and purpose of the guideline, and who is meant to use it.
• How available evidence was identified (search process and selection of evidence that was reviewed) and over what period of time.
• How the available evidence was evaluated (e.g., quality of the evidence using an established grading system).
• Who formulated the guidelines from the available evidence. This includes their selection process, determination of their potential conflict of interests, and the areas of expertise and interest they represent. This also includes clearly indicating contributing organizations for collaborative guidelines.

• What was considered in addition to the available scientific evidence in the formulation of recommendations (e.g., clinical judgment; expert opinion; risks versus benefits; practicing clinician or consumer views, values, or preferences) and the process for establishing consensus within the guideline development group.
• The external review process.
• How the guideline is to be disseminated.
• The process and timeline for updating the guideline.

Additional steps can be taken to facilitate routine, timely and meaningful updates of feeding guidelines, such as:

• Including a systematic way to identify and prioritize research needs during the guideline process, listing them in the report in an easily identifiable way, and including steps within the communication and dissemination plan to communicate these research needs to researchers and funders;
• Planning how to integrate important new research findings from rapidly evolving areas (e.g., allergenic foods) into guidelines in a timely way; and
• Including the date of publication on all guideline documents and associated materials, and indicating if current guidance is meant to supersede previous guidance from the same organization.

If multiple guideline documents are released from one source for different audiences detailing different aspects of the development process, it would be useful to explicitly outline within each document how they fit together and provide a link to access each of the individual documents (e.g., in an appendix). In addition, the readability of future feeding guidelines can be improved by making the recommendations “specific,” “unambiguous,” and clearly identifiable in the text (AGREE, 2020; iCAHE, 2020). Supporting evidence, and any rating or ranking of the supporting evidence, ought to be clearly linked to each recommendation (AGREE, 2020; iCAHE, 2020). The system used to rate or rank the supporting evidence for a recommendation needs to be clearly identified and explained. In addition, guideline developers need to consider systematically communicating other considerations related to the recommendations, such as potential benefits versus harms, side effects, or risks (AGREE, 2020).

Finally, feeding guideline development may require additional consideration and consensus around some challenging aspects common to the field of nutrition, such as:

• The most appropriate tools to assess the quality of the available evidence (NASEM, 2017);
• The best way to establish and communicate recommendations when the systematic evidence review does not include randomized controlled trials, which may be challenging to conduct for some nutrition topic areas, such as breastfeeding (Hébert et al., 2016);
• The best way to establish and communicate recommendations when only expert consensus is available; and
• Whether there are any additional guideline criteria or process steps that should be included when consumers are the primary target audience of a guideline (as compared to health care professionals).

Topics for which there is little scientific evidence to support a feeding recommendation are very challenging not only for those making the recommendation, but also for health professionals, who need to provide some sort of guidance, and for parents who need to decide what to do. Health care providers need to know how to effectively communicate that there is uncertainty, guide consideration of risk versus benefit, and communicate their clinical judgment and experience to caregivers. Caregivers also need guidance on how to judge credible versus less credible sources of information, which recommendations are strongly supported by scientific evidence, and which practices allow more flexibility to incorporate family preferences and values without risking infant or child health.

Planning for Dissemination and Implementation of Guidelines

Effective and timely strategies are needed to support the communication, dissemination, and implementation of feeding guidance, and to promote the systematic uptake of guidelines (Brownson et al., 2013; Koh et al., 2020; Shelton et al., 2020). However, the effectiveness of guideline implementation is often limited because of a lack of resources and the numerous challenges at the individual, institutional, and system levels. Dissemination and implementation (D&I) science has recently been defined as a field that “seeks to inform how evidence-based interventions can be successfully adopted, implemented, and maintained in health care delivery and community settings” (Holtrop et al., 2018). For example, organizations can use D&I science in planning and practice to address how feeding guidelines are perceived by health care providers and parents (e.g., relevance, level of difficulty, and time required). Subsequently these perceptions can be used for tailoring the message and the implementation strategies for the specific social and demographic characteristics of the target audience. Organizations can also assess the contextual factors that

may influence adoption of feeding guidelines (e.g., policies of the health care system; existing knowledge, attitudes, or practices; messages disseminated by industry) to help select implementation strategies that will most effectively fit within that organization or community (Brownson et al., 2018a,b). Finally, systematic evaluation of the actions taken to implement feeding guidelines will facilitate identification of the most effective strategies in reaching health care providers or parents (Proctor et al., 2011; Rabin et al., 2016). Ultimately the use of D&I science will reduce the time required for feeding guidelines to be adopted into policy or practice by providers, parents, and organizations (Koh et al., 2020).

There are several D&I frameworks that can be used to guide (1) the process of translating dietary guidelines into practice, (2) understanding what influences implementation outcomes (barriers and facilitators), and (3) assessing process (e.g., coverage, quality) and impact outcomes with a focus on effectiveness (Nilsen, 2015; Tabak et al., 2012; Tumilowicz et al., 2019). A comprehensive review of existing frameworks to evaluate which is best suited for the purposes of feeding guidelines was outside of the scope of this report. However, it is important to note that these frameworks generally recognize several common principles or steps that are used in D&I planning and practice (see Box 7-2).

A continuous and participatory iterative system, providing feedback on D&I plans and practices, will ensure that timely adaptations can be made to ensure successful implementation of feeding guidelines (Brown et al., 2017b). Adaptations to D&I activities, as a result of socioeconomic, cultural, demographic, and health care and food systems differences across contexts, are expected and need to be carefully documented (Chambers and Norton, 2016). Documentation of any adaptations that are made can guide future D&I efforts for feeding guidelines, taking equity principles into account (Chambers and Norton, 2016).

EVIDENCE GAPS

Based on its review of the existing feeding guidelines for infants and young children, the committee identified key evidence gaps. These gaps are part of the reason for inconsistencies in feeding recommendations, and thus research to fill these gaps is necessary to facilitate harmonization of guidance development across organizations. The areas of research highlighted

below are directed at addressing inconsistencies in the feeding recommen dations and improving understanding of best practices for communicatio and dissemination of feeding guidelines. The suggestions that follow ar by no means exhaustive, but rather they are the committee’s reflection o the materials it reviewed within its charge. Filling these evidence gaps wil require careful consideration of study design, feasibility, and resources, bu it was beyond the scope of this study to delve into such issues.

Evidence Gaps Related to Inconsistencies in Feeding Recommendations

The list below represents the committee’s expert judgment on key evidence gaps that arose after reviewing the topics for which there was inconsistency or lack of evidence in the guideline documents. There are other issues of interest to caregivers, stakeholders, researchers, and others, such as “baby-led weaning” (Brown et al., 2017a), that may also warrant further research, but those are not included below because the committee did not find mention of them in the recommendations abstracted from the eligible guideline documents.

• Recommended duration of exclusive breastfeeding and the age of introduction of complementary foods: Most recommendations stated “about” or “around” 6 months of age, but a few indicated an age range of 4–6 months and one Robert Wood Johnson Foundation-Healthy Eating Research specified “when the child is developmentally ready, which usually happens between 4 and 6 months” (Pérez-Escamilla et al., 2017). There is only one published randomized controlled trial on this issue among breastfed infants in a high-income country (Jonsdottir et al., 2012). Additional trials would be useful to understand the potential effects of introducing complementary foods at 4–5 versus at about 6 months of age.
• Continuation of breastfeeding: Most recommendations stated that breastfeeding should continue until at least 12 months of age, but several stated that breastfeeding may continue for 2 years and beyond. Further research on the effects (child and maternal) of breastfeeding beyond the first year of age in high-income contexts is needed, with adequate control for the baseline determinants of health (WHO, 2020).
• Recommended duration of use of infant formula for formula-fed infants: Most recommendations stated that infant formula should be the only breastmilk substitute until 12 months of age, but one guideline document gave an age range of 9–12 months. Similarly, some recommended delaying any cow milk until after 12 months, while others stated that cow milk could be introduced at 9–12 months or even as early as 6 months. Research would be useful to evaluate the consequences of replacing infant formula with cow milk at 9 months versus 12 months among formula-fed infants and partially breastfed infants.
• Duration of need for iron-fortified formulas, among formula-fed infants and partially breastfed infants during the first year of life: Recommendations differed with regard to whether iron-fortified formula is advised until 6 months, 9 months, or 12 months of

• age. Research on this issue would need to take into account the relatively large amounts of iron currently added to infant formula (10–12 mg/L) in the United States, and the role of iron-fortified complementary foods. Such research would likely require the use of biomarkers specific to iron deficiency, not just hemoglobin.

• Recommended age of introduction of potentially allergenic foods: Recommendations were generally consistent for introduction of allergenic foods in the second 6 months of life and not delaying introduction beyond 12 months of age. The evidence for complementary feeding practices to prevent food allergy is strongest for peanuts; additional research is needed for other allergenic foods, particularly eggs (including the preparation of egg used, as in whole, egg white, egg yolk). Additional research would also be useful to establish recommendations for how to introduce allergenic foods to infants at different levels of risk for food allergies, as this more nuanced guidance has only been established for peanuts, with the peanut recommendations for infants at high risk for peanut allergy being based on high-quality trials (Du Toit et al., 2015; Perkin et al., 2016). More evidence is needed for timing of peanut introduction for infants at moderate or low risk for the development of peanut allergy.
• Vitamin D supplementation: Most recommendations advised vitamin D supplementation for breastfed infants only, but some recommended it for all infants including nonbreastfed infants. There are also unresolved questions regarding the extent to which adequate vitamin D status of breastfed infants can be maintained via maternal high-dose vitamin D supplements during lactation, and if so, the risks and benefits of such an approach. Clarifying which infants do or do not need vitamin D supplements would focus the recommendations related to vitamin D and tailor them for populations with the highest need.
• Iron supplements for breastfed infants: The most recent recommendations did not advise routine use of iron supplements for exclusively breastfed infants, but they mentioned that some infants may be at risk for iron deficiency prior to 6 months (after which iron-rich complementary foods can meet iron needs). Research is needed on strategies to identify and appropriately treat such infants, such as those born with low iron stores (e.g., due to lower birth weight, shorter duration of gestation, immediate clamping of the umbilical cord, and maternal iron deficiency during pregnancy).
• Use of supplements or fortified products for vegetarian or vegan infants and young children: Most recommendations emphasized the need for extra attention to certain nutrients in the case of veg-

• etarian or vegan diets, but only some of them explicitly mentioned a need for fortified products or nutrient supplements. Research is needed on how to achieve nutritional adequacy for infants and young children who are fed vegetarian or vegan diets.

• Consumption of iodine from foods: There was consistency in recommending that foods for infants and young children be prepared without adding salt, but some guideline documents pointed to the need for adequate iodine intake and recommended using iodized salt, if salt was to be used in the food provided. Research is needed on iodine requirements for children under 2 years of age, and how to meet those requirements.

Evidence Gaps Related to Communication and Dissemination of Feeding Guidelines

The evidence gaps listed below were identified through the committee’s review of communication and dissemination strategies provided in existing feeding guideline documents.

• Identify the best ways to communicate to parents and other caregivers, health care providers, early care and education providers, program administrators, and policy makers regarding levels of evidence to support different recommendations for feeding children under 2 years of age, and implications for implementation.
• Design and evaluate protocols for communication and dissemination strategies specific to infant and young child feeding and tailored to the specific needs of diverse audiences, including low-resource and low-literacy populations, historically marginalized groups, and first-time parents. Include evaluation of the role of community health workers and peer counselors to facilitate effective dissemination of feeding guidelines.
• Evaluate the effect of telehealth use by health professionals, community health workers, and peer counselors to facilitate effective dissemination of feeding guidelines among socioeconomically vulnerable, rural, and other hard-to-reach populations.
• Evaluate the effect of different pathways for dissemination of guidelines via applications and social media. This could include development and validation of social network analysis methods for identifying key influencers and optimizing dissemination approaches.
• Evaluate effectiveness of dissemination strategies, based on D&I science, that target diverse audiences including parents and other caregivers, health care providers, early care and education provid-

CLOSING REMARKS

In spite of all of the differences in the ways that guideline documents were developed, it is encouraging that there was consistency in many of the feeding recommendations across a variety of authoritative organizations. Moving forward, however, it is important to encourage more rigorous, systematic, and ideally harmonized methods for developing recommendations for feeding infants and children from birth to 24 months of age, and for disseminating and implementing the recommendations using the principles of D&I science.