2Methodology

Targeted Website Searches

Recognizing that guideline documents and resources likely exist in the gray literature and may not be cataloged in databases, the primary approach for identifying guideline documents consisted of targeted website searches. The committee first created a list of potentially relevant government agencies, professional organizations, nonprofit organizations, and international groups. Using a snowball search approach, other agencies and organizations were added to the list when, over the course of a targeted website search, they were noted as a source of additional information. Agencies, groups, and organizations that ultimately did not qualify as authoritative were still searched, as their websites often provided links to primary guideline documents and resources within and beyond their organizations.

Website searches began on the homepage for each organization. The search strategy depended on the website structure; dropdown menus, topic area groupings, and document search options were used to find potentially relevant guideline documents. Staff conducted targeted website searches during the last 2 weeks of November 2019. Given resource constraints, duplicate searches were only conducted when the first search did not yield any results. The committee acknowledges that this approach may have led to the inadvertent omission of some eligible documents. The other search approaches (i.e., food-based dietary guideline search, expert input, and database searches) were intended to supplement the targeted website searches, and help to fill in potential gaps.

Food-Based Dietary Guideline Search

The Food and Agriculture Organization of the United Nations maintains a website that catalogs national food-based dietary guidelines and food guides (FAO, 2019). This catalog was used to search for English-language documents from high-income countries that included feeding recommendations relevant to infants and young children. National food-based dietary guidelines and related resources from eight countries were included for screening (see Figure 2-2).

FIGURE 2-2

Flow diagram of how the Food and Agriculture Organization of the United Nations' catalog of national food-based guidelines was used to identify potentially relevant guideline documents. ^a As cataloged by the Food and Agriculture Organization of the United (more...)

Expert Input

Members of the committee contacted international experts about English-language versions of national feeding guideline documents and supporting materials from Denmark, Germany, Italy, and Norway. Documents available in English were included in the committee's screening process. The committee also took advantage of its expertise to identify potentially relevant resources not identified through other search approaches.

Database Searches

As an adjunct to its targeted website searches, the following databases were used to search for relevant guideline documents: CPG Infobase, ECRI Institute Guideline Trust, International Guideline Library, GuidelineCentral, NICE Evidence Search, and PubMed. The search strategy was developed with the assistance of a senior librarian and was carried out by staff (see Table 2-1). Filters were used for language, publication type, age group, and publication date as available within each database. Searches included documents published through the first week of December 2019. Title and abstract screening from these searches was not conducted in duplicate owing to resource limitations. The committee acknowledges that this approach may have led to the inadvertent omission of some eligible documents.

TABLE 2-1

Summary of Database Searches.

Document-Level Eligibility Criteria

Eligibility criteria used to screen the identified guideline documents are presented in Table 2-2. Screening focused on whether or not the identified document contained any recommendations on what or how to feed infants and children younger than 2 years of age; feeding recommendations encompassed human milk, infant formula, other foods and beverages, and supplements. Guideline documents that included a summary of the state of science in an attempt to make recommendations, but which found inadequate evidence to do so, were considered eligible. Given the committee's charge of reviewing “public-facing” guidelines, eligible documents were readily accessible on the Internet. As noted previously, the committee's search, screening, and selection focused on guideline documents relevant to healthy mother–infant dyads in the general population of high-income countries. Guideline documents specific to preterm infants were excluded owing to challenges of drawing comparisons (e.g., related to the “age” of the infant) and the special considerations related to medical management.

Recommendation-Level Eligibility Criteria

Guideline documents that screened eligible differed in span and scope. While some focused on a specific topic (e.g., vitamin D supplementation), others provided extensive guidance on a wide range of topics. To identify a collection of feeding recommendations that could be compared for consistency, the committee established recommendation-level eligibility criteria (see Table 2-3).

TABLE 2-2

Guideline Document-Level Eligibility Criteria.

Challenges Encountered During Screening

Guidelines and their underlying evidence base exist in a wide range of formats. Government agencies, for instance, may publish a set of public-facing guideline documents informed by a separate scientific report. Professional organizations often publish position papers, policy statements, or practice papers or guidelines accompanied by a narrative review. Other organizations may release guideline reports with different structures and content. The format, terminology, target audience, and approach to developing guideline documents may also change over time within an organization.

Even with detailed eligibility criteria, screening the identified materials was not straightforward. As the committee's charge was to not only compare the consistency of guidance but also to document the type of evidence used to support each recommendation, it was imperative to find the source document of the feeding recommendations. One substantial challenge the committee faced was determining what constituted guideline documents and resources, as opposed to communication or dissemination materials. Communication materials typically include lay-friendly feeding guidance, but do not typically provide the methodological or scientific underpinning of each recommendation. Although these types of materials often provide specific feeding recommendations, the committee did not consider these resources to be eligible guideline documents. The dividing line between a guideline document and a communication product, however, is not clear-cut, and the determination is somewhat subjective. Although not a formal criterion, the committee considered whether the document provided information about the guideline development process or evidence supporting the recommendation, rather than being a simple, lay-friendly resource, in distinguishing between guidelines and communication materials.

TABLE 2-3

Recommendation-Level Eligibility Criteria.

A major resource that provides guidance on feeding infants and young children is the American Academy of Pediatrics' Pediatric Nutrition (AAP Committee on Nutrition, 2019). This comprehensive book includes a number of sections related to what and how to feed young children. Despite its preeminence, this handbook was determined not to meet the committee's eligibility criteria because the guidance in the book is intended to align with current AAP policy statements. Accordingly, the policy statements were considered the origin document of the recommendations. The committee acknowledges the importance of this resource to the field, but has opted to include and abstract policy and clinical practice position statements from AAP.

Even scientifically oriented documents with clear statements of guidance on what or how to feed infants and young children posed challenges. For example, in its 2019 report, Appropriate Age Range for Introduction of Complementary Feeding into an Infant's Diet, the European Food Safety Authority Panel on Nutrition, Novel Foods and Food Allergens noted that the purpose of the document was to assess the scientific evidence, and that providing public health recommendations was considered out of scope (EFSA Panel on Nutrition, Novel Foods and Food Allergens, 2019). Nevertheless, the document provides clear statements of when to introduce complementary foods into infants' diets and was ultimately considered eligible. In another example, the Academy of Breastfeeding Medicine published a protocol related to supplementary feedings for healthy, term breastfed newborns (Kellams et al., 2017). This guideline document provided guidance related to what and how to feed infants; however, the document was ultimately excluded because the recommendations were focused on clinical management and did not align with the committee's interpretation of its charge.

Another challenge the committee encountered was assessing documents consistently, despite inconsistent formats. For instance, the Australian government has a suite of publications including the guideline document, the supporting literature review, public consultation on the feeding guideline document, and a summary of the recommendations (NHMRC, 2012). Taken in isolation, the guideline document does not provide the full context of the scientific review supporting each recommendation. Accordingly, when guideline documents were supported by a separate literature search or documentation of the evidence, the committee considered these publications collectively, as necessary. Some guidelines from international groups and organizations had an English guideline document or summaries, but the corresponding evidence base was not available in English (Bührer et al., 2014; Davanzo et al., 2015). These documents were ultimately excluded because the committee could not determine the type of evidence underlying the feeding recommendations. The heterogeneous formats of guidelines and associated evidence reviews posed challenges in identifying and assessing the documents in a systematic and consistent way.

Document-Level Abstraction

The following information was abstracted from each of the eligible guideline documents and resources:

• Organization(s)—The agencies, organizations, or groups that published, released, or otherwise directly participated in the development of the guideline document were recorded.
• Year—The year the guideline was published or otherwise updated was recorded.
• Target country or region, as specified in the guideline document—Unless otherwise stated, it was assumed that professional organization guidelines were for the country or region specified in the organization's name (e.g., the target region for a guideline document from the European Society for Paediatric Gastroenterology, Hepatology and Nutrition was Europe).
• Target audience of the document, as specified in the guideline document—Document target audiences may include, but are not limited to, health care providers (e.g., physicians, nurse practitioners, nurses, dentists, registered dietitian nutritionists, and other nutritional professionals), parents and guardians, early care and education providers, program administrators, and policy makers.
• Inclusion of conflict-of-interest statement for guideline authors—This included any statement about the guideline authors' potential conflicts.
• Guideline development methodology—If available, the statement of methods for arriving at the recommendations was recorded as described in the document.
• Inclusion of guidance on communication and dissemination—The abstractor made note of any guidance contained within each eligible document related to changing knowledge, attitudes, and/or behaviors, and the channels by which to spread feeding guidance.
• Inclusion of research recommendations—The abstractor noted when a guideline document included a specific section or statement of research recommendations, future work, knowledge gaps, future research needs, or similar text. The content of those research recommendations, however, was not abstracted.

Recommendation-Level Abstraction

The committee sought to capture the following information about each eligible recommendation:

• Recommendation—The statement of recommendation was recorded verbatim. Statements of recommendation were captured, as presented in the guideline document (e.g., in paragraph form, as a stand-alone sentence, or a statement accompanied by a bulleted list).
• Type of evidence that directly maps to the recommendation—Based only on the information provided in the guideline document itself or, in select cases, in the suite of guideline-related publications, the abstractor noted whether the type of evidence used to support the recommendation could be determined and, if so, what type was used. In an attempt to standardize responses, options for type of evidence included

  • Systematic review, included as part of the guideline document: This option was selected when the recommendation could be mapped to a systematic review question included in the guideline document.
  • Systematic review, specifically prepared for the guideline document, but published elsewhere: This option was selected when the guideline document was part of a suite of guideline-related publications. This option was not used when a guideline document only cited a systematic review as its evidence within a narrative review of the evidence.
  • Narrative review, included as part of the guideline document: This option was selected when the recommendation could be mapped to a section of the guideline document that discussed the evidence in a narrative way, and cited at least two publications (regardless of publication type).
  • A single citation: This option was selected when the guideline document stated that the recommendation stemmed from one publication (e.g., a guideline document noting that the recommendation was drawn from a government website). The resource was recorded.
  • Other: This option was selected when the type of evidence that mapped to the recommendation did not fall into any of the above categories (e.g., a guideline document noting that the recommendation was informed by reports from other organizations; a guideline document describing recommendations based on a “systematic literature search” rather than a “systematic review”).
  • Could not be mapped: This option was selected when the type of evidence used to support the recommendation could not be discerned, and could include instances when no evidence was presented to support a recommendation, or when the link between the evidence and recommendation was unclear.
• System used to determine strength of evidence—The specific system used to evaluate the strength of the evidence (e.g., Agency for Healthcare Research and Quality, Grading of Recommendations Assessment, Development and Evaluation) supporting the recommendation was noted.
• Strength-of-evidence rating—The verbatim word or phrase used to characterize the strength of evidence supporting the recommendation (e.g., Low, Grade B) was recorded.