Parental Decision-Making Around Introducing Complementary Foods: An Integrative Review

Parents assume many responsibilities when they have a new infant. One of the most essential early parenting roles is feeding their infant to ensure they survive and grow (Slaughter & Bryant, 2004). In fact, much of parents’ interactions with their infant in the first year are centered around feeding (Hofferth & Sandberg, 2001; Negayama et al., 2012). Because feeding has important health implications for the infant, parents may feel anxiety and stress around their infant’s feeding (Fletcher et al., 2017; Howe et al., 2019; Sleigh, 2005). This stress can continue throughout the first year as the parent’s and infant’s roles and relationships around feeding change.

A major shift in the parental feeding role is the introduction to CF. CF, or complementary feeding, refers to the introduction of foods other than breast milk or formula to an infant (also known as the introduction to solid foods). Parents navigating the transition to CF encounter decisions about when and how to introduce CF to support their infant’s feeding skills and nutrition. For example, parents must make decisions at every mealtime about what, when, and where to offer CF (Satter, 2000). Further, feeding is a relational endeavor, and this is only emphasized during the transition to CF as parents and infants renegotiate their roles in feeding (Henton, 2018; van Dijk et al., 2012, 2018). Critically, the parent’s initial decision about when to start CF with their infant can have long-term effects on the infant’s health and eating skills (e.g., Palmer & Makrides, 2012; Spill et al., 2019). Therefore, parents face an important initial decision about when to introduce CF, as well as downstream implications and decisions, during this critical shift in their feeding role as a parent.

Given the importance and complexity of the introduction to CF, parents actively seek information about how to navigate this transition with their infant. For example, parents of low-risk infants report using sources like health care providers, the internet, friends, and their own upbringing to gain information about the introduction to CF (Boulanger & Vernet, 2018; A. L. Garcia et al., 2019; Moore et al., 2012; Walsh et al., 2015). In addition, multiple family members may offer their varying opinions on the best way to introduce CF, potentially creating conflict and stress for parents during mealtimes (A. L. Garcia et al., 2019). With this growing volume of input, parents are confused by mixed information across sources and desire more support (Begley et al., 2019; Moore et al., 2012). Parents may be stressed and challenged by navigating this transition given the overwhelming and conflicting information.

Both parents and health care providers may look to professional medical organizations for guidance on navigating the transition to CF. Currently, the Centers for Disease Control and Prevention (CDC), the American Academy of Pediatrics (AAP), and the Dietary Guidelines for Americans recommend that parents introduce CF to their infants at around 6 months of age and explicitly state that they do not recommend introduction before 4 months (AAP, 2020; CDC, 2019; United States Department of Agriculture [USDA], 2020). These guidelines are based on research touting the risks of introducing CF too early due to an increased long-term risk of obesity, allergies, and picky eating as well as short-term nutritional deficits (Gingras et al., 2019; Morgan et al., 2004; Palmer & Makrides, 2012; Shim et al., 2011; USDA, 2020) or too late due to influences on oral motor skill development, infant food repertoires, and feeding problems (da Costa et al., 2017; Emmett et al., 2018; Simione et al., 2018; Spill et al., 2019). Based on these risks parents must strike a balance of introducing CF within a narrow time frame to reap the optimal benefits for their infant.

It may not be surprising that parental feeding practices often do not match the clinical ideal. In the United States, a recent CDC report stated that 31.9% of parents introduce CF before their infant reaches 4 months of age (Chiang et al., 2020). High rates of early introduction have been observed in other developed countries as well, including Great Britain, Italy, and Australia (Costantini et al., 2019; Scott et al., 2009). Given the long-term health implications for infants, public health interventions have attempted to address this discrepancy. For example, in 2022 the United Kingdom launched an initiative to educate families around when and how to introduce CF (Department of Health and Social Care, 2022). However, in the United States, updated supports for parents are lacking. Notably, the AAP has a “Healthy Active Living for Families” project for obesity prevention in children. However, it is informed by parent interviews from a single U.S. city in 2010 and 2011 and does not focus on the introduction to CF specifically (Radecki et al., 2012). Current U.S. guidelines and interventions do not appear to be influential in parental decision-making around CF introduction. Therefore, to effectively support families in navigating this transition it is essential to identify the most up-to-date factors and information sources parents are using to inform the introduction to CF.

Given the importance of the introduction to CF, multiple researchers have summarized literature on this topic. However, the existing reviews have focused on a specific methodology (e.g., only qualitative studies) and/or factor (e.g., cultural norms; A. H. Garcia et al., 2016; Spyreli et al., 2021). By limiting the research by these criteria, the potential applicability is lessened. In contrast, this integrative review examines quantitative, qualitative, and mixed methods literature together to holistically examine parental decision-making related to introduction to CF. Further, other comprehensive reviews have focused on specific geographical regions (e.g., developing countries), but none have focused on the United States. Given the probable influence cultural and societal factors can have on parental decision-making, we examine the United States specifically to address the gap in the literature regarding this country, with a focus on implications for future interventions.

The Social Ecological Model (SEM) is a theoretical framework that can guide the organization and synthesis of the factors influencing parental decision-making around CF introduction to provide guidance for developing successful health interventions (Glanz & Bishop, 2010). Importantly, SEM recognizes that influences on parental decision-making are layered and multifaceted, including individual, interpersonal, community, and societal levels. The individual level refers to factors about the parent and their beliefs, knowledge, and attitudes, as well as characteristics like race, educational attainment, and weight (McLeroy et al., 1988). The interpersonal level refers to influences from the parent’s relationships, such as their interactions with extended family, friends, and health care providers, but also the reciprocal relationship between them and their infant. At the community level, the parent may be influenced by broader factors like social institutions and the setting in which they live. The societal level represents more global influences, such as the economic climate, the parent’s cultural norms, and health care policies.

SEM was chosen as a guiding framework for this work because it acknowledges the multiple spheres of influence in an increasingly connected world. Today parents making health care decisions for their infant are exposed to broad influences through the internet and social media (Pretorius et al., 2019). SEM considers this broad network to evaluate the importance and interaction between spheres of influence. In addition, SEM has been used extensively around parental feeding decisions, such as parental decision-making around breastfeeding (Dunn et al., 2015; Reeves & Woods-Giscombé, 2014; Roll & Cheater, 2016) and infant feeding among immigrant women in Australia (Joseph et al., 2019). SEM is appropriate as a theoretical framework for this work due to its acknowledgment of broad influences as well as its prior use in feeding-related work.

Parental decision-making around feeding their child, particularly at the time of introduction to CF, has important implications including risks associated with nutrition, obesity, and feeding problems. Parents are faced with a shift in their feeding role and multiple spheres of influence and are seeking support around these transitions. Understanding the current influences on parental decision-making can provide evidence for designing effective interventions to aid parents and reduce stress. The purpose of this review is to provide a modern, integrative synthesis of the available literature in the United States to address the following aims:

Studies were included if they (a) were conducted in the United States, (b) utilized qualitative, quantitative, or mixed methods research methods, (c) examined parental decision-making, perceptions, or influences regarding introduction of CF, (d) included at least some infants aged 0 to 24 months, and (e) were published as full-text articles in the English language. Studies were excluded if they (a) did not report any factors or did not connect factors to initiation of CF, (b) only examined CF outcomes related to nutrition or hygiene, (c) only reported health care providers’ perspectives, or (d) were published as commentaries or reviews. Interventions were also excluded to avoid bias around influences on introduction.

A health sciences librarian searched through the final search date of May 26, 2022 in the following databases: PubMed, Cumulative Index of Nursing and Allied Health Literature Plus with Full Text (EBSCOhost), and Scopus (Elsevier). After some initial pilot searches, we opted to use two search strategies. Both strategies included a combination of subject headings and keywords. The first strategy included these four key concepts: infants, caregivers/parents, solid foods, and decision-making. The second strategy included these five key concepts: infants, caregivers/parents, solid foods, factors, and the United States. To be more comprehensive, we did not include terms for the United States in the decision-making strategy because many studies that occur in the United States do not state so in the titles or abstracts. However, we found that quantitative studies around associated factors for CF make use of national or regional data sets and do indicate a U.S. location in the abstract, and therefore this was included in the second strategy. The searches were filtered to English language and to publication dates of 2012 to 2022. This date filter was applied due to the advent of social media. The complete, reproducible search strategies can be found in the Supplemental material.

We used Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia, available at www.covidence.org) to complete the screening and data extraction processes. Two researchers independently screened the titles and abstracts of all references for eligibility criteria and resolved conflicts through discussion. We repeated the process for the full-text screening stage.

Two researchers independently assessed the quality of the studies using the Mixed Methods Appraisal Tool (MMAT; Hong et al., 2018). We chose this tool due to its flexibility in assessing qualitative, quantitative, and mixed methods research. We assessed each included study for two items: (a) clear research questions and (b) collected data that address these questions. Then, each included study was assessed on five design-specific criteria. Overall agreement between raters was 83.25%. We resolved discrepancies around quality assessment through discussion.

We created a data extraction template within Covidence that included the following information: article title, first author, publication year, study aim, study design, participant type, age range of infant participants, inclusion and exclusion criteria, number of participants, cultural group (if applicable), research instruments, percentage of infants introduced to CF by age, factors associated with and not associated with timely, early, and late introduction of CF, sources of information, other key findings, and implementation tips. We excluded missing or unclear data from the data extraction tool.

Two researchers independently extracted data for each study. We resolved discrepancies through discussion. We examined three subgroups based on the population studied: (a) studies that sampled on a national level and/or with no specific sub-population (hereafter referred to as the general sample), (b) studies that sampled by a parental or infant health characteristic, and (c) studies that sampled by race, ethnicity, or socioeconomic status. Within each subgroup we examined three categories: (a) age of introduction to CF, (b) factors associated and not associated with age of introduction, and (c) sources of information. We created data displays for each subgroup and organized using the SEM (Figures 1-3). We extracted themes both within and between subgroups.

The flowchart of the screened and included studies is shown in Figure 4. A total of 29 studies met eligibility criteria and were included in the final review. See Tables 1 to 3 for study characteristics. The quality of the studies was assessed using the seven MMAT criteria. No studies were excluded on the basis of poor quality; all studies rated fulfilled at least five of the seven quality criteria and none of the deficits were deemed severe enough to disqualify a study from exclusion. Having a representative sample and having a low risk of nonresponse bias were most commonly missing.

The majority of studies targeted specific populations within the United States including a particular racial, ethnic, or socioeconomic group (Table 3) or a particular maternal or infant health characteristic (Table 2). Most studies sampled only mothers, although six sampled caregivers or guardians, one sampled co-parenting mothers and fathers (Thullen et al., 2016), and one sampled pairs of mothers and caregivers (Johnson et al., 2019). Six studies utilized nationally representative samples (Barrera et al., 2018; Braid et al., 2015; Clayton et al., 2013; Gaffney et al., 2014; Kim & Gallien, 2016; Roess et al., 2018), but most studies (23/29) recruited from a single geographic area. Geographic regions of the United States were not equally represented with more studies taking place in the Northeast and the South.

Of note, Clayton et al. (2013) and Gaffney et al. (2014) both utilized the same sample (the Infant Feeding Practices Study II). However, Gaffney et al. (2014) limited their sample to matched individuals with and without postpartum depression (PPD) and therefore both studies were included and analyzed as unique studies. Khalessi and Reich (2013) and McManus et al. (2017) both utilized samples from the Baby Books study but used different exclusion criteria; therefore, the samples utilized for data analysis were different and these studies were also independently analyzed.

The first category examined across subgroups was the age at which infants were introduced to CF. First, it is important to note that the definition of the appropriate age to introduce CF varied across studies (see Table 4). However, despite the inconsistency in defining an appropriate age to introduce CF, most studies agreed that introducing CF before 4 months was early.

Overall, across subgroups, most infants were introduced to CF between 4 and 6 months of age, with an overall mean age of introduction of 5.3 months (SD: 2.1 months) across studies that reported an average age of introduction (see Figure 1). Specific sub-populations may be at particular risk of introducing CF before 4 months and broadly, rates of introduction prior to 4 months were high (M: 29.6%; SD: 18.6%). For example, a majority of preterm infants (64.5%) were introduced to CF before 4 months adjusted age (Braid et al., 2015). In contrast, only 4.6% of infants with Down syndrome and 7.0% of infants whose parents had a history of food allergy were introduced to CF before 4 months (Cochran et al., 2022; McKean et al., 2015). While introduction after 6 months was rarely defined or measured, it was rare, with an average of only 9.6% of infants introduced at 6 months or later. However, introduction after 6 months was common among a Marshallese (from the Marshall Islands) sample living in Arkansas where CF was introduced at an average of 10.2 months by mothers and 11 months by caregivers (Johnson et al., 2019).

Across studies, there was little consistency in the variables examined in relation to the age of introduction to CF; therefore, themes across subgroups are presented together below according to the SEM framework.

Sources of information about when to introduce CF were primarily explored in qualitative studies (Figure 3). The influence of a health care provider, such as a pediatrician, was cited across studies as a common source of advice (Clayton et al., 2013; Cochran et al., 2022; Hartman et al., 2016; Horodynski & Mills, 2014; Lindsay et al., 2017; McFarren et al., 2020). Friends and family, particularly female relatives like the mother’s own mother, were also frequently cited (Horodynski & Mills, 2014; Johnson et al., 2019; Lindsay et al., 2017; Thullen et al., 2016; Wasser et al., 2013; Woo Baidal et al., 2015). However, it was not uncommon for the advice of friends and family to conflict with that of a health care provider (Horodynski & Mills, 2014; Lindsay et al., 2017). In these instances, the weight placed on different sources of information varied by population. For example, Brazilian immigrant mothers felt more comfortable taking the advice of friends and family over that of a health care provider (Lindsay et al., 2017). Similarly, low-income adolescents received conflicting advice from health care professionals and family members and reported the advice from their mothers was ultimately most impactful. (Hartman et al., 2016; Horodynski & Mills, 2014). Alarmingly, in a nationally representative sample, Clayton et al. (2013) reported that 55% of mothers who introduced CF before 4 months did so at the recommendation of a health care provider. Finally, WIC was cited among Latinx and Brazilian immigrant mothers as valuable sources of CF information (Lindsay et al., 2017; McFarren et al., 2020).

A consistent theme across studies was the need for clear, consistent, evidence-based guidelines on the appropriate age to begin CF (Barrera et al., 2018; Braid et al., 2015; Cochran et al., 2022; Hartman et al., 2016; McKean et al., 2015). This was in part due to the confusion among parents and providers about the best age for CF introduction (Barrera et al., 2018; Hartman et al., 2016; McKean et al., 2015). In addition, providing families culturally sensitive counseling, acknowledging maternal beliefs, and helping families establish and reach goals around infant feeding were recommended (DiSantis et al., 2013; Doub et al., 2015; Kim & Mathai, 2015; Lee & Brann, 2015; Lindsay et al., 2017; McFarren et al., 2020; McKean et al., 2015). Given the influence family members can have on feeding decisions, using intervention methods appropriate for both mothers and family members was suggested (Johnson et al., 2019; Wasser et al., 2013; Woo Baidal et al., 2015). Methods for disseminating information about CF to families included counseling during well-child visits, motivational interviewing, text messaging, social media, print newsletters, home visits, the internet, and/or parenting groups (Horodynski & Mills, 2014; McFarren et al., 2020; Woo Baidal et al., 2015). In addition, mental health support, provider training, and screening were recommended practices to support parents in introducing CF appropriately (Cochran et al., 2022; Hurley et al., 2015; McKean et al., 2015; McManus et al., 2017). There were also suggestions made that were specific to the population studied. Most notably, guidelines specific to the needs of preterm infants and infants with Down syndrome were recommended to help providers guide families (Braid et al., 2015; Cochran et al., 2022). Mothers who are obese or overweight were also included as a population that would benefit from tailored intervention (Mehta et al., 2012; Vadiveloo et al., 2019).

This review highlights the need for additional research efforts focused on the introduction to CF from diverse perspectives. For example, from the lens of the SEM framework, most of the research has focused on variables at the individual level. Further, even at the individual level there is little consistency in the factors examined, making it challenging to compare findings across populations. While most studies examined some maternal demographic characteristics, the characteristics chosen and how they were defined varied. In addition, other individual-level factors like beliefs, knowledge, infant characteristics, and maternal mental health were inconsistently incorporated into statistical models. Future research would benefit from more consistent examination of factors across study populations.

In addition to deficits at the individual level, the interpersonal level received surprisingly little attention given the overwhelming research on the impact family members can have on breastfeeding (e.g., Ogbo et al., 2020). In those that reported it, health care professionals and friends and family were the most common sources of information about introducing CF (Clayton et al., 2013; Cochran et al., 2022; Hartman et al., 2016; Horodynski & Mills, 2014; Lindsay et al., 2017; McFarren et al., 2020; Thullen et al., 2016; Wasser et al., 2013; Woo Baidal et al., 2015). However, there was often conflict between the advice of health care professionals and friends or family, with mothers relying on friends or family because they understood their cultural beliefs and practices (Horodynski & Mills, 2014; Lindsay et al., 2017). The influence of family members may only be increasing with the increasing rates of multigenerational households in the United States (Cohn et al., 2022). Therefore, health care providers, like pediatric nurse practitioners and pediatricians, would benefit from discussing familial and cultural practices and beliefs for introducing CF with families (Dickin et al., 2021). Within a culturally sensitive mindset health care providers can then assist families in incorporating healthful practices, such as delaying introduction to CF past 4 months, while honoring cultural traditions, such as specific foods that families value introducing first. Finally, this review was narrowed to the preceding 10 years to account for the influence of the internet and social media; however, these resources were seldom inquired about in the included research. Given the growing influence of social media on a variety of parenting behaviors, future inquiry into the impact of social media on CF patterns is warranted (Doub et al., 2016; Griauzde et al., 2020; Kallem et al., 2018; Pretorius et al., 2019). This can be integrated into research that examines the community and societal levels to inform researchers and practitioners about the broader factors influencing parental decision-making around CF introduction.

Before interventions to support parents can be developed, defining appropriate guidelines is critical. One area with a dearth of evidence that is not addressed in current AAP or CDC recommendations is the late introduction of CF. Late introduction to CF (after 9 or 10 months) can lead to feeding problems, oral motor delays, and picky eating (Cameron et al., 2012; Coulthard et al., 2009; Emmett et al., 2018; Northstone et al., 2001). However, in the reviewed articles, the number of infants who were introduced to CF any time after 6 months of age were reported as one group. By doing this, critical differences in age of introduction are washed out. For example, Johnson et al. (2019) found that Marshallese mothers didn’t introduce CF until an average of 10.2 months, an important effect that could be missed by averaging across a sample. Therefore, late introduction, and why parents decide to delay introducing CF, should be carefully measured in future studies.

When deciding when to introduce CF, the CDC and AAP state that it is appropriate to do so around 6 months based on the infant’s developmental readiness signs. However, only one of the studies reviewed reported on parental awareness, understanding, or use of signs of developmental readiness for CF in their decision-making process. Specifically, only parents of infants with Down syndrome referred to developmental readiness as a factor in their decision-making (Cochran et al., 2022). Developmental readiness may be more important than the specific age an infant is introduced to CF (Chung et al., 2014). Therefore, future research should examine parents’ knowledge of and identification of infant developmental readiness signs at the age of CF introduction and how this relates to long-term outcomes to offer a more accurate and individualized look at the impact of age of introduction.

Finally, while most studies excluded populations of infants with medical complications, those studies that reported on infants with health complexities emphasized the need for guidelines specific to these populations, which currently do not exist (Braid et al., 2015; Cochran et al., 2022; Elfzzani et al., 2019; Liotto et al., 2020). The lack of specific guidelines puts infants at risk because high-risk infants may not follow the same developmental timeline as other infants and may exhibit different readiness cues, making deciding when to introduce CF challenging for families and health care providers (Cochran et al., 2022). The lack of guidelines also puts parents at risk because parents of high-risk infants already experience greater feeding-related concern and the lack of clarity can add stress to this transition (Howe et al., 2019). Future research should develop evidence-based guidelines around introducing CF to high-risk infants to guide practitioners and families in successfully introducing CF.

This study has several limitations. First, we may have missed relevant articles due to our search strategy and our selected databases. Specifically, we only included studies that directly related influencing factors to parental practices. Therefore, studies that examined factors more broadly were not incorporated into this review but could be used to inform future studies. Second, our review includes studies that ranged in quality and rigor. Some included studies were not generalizable to larger populations, which should be considered when applying to future work. However, these studies do add valuable information specific to sub-populations that can be applied to future targeted interventions. In addition, the representation of regions of the United States was varied and therefore was not compared in this study. Future reviews may consider whether varying geographical regions of the United States contribute to CF practices. Finally, there was variation in the definitions of early and appropriate introduction to CF across studies, as well as some studies that did not clearly define these concepts. This made comparison across studies challenging. Future studies should include clear definitions and mirror current guidelines so that studies can be directly compared.