A Qualitative Analysis of a Caregivers’ Experience of Complementary Feeding in a Population of Native Hawaiian, Other Pacific Islander and Filipino Infants: The Timing of the Introduction of Complementary Foods, and the Role of Transgenerational Experience

The aim of this study was to investigate caregivers’ experiences of complementary feeding (CF) among the Native Hawaiian and Other Pacific Islander (NHPI), and Filipino populations. Research focused on the timing of CF commencement, and the influence of transgenerational experience on feeding practices. The experiences and practices of those who fed human milk exclusively (HME), were compared to those who included infant formula (F&HM). Caregivers of a subset of 32 infants who were participating in a larger longitudinal study relating to CF and diet diversity, took part in voluntary in-depth interviews relating to CF practices. Interviews were recorded and transcribed. Two researchers analyzed interview transcripts. Interrater reliability and saturation were established. Institutional Review Board exemption was confirmed prior to study commencement. Interviews with 29 caregivers of infants were included in this study. Only infants of the F&HM group had an early introduction to complementary foods (<4 months of age). Caregivers reported receiving conflicting advice from healthcare professionals (HCPs) in relation to timing of the introduction of complementary foods. Nonetheless, the majority of caregivers reported following the advice of HCPs. Extended family (including grandparents) played less of a role in infant feeding, compared to previous generations. While transgenerational practices were valued and included, ultimately, the perceived health and safety of the practice for infants influenced decisions.


Introduction
Complementary foods, or any solids or liquids other than human milk or infant formula, are needed in order to meet nutritional requirements from approximately 6 months of age [1,2]. Behavioral cues of "readiness to feed" are often observed at the 4-6 month age range, correlating with when infants have physiologically developed to safely tolerate the introduction of complementary foods [3][4][5][6]. Although the World Health Organization (WHO) and the American Academy of Pediatrics (AAP) recommend exclusive breastfeeding from birth to 6 months due to its numerous benefits for both maternal and infant health [7][8][9].
CF practices of those who exclusively breastfed were different from the practices of those who also included infant formula.

Materials and Methods
Prior to commencement of data collection, all participants provided written consent. The University of Hawai'i Institutional Review Board deemed the study exempt.
This analysis was part of a larger, quantitative study exploring CF among NHPI and Filipino infants living in O'ahu, Hawai'i [36]. The study took place between Spring 2018 and Spring 2019. The longitudinal study investigated CF (timing and the types of food offered) among infants 3-6 months of age, and diet diversity among those 6-12 months of age. A total of 70 infants were included in the longitudinal study [36]. Inclusion criteria were: age (3-12 months), engagement in CF, and race/ethnicity (at least one of NHPI and/or Filipino). Convenience sampling, through community programs, community events as well as personal and professional networks, was used to recruit participants. The methods used for data collection in this analysis were semiqualitative.
For this study, the infant feeding method was quota-sampled from a subsample of participants from the larger study. The participants were divided into two groups based on method of infant-feeding: feeding human milk exclusively (HME) or infant formula and human milk combined (F&HM). The goal was to interview 32 participants (two groups of 16 participants each). Caregivers self-reported information on infant feeding. Caregivers were compensated with a gift card after the interview. Three foster care participants were excluded as their distinct situation and small sample size meant that analysis of their interviews would not elicit meaningful results for this sub-group. Therefore, only the interviews of caregivers (29 mothers and 2 fathers) of 29 infants were included in this study. Data replication and redundancy, indicating data saturation, occurred during the analysis of interview 8 [37]. Researchers agreed that caregivers of 29 infants would elicit sufficient data through the interview responses.
A researcher-designed survey administered through a secure online web application collected demographic details of the infants and caregivers, including infant age, sex, race/ethnicity, benefits received (Special Supplemental Nutrition Program for Women Infants and Children (WIC) [WIC]/Supplemental Nutritional Assistance Program [SNAP]), feeding since birth (HME or F&HM), household size, and cultural identity. The survey responses from caregivers of the 29 infants involved in this study were descriptively analyzed.
In-depth, semi-structured interviews provided the qualitative data used in this study. The research question guiding the qualitative data collection was: what influences decisionmaking on first foods of NHPI and Filipino infants? The protocol used by interviewers was adapted from research by Fialkowski et al. [35]. This protocol facilitated the interviewers' ability to guide conversations in order to obtain relevant information and gave flexibility to acknowledge and respond to the matter at hand (such as emerging ideas). The interview involved 10 leading questions and associated follow-up questions. Participants consented for the interviews to be audio-recorded, and interviews took place at a time and private location convenient for participants.
Caregivers' responses to a subset of 3 of the 10 leading questions related to the timing of the introduction of complementary foods and the caregivers' transgenerational experience of CF practices were analyzed for the purpose of this study (Table S1). Responses from caregivers who fed HME were compared to those who fed F&HM. The researchers agreed that these questions would reveal when complementary foods introduction occurred, and whether this was what caregivers were advised to do. For the purpose of this study, the CF guidelines of the WHO and AAP are used as reference for the appropriate timing of the introduction of complementary foods (from 6 months) [7,9]. The remaining 2 questions would provide insight into transgenerational CF practices among the NHPI and Filipino populations.
All interview transcripts were coded first, then key ideas and recurring themes were compared within and between groups. To ensure reliability, 4 interview transcripts were chosen at random, and 2 researchers analyzed and coded these according to a set codebook, using appropriate themes. A Cohen K of 0.72 was achieved after this initial round of analysis, signifying a level of moderate agreement [38]. Discussions occurred, and the codebook updated accordingly. Another round of analysis of 4 additional transcripts selected at random was completed by the same two 2 researchers using the updated codebook. This second round of coding resulted in a Cohen K of 0.98, which signified a strong level of agreement [38]. One researcher, using the established codebook, then analyzed the remaining 21 transcripts. The analysis of transcripts reduced the data and identified core themes and meanings [39]. Representative quotes were used to indicate transferability and qualitative trustworthiness [40]. Comparison of frequency of occurring themes between groups facilitated exploration of differences in infant feeding practices and behaviors. All participants were identified using study identification numbers with qualitative analysis occurring in Google Sheets.
Descriptive analysis of participants' demographic data was carried out in IBM SPSS Statistics Version 27.0 (SPSS Inc.: Chicago, IL, USA). Although no statistical tests were used in analysis as this was not the intent of the study, findings are presented semi-quantitatively for the purpose of comparison of the feeding practices and experiences of the F&HM and HME groups.

Participant Demographics
Caregivers of 29 NHPI and Filipino infants took part in this study (Table 1). In total, 13 infants were fed F&HM, and 16 were fed HME. Over half of the F&HM were fed formula since birth. Half of the caregivers in the HME group had an education attainment of an undergraduate-level degree or more, compared to under one-third of F&HM caregivers. More of the HME group were stay-at-home parents, while the F&HM group had a higher proportion of caregivers in part-or full-time employment. There was a greater variation between caregivers' annual household income in the F&HM group. More HME caregivers lived with extended family members (including grandparents). The majority of caregivers in both groups reported Native Hawaiian as the race/ethnicity of the infant, which was similar to the breakdown of caregivers' self-reported race/ethnicity.

Overview of Semi-Structured Interview Themes
The principal findings from the interviews are semi-quantitatively and thematically presented, using verbatim exemplifying quotes from the interview transcripts. A total of 25 themes and 25 subthemes were identified. Themes were grouped into two categories and their sub-categories. The first category focuses on the introduction of complementary foods, with two sub-categories: timing of the introduction of complementary foods, and principal sources of advice. The second category is transgenerational and contemporary CF practices, with three subcategories: first complementary foods, cooking and feeding roles in CF, and CF feeding behaviors and environment.

Timing of the Introduction of Complementary Foods
A small number of caregivers in the F&HM group reported introducing foods to their infant's diet before 4 months of age, compared to none of the caregivers in the HME group ( Table 2). The majority of both groups reported commencing CF between 4-6 months of age. A higher proportion of caregivers in the F&HM group withheld CF until 6 months. One caregiver in the HME group reported late commencement of CF (>6 months).

Principal Sources of Advice
A variation was observed in the advice, which caregivers received from healthcare professionals (HCPs; e.g., pediatricians and WIC professionals). Approximately half of such caregivers reported being advised to introduce complementary foods between 4-6 months (or earlier), while others reported being advised to withhold introduction until 6 months ( Table 3). The majority of caregivers from both groups reported following the advice given by healthcare professionals (HCPs), irrespective of the timeframe advised. Conversely, fewer caregivers from both groups reported following their families' advice. Some were advised by older family members to commence CF earlier than was recommended by HCPs. A minority of caregivers (mostly of the HME group) reported relying on their own knowledge (own research or experience raising other children), rather than other sources.

First Complementary Foods
Most caregivers in both groups recall traditional/cultural foods being offered to infants by previous generations. Caregivers frequently reported poi, (steamed and mashed taro root) as the first complementary food offered, as was traditional for their families (Table 4). Several caregivers that fed HME, who do not self-identify as Native Hawaiian, offered poi as a first food. Some reported being encouraged by HCPs or peers, that poi was a 'healthy' and 'hypoallergenic' first food. Other caregivers also provided foods specific to their traditions and culture (Japanese, Chinese, Filipino, Tahitian and Italian). Several caregivers recalled rice cereal or premade foods being given as the first complementary foods. A higher proportion of caregivers in the F&HM group adopted the practice of providing pre-made foods for their infants. Conversely, more of the HME group described home-preparing foods. You know I think she (her mother) just forgot-she was like "oh" like, she didn't remember you know? I would tell her "Oh you know she's going to start solids like at six months or whatever" and she's like "Really? I don't remember it being like that."  While I don't have personal recollection of this, I know she firstly told me that with me, she would make her own baby foods, and puree everything, because I was the first child, so everything had to be so perfect.

Roles in Cooking and Feeding Complementary Foods
Most participants recalled females (mothers and grandmothers) in their families previously undertaking the majority of such activities (Table 5). Only a small number of caregivers expressed that females (mothers) were responsible for the majority of these roles in their families today. A large number recalled males (fathers and grandfathers) having some level of responsibility in cooking and feeding in the past. Furthermore, a third of caregivers reported that, when compared to previous generations, males have a more significant role in cooking and feeding, particularly among the HME group. Grandparents in previous generations cooked and fed in recollections of nearly all of the caregivers in the F&HM group, compared with only two caregivers in the HME group. Only a small number from both groups reported that grandparents' involvement has transcended to this generation. The same trend was seen in relation to the role of extended (non-parent) family members.

CF Behaviors and Environment
Caregivers from both groups remembered solid food (e.g., rice cereal, poi) being added to infants' bottles (Table 6). One caregiver explained that the belief was it helped infants to sleep for longer periods. Only one caregiver (of the F&HM group) continued this tradition. Others chose not to adopt this practice as they were advised by HCPs that it was unsafe. Pre-mastication of foods for infants was a recollection shared by a third of caregivers (equally among groups). Over half of these caregivers (mostly of the HME group) reported continuing this tradition. The main reason cited for discontinuing this practice was concerns about infant safety with several caregivers (mostly of the HME group) expressing that they were being more cautious in their decision-making relating to infant feeding, compared to previous generations. Regarding the CF environment, a third of caregivers reported feeling that poi, in particular, is increasingly more available, facilitating its inclusion in their infant's diet. A small number of caregivers felt that the price of certain products (formula and traditional foods-including poi) has increased. Others described how poi, for them, has become more accessible thanks to programs such as WIC alleviating the burden of cost. My eight year-old, he kind of does a lot too for us (when it comes to feeding the baby). Both of our parents said they put rice cereal in our bottles when we were young. 3 Yes. So, they'd put a little bit of poi in the bottle and make it thicker because the thought was that feeding the baby poi will help the baby to sleep longer. That is the 'old wives' tale' behind the poi.

Discussion
This study is the first to investigate primary caregivers' self-reported CF experiences among NHPI and Filipino populations. The findings on the timing of CF commencement agree with existing research which suggests that the practice of EIOCF is more prevalent among formula-fed infants than among infants fed human milk [18]. Although evidence regarding the relationship between EIOCF and later obesity is conflicting [10,11,14], research suggests that adverse effects of EIOCF most strongly impact formula-fed infants, likely due to lack of exposure to the anti-obesogenic properties of human milk [10,21]. Current research suggests that caregivers' understanding of CF guidelines is associated with more appropriate timing of the introduction of complementary foods [21]. A concentrated effort is needed to ensure caregivers' understanding of the potential health implications and safety concerns regarding EIOCF, particularly among those who feed infants formula.
The CF guidelines of the WHO and AAP recommend that CF should commence from 6 months of age [7,9]. However, half of the caregivers reported receiving advice from HCPs that CF can commence earlier than this. Other qualitative studies have reported similar findings regarding the inconsistency of CF advice among HCPs [41,42]. Mothers in one study had less trust in HCPs after advice given regarding the timing of CF commencement varied from the WHO guidelines [41]. Despite the conflicting advice, the majority of caregivers in both groups reported following the advice of HCPs regarding the timing of CF commencement. This finding was unexpected based on research among other populations [19,21]. However, in order to safeguard caregivers' trust in the advice of HCPs, consistency and clarity of CF advice is necessary [41].
Among other traditional and cultural first complementary foods, caregivers frequently introduced poi. Poi was the most frequently offered first complementary food among infants in the larger study, of which these participants were a subsample [36]. The introduction of poi has a traditional basis. Poi is a traditional first food offered by the Native Hawaiian population [32]. Similarly, taro is a staple in the diet of many Other Pacific Islanders [33]. The WHO also encourages the inclusion of culturally appropriate foods during the CF period [7]. However, this study conveys that poi has grown in popularity across different cultural groups. This is likely because, as caregivers described, poi is readily available, and is promoted widely in the community and by HCPs. Poi is hypoallergenic due to its low protein content, easily digestible, texture-appropriate, and a source of several micronutrients (e.g., several B-vitamins, calcium, magnesium, potassium) of specific importance during the CF period [43,44]. The widespread use of poi as a first complementary food among the NHPI and Filipino populations can be considered an appropriate, if not a beneficial infant feeding practice that has transcended from previous generations. Some caregivers expressed that poi has become more accessible to them as a result of governmentfunded programs such as WIC. Recent findings by Campbell et al. similarly suggested that WIC/SNAP participation among NHPI and Filipino infants facilitates the inclusion of foods from 'healthy' food groups [45].
As was anticipated based on current research, males (fathers) were reported to have a more prominent role in cooking and feeding activities than in previous generations [46]. However, grandparents' role in cooking and feeding was much less prevalent than is traditional for families of the NHPI population [35]. Although, the small sample size in this study is acknowledged. This observed change in feeding roles could be related to caregivers' frequent expressions of disagreement with CF advice offered by family members, which often contradicted that of HCPs. This is in line with current research, which suggests that grandparent involvement in feeding is associated with earlier CF [20,21].
Another theme that emerged was a belief among some caregivers to be more cautious when making decisions regarding CF, compared to previous generations. This belief was also identified among caregivers in another study [6]. Likewise, some caregivers expressed distrust in pre-made infant foods and reported exclusively feeding "organic" varieties of foods, although, such beliefs are not supported by current evidence [47][48][49][50]. Similarly, most caregivers in both groups were receptive to the advice of HCPs relating to unadvised feeding practices, including pre-mastication. Pre-mastication is a transgenerational feeding practice, which holds cultural significance in the Native Hawaiian population [35]. This practice nourishes the infant's soul [51]. However, due to concerns regarding links with communicable diseases [52], some caregivers chose not to adopt this CF practice. The discontinuation of this traditional practice based on the advice of HCPs suggest that participants' CF choices are becoming more cautious.
The limitations of the study may be that convenience sampling may have contributed to selection bias. However, the inclusion criterion minimized recall bias, as infants included were engaged in CF at the time of the study. Demographic characteristics of participants (race/ethnicity, sex, education level) were not equally represented so there is the potential for confounding results, as such demographic factors have previously been associated with the timing of solid food introduction [10,18,20,21]. Similarly, going forward, quantitative analysis of outcomes including diet-related disease and anthropometric measures (e.g., weight, length, and other body composition indicators) would allow for more definite conclusions to be drawn from research in this area. Conversely, a strength of this study is that this is the first study to investigate caregivers' CF experience in NHPI and Filipino populations. This investigation of under-researched racial/ethnic groups will add to the current literature surrounding NHPI and Filipino populations. Investigation of other confounding factors is warranted in future larger studies among these populations. Similarly, our findings will add to existing research regarding infant feeding methods (breastfeeding and the inclusion of formula) and CF practices. The presence of infants who were exclusively formula-fed in this sample would have been beneficial, as the practices of this group could be compared to the HME and F&HM groups. Finally, the participants may have been influenced by the interviewer, by feeling compelled to give certain responses. Open-ended and exploratory questions minimized bias by generally prohibiting simple agreement or disagreement responses. A notable strength of the study was that the interviews were conducted by researchers who self-identified as the same race/ethnicity as the interviewees. This may have put participants at ease and facilitated reassurance and comfort in sharing personal experiences.

Conclusions
Similar to other populations, EIOCF among NHPI and Filipino infants was more prevalent among those who were fed infant formula. Despite this, NHPI and Filipino caregivers were more receptive overall to the advice of HCPs relating to the timing of CF commencement than was expected based on existing research [19,21], with the advice of HCPs often overruling the advice from family. Caregivers also expressed the health and safety of their infant to be their priority during the CF period and took caution in making CF decisions. Importantly, the desire to uphold transgenerational feeding practices, such as the inclusion of poi, was evident among NHPI and Filipino caregivers. Ultimately, these findings provide a richer understanding of NHPI and Filipino caregivers' chosen CF practices, and insight into the nutritional intake during these infants' first 1000 days of life, which is a critical nutritional window [28]. However, in order to promote appropriate CF practices, HCPs must provide clear, consistent CF advice, that is in compliance with current CF guidelines. Further research into NHPI and Filipino caregivers' experiences in relation to wider early feeding practices is necessary to fully assess the role of caregivers' chosen feeding practices in the development of childhood obesity among these populations.