Understanding feeding practices of Latinx mothers of infants and toddlers at risk for childhood obesity: A qualitative study

Abstract Infancy and toddlerhood are critical stages for the development of habits that can lead to future obesity, and caregivers have an important influence on these habits. We conducted this qualitative semistructured interview study to explore the feeding practices of Latinx mothers of young children who are at risk for childhood obesity in order to identify targets for obesity prevention. We interviewed Latinx mothers (N = 14) of a child ages 6–18 months with a weight‐for‐length ratio > 85th percentile at the time of recruitment. Two researchers independently read through the interviews, identified sections of the interviews pertaining to feeding, and used constant comparative methods to identify the following common themes: mothers overwhelmingly reported permissive feeding styles, driving overfeeding and frequent night‐time feeding. Mothers expressed some difficulty with transitioning to solid foods and reported desiring to feed their child healthy foods by minimizing juice and giving vegetables. Paediatricians and WIC staff were viewed by mothers as trustworthy sources of nutrition information. Most identified a connection between their child's weight and diet, but many lacked the insight or capacity to change their current practices. The mothers in our study provide insight into factors that may predispose young children to obesity and thus potential avenues to support these families. Healthcare providers can better serve them by giving clear, actionable advice on healthy feeding practices for their child, while understanding constraints that may make healthy habits difficult to implement. Paediatricians should be honest about their child's weight status early on to allow time for intervention.

Racial and ethnic minorities experience large disparities in childhood obesity rates. Latinx children are two times more likely than black children and three times more likely than white or Asian children to be obese (Ogden et al., 2016). Additionally, Hispanic children are now more than one fourth of the population under 1 year of age in the United States (Gaffney et al., 2018). Prior obesity research has primarily focused on understanding, preventing, and treating childhood obesity among older children and youth, and obesity studies regarding infants have focused mostly on breastfeeding versus formula feeding.
However, a growing body of evidence shows that risk factors for the development of childhood obesity begin at the prenatal, infant, and toddler stages (Weng, Redsell, Swift, Yang, & Glazebrook, 2012;Geserick et al., 2018). In a systematic review of studies assessing rapid weight gain in the first 2 years of life, 45 out of 46 studies found an association between rapid weight gain in infancy and subsequent overweight status in childhood (Woo Baidal et al., 2016).
Additionally, evidence suggests that eating habits are established as young as 2 years of age, and, like weight, have been shown to track into adulthood (Harrison, Brodribb, & Hepworth, 2017). Meal choices, timing and quantity of feeding, feeding styles, and perception of healthy child weight have been identified as key factors for the development of childhood obesity (Gross, Mendelsohn, Fierman, Racine, & Messito, 2012;Harrison et al., 2017;Li, Scanlon, May, Rose, & Birch, 2014;Azad et al., 2018). Focusing on early childhood may help to get ahead of childhood obesity by setting a firm foundation for the child and the parent to learn and establish healthy habits (Helle, Hillesund, Wills, & Øverby, 2019).
Previous investigators have suggested that parental feeding styles, or the structure and responsiveness of the parent to the child's hunger and satiety cues , shape the early feeding environment (Horodynski et al., 2018). Among the four parenting styles described in Baumrind's theory of feeding styles, an authoritative feeding style is associated with a healthier dietary pattern (more vegetables and dairy), compared to indulgent, uninvolved, and authoritarian styles . However, Hispanic parents are more likely to practice indulgent feeding styles , which are correlated with higher body mass index in longitudinal studies (Shloim, Edelson, Martin, & Hetherington, 2015).
These feeding styles are also associated with feeding to soothe, and it has been shown that children who do not learn to self soothe may also continue to cry throughout the night (McGeorge, Milne, Cotton, & Whelan, 2015), establishing a cycle of night-time feeding. Some studies also suggest that Latinx families have different standards and ideals for what constitutes a healthy child (Isasi, Rastogi, & Molina, 2016) and may consider higher infant weight a "safety net" and a sign of good mothering (Harrison et al., 2017).
To our knowledge, there are no studies to date focused on feeding practices among Latinx mothers of children under 2 years of age who are at risk for childhood obesity in the southern United States region. The purpose of this research was to explore parenting practices and experiences generally. However, this particular study focuses on feeding practices and explores how practices affect and are affected by soothing practices, the child's sleep patterns, routines, and social stressors among Latinx mothers of children at risk for childhood obesity. In doing so, we hope to identify potential targets for successful, culturally sensitive obesity prevention counselling and programming.

| Recruitment and eligibility
Parents were eligible to participate in interviews regarding parenting practices and experiences if they had a child aged 6 to 18 months with a weight-for-length ratio greater than or equal to the 85th percentile within the past 3 months at time of recruitment, spoke English or Spanish, and if the parent/carer age was over 18 years. A nonbiological parent was included as part of the sample. See Table 1 below for a description of the participants' demographic data. Participants were recruited at a Federally Qualified Health Center in Central Texas. This health centre primarily serves Latinx patients on Medicaid, Children's Health Insurance Program, and local insurance plans. A list of potential participants was created ahead of time based on these criteria by searching the health record for upcoming check-up appointments, at which the child's growth is measured.
The researcher went to the clinic on the days of potential participant appointments and informed the paediatricians of which patients were eligible. Caregivers were recruited in person with the permission of the paediatrician after the paediatrician had already discussed the child's weight status with the caregiver. Informed consent was obtained, and the interview was scheduled at a time and

Key messages
• There are several potential obesity prevention intervention targets regarding early childhood feeding including timing and quantity of meals. Healthcare delivery methods and models should be designed to address these target for intervention to help mothers establish healthier feeding practices.
• In this study, mothers tended to practice permissive feeding styles, driving overfeeding. Many mothers did not perceive their child's weight as an issue of concern for a number of reasons.
• Further research is needed to investigate the relationship between feeding styles and childhood obesity as well as the impact of nutrition interventions on feeding style.
location that was convenient for the caregiver. A $10 grocery store gift card was given both at the time of recruitment and after completion of the interview.

| Study procedures
Semistructured interviews were conducted in the caregiver's language of preference. Most interviews took place at the participant's home, whereas others took place in a clinic conference room or in a neutral location such as a coffee shop. We used a semistructured interview protocol that was developed with input from a paediatrician and from qualitative researchers. The protocol included 22 questions that were intended to elicit the overall experience that parents have with raising a small child, with a focus on key health indicators such as family support and neighbourhood, feeding, play, sleep, and experience with the clinic (see Appendix Table A1). After three initial interviews, there were emerging themes, such as feeding to soothe and justification of weight. To further explore these themes, we added nine questions to the interview protocol for a total of 31 questions. We conducted interviews until thematic saturation was reached and chose to focus on feeding themes for this manuscript. This study was reviewed by an ethics committee and approved by the University of Texas Institutional Review Board; participants signed consent forms before participating and were given assurance of confidentiality.
The research team, comprised of physicians, graduate students, and an undergraduate student, practiced reflexivity by writing memos and engaging in debriefing in order to set aside, as much as possible, any preconceived notions so that the analytic, substantive theory could emerge (Creswell, 2007). Due to all of the participants being patients at the same clinic, the research team reassured the participants that their participation in this qualitative study was voluntary and did not affect any other services or programmes they were receiving at the clinic. The research team also had to manage the possibility of having insider knowledge from previous interactions with the participants; this was done by being intentional about how questions were asked and probed so as to not assume to know anything that the participant meant, especially when using slang or saying "you know."

| Data analysis
The interviews were audio recorded, transcribed verbatim in the language in which they were conducted, and translated with verification from two bilingual researchers. The transcripts were uploaded to NVivo 12 for analysis. Two researchers read through the transcripts independently, following Creswell's analysis process, consisting of a first read-through of transcripts to identify initial codes, then comparing initial codes between transcripts, expanding these codes, and condensing to final themes (Creswell, 2007). The researchers met periodically to discuss emerging themes, illustrative quotes, and areas of particular interest that appeared repeatedly.
The final identified themes were confirmed by all four members of the research team.

| Ethical Considerations
This study was approved by the University of Texas Institutional Review Board.

| RESULTS
Interviews took place from December 2018 to February 2019 and lasted 10 to 76 min with a median time of 25 min. A total of 15 interviews were conducted. Eleven of the interviews were in Spanish. The interview of a mother who identified as non-Hispanic was not included in the analysis, and theme saturation was confirmed without her interview. Demographic data are available in Table 1 themes emerged regarding feeding practices, which we explore here.
There are three major themes: feeding, food choices, and food and weight, which are divided further into subthemes, which can be found in Feeding in response to crying or in anticipation of crying generally led to feeding the child large quantities of milk or formula above the American Academy of Pediatrics' recommendations for each developmental stage and for the weight of the child (Hagan, Shaw, & Duncan, 2017). Some mothers such as Mary reflected on their own emotions having a role in overfeeding. "My husband tells me not to give them more. That I'm going to get him into the habit of eating too much.
Then he starts to cry and cry. As a mom it hurts me to see them cry.
Sometimes my heart wins." Although some mothers recognized the need to change their habit of feeding in response to crying and opt for scheduled feedings, they feared that the child would cry if they did not receive the milk that they were expecting, and they would have no other way to successfully soothe the child. For example, Alicia recognized that "he drinks too much milk throughout the day, that's not normal. But then again if he didn't get the bottle he would cry". Interestingly, the foster mother in our study, Nicole, reflected that her child's birth parents "were feeding him whenever he cried. 3.1.2 | Loose meal structure, feeding when the child requests food, especially at night Most mothers did not have a set schedule for feeding their child; permissive feeding, or feeding in response to the child's cries, was the norm. Although many discussed that they fed the child a set number of meals in a day, the times varied from day to day, and snacks or milk were given at random at the request or cries of the child. A few mothers did have a feeding schedule, but it consisted of a set number of meals that were too frequent for the child's age (Altmann, Hill, Shelov, & Hannemann, 2019 So it's also helped me to know and like learn." Speaking with physicians also helped some of the mothers better understand their child's weight status, as Andrea continued, "Because the doctor had talked to me that it's not so much about her physical state, but it's for her health." Some also mentioned specific nutrition questions that they would ask at future visits. For example, Sofia stated, "I'm going to speak with the doctor to see what they recommend for me to feed him and all of that." The clinic was seen as a reputable and respectable source of information. One mother noted that living in the US and having access to WIC made it much easier to raise a child and not worry about access to formula or nutritious foods for their child.

| Use of juice
Mothers mostly stated they did not give their children large amounts of juice; many mentioned that they were worried about the high sugar content of juice. For example, Ana noted, "It's mostly water because I don't want him getting with a lot of sugar. I don't like sugar." However, juice was still a part of many of the children's diet. Mothers who did say they gave juice mostly gave the child only small amounts of juice, and some mixed juice with water to reduce the sugar content.
This was clear from Andrea's comments: "Yes, but I give her the juices for babies. I add a little bit of water because I think that they're a little sweet… If it was up to her all she would drink is water". Others that gave juice noted that natural juice or juice that is marketed as "baby" juice is preferable.

| Justification and perceptions of weight
Many mentioned the child's weight when discussing the content of meals or the number of bottles given in a day, making the connection between feeding and weight. Mothers either came to this realization on their own, or were told by their partner or their physician, as evidenced by Mary's reflection: "Well since she was small I had some problems with her weight because since she was about six months she was already eating. She ate a lot. I've also had issues with her because even now she's short. They always tell me at the doctor's that she weighs too much." Among those mothers who mentioned their child's weight and diet, only a few spoke of practical steps to take to adjust their feeding routine, whether decreasing the amount of milk being given, or introducing a feeding schedule. For example, Elena shared that "the doctor called me and said that she was a little heavy. But it's because she loves to eat … So, because the doctor told me that she's a little heavy, today I've been organizing… I was thinking that it's better to give her one daily." Even among these mothers, they noted that the child's behaviour was a barrier to making these changes, as they often cried due to a low tolerance for change in the amount of food or formula given.
Mothers were mixed in their responses to their child's weight status. Most understood that it was an issue of concern, and some minimized or justified their child's weight. Several were not worried about their child's weight, noting that their child was born large and consequently have always wanted to eat more. Among these mothers was Rachel who reflected, "So it's not that I am over-feeding her, I think that she is just a large child. I give her the recommended portions. It's just that she has that body-type, so tall, so large." Some felt that it was better for the child to be overweight rather than underweight, including Ana: "Honestly, I feel good about it because I know he's eating good. But then I don't want him gaining weight like… because my nephew like I said, he was thinner and because of the same thing he wouldn't really eat good. And I'm glad my baby is not thin." Some mothers reported their children were underweight or not gaining sufficient weight early in life and compensated for this appropriately by increasing the amount of milk and formula given per physician instructions. However, they had difficulty adjusting back to a normal quantity of milk and formula for the child's age and current weight. A few did not mention weight as an issue.

| DISCUSSION
The Latinx mothers in our study shared important information about how they approach feeding. Overall, their experience is shaped by a strong desire to care and provide for the child. Mothers desire for their child to grow well, and they are generally open to advice and insight from healthcare professionals on healthy feeding practices. At the same time, mothers' desire for the child to be happy and cared for drives permissive feeding styles in which parents respond to or prevent a child's cries or requests by providing food or milk (Horodynski et al., 2018). This phenomenon is well documented and linked to unhealthy weight. Feeding was found to be very emotionally tied; it made mothers sad and frustrated to allow their child to cry instead of feeding them more. Feeding seems largely driven by the child's requests among our mothers interviewed, which is consistent with previous research in Latinx populations Davis, Cole, Blake, McKenney-Shubert, & Peterson, 2016).
Overfeeding at night was especially prominent in our discussions with mothers. This pattern is concerning because mothers consistently feeding a child at night past 6 months of age can lead to overfeeding and also can pose a problem for the child learning to self soothe without food. Feeding with the intention to calm crying predisposes infants to develop a learned response, thus "training" the child to expect a feeding when they cry or are upset (Altmann et al., 2019). Mothers, however, often felt that they had no other option but to continue this style of feeding, continuing a cycle of night-time feeding.
Although mothers found it difficult to accept feeding advice from healthcare providers in its entirety, some recommendations from healthcare staff did resonate with them. For instance, mothers reported being advised that a diet higher in vegetables and lower in juice and sugar was preferable; however, problems with implementation arose due to child refusal of food or the mother's uncertainty of how to make and provide healthier foods. One can speculate that this is an issue of low food literacy. In addition, the benefits of organic foods may be overemphasized via marketing or messaging to carers, creating a perceived cost barrier to eating healthy foods that does not necessarily need to exist.
As previously observed, other messages from healthcare professionals such as adhering to a feeding schedule and not overfeeding the child milk or formula or introducing solids early were the most difficult for carers to implement (Harrison et al., 2017). It is possible that other life stressors such as living in a new country, financial stressors, lack of social support, as well as culture and family norms may prevent mothers from following through on advice given by healthcare providers (Harrison et al., 2017). A few recent immigrant mothers expressed stress or confusion due to adjusting to feeding recommendations in the US versus their home country, particularly when initiating solid foods, as they had never been given advice on feeding from a physician prior to immigrating to the US. It is important for healthcare providers to ask about norms of feeding in the carer's home country to better understand and address hesitations and questions about a solid food initiation and feeding schedule that promotes healthy weight. (Chaidez, Townsend, & Kaiser, 2011;Harrison et al., 2017) Overall, most mothers did not appear to change their feeding practices in light of their child's weight status. Although some spoke of plans to change their practices, many had vague plans or no plans to reduce the child's weight; most were happy that their child at least was not underweight. This is consistent with previous studies that have shown that Latinx mothers view excess weight as protecting the child from being underweight (Isasi et al., 2016;Lindsay, Sussner, Greaney, & Peterson, 2011). One can speculate that cultural norms concerning weight and historical fear of not having sufficient access to food contribute to mother's perception of food and weight, although not explicitly expressed.
One possible solution for helping promote healthy feeding habits among this population is offering group clinic visits, during which a facilitator and physician could spend much more time with a group than they could with individual patients, relaying important feeding and nutrition guidance for each developmental stage. Most mothers did not have a robust support system for raising their child; many had family living in different countries or not available to help, and most worked or stayed at home where they were mostly isolated from social networks. Placing the parent in a cohort of other new parents with children of the same age could help to build community and a space to share ideas, frustrations, and successes, helping parents make practical steps toward behavioural change (Bloomfield & Rising, 2013).
Motivational interviewing may also help mothers to identify their own priorities and reasons to make improvements to their child's diet. This style helps to align the healthcare provider with the parent as an ally because both are invested in the child's wellbeing, opening communication and allowing the mother to set goals that are reasonable for her and her child (Borrelli, Tooley, & Scott-Sheldon, 2015).

| Strengths and limitations
Our study was small and only focused on mothers at one clinic in Central Texas; we recruited at the clinic, which may have captured a more motivated or resourced group of mothers. This means that our findings may not be generalizable to all Latinx mothers of young children.
We collected self-reported data, which can be affected by social response bias. However, our study had several strengths as well. Our sample included mothers with children at various stages of development and eating practices (e.g., some children were walking or eating solid foods vs. others who were not). Additionally, the data showed that the mothers represented at least four different Latin American countries (e.g., Mexico and El Salvador). Lastly, because these families attend the same clinic, we may be able to conduct follow-up studies with the same families to understand changes over time.

| CONCLUSIONS
Our study contributes to the existing literature by advancing the understanding of feeding practices and experiences of Latinx mothers with young children at risk for childhood obesity. Mothers' desire to provide for their child and prevent their child from crying can drive permissive feeding practices that are counter to maintaining a healthy weight. There are opportunities to intervene to potentially prevent obesogenic feeding habits by coaching mothers on culturally appropriate alternative soothing techniques, including swaddling, singing, and bathing, appropriate feeding timing and portions, and on healthy weight. Although mothers value the opinion of healthcare professionals, they have difficulty implementing their recommendations.
Alternative care models and methods of communication including group clinic visits and motivational interviewing may be better strategies for helping mothers implement recommendations. Further research is also needed so providers can better understand and address the effects of cultural norms, food insecurity, and sleep and routines on feeding practices and future childhood obesity.

ACKNOWLEDGMENTS
Keli Hawthorne, RD and Steve Abrams, MD who aided in the inception of this study and in obtaining funding for this study.
Department of Pediatrics at Dell Medical School at the University of Texas at Austin.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

CONTRIBUTIONS
HM and GDT recruited participants. HM and MG designed the study.
EAJ provided input on the design and execution of the study as an expert in qualitative research. HM, CV, and GDT conducted interviews. HM transcribed English interviews. GDT translated Spanish interviews with input from MG and CV. HM and CV analysed the data. HM, CV, and GDT wrote the paper. EAJ and MG provided input on the paper. What other support services would you like to have?

ORCID
Topic: Other Is there anything else I missed that you think is important? ‡ These questions were added after analysis of the first three interviews.