What children bring to the table: The association of temperament and child fussy eating with maternal and paternal mealtime structure

Abstract

Fussy eating is a nuanced, mealtime-specific behaviour associated with difficult temperament but has been rarely examined within the context of mealtime structure. The aim of this study was to a) examine associations between child temperament, and mothers' and fathers' structure-related feeding practices and b) explore whether these associations were mediated by child fussy eating. Cohabiting mother-father pairs (N = 205) of children aged between 2- to 5-years residing in a socioeconomically disadvantaged Australian city completed self-reported, validated measures of child temperament, food fussiness and structure-related feeding practices (structured meal timing, structured meal setting and family meal setting). Child temperament was associated with maternal and paternal structure-related feeding practices, such that more difficult temperament was associated with less mealtime structure. Mothers' perception of child food fussiness mediated the relationship between difficult temperament and increased provision of alternative meals to the child from the rest of the family. Additionally, mothers' and fathers' perception of child food fussiness mediated the relationship between difficult child temperament and lower frequency of sitting at a table together for family meals. Therefore, perceptions of child food fussiness may explain why mothers and fathers use less structure at mealtimes with children who have more difficult temperaments. These results suggests that similar intervention approaches can be used for both mothers and fathers from socioeconomically disadvantaged families to target fussy eating and structure the mealtime environment. Promoting mealtime structure to facilitate parents’ appropriate responses to food refusal or difficult behaviour at mealtimes is indicated.

Introduction

Consuming a wide variety of healthy foods, such as vegetables, reduces the risk of chronic diseases such as some cancers (Boeing et al., 2012) and cardiovascular disease (Liu, 2013). Dietary patterns are established in early childhood (Mikkilä, Räsänen, Raitakari, Pietinen, & Viikari, 2005), yet children from socioeconomically disadvantaged backgrounds have an increased risk of poor nutrition and attendant health consequences (Hall et al., 2011; Nackers & Appelhans, 2013). Therefore, it is critical to expand our understanding of modifiable factors related to children's eating in early life, particularly in the context of socioeconomic disadvantage. A potentially modifiable factor related to children's early eating experiences is through parents' feeding practices. However, previous research has generally focused on associations between parents' ‘non-responsive’ feeding practices and child outcomes, such as weight, food preferences and eating behaviours (Hughes, Power, O’connor, Orlet Fisher, & Chen, 2016; Yuan et al., 2016). Non-responsive feeding practices, such as pressuring, restricting and using food rewards are thought to be inappropriate responses to children's hunger and satiety cues (Birch & Fisher, 1998). However, less attention has been paid to feeding practices which may support children's sensitivity to internal appetite cues and promote eating autonomy, such as providing structure to the child's mealtime environment (Black & Aboud, 2011).

Structured, regular and predictable family meals have been associated with positive child eating outcomes, such as increased vegetable intake (Neumark-Sztainer, Hannan, Story, Croll, & Perry, 2003). Jansen and colleagues have conceptualised three broad structure-related feeding practices in the Feeding Practices and Structure Questionnaire (FPSQ) (Jansen, Mallan, Nicholson, & Daniels, 2014, 2016). Firstly, structured meal setting measures whether the family sits down at a table to eat a meal together. The underlying assumption is that a child's hunger and satiety cues can be better recognised and role modelling and positive communication about food can more easily take place in the context of a shared family meal. Secondly, structured meal timing describes the extent to which children's mealtimes follow predictable schedules. If the child mostly decides meal timing, this may reflect feeding on demand; however, parents who feed at regular intervals may provide the child with concrete expectations for appropriate eating times and encourage them to enter mealtimes feeling hungry. Finally, family meal setting assesses whether the whole family eats the same meal rather than offering an alternative meal to children. Providing children with the same meal encourages exposure to variety of family foods rather than allowing the child to choose a preferred, already accepted food. Child characteristics may contribute to parents' positive feeding practices (Holley, Haycraft, & Farrow, 2020), but how these characteristics relate to mealtime structure is not well understood. A better understanding of these interrelationships could inform future ‘strengths-based’ feeding interventions and highlight which factors are effective targets for intervention. Starting with two potential child characteristics, the current research explored how children's temperamental and eating (fussy eating) characteristics may contribute to mothers' and fathers' structure-related feeding practices in the context of families living in socioeconomically disadvantaged communities.

Parent-child mealtime interactions are an exchange of complex transactional (Kuczynski & De Mol, 2015; Walton, Kuczynski, Haycraft, Breen, & Haines, 2017) and bidirectional processes (Jansen et al., 2017a, Jansen et al., 2017b; Jansen, Williams, Mallan, Nicholson, & Daniels, 2018), with both child and parent playing an active role in food socialisation. Child temperament (Sanson, Hemphill, & Smart, 2004) has been identified as a central factor in how children approach the feeding environment, interact with parents during feeding and influence parents' feeding practices (Stifter & Moding, 2019). Temperament emerges early in life, and describes children's emotionality, attentional reactions and self-regulation (Sanson et al., 2004) While there is little explicit research, temperament theory suggests a biological and neurological underpinning of individual difference. In terms of fussy eating this may manifest behaviourally as unwillingness to try new foods or biologically as higher sensitivity to taste, sensation and texture that may lead to food rejection. Although a large variety of tools have been used for measurement (Bergmeier et al., 2014a, Bergmeier et al., 2014b), ‘difficult’ temperaments are broadly characterised by high levels of negative affect, social withdraw and rigidity (Sanson et al., 2004), and have been associated with increased non-responsive feeding practices (Bergmeier et al., 2014a, Bergmeier et al., 2014b; Horn, Galloway, Webb, & Gagnon, 2011). It is possible that children with more difficult temperaments elicit non-responsive feeding practices from their parents. For example, parents of highly irritable infants have been shown to use more ‘food to soothe’, possibly owing to the effectiveness of food in calming a distressed infant (Stifter, Anzman-Frasca, Birch, & Voegtline, 2011). However, little is known about the association between child temperament and structure-related feeding practices in the family. In a cross-sectional study, Vollrath, Tonstad, Rothbart, and Hampson (2011) showed that mothers of toddlers (N = 40 266) with more difficult temperaments were more likely to offer sweet foods and drinks (family meal setting) during the day and night (structured meal timing). In this paper we extend this body of research by examining the role of temperament in parental feeding practices with preschool children aged 2- to 5-years, when increasing expression of autonomy is a developmental phenomenon. Our aim is to inform future child feeding interventions.

Within the feeding context, parents may interpret the child's expressions of a difficult temperament as fussy eating. Child fussy eating, defined as the rejection of both familiar and unfamiliar foods (Taylor, Wernimont, Northstone, & Emmett, 2015), is thought to be a developmental expression of emerging autonomy (Harris, Searle, Jansen, & Thorpe, 2018; Walton et al., 2017). From an evolutionary perspective, rejection of unfamiliar foods is thought to serve as a protective mechanism against the ingestion of potential harmful material (Birch & Marlin, 1982; Darmon & Drewnowski, 2008). Although conceptually distinct, temperament and fussy eating are associated with one another. More difficult temperaments have been associated cross-sectionally (Haycraft, Farrow, Meyer, Powell, & Blissett, 2011; Pliner & Loewen, 1997; Tate, Trofholz, Rudasill, Neumark-Sztainer, & Berge, 2016) and longitudinally with greater fussy eating (Hafstad, Abebe, Torgersen, & von Soest, 2013), although effect sizes are small when controlling for baseline fussy eating (R² = 0.06) (Kidwell, Kozikowski, Roth, Lundahl, & Nelson, 2018). Additionally, consistent rejection of food at mealtimes may elicit non-responsive feeding practices (Harris, Fildes, Mallan, & Llewellyn, 2016) and result in parent-child conflict (Harris et al., 2018). Parents of fussy eaters report high levels of food waste, additional meal preparation and guilt concerning child hunger (Pescud & Pettigrew, 2014; Trofholz, Schulte, & Berge, 2017). This is of particular concern for families from socioeconomically disadvantaged backgrounds, who have greater constraints relating to food wastage (Daniel, 2016) and are more likely to report their child as fussy (Cardona Cano et al., 2015; Dubois, Farmer, Girard, Peterson, & Tatone-Tokuda, 2007; Flight, Leppard, & Cox, 2003). While more research is needed to fully understand the relationship between socioeconomic status and fussy eating, higher reported rates of fussy eating in socioeconomically disadvantaged families may be explained by factors such as limited access to and therefore lower exposure to fruits and vegetables (Darmon 2008).

Most child feeding research has focused on mothers, with very little exploration of the role of fathers, despite emerging evidence that fathers have moderate-high involvement in child feeding (Mallan et al., 2014). However, there is some evidence that child temperament characteristics are associated with fathers' and mothers' non-responsive feeding practices. One cross-sectional study found that mothers' feeding practices are more likely to be explained by child temperament and eating behaviours, while fathers' feeding practices appear to be consistently explained by child eating behaviours (Haycraft & Blissett, 2012). For example, lower child sociability was associated with greater maternal monitoring, whereas children's slowness in eating and emotional undereating was associated with fathers' restriction and monitoring (Haycraft & Blissett, 2012). A holistic exploration of the family, including mother, father and child, would better inform the development of inclusive family feeding interventions.

To date, evidence remains unclear if parents are responding in the feeding context directly to general child temperament, or a nuanced, mealtime-specific component of temperament, fussy eating, or both. A deeper understanding of how child temperament influences parents' mealtime structure, and whether this is filtered through parents' perceptions of child fussy eating, can be useful for framing positive feeding interventions. Therefore, this research aimed to examine associations between child temperament and mothers and fathers’ structure-related feeding practices in a sample of parents residing in a socioeconomically disadvantaged Australian community. The second aim was to examine if fussy eating mediated the relationship between child temperament and structure-related feeding practices. We expected that mothers and fathers who perceive their children as more difficult would be less likely to use structure-related feeding practices. However, owing to the lack of research related to mechanisms underlying this relationship, the second aim was exploratory and therefore we did not have hypotheses.