Parental Feeding Practices in Families Experiencing Food Insecurity: A Scoping Review

Parental feeding practices and styles influence child diet quality and growth. The extent to which these factors have been assessed in the context of disadvantage, particularly household food insecurity (HFI), is unknown. This is important, as interventions designed to increase responsive practices and styles may not consider the unique needs of families with HFI. To address this gap, a scoping review of studies published from 1990 to July 2021 in three electronic databases was conducted. A priori inclusion criteria were, population: families with children aged 0–5 years experiencing food insecurity and/or disadvantage; concept: parental feeding practices/behaviours/style; and context: high income countries. The search identified 12,950 unique papers, 504 full-text articles were screened and 131 met the inclusion criteria. Almost all the studies (91%) were conducted in the United States with recruitment via existing programs for families on low incomes. Only 27 papers assessed feeding practices or styles in the context of HFI. Of the eleven interventions identified, two assessed the proportion of participants who were food insecure. More research is required in families outside of the United States, with an emphasis on comprehensive and valid measures of HFI and feeding practices. Intervention design should be sensitive to factors associated with poverty, including food insecurity.


Introduction
Parental feeding practices and styles play an important role in the development of child diet quality, eating behaviours and healthy growth [1]. Children are born with an innate ability to self-regulate their energy intake, which allows them to follow their own hunger and satiety cues [2]. This can be easily overridden by parental practices such as pressure to eat or the use of rewards for eating. These parent behaviours, referred to as 'coercive control' or non-responsive feeding practices, "teach" children to eat for reasons other than hunger [3]. Conversely, responsive feeding refers to prompt, emotionally supportive, contingent, and developmentally appropriate reciprocity between the child and their caregiver in relation to feeding and food intake [4]. Responsive practices fall broadly under the higher-order constructs of 'structure' and 'autonomy support or promotion' [3], whereby parents provide safe, nutritious, and developmentally appropriate foods and the child decides how much is eaten [1,5]. While practices are the specific goal-oriented actions a parent takes in relation to child feeding and eating, these sit within a broader construct known as feeding styles. Feeding style refers to the general way that parents interact with a child during meal and snack times [6]. An authoritative style is considered most appropriate, characterized by high levels of warmth and responsiveness to a child's needs, along with high levels of age-appropriate reasoning and structure [7].
Over the last three decades, the global rise in the prevalence of childhood overweight and obesity prompted extensive research into the associations between parental feeding practices and styles with child outcomes. Responsive feeding practices are considered a protective factor in the prevention of excess weight and obesity [8,9], via the impact on a child's ability to self-regulate their appetite and intake. Feeding practices also influence diet quality, for example, a pressure to eat has largely been associated with poorer quality diets in children, while parental modelling and encouragement are associated with improved diet quality, such as increased vegetable intake [10]. Such findings have led to the development of interventions aimed to modify feeding practices. Indeed, systematic reviews of randomized controlled trials of interventions found that promotion of responsive feeding is the most promising avenue for obesity prevention for children under two years [11,12]. However, exactly what components of interventions are most effective, and what components are most appropriate for different populations remains unclear [13]. This is particularly true for families experiencing socioeconomic disadvantage, who are disproportionally impacted by poor diet, suboptimal nutrition, and poor growth, including obesity [14].
Disadvantage, which includes financial and material hardship (low income, poor living conditions) and/or social isolation [15] has been strongly linked to poorer physical, cognitive, and social development in children [16]. The environmental conditions and adversity children experience during critical periods is known to impact on both immediate and long-term health. This has led to the nurturing framework linked to the sustainable development goals that posits that early child development is supported by seven key dimensions: good health, adequate nutrition, safety and security, responsive caregiving and learning and stimulation [17]. Within the context of responsive feeding, the nurturing framework is relevant; however, two circumstances may have particular significance for families living with disadvantage, that is, food insecurity and household chaos. Food insecurity is defined as the limited financial, physical, and social access to food of sufficient quality and quantity for a healthy and active life [18] and has been linked to poor child outcomes [19]. Food insecurity has a prevalence of around 12% at a population level in high income countries [20], with much higher rates in more disadvantaged communities. For example, in the USA 35.3% of households with incomes below the Federal poverty level were food insecure in 2020 [21], and in Australia up to 25% of households in low-income areas are affected [22,23]. A recent review of the literature by Gallegos et al. (2021) found that both persistent and transient household food insecurity were associated with suboptimal child development outcomes [24]. Chaotic households that are prone to high noise and crowding, with low levels of routine, organisation and overall stability have been linked to poorer child development, overweight and obesity and food insecurity [25]. Household chaos and a lack of meal planning are potential mediating factors in food insecurity [26]. In contrast, responsive feeding is contingent on environments being pleasant, structured and without distractions, such that parents can recognize and respond to child cues in a prompt, developmentally appropriate way [4].
A narrative review by Arlinghaus and Laksa (2021) [27] argued that there are considerable structural constraints, such as the ability to access food and the cost of food, which influence how parents experiencing food insecurity feed their children. Those experiencing food insecurity have significantly more time constraints, particularly if they are single parents [27]. One of the benefits of responsive feeding, is that it promotes the development of healthy food preferences. Often, repeated exposure to novel foods is required before the child gains acceptance of a new food, but parents who are food insecure, may not offer foods that are not accepted immediately, particularly if they are expensive. The authors noted that low fruit and vegetable consumption may be the result of trying to prevent food wastage and the higher cost of such foods.
Food insecurity can also be experienced intergenerationally, where chronic food insecurity shapes the way in which children learn about, acquire, and prepare food. There may be an emphasis on consuming foods with a high satiety value (that is, energy dense) over foods that are of higher quality (nutrient dense). Thus, interventions designed to support responsive feeding in households experiencing food insecurity, who may also have high levels of chaos, may require a different approach to commonly promoted strategies, such as repeated exposure to foods [28].
Therefore, the aim was to undertake a scoping review of the evidence related to parental feeding practices in families experiencing socioeconomic disadvantage-and food insecurity-in high income countries. The scoping review methodology was deemed appropriate to map the evidence and synthesise the key concepts given this diverse topic [29]. The objectives were to describe what and how parental feeding practices and styles have been assessed amongst families experiencing disadvantage, understand the characteristics of studies examining parent feeding practices in families with household food insecurity (HFI); and to identify and describe the key components of interventions that aim to modify feeding practices in families living with disadvantage and/or HFI.

Materials and Methods
This review was compliant with the PRISMA checklist for scoping reviews [30] and the Joanna Briggs Institute (JBI) approach to scoping reviews [31]. The protocol was registered with the Open Science Framework (OSF) (doi:10.17605/OSF.IO/Q47VP) (created on 9 June 2021).

Inclusion and Exclusion Criteria
A priori eligibility inclusion and exclusion criteria were developed as follows: • Population: families with children aged 0-5 years experiencing HFI or disadvantage. Disadvantage could include a measure of HFI, poverty, low income, low education attainment, receiving welfare/food assistance or other indicators of socioeconomic disadvantage. • Concept: Parental feeding practices or styles. Papers were included if a measure of parental feeding practices and/or styles was used or identified as a theme in qualitative research. • Context: high income countries according to the World Bank definition [32].
Full-text, peer-reviewed articles that were published in English were included in this scoping review according to the above criteria between the years 1990 and 2021 (database searches conducted on 2 September 2020 and updated 12 July 2021). Articles were excluded if the population group had a diagnosed illness/disorder that would impact feeding (e.g., cystic fibrosis, premature birth), or the focus was on infant feeding practices exclusively (i.e., breastfeeding, use of formula, age of introduction of solid foods). Opinion pieces, editorials, reviews, conference abstracts or protocol papers were also excluded.

Search Strategy
A search strategy was developed by KB and SNM in consultation with an experienced academic librarian. The search was run in three electronic bibliographic databases by KB (CINAHL, Medline and PsycInfo). Key words for the search strategy used in each database are shown in Appendix A. Citations were exported into EndNote and then imported into Covidence; a web based systematic review production tool [33]. The reference lists of included sources and relevant reviews were also checked.

Selection of Included Articles
The title and abstract of each article were screened in Covidence using a priori eligibility criteria. All authors were involved in the screening process. Two authors screened citations for inclusion independently, with inter-rater conflicts resolved by another reviewer, and this task was shared across authors (KB, SNM, RB, DG, JS). This process was repeated to screen full-text articles. The final list of included articles can be found in Appendix B.

Data Extraction
Data extraction was completed in Covidence using a modified version of their data extraction form. Extraction was done by one author and checked by a second author for completeness.

Data Synthesis and Analysis
Descriptive statistics were used to describe the characteristics of included papers, namely, those that directly measured and reported household food insecurity (HFI) using a specific tool and those that did not, country of origin, study design, and assessment of feeding styles or practices. The number of different feeding practices assessed across all papers were tallied, using the Vaughn content map of food parenting practices [3] as a guide and a count made of the most frequently used tools to assess styles and practices.
Data from those papers that measured HFI were described in more detail including study design, primary objective, country of origin, sample characteristics (age, gender, recruitment details), measures and tools used and key findings. Similarly, a table describing intervention studies designed to modify feeding practices amongst families experiencing food insecurity was included. Given the search identified only two intervention studies with families that reported HFI, this table was expanded beyond the original objective, to also include interventions for families experiencing disadvantage. Findings were also synthesised descriptively to map the relevant aspects of the literature as related to our research question. Results of the review are presented in narrative form. Quality appraisal was not conducted as this was not deemed necessary to meet the objectives of the review.
One hundred and six papers examined feeding practices (81%). There was considerable heterogeneity in the types of practices assessed ( Figure 1) and the tools used to assess these. Practices were categorised under the three higher-order food parenting constructs defined by Vaughn et al. (2016)-coercive control, structure, and autonomy support [3]. 'Other' practices included feeding practices that do not fall within the above known classification systems, such as laboratory eating protocols and food exposure practices.
Practices representative of coercive control such as a pressure to eat and restriction were most often assessed, in 46% and 42% of papers, respectively. Meal and snack routines were the most frequently assessed practice under the construct of 'structure' at 28% of studies, followed by the practice of modelling. Practices that aligned with 'autonomy support and promotion' were assessed least often. Another 29 studies (27%) were classified as other, representing a disparate set of practices that parents used to influence child intake or eating behaviour, but could not be easily categorised within the Vaughn framework. More than thirty different questionnaires were used to assess feeding practices within the studies included in this review, the most frequent being the Child Feeding Questionnaire (n = 26 studies) [34], followed by the Comprehensive Feeding Practice Questionnaire (n = 7) [35] and the Feeding Practices and Structure Questionnaire (n = 5) [36]. Forty papers assessed feeding styles within a population experiencing disadvantage, with the most used questionnaire being the Caregiver Feeding Style Questionnaire (CFSQ) [7] in 25 papers, while another 10 papers used the Infant Feeding Style Questionnaire (IFSQ) [37].
Validation studies identified in this review provide evidence that the psychometric properties of the Child Feeding Questionnaire (CFQ), Caregiver's Feeding Practices Questionnaire (CFPQ) and the Infant Feeding Style Questionnaire (IFSQ) have been assessed in disadvantaged populations in the United States, in particular Hispanic and African American populations; however, no specific methodological studies assessing the use of tools outside of the US were found.

Studies Examining Household Food Insecurity and Parental Feeding Practices/Styles
The 27 papers identified are described in detail in Table 2. Twenty-three were conducted in the United States while the remaining four were in Australia.

Household Food Insecurity
In those studies that reported HFI (n = 27), a variety of tools were used to define HFI in their participant cohorts. Most studies (17/27, 63%) used a variation of the USDA Household Food Security Survey Module (HFSSM), namely, either the 6-item [39][40][41][42][43][44], 10item [45][46][47], or 18-item measure [48][49][50][51][52][53][54][55]; followed by a 2-item measure by Hager et al.  [59][60][61]. The Radimer/Cornell Scale was also used in one paper [62], along with the Household Food Insecurity Access Scale (HFIAS) in another one paper [63]. Lastly, the remaining two papers used less rigorous methods with one paper using a study specific question, 'Do you ever feel that you don't have enough food for your family?' (no evidence of validity or reliability provided) [64] and one paper describing food insecurity as a theme from focus group discussions with low-income parents [65]. Validation studies identified in this review provide evidence that the psychometric properties of the Child Feeding Questionnaire (CFQ), Caregiver's Feeding Practices Questionnaire (CFPQ) and the Infant Feeding Style Questionnaire (IFSQ) have been assessed in disadvantaged populations in the United States, in particular Hispanic and African American populations; however, no specific methodological studies assessing the use of tools outside of the US were found.

Studies Examining Household Food Insecurity and Parental Feeding Practices/Styles
The 27 papers identified are described in detail in Table 2. Twenty-three were conducted in the United States while the remaining four were in Australia.
There was wide variation in the reported proportion of HFI experienced between the groups described in each of the papers, ranging between 0-80%.

Feeding Practices and/or Styles
The relationship between feeding practices and/or styles was most often examined within the context of child weight and obesity prevention [40,41,44,49,50,54,62,64]. The relationship between HFI and practices varied with HFI being associated with non-responsive practices in twelve [39,40,44,46,[49][50][51]54,56,57,62,64] and non-responsive feeding styles in three [45,48,55] studies, respectively, with null findings in two others [41,64]. Interestingly, , who found no relationship between feeding practices and styles, concluded that food insecurity may have a protective effect on dietary quality due to the adoption of coping mechanisms by mothers and grandmothers [41].

Intervention Studies to Modify Feeding Practices in Families Living with Disadvantage and/or HFI
Twelve studies described an intervention study that sought to modify early feeding practices amongst families who were categorised as low income, experiencing disadvantage and/or food insecure, these are summarised in Table 3. Only two of the interventions sought to assess and report the proportion of participants who were food insecure [50,58]. All the intervention studies identified originated from the US. Most of these research studies recruited participants via established programs for families on low incomes such as Head Start, Early Head Start or the Supplemental Nutrition Assistance Program (SNAP), with many research groups then utilising these existing programs and infrastructure to deliver the intervention.
Length of the interventions ranged from a one-off video to three years (although the paper describing the 3-year intervention reported early outcomes at 10 months [50]). Interventions were largely aimed at mothers (10/12, 83% exclusively targeted mothers). Within one paper that included both mothers and fathers as participants, 92% were mothers [66] while the other paper reported participants as 'parents' and did not report the split of mothers to fathers [67].
Mode of delivery ranged from intensive multiple face-to-face appointments to remotely provided content via mail or phone and a computer tablet-based intervention in one case. Visual media content was a commonly used mode to deliver messaging in the interventions, with video described in several studies (n = 6, 50%) [58,[68][69][70][71][72] as well as picture-based messaging [50]. In those papers using videos, these were described as short, curriculum-based videos, which included animation [72], real footage of mothers feeding their children in a home environment [68] and were tailored for the ethnicity of the target audience [68][69][70]72].
With the exception of Horodynski et al. 2005 [66], all the interventions described positive impacts on the intervention group in terms of the target feeding practices. Interventions largely targeted parental behaviours (feeding practices/styles), although Fisher et al. (2019) primary outcome was a reduction in calories from solid fat and added sugars (which was reduced by 23% at 12 weeks). Although many interventions had the underlying intention to prevent unhealthy weight gain among children, only Hughes et al. (2021) reported reduced child overweight/obesity compared to the control group [70].  showed a reduction in BMI among mothers in the intervention group compared with the control [72].
Of the two intervention papers that reported HFI,  found that HFI was significantly different at baseline between the intervention (HFI = 26%) and the control group (HFI = 60%) and, therefore, HFI was tested as a factor in their intention-to-treat analysis for health outcomes, with unchanged results.  also reported the HFI rate of the participant cohort with 30.2% in the intervention and 34.5% in the control, which was found to be not significantly different at baseline.  described tailoring content in the intervention to be sensitive to factors associated with poverty, including food insecurity [50]. TFBQ [74] Relative increases in HFI were indirectly related to increases in restrictive and decreases in responsive child feeding practices, mediated through increases in mothers' own restrained eating.

Barroso et al. 2016 [40]
To determine the association between measures of HFI, maternal feeding practices, maternal weight, and child weight-for-length in low-income Mexican Americans. CFQ [34] Parents' weight, perceptions of child's weight, adherence to the Hispanic culture, and food insecurity appear to impact parental concerns and behaviours, particularly restrictive and pressure-to-eat behaviours. IFSQ-10 items [37] A social media intervention resulted in high engagement and modestly improved feeding behaviours. Intervention reported significantly healthier feeding behaviours.

Gross et al. 2018 [45]
To determine the differential and additive impacts of HFI during the prenatal and infancy periods on obesity-promoting maternal infant feeding styles and practices at infant age 10 months. IFSQ [37] Prolonged HFI was associated with greater pressuring, indulgent and laissez-faire styles. Prenatal food insecurity was associated with less vegetable and more juice intake.

Harris et al. 2018 [59]
To examine the role of parent concern in explaining nonresponsive feeding practices in response to child fussy eating in socioeconomically disadvantaged families.

Horodynski et al. 2018 [48]
To test the interactive effects of caregiver feeding style (CFS) and familial psychosocial risk in the association BMI-score in pre-schoolers from low-income families CFSQ [7] HFI was correlated with caregiver depressive symptoms and dysfunctional parenting. Uninvolved feeding styles intensified the risk, and an authoritative feeding style muted the risk conferred by living in a poor, food insecure and depressed family. Higher food resource management skills and greater maternal presence when the child ate was significantly associated with lower child BMI z-scores Variables positively associated with child overweight were income, mother's BMI, child birth weight and juice intake. Biological and socioeconomic factors are more associated with overweight than self-reported child-feeding strategies.

Messito et al. 2020 [50]
To determine the impact of a primary care-based child obesity prevention intervention (StEP) beginning in pregnancy on maternal-infant feeding practices, knowledge, and styles at 10 months. StEP reduced obesity-promoting feeding practices and styles, and increased knowledge at 10 months. Integration into primary health care helped to reach high-risk families.

Na et al. 2021 [51]
To explore relationships between HFI, food resource management skills (FRM) and child feeding practices of low-income parents.  2-item household food security screener [76] IFSQ-15 items [37] Feeding practices differed by HFI status.
Food-insecure households had increased odds of agreeing with some obesity promoting practices such as immediately feeding a baby when they cry.

Orr et al. 2020 [57]
To examine associations between HFI status and parental feeding behaviour, weight perception, and child weight status in a diverse sample of young children CFQ 28 [34] Both measures need continued psychometric work; group comparisons using some subscales should be interpreted cautiously. Subscales such as food responsiveness and restriction may be assessing behaviours that are less applicable in the context of HFI. Mothers of children with obesity may alter their feeding behaviour differentially based on food type.

Searle et al. 2020 [61]
To examine associations between child temperament and parents' structure-related feeding practices in a socioeconomically disadvantaged community.   Mothers may perceive snacks as more important in managing children's behaviour than providing nutrition. Snacks have a powerful hedonic appeal for mother and child.

Gross et al. 2019 [46]
To learn more about the financial pressures and perceived effects on infant and toddler feeding amongst low-income Hispanic mothers with children in infancy and toddlerhood.  Mothers' aspirations in feeding were compatible with obesity prevention strategies to limit portion size and intake of fats/sugars. Mothers faced many feeding challenges.

Tartaglia et al. 2021 [65]
To explore parents' experiences of feeding 0-5-year-old children and food literacy behaviours.   women from a similar background.  "Strategies for Effective Eating Development (SEEDS)" RCT 6-and 12-month results As above 7 weeks As above As above As above As above INV had significant improvements in repeated exposure of new foods, measured portion sizes, child involvement in food prep, feeding responsiveness, knowledge of best feeding practices, and feeding efficacy, reduced feeding misconceptions and uninvolved feeding. Effects on child eating behaviour were minimal. At 12 months, children were less likely to be overweight/obese. Outcome data at 6 and 12 months showed maintained improvement in key outcomes. Facilitators promoted a learner-based approach rather than a didactic one. Group session were pilot tested. Videos showed diversity  INV showed greater breastfeeding, reduced juice and cereal in the bottle, and increased family meals than controls.
INV had higher knowledge and lower nonresponsive feeding styles. High attendance at sessions.
Utilising primary care provided access to high-risk families; built on-existing provider relationships; reduced costs; saved time Mothers reported improvements in food parenting practices following the INV. INV had a decrease in controlling practices, 'pressure to eat' and 'food as a reward' and an increase in supportive practices, 'involvement', 'environment' and 'modelling'. 93% of mothers 'strongly agreed' it was worth their effort to participate.
Most mothers found that watching themselves on video was informative and applicable to their own lives. Childcare was provided; INV conducted at times convenient to the mother

Discussion
This scoping review examined the evidence related to parental feeding practices and styles in families with a young child (aged 0-5 years) experiencing socioeconomic disadvantage (with and without food insecurity)-in high income countries. After using broad search terms of socioeconomic disadvantage, of the 131 papers identified, only 27 (21%) papers were found to address the issue of household food insecurity (HFI), and only two of these papers described an intervention to support responsive feeding in families experiencing HFI. Whilst the evidence on the direct impact of food insecurity on parental feeding practices is scant, the literature suggests that it does likely influence how and what parents feed their children. Parental feeding practices are sensitive to factors which influence the feeding environment such as food insecurity and, therefore, such factors are important to consider in parental feeding practice research and intervention design.
This review identified the most common measures used to assess feeding practices and styles, though there was little evidence that the validity and reliability of these tools have been assessed amongst families experiencing HFI. The practices most frequently assessed-pressure to eat and restriction-fall within the higher order construct known as 'coercive control', while fewer studies assessed 'structure' related feeding practices. In the future, studies could assess the aspects of structure to better elucidate the relationship between HFI, household chaos and a family's ability to implement responsive feeding practices. Very few papers examined practices related to 'autonomy support or promotion'. While the reasons for this cannot be determined from the review, it may be that practices such as educating children about the benefits of healthy eating or child involvement in meal planning and preparation may be considered less applicable in children under the age of five years.
Variation in the tools used to measure HFI makes describing and comparing HFI amongst populations challenging and there are calls for greater consistency in measuring food insecurity [24,104]. This was reflected in this review, which found significant variation in the measures used to describe HFI. Several studies used short 1-or 2-item measures (7/27, 36%). Whilst these measures provide an indication of HFI levels, they may be less reliable and may also underestimate HFI by 5-8% points when compared to more rigorous, multi-item tools [104,105]. The most used HFI measure was the 18-item United States Department of Agriculture Household Food Security Survey Module (USDA HFSSM), which was the predominant tool cited in the literature [105,106]. The 18-item USDA HFSSM includes eight child-related items and therefore may be the most relevant in the context of parental feeding practices and HFI research which focuses on child-related outcomes. In this review 8/27, 30% of the papers used the 18-item USDA HFSSM which includes the child specific items. The short form (6-item) and 10-item form USDA HFSSM were also found to be used among 9/27 (33%) of the included papers. Studies balance the burden of administering tools and surveys to their participant group and therefore may opt for shorter measures of HFI; however, choosing measures that account for HFI severity and allow for child specific measures may be advantageous in parenting feeding practice research, especially in the context of socioeconomic disadvantage where the prevalence of HFI is likely to be high. In addition, the degree of severity of HFI may influence the type and frequency of feeding practices used at any given time.
Another strong feature of the parental feeding practices and socioeconomic disadvantage/HFI literature summarised here is the heavy representation of US populations, which commonly draw on Head Start/Early Head Start and SNAP programs for recruitment. Studies conducted in the United States also tend to have a high proportion of Hispanic, Latina and/or African American participants. Perceptions of ideal body size, appropriate meal-time practices and family traditions vary across culture, and conceptualisations of "ideal" feeding practices in the scientific literature may clash with culture and community [107]. This may reduce the applicability of research findings to other countries or social and government assistance contexts outside of the US. Given that high-income countries, outside of the US, have evidence of significant HFI among their population, particularly in disadvantaged groups, this is of note and indicates the need for further research into HFI in other high-income countries. Whereas the US has readily identifiable groups among their population to recruit for research purposes (e.g., SNAP and Head Start), recruitment for such studies can be challenging in other countries due to the difficulty in identifying and successfully recruiting socioeconomically disadvantaged groups. In addition, food insecurity is monitored annually in the USA and has been identified as a significant public health issue, thus potentially highlighting it as an area of concern [108]. Further research may therefore also be warranted identifying successful avenues to recruit disadvantaged and HFI groups, which may also facilitate further research in this area.
A recent narrative review of parent feeding practices in the context of food insecurity identified no existing interventions that target parent feeding practices specifically addressing the context of food insecurity [27]. Our scoping review of the literature supports this finding and whilst two interventions were identified which reported HFI, only one of those appeared to take into account the poverty related challenges of food insecurity [50]. This review adds to the evidence by identifying some of the key features and characteristics of interventions targeting feeding practices in disadvantaged groups. The intervention studies identified in this review showed largely positive improvements in the parent and child outcomes measured subsequent to participation in the intervention.
A key feature identified in the interventions summarised was the high use of visual media content. Video and/or images are often used to convey messages to low health literacy groups. A systematic review has identified that pictorial information improves understanding and recall and is most impactful in the lowest health literacy groups [109].  developed a video which featured mothers from their target population, i.e., low-income adolescent African American mothers [68]. The video content, messaging and music was developed by an advisory panel of six African American adolescent mothers who were featured in the footage in their own homes feeding their babies. This culturally sensitive approach enhances the relatability of the messages. Other studies also adapted intervention content for their specific audience, including  who developed an intervention for Chinese immigrant mothers and included videos in Cantonese featuring Chinese mothers with their children, including images, sample menus and foods which were also tailored to the Chinese culture [72].  reporting on the intervention, 'Strategies for Effective Eating Development (SEEDS)', also utilised short videos in their face-to-face group sessions [70]. Videos can also be used in interventions to moderate the content and direct the conversation to targeted positive parent behaviours, such as in the 'Grow2Gether' intervention by -an online social media group-based intervention that encouraged participation and discussion among peer mothers [58]. Videos were posted on closed social media groups, which acted to deliver positive feeding messages as well as to be a catalyst for productive discussion among participants around the content. Short, realistic, and relatable videos and media may be a successful feature to incorporate into interventions targeting parents from low income, disadvantaged backgrounds.
The summarised interventions also demonstrated that a range of modes of delivery can be successful in this group, including traditional approaches of intensive face-toface individual or group delivery of nutrition-based information, to remote modes of intervention delivery (i.e., video, mailed content, social media, and technology-based interventions). This is important given the context of COVID-19 impacting health service delivery and the engagement with families of young children [110]. Traditional, intensive, face-to-face interventions may not be practical or feasible in a post-COVID-19 environment and it may take some time until families are willing or able to attend such intensive faceto-face interventions. It is also important to note that the one intervention that showed no positive impact on parent behavior, , was the most intensive of the interventions described with 4 group sessions and 18 individual home visits over 6 months [66]. This suggests that interventions need to move beyond intensive face-to-face sessions and instead implement multi-modal strategies to engage families. This scoping review also identified aspects from the summarised papers that reported HFI (n = 27) that may be potential areas to explore or target in interventions. Some of the studies highlighted different strengths within families that could potentially protect parental feeding practices from the negative impact of HFI. Food resource management (FRM) skills is one area that could be further explored.  showed that better FRM skills and parental presence at meals was associated with healthier weight among 2-5-year-old children in low-income families. The potential pathway between FRM skills and healthier child weight needs to be further elucidated, but the mechanisms suggested by McCurdy et al. (2014) may reduce takeaway consumption due to more home cooking, parent modelling of healthy eating, as well as an increased structure in feeding practices, e.g., more family meals and parent presence at mealtimes. The potential role of FRM skills was also described in , which reported that low FRM skills were associated with suboptimal child feeding with and without HFI. In this paper, parents in food insecure households who had high FRM skills used similarly positive feeding practices as parents from food secure households with high FRM skills [51].  also suggests that families may use coping strategies which may mitigate the negative consequences of HFI. This paper found that dietary quality improved over 18 months in HFI families which was unexpected and needs further research but may indicate the adoption of coping strategies among families [41]. These findings, although requiring further exploration and research, may suggest how interventions can be designed to incorporate the strategies and coping mechanisms families who are at high risk of HFI already use to mitigate the negative impact of HFI on their feeding practice.
It is also important to note that all the interventions identified within this review focused on individual behaviour change strategies, particularly that of mothers. This approach has been criticised for placing the responsibility for a child's health solely on the mother and failing to advocate for structural interventions (e.g., policy change) to support parent feeding practices [111]. Researchers and practitioners are encouraged to utilise a socioecological model to intervene across systems for maximum impact [24]. This review has several strengths. It followed best practice guidelines using an a priori protocol. Due to the inconsistency of terminology used in the literature to describe feeding practices and styles, a deliberate decision was made to use broad search terms to identify as many papers as possible; however, given that some included studies (e.g., qualitative studies employing interview or focus group methodologies) did not set out to assess or describe HFI and feeding practices or styles, but these issues were raised by participants and reported in the results, it is possible that similar papers were not identified and included. This should be considered as a limitation.

Conclusions
This scoping review highlights the lack of research at the crossover of parental feeding practices and food insecurity, especially in terms of interventions that target feeding practices among groups likely to have a high prevalence of food insecurity. More research is needed outside of the United States, with an emphasis on comprehensive and valid measures of HFI and feeding practices. Intervention design should be sensitive to factors associated with poverty, including food insecurity. Acknowledgments: The assistance of Peter Sondergeld, Liaison Librarian for the School of Exercise of Nutrition Sciences at the Queensland University of Technology is greatly appreciated.

Conflicts of Interest:
The WCCNR is funded by Woolworths through the Children's Hospital Foundation. Woolworths has not been involved in the design or conduct of the research or in the evaluation of the scientific quality of the research projects or in the establishment of the Centre governance.