Grandparental dietary provision, feeding practices and feeding styles when caring for preschool‐aged grandchildren: A systematic mixed methods review

Grandparents are frequently relied upon to care for their preschool‐aged grandchildren. These early years are a crucial age in the development of dietary habits and preferences. This review aims to determine grandparental dietary provision, feeding practices and feeding styles when caring for their preschool‐aged grandchildren. Medline, PsycInfo and Web of Science were searched in January 2020. A systematic mixed methods approach was used to synthesize the qualitative (n = 13) and quantitative (n = 7) articles describing grandparents' feeding styles (n = 9), feeding practices (n = 14) and dietary provision (n = 18). Grandparents are serving large portion sizes and encouraging their grandchildren to eat frequently. Results are mixed for the types of foods provided; grandparents provide discretionary foods high in sugar and fat, and some also choose more expensive core foods as treats (e.g., berries). Grandparents engage in feeding practices that promote healthy eating (e.g., creating a healthy feeding environment) and promote autonomy and independence (e.g., considering their grandchild's preferences). However, they also use some coercive feeding practices (e.g., using food as a reward) and may be using indulgent feeding styles, which can be conducive to obesogenic dietary intakes. Interventions targeting grandparents could be an effective way to improve dietary‐related health outcomes in young children.

with intake and variety at aged 22. 11 The food environment during these formative years can therefore have a significant impact on childhood and adulthood obesity.
Studies with parents have demonstrated that there may be three key mechanisms by which grandparents might influence children's food preferences and dietary behaviour: their dietary provision, the feeding practices they use and their feeding style. 12,13 Dietary provision, defined here as the types, quantity and frequency that food is provided by caregivers, can have a direct influence on children's dietary behaviours. Studies have shown that caregivers are frequently offering young children larger than recommended portion sizes for meals, drinks and snack food items, 14 which may be problematic because the more young children are offered, the more they will consume. 15,16 Nevertheless, very little is known about grandparents' dietary provision to their grandchildren.
Feeding practices, on the other hand, influence dietary behaviours in a more indirect manner and are defined as "the specific goal directed behaviours used by caregivers to directly influence their children's behaviour". 17 Although feeding practices are often used to influence children's eating behaviour positively, some practices have the opposite effect and lead to unintended negative outcomes. For example, caregiver restriction of food items, often used to limit high fat and high sugar foods, is associated with increased desire for, and consumption of, those restricted foods. 18 Pressuring children to eat, most frequently used to increase consumption of low energy dense food items such as fruit and vegetables, is a predictor of food refusal and picky eating, 19 and using food for emotion regulation can lead to children eating in the absence of hunger. 20 Nevertheless, some feeding practices such as ensuring that healthy foods are available in the home, modelling healthy food intake and encouraging balance and variety lead to positive dietary behaviours. 5 A recent study of 1,076 Australian grandparents found that grandparents of grandchildren aged 3-14 years old were more likely to use more "positive" feeding practices, with encouraging balance and variety the most frequently reported positive practice. 21 Whether this finding is true for grandparents of children younger than this and from other countries is something that requires further exploration.
Finally, caregivers' feeding style, defined as the emotional climate in which feeding interactions take place, 5 can influence young children's eating behaviours. Four feeding styles (indulgent, authoritarian, authoritative and uninvolved), across two dimensions (responsiveness and demandingness), have been identified previously in caregivers of young children. 5 These impact upon the types of food children consume, for instance in a study of 750 3-5 year olds, indulgent or uninvolved parenting styles were significantly associated with consuming higher energy dense foods and lower intakes of fruits and vegetables compared with authoritarian feeding styles. 22 From a sociological perspective, being indulgent is integral to the role of being a grandparent, 23 and it is unknown if this translates to their feeding style.
Three reviews have been published previously investigating the relationship between grandparental behaviours and the diets of their grandchildren. [24][25][26] Two of the reviews concluded that grandparents were predominantly exhibiting negative feeding behaviours towards their grandchildren such as spoiling their grandchildren with food, providing energy dense, nutrient-poor food and drinks, using food to reinforce behaviour and providing too much food. 24,26 In contrast, one review found mixed evidence for grandparental influence on children's eating behaviours. 25 For instance, whether grandparents had a positive or negative influence on mothers breastfeeding initiation and duration varied across cultures. 25 Nevertheless, several limitations of these reviews are of note. (1) The searches were conducted over 2 years ago and would benefit from being updated. (2) All of the reviews looked at a broad age range (0-18 25,26 or 2-12 24 ); therefore, it is difficult to assess the impact of grandparents on the diets of children in the early years. Considering that grandparental feeding practices may differ depending on the grandchild's age, 21 it is necessary to look at different age groups separately. (3) Two of the reviews focused on multiple lifestyle risk factors 25,26 and did not provide an in-depth review of the relationship between grandparents' feeding behaviours and their grandchildren's eating behaviours. (4) The third review, that did focus specifically on dietary influences, used a limited search strategy and may have missed key papers. 24 Consequently, no review has provided a complete synthesis of evidence exploring grandparental feeding behaviours when caring for their preschoolaged grandchildren. Additionally, no review has attempted to synthesize the evidence according to these three feeding constructs (dietary provision, feeding practices and feeding style), and doing so may highlight the specific areas of feeding that grandparents need support with.

| AIM
The aim of this mixed methods review was to explore grandparental dietary provision, feeding practices and feedings styles when caring for their preschool-aged grandchildren.

| METHODS
A systematic mixed method approach was used to synthesize articles of both qualitative and quantitative methodology. The advantage of this design is that it can incorporate a wider pool of research to develop a broader understanding of a topic. [27][28][29][30] This design is also particularly useful when there are limited number of studies conducted in the field.

| Search strategy
A search strategy was developed by identifying synonyms related to the population (grandparents and grandchildren) and outcome (diet, food and etc.) of interest. Literature and previously conducted reviews [24][25][26] in the field were also scoped to identify key terms.
Search terms within the population and within the outcome were combined using the Boolean term OR. Population and outcome terms were combined using the Boolean operator AND. See supporting information for full search terms.
Using the defined terms, Medline, PsycInfo and Web of Science were searched in January 2020, and no limit on publication date was applied. Databases were searched via title, abstract and keywords.
The reference list of included papers were hand searched for additional relevant titles. A Google Scholar search was conducted using modified search terms; the first 25 results pages were searched for further studies; after which, it was deemed that no articles were relevant for inclusion.

| Selection of studies
The first author (C.M.) screened titles, abstracts and full texts against the inclusion and exclusion criteria. Papers were included if they met all of the inclusion criteria and none of the exclusion criteria. A second independent reviewer (S.R.) screened 10% of the initial search results via title and abstract, and 10% of the full texts to ensure no relevant papers were excluded. The review was registered with Prospero, registration number: CRD42019121418.

| Inclusion and exclusion criteria
The review included papers that reported on grandparental feeding behaviours towards their grandchildren aged 2-4 years old. A full list of inclusion and exclusion criteria can be seen in Table 1.

| Quality appraisal
Study quality was assessed using methodologically appropriate tools.
Qualitative, case-control and cohort studies were assessed using relevant critical appraisal skills programme checklists, [31][32][33] and crosssectional studies were assessed using the AXIS tool. 34 These tools assess key aspects of the design including aims, methodology and results to allow appraisal of study internal and external validity. Two reviewers (C.M. and S.R.) critically appraised all included papers. The heterogeneity across appraisal tools made it difficult to compare quality, and consequently, the reviewers assigned a scoring system to categorize the papers as low (meeting below 50% of appraisal tool criteria), moderate (50-75%) or high quality (>75%) to allow for comparison. 24 Disagreements regarding study quality were resolved by consensus.

Variable Inclusion Exclusion
Population Children aged 2-4 years old.
Studies with an age range including children below aged 2 were included if they met the following criteria: 1) A mean or median age range within 2-4 years old Studies with an age range going beyond the age of 4 were included if they met both of the following criteria: 1) The maximum age was not beyond age 6 2) The mean or median age was within 2 to 4 years, 11 months quantitative data, key results relating to dietary provision, feeding style and feeding practices (see supporting information). For qualitative data, participant quotations and author discussions related to dietary provision, feeding style and feeding practices from the results section of the paper were extracted into NVivo for analysis. The data extraction process was verified in two ways; a second reviewer (S.R.) completed the data extraction for 10% of the papers, and no disagreements were noted. A third reviewer (S.C.) cross-checked all data extraction entries by the first reviewer (CM) for accuracy and completeness.

| Data synthesis
Data from both quantitative and qualitative studies were synthesized under the three feeding constructs of dietary provision, feeding style and feeding practices using a results-based convergent synthesis design. This allows the qualitative and quantitative data to be analysed separately but integrated in the results section to address the three feeding constructs. 35 The qualitative data were analysed using thematic analysis of participant's quotes and relevant text from the results section of the papers. 36 A framework method of thematic analysis was adopted using a deductive approach 37 to identify codes and themes within the three identified feeding constructs (dietary provision, feeding style and feeding practices). The six phases of thematic analysis described by Braun and Clarke were followed closely for this synthesis, 38 which allowed the reviewer to identify, analyse and report patterns within the data. All texts were coded in NVivo, and themes were identified. The second reviewer (S.R.) cross-checked 10% of the results,  42 Influences on what and how they feed their child ✓ ✓ High Jiang et al. 43 Who is generally responsible for looking after the young child in your family? Who is responsible for family meals? What is your own view on how a small child should be fed?
Rogers et al. 44 Views, perceptions and practices regarding dietary care, provision of treats, roles and responsibilities and challenges regarding feeding Toftemo et al. 45 Perceptions of their child's weight and the family's relationship within their extended family and with kindergarten ✓ ✓ ✓ High Dwyer et al. 46 What experiences and challenges have you or other parents had in supporting healthy eating for preschoolers?
Lindsay et al. 47 Perceptions of child's weight status and definition of overweight, child feeding practices related to overweight, and the role of socio-demographic, sociocultural and environmental influences on mothers' child feeding.

Moderate
Chung et al. 48 Perceptions of factors that influence child sweet drink consumption

| RESULTS
The . Of the included 20 articles, three stemmed from a single study. [39][40][41] Although these three articles used the same participants and methods, the study aims and outcomes were different and therefore are treated as separate studies. Reasons for exclusion at the full-text screening stage include the following: article is a review (n = 6), no access to full text or English version (n = 4), the age range of the children (not specified, too young or too old) (n = 22), inability to differentiate between carers influence (n = 5), no measure/discussion of grandparent direct influence on child diet (n = 15) and grandparents were primary caregivers not informal caregivers (n = 1), and weight as an outcome (n = 1).

| STUDY CHARACTERISTICS
The characteristics of the included qualitative studies are presented in Table 2 and those of the quantitative studies in Table 3.

| Location
Eight of the studies were conducted in North America: in the United States [39][40][41][42]47,49,54 and Canada. 46 Six of the studies were in Asia: in China, 43 The remaining three qualitative studies used focus group design. 42,46,50 Of the quantitative studies (n = 7), four studies used cross-sectional design, 52-55 one was a case-control study, 56 and two used data from a cohort study. 57,58

| Quality appraisal
Based upon the results of the relevant quality appraisal tools, 13 studies were rated as high quality, [39][40][41][42][43][44][45][46][48][49][50][51]58 five as moderate quality 47,52,53,55,56 and two as low quality. 54,57 The low quality ratings were mostly due to methodological weaknesses; not measuring outcomes using validated tools, 57 failing to consider confounding variables 57 and methodological reporting, poor reporting of statistical methods, 54 neglecting to justify chosen methods 54 and unclear ethical procedures. 54 As few studies have been published exploring grandparent's feeding behaviours when caring for their 2-to 4-yearold grandchildren, these studies were included in the review.
Nevertheless, the results of those studies assessed as low quality 54,57 should be interpreted with caution.

| Dietary provision
The first feeding construct relates to dietary provision. Within this category three main subthemes emerged: frequency of food provision, the types of food and drinks and the portion size of food and drinks provided.
Two studies in Japan reported that three-generational households (where grandparents live with their grandchildren) were associated with unfixed 55 or irregular snack intake in children 57 but also regular meal times. 55 Similarly, in another Japanese study, the odds of 3-yearold boys and girls eating between meals three or more times per day was 2.9 and 2.5 times the odds when cared for by grandparents compared with when cared for by parents. 58 Likewise, two qualitative studies from the United Kingdom also suggested grandparents provide frequent snack foods 39,50 and are more likely to provide confectionary than parents, 39 even though grandparents may discourage snacking before meal times. 50 In another study, in the United States, Latina mothers complained that their parents were frequently serving their children junk food before meals. 42 However, Australian grandparents described how they ensured their grandchildren did not eat too close to a meal so as to not spoil their appetite, or if their grandchildren were really hungry they would only provide a healthy option like fruit. 51 Qualitative interviews with parents and grandparents also described how grandparents focus their attention on providing set meals 39,51 but often this was to their grandchild's detriment as they may encourage children to eat more frequently. For instance, it was noted in two studies that grandparents often served an additional meal, even if their grandchild had eaten that meal previously, while in childcare, for example, lunch. 43,45

| Types of food and drink served
The review identified 14 studies (quantitative n = 2, 53,54 qualitative n = 12 [39][40][41][42][43][44][45][46][48][49][50][51] ) that reported on the types of food that grandparents provide to their grandchildren. Two out of 12 studies focused specifically on fruit and vegetable consumption. For example, a US observational study compared the food served to children by parents and grandparents. 54 The authors reported that more grandparents served fruit in an evening meal than parents (39% vs. 25% respectively) and that a similar number of parents and grandparents served vegetables (75% vs. 73%, respectively such as cheese and fruit and buy novel or more expensive but healthy items that the parents found too costly. 51 Nevertheless, grandparents did provide discretionary items as treats occasionally, believing that exposing their grandchildren to some discretionary items would allow them to develop self-control and learn to moderate their own intake. 51 Evidence on grandparents drinks provision was also mixed, despite grandparents and parents agreeing on which beverages are healthy/unhealthy. 40 Grandparents in one study had stricter rules than parents on sugary drinks provision. For example, not using beverages as treats and using low fat or low sugar substitutes for more unhealthy drinks. 40 This finding is in contrast to what was reported in four other studies, which described grandparents as providing frequent sugary drinks to their grandchildren. 39,42,45,48

| Portion size of food served
Three qualitative studies 43,46,47 reported on the portion size of food that grandparents provided to their grandchildren; no quantitative evidence was found. When parents in the United States were interviewed, they described grandparents serving "huge quantities of chips and candy" to their grandchildren. 46 Another study interviewing Latina mothers described grandparents who feed their grandchildren as if they were "stuffing a bag". 47 Chinese grandparents were also accused of overfeeding their grandchildren, providing large portion sizes and encouraging their grandchildren to eat more even if the grandchild is full. 43 In Latina and Chinese grandparents, this feeding behaviour stemmed from grandparents preference for overweight in children, 43,47 and some Chinese grandparents held the belief that overweight children will be thin when they grow up, or that eating more will help their grandchildren to grow taller. 43

| Structure
Grandparents may engage in some feeding practices that provide structure to children's eating environment and promote healthy eating behaviours in children. For instance, in a UK study comparing parent and grandparent feeding practices, grandparents were more likely than parents to create a healthy feeding environment. 52 Similarly, when grandparents from the United States were interviewed, grandparents discussed how they ensured that their grandchildren had access to healthy choices in their home, with specific reference to the choice of beverages. 40 Grandparents may also model appropriate food and drink consumption, for instance when Australian grandparents were interviewed about their treat provision, it was evident that grandparents were mindful of what they were consuming in the presence of their grandchildren and understood the importance of modelling healthy behaviours. 51 Similarly, grandparents from the United States described how they modelled healthy beverage consumption after observing the mother's modelling techniques regarding sugary drinks. 40 However, two studies comparing the feeding practices of parents and grandparents suggest that grandparents may be less likely than parents to engage in some feeding practices that promote healthy eating behaviours. In these studies, grandparents were less likely to model healthy eating behaviours, 52,56 promote variety and a well-balanced food intake 52,56 and less likely to monitor intake 56 than parents.

| Autonomy
Grandparents may also engage in feeding practices that help to promote autonomy and independence in children. For instance, in two studies, grandparents described how they spend time to teach their grandchildren about nutrition and communicate about healthy eating to allow their grandchildren to make healthy choices themselves, 40,44 although not all grandparents saw this as their role or responsibility. 44 Moreover, five qualitative studies demonstrated how grandparents consider their grandchildren's food and drink preferences, [42][43][44][45]50 for meal choice, 43 drinks and snacks. 42,44,45 Nevertheless, this was often done in a permissive manner 42,44,45,50 and in one study was framed as grandparents being unable to say no to their grandchildren's requests. 50 Similarly, although grandparents from the United States described how they wanted to create a balanced eating environment and prevent their grandchildren from becoming fixated on unhealthy foods, they framed this as a justification for their excessive provision of treat foods and snacks. 39

| Coercive control
Grandparents were also described as using coercive feeding practices.
Coercive practices are often used to impose the caregiver's own desires and influence a child's socialization around food. They include pressure, restriction and using food to control behaviour or emotions, and they can have a negative impact on eating behaviours and preferences. 59 Restriction was the most commonly reported of these feeding practices, whereby five qualitative studies 40,[42][43][44]51 and two quantitative studies 52,56 highlighted that grandparents use restriction practices. Grandparents in one study restricted carbonated sugarsweetened beverages (SSBs) for health reasons, 40 whereas in another study, grandparent's restriction of food items was more likely to be for weight concerns 52 than parents. Australian grandparents, in another study, chose to restrict items based on sugar and foodadditive content rather than restricting items based upon salt and fat content. 44 Their grandchild's behaviour was also a motivating factor for restriction, rather than nutritional concerns. 44 Grandparents may also choose to restrict food items when a main meal is imminent for fear of spoiling their grandchildren's appetites. 44 Interestingly, although grandparents do use restriction practices, some grandparents did acknowledge that prohibiting food and drink items entirely can have negative consequences. 40,51 Four studies reported grandparents using food or drinks as a reward for behaviour, 40,43,44,51 whether this was withholding treat foods (e.g., fast food items) due to negative behaviours 43 or rewarding good behaviours with snack foods (e.g., ice cream and chips/ crisps). 43,44,51 Some grandparents only used food to reward their grandchildren's behaviour to align with parental practices, particularly to reinforce toilet training. 51 In the same study, grandparents were also opposed to using food to regulate their grandchildren's emotions, yet in other studies, grandparents were more likely than parents to use food to regulate their grandchildren's emotions. 52,56 Pressuring grandchildren to eat was also a theme that emerged from one qualitative study in China. 43 This included encouraging grandchildren to eat in the absence of hunger, to finish food in the bowl and to have second servings of food. Grandparents reportedly used pressure to prevent food waste and encourage greater consumption. Other reasons for pressuring their grandchildren to eat include having painful memories of hunger themselves and not wanting their grandchildren to experience this. Also, some grandparents had inaccurate perceptions of both child weight status (strong not fat) and consequences of eating more (becoming tall) and therefore pressured their grandchildren to eat.
Grandparents rationalized their indulgent feeding style as a way of counteracting parents strictness, introducing a sense of fun and creating a bond with the child 39 as well as a way to transmit love. 43 Grandparents believed that being indulgent was inconsequential due to the limited time they spent caring for their grandchild. 40,44 Although grandparents in one study were often indulgent, providing their grandchildren with food items they requested, they did not want to be manipulated by their grandchildren. 51 To overcome this, grandparents had rules and boundaries, particularly regarding the number of treats their grandchildren could have in one go, 51 suggesting that a more authoritative feeding approach is adopted by some. 51 Three other studies referred to grandparents having rules and being strict with food and drink provision. 40,41,48 This could also reflect either an authoritative or an authoritarian feeding style, but it was unclear if these rules were accompanied by a responsiveness to their grandchild's preferences and needs or not, which would clarify this.

| DISCUSSION
This review identified 20 studies that reported grandparental feeding behaviours when caring for their 2-to 4-year-old grandchildren. This included both qualitative and quantitative studies of mostly high and moderate quality from a wide range of countries. The findings are described using three concepts: dietary provision, feeding practices and feeding styles.

| Dietary provision
The evidence suggests that grandparents might be serving large portion sizes and encouraging their grandchildren to eat frequently, offering snacks and additional meals to their preschool-aged grandchildren.
However, the evidence is less clear regarding the types of food and drink provided. Although a majority of the studies indicated that grandparents are typically providing unhealthy food and drink items high in fat and sugar, there was also evidence to suggest grandparents provide healthy options to their grandchildren, using core foods as treat items and providing fruit and vegetables with meals. Discrepancies may result from differences between parental and grandparental report; for instance, within a single study in Australia, grandparents described themselves as being strict with sugary drink consumption, whereas parents of the same children described grandparents as regularly offering fruit juice and soft drinks. 48 Another finding is that grandparents value their role as a provider of meals for preschoolaged children. Sometimes, this led to grandparents providing a meal that the child had already eaten elsewhere so, for example, the child would eat two evening meals. Even though none of the included studies assessed the content of meals provided by grandparents, consuming regular meals is associated with higher fibre intake, lower added sugar intake and better diet quality. 60 Consequently, this behaviour may have a positive effect on the diets of preschool-aged children.

| Feeding practices
The results for feeding practices were also mixed. Grandparents demonstrated some feeding practices that provide structure to a child's eating environment such as creating a healthy food environment, ensuring healthy food is accessible to their grandchildren and modelling healthy eating, which are all conducive to healthy food intake in children. 61,62 This is particularly important because in a study of Australian grandparents, creating a healthy food environment and having healthy food available and accessible in the home was the strongest predictor of child diet quality compared with other feeding practices. 21 Moreover, the results of this review suggest that grandparents use some feeding practices that may help to support their grandchildren's autonomy and independence with food and drink, such as teaching their grandchildren about nutrition and considering their grandchildren's preferences and choices. 59 Nevertheless, it is difficult to distinguish between grandparents considering their grandchildren's preferences and being permissive. Evidence suggests that grandparents are also engaging in some feeding practices that are coercive in nature such as restriction, using food as a reward or for emotional comfort and pressuring grandchildren to eat, which correlate with children's unhealthy food consumption. 61,62 Studies of parental feeding practices suggest that these coercive practices are also adopted by parents of children this age but may vary depending on child weight status, parental ethnicity and child sex. [63][64][65] It may therefore be useful to explore the variation in the use of different feeding practices across distinctive characteristics of grandparents and grandchildren. It is also worth noting that there is not always a straight forward relationship between feeding practices and eating behaviour. For instance, the ways in which caregivers choose to restrict food and drink items may influence children's eating behaviour positively or negatively. Covert restriction, whereby children are unaware that caregivers are restricting food and drink items, can promote the consumption of healthy foods in children, whereas overt restriction, whereby the child is aware of the food or drink being restricted, can increase child consumption of these restricted items. 66,67 None of the studies in this review explored these constructs explicitly, and therefore, it is unknown if grandparents use overt or covert methods when they restrict food and drink items.
However, in a few studies, grandparents understood that forbidding food entirely, a key characteristic of overt restriction, can have negative consequences. 66 Another oversight of the literature reviewed here is that none of the studies considered how the behaviour of the grandchild might shape the feeding behaviours of grandparents. Some child eating behaviours, for example, satiety responsiveness and slowness in eating, are, at least in part, genetically driven, 68,69 and evidence indicates that child eating behaviours can precede and drive caregiver feeding behaviours. 70,71 For instance, fussy eating in young children is predictive of caregivers use of coercive feeding practices such as pressuring to eat, using food as a reward and overt restriction. 71,72 No study in this review captured child eating behaviours, and therefore, we do not know if grandparents feeding practices are used in response to child eating behaviours or vice versa. Exploring the direction of the relationship between grandparents feeding practices and their grandchildren's eating behaviours may be a key future direction for research.

| Feeding style
Finally, this review found that grandparents predominantly use an indulgent feeding style with their preschool-aged grandchildren, confirming that being indulgent is a common part of the grandparenting role. 23 Indulgent feeding styles are characterized by being responsive to children's needs and preferences but having few rules and demands on their grandchildren's needs. This is important because compared with the authoritarian feeding style, an indulgent feeding style is associated with increased intake of energy dense, nutrient poor foods 73 and a decreased intake of fruit and vegetables. 22 Consequently, an indulgent feeding style is negatively associated with preschoolers' weight status. 74 Other reviews have found that grandparents predominantly have a negative influence on the diets of children aged 0-18 years old 24,26 ; however, this review suggests that the relationship between grandparental care and preschool-aged children's eating is less clear. The age of the grandchild may explain these differences, whereby grandparents feeding mechanisms differ for preschool-aged children compared with older children. This argument is supported by a recent study which found that grandparents of younger grandchildren were more likely to give praise to their grandchildren for healthy eating, set limits of unhealthy food consumption and use food as a reward than grandparents of older grandchildren. 21  Similarly, no study has assessed grandparental feeding styles using a validated feeding styles tool, and as discussed, this is needed to fully examine their feeding styles and how they might influence their grandchildren's dietary intake. Doing so may also help to establish whether grandparent's indulgent feeding styles are analogous to parent's indulgent feeding styles.
Few studies have explored the relationship between the amount of time that grandparents spend with their preschool aged grandchildren and their feeding behaviours. A recent study found that 3-to 14-year-old grandchildren cared for regularly by their grandparents were more likely to consume both healthy (fruits and vegetables) and unhealthy (sugary and savoury snacks and sugary drinks) foods compared with those cared for by their grandparents less frequently. 21 This suggests that increased time in grandparental care may have both a positive and negative effect on children's eating behaviours. Moreover, further research is needed to explore how the amount of time grandparents spend caring for their preschool-aged grandchildren impacts upon their dietary provision, feeding styles and feeding practices.
Further exploration is also needed into the intergenerational nature of feeding behaviours. No study explored how feeding styles may be similar between related parents and grandparents, but there is some evidence to suggest this is the case for feeding practices. For instance, one study explored the correlation between related mother's and grandmother's feeding practices and found significant correlations between their use of pressure to eat, restriction and encouragement of energy balance, 56 suggesting that these behaviours may be passed down generationally from grandparent to parent. In another study, although similarities were found between parent and grandparent feeding practices, these parents and grandparents were independent and unrelated. 52 Eli et al. 39 found that grandparents residing in intergenerational households had similar feeding attitudes to parents, but this may be a consequence of residing in the same home. Similarly, grandparents often described conforming to parental rules and practices. 39,44 It is evident that further work is needed to characterize the intergenerational nature of feeding behaviours and explore whether they are learnt behaviours or if these similarities are a result of cooperation and agreement in approaches. Doing so would also help to establish whether grandparents used similar feeding practices or adopted an indulgent feeding style with their own children or if it is unique to the grandparenting role.
Considering the results of this review and the gaps within the literature, the authors propose a set of recommendations ( Figure 2) for future research exploring the relationship between grandparents feeding behaviours and their preschool-aged grandchildren's eating behaviours.
It is clear from this review that grandparents play an important role in caring for their grandchildren in the early years and they may need support to modify their feeding styles, practices and food provision. Being indulgent may be embroiled within what it means to be a grandparent, and campaigns could encourage grandparents to be indulgent with their time, attention and activities rather than with food and drinks. Similarly, strategies could be developed to encourage grandparents to use non-food-based rewards and to refrain from using food as an emotional comfort. However, it is important to note that this review focuses specifically on feeding and does not account for the potentially positive and supportive role that grandparents play in their grandchildren's social and emotional wellbeing 3,75-77 and this should not be ignored. The results of this review do not suggest that grandparents care should be limited, rather that healthy eating initiatives targeting parents should be widened to ensure grandparents also receive support with their feeding.
In conclusion, grandparents use a range of feeding practices to influence their preschool-aged grandchildren's eating, and some of these, such as creating a healthy eating environment, may lead to healthful eating in young children, but some of these, such as using food as a reward, may lead to negative eating outcomes. An indulgent feeding environment was also documented widely in grandparents caring for their preschool-aged grandchildren as well as evidence that grandparents offer large and frequent snacks and meals. Grandparents are therefore an important group to target to improve dietary-related health outcomes in young children. Nevertheless, there is still a need to use validated tools to assess grandparental feeding styles and to directly measure grandparent's overall dietary provision. Only then can interventions be designed that are sensitive to both the positive and negative influences that grandparents have on the diets of their 2-to 4-year-old grandchildren.

ACKNOWLEDGEMENTS
This work was supported by funding from Wellcome Trust (108903/ B/15/Z) and the University of Sheffield.
F I G U R E 2 Recommendations for future studies assessing the relationship between grandparent care and preschoolers' diet