Home environment predictors of vegetable and fruit intakes among Australian children aged 18 months

Appetite 139 (2019) 95–104

Contents lists available at ScienceDirect

Appetite journal homepage: www.elsevier.com/locate/appet

Home environment predictors of vegetable and fruit intakes among Australian children aged 18 months

T

Kathleen E. Lacya,∗, Alison C. Spencea, Sarah A. McNaughtona, David A. Crawforda, Rebecca J. Wyseb,c, Luke Wolfendenb,c, Karen J. Campbella a

Deakin University, Institute for Physical Activity and Nutrition (IPAN), School of Exercise and Nutrition Sciences, Geelong, Australia Hunter New England Population Health, Hunter New England Local Health District, NSW, Australia c The University of Newcastle, School of Medicine and Public Health, Callaghan, NSW, Australia b

A R T I C LE I N FO

A B S T R A C T

Keywords: Vegetable Fruit Child Toddler Predictor

Suboptimal vegetable and fruit consumption by young children is common. Identifying predictors of vegetable and fruit intakes is important for informing strategies to promote sufficient intakes of these foods from early life. The aim of the present study was to examine predictors of toddlers’ vegetable and fruit intakes at age 18 months. This study involved secondary analysis of data from 361 child-mother dyads participating in the Melbourne Infant Feeding, Activity and Nutrition Trial in 2008–2010 at child ages four, nine and 18 months. Children's vegetable and fruit intakes were assessed at age 18 months using multiple 24-h dietary recalls. Data on potential predictor measures were collected via parent-completed questionnaires when children were four or nine months of age. Bivariate and multivariable linear regression models were used to test associations between children's average daily vegetable or fruit intake and potential predictors controlling for treatment arm and clustering by parent group. Multivariable models also controlled for covariates and potential confounders. Home availability of vegetables at age nine months was found to predict children's vegetable intake at age 18 months and remained significant (β = 20.19, 95% CI:7.23, 33.15, p = 0.003) in the multivariable model. Children's average daily fruit intake at age 18 months was predicted by maternal education at child age four months and the availability of fruits in their home at child age nine months. Maternal education remained significant (β = 30.83, 95% CI:12.17, 49.48, p = 0.002) in the multivariable model. Strategies to promote adequate vegetable and fruit intakes among young children should address known barriers to the availability of vegetables and fruits in the home from early in life. Additionally, messages encouraging fruit consumption may need to be tailored to mothers with lower levels of education.

1. Introduction Vegetables and fruit are key components of a healthful diet (National Health and Medical Research Council, 2013). Low consumption of vegetables by children is common worldwide (Australian Bureau of Statistics, 2016; Banfield, Liu, Davis, Chang, & Frazier-Wood, 2016; Black & Billette, 2013; Krebs-Smith, Guenther, Subar, Kirkpatrick, & Dodd, 2010; Public Health England and Food Standards Agency, 2014), even among very young children (Miles & Siega-Riz, 2017; Spence, Campbell, Lioret, & McNaughton, 2018). For example, fewer than one in five children aged 2–3 years are meeting vegetable intake recommendations in industrialised countries such as Australia

and the United States (Australian Bureau of Statistics, 2016; KrebsSmith et al., 2010). Although a higher proportion of children meet dietary guidelines for fruit intake compared with vegetable intake, an estimated 22–32% of Australian and American children aged 2–3 years do not meet their daily recommended number of fruit servings on a usual basis (Australian Bureau of Statistics, 2016; Krebs-Smith et al., 2010). A recent Australian longitudinal study has shown that even fewer children meet vegetable and fruit intake guidelines as they move across the first five years of life (Spence et al., 2018). As early as 18 months of age, children's vegetable and fruit consumption mirror inadequate intakes of adults, highlighting the importance of developing healthful eating behaviours in the first two years of life (Saavedra,

∗

Corresponding author. Deakin University, Locked Bag 20000, Geelong, VIC 3220, Australia. E-mail addresses: [email protected] (K.E. Lacy), [email protected] (A.C. Spence), [email protected] (S.A. McNaughton), [email protected] (D.A. Crawford), [email protected] (R.J. Wyse), [email protected] (L. Wolfenden), [email protected] (K.J. Campbell). https://doi.org/10.1016/j.appet.2019.04.009 Received 30 October 2018; Received in revised form 8 April 2019; Accepted 10 April 2019 Available online 13 April 2019 0195-6663/ © 2019 Elsevier Ltd. All rights reserved.

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

Groups from within a 60-km radius of the research centre with at least eight eligible families (English-speaking; or six families from groups in low socio-economic position areas) who provided informed consent (n = 542; 86% of the eligible participants) were randomised to either the intervention or control arm of the Melbourne InFANT Program. Assessments of children's dietary intake, home food environment, parental child-feeding behaviours and demographic characteristics were collected when children were approximately four, nine and 18 months of age. The primary outcomes of the Melbourne InFANT Program have been previously reported (Campbell et al., 2013a). The current study reports children's vegetable and fruit intakes at 18 months of age and the predictors (assessed when children were four or nine months of age) of these intakes. The Melbourne InFANT Program was approved by the Deakin University Human Research Ethics Committee (ID number: EC 175–2007) and by the Victorian Office for Children (Ref: CDF/07/ 1138).

Deming, Dattilo, & Reidy, 2013) especially given evidence that dietary patterns established in early childhood track into later childhood (Madruga, Araujo, Bertoldi, & Neutzling, 2012) and possibly beyond (Craigie, Lake, Kelly, Adamson, & Mathers, 2011). Additionally, good nutrition during the first 1000 days (conception through age 24 months) of life is thought to be crucial for non-communicable disease prevention later in life (Mameli, Mazzantini, & Zuccotti, 2016). Effective interventions promoting adequate intakes of vegetables and fruits from early in life are needed (Hodder et al., 2017). In order to develop effective interventions, it is important to understand the modifiable and non-modifiable predictors of vegetable and fruit intakes. The home environment is a particularly important setting influencing the development of young children's eating behaviours and food preferences (Birch, 1999; Birch & Fisher, 1998; Harrison et al., 2011). Rosenkranz and Dzewaltowski have proposed a model that conceptualises the child's home food environment as three overlapping domains (Political and Economic Environments, Built and Natural Environments, Socio-Cultural Environments), each with macro- (existing at the larger community level) and micro-level components (most proximal to a child's home life) (Rosenkranz & Dzewaltowski, 2008). The micro-level components within these domains may be useful in determining groups that need the greatest support or identifying factors that could be most amenable to intervention. The existing literature suggests that several components of the child's home environment, including socioeconomic position (Rasmussen et al., 2006), home availability of vegetables and fruits (Pearson, Biddle, & Gorely, 2009; van der Horst et al., 2007), parental eating behaviours (van der Horst et al., 2007) and parenting practices (Burnier, Dubois, & Girard, 2011; Möller, de Hoog, van Eijsden, Gemke, & Vrijkotte, 2013; Perrine, Galuska, Thompson, & Scanlon, 2014) are related to the vegetable and/or fruit intakes of older children, but little is known about whether these or other components are predictors of vegetable and fruit intakes of toddlers (Carletti et al., 2013). In addition to the potential predictors identified through the model (Rosenkranz & Dzewaltowski, 2008), more recent research has indicated the importance of maternal self-efficacy (the confidence with which a behaviour is performed) in influencing early childhood diets (Campbell, Hesketh, Silverii, & Abbott, 2010; Spence, Campbell, Crawford, McNaughton, & Hesketh, 2014; Wyse et al., 2012). For example, maternal self-efficacy for promoting healthy eating has been shown to be associated with young children's vegetable and fruit consumption (Campbell et al., 2010). Maternal self-efficacy around child feeding could be considered as part of a child's socio-cultural environment at home and should also be considered as a potential predictor of toddlers' vegetable and fruit intakes. The aim of the present study was to examine modifiable and nonmodifiable home environment predictors of vegetable and fruit intakes among a sample of Australian toddlers at 18 months of age.

2.2. Dietary intake Children's dietary intakes were assessed using telephone-administered, five-pass, 24-h dietary recalls completed by parents when children were 18 months of age. This method was based on those used by the United States' Feeding Infants and Toddlers Study (Devaney et al., 2004), the Australian National Children's Nutrition and Physical Activity Survey (Department of Health and Ageing, 2008), and the United States Department of Agriculture (Conway, Ingwersen, & Moshfegh, 2004; Moshfegh, Vinyard, Ingwersen, & Conway, 2003). Trained nutritionists administered dietary recalls on three non-consecutive days, including one weekend day. This method of dietary data collection is considered the most accurate for estimating total energy intake in slightly older children (4–11 years) (Burrows, Martin, & Collins, 2010) and has also been used in multiple national surveys with infants and toddlers (Anater et al., 2018; Australian Bureau of Statistics, 2014; Foster & Bradley, 2018; National Center for Health Statistics, 2018). To assist with the classification of mixed dishes parents gave specific information about the ingredients and proportions used in homemade recipes, and purpose-designed food measurement booklets were provided to assist parents with quantity estimates. The booklets were purpose-made for this population and incorporated original photographs of measured food quantities, together with pictures from the food model book used in the 2007 Australian National Children's Nutrition and Physical Activity Survey (Department of Health and Ageing, 2008). Foods included were based on previous report of frequently consumed foods in this age group (Webb, Rutishauser, & Knezevic, 2008), and portion sizes photographed (three for each food) were based on the 25th percentile, median, and 75th percentile of reported intake per eating occasion (Webb et al., 2008). A detailed description of all methods associated with dietary assessment is available elsewhere (Spence, 2012). Data were coded by trained researchers using an in-house, purposedesigned database incorporating the Australian Food, Supplement and Nutrient Database (“AUSNUT 2007”) (Food Standards Australia New Zealand, 2018). New food items, particularly infant-specific products, were added to the study database when reported foods or beverages did not correspond to an item in AUSNUT 2007 and nutrient composition data for foods and beverages were obtained from the product's nutrition information panel or from the manufacturer. The coding of each recall was examined for accuracy and completeness by a dietitian. Vegetable and fruit intakes were calculated based on the standard food group classifications developed by Food Standards Australia New Zealand (Food Standards Australia New Zealand, 2018). The calculation for the amount of vegetables consumed included all raw and cooked vegetables, legumes and the relevant proportion of mixed dishes that contained vegetables (i.e. “disaggregation” using recipe calculations (Fitt et al., 2010)). Potatoes were not included in the amount of

2. Methods 2.1. Study design and participants The present study utilised data from the intervention and control arms of The Melbourne Infant Feeding, Activity and Nutrition Trial (InFANT) Program. The Melbourne InFANT Program was a clusterrandomised controlled trial to develop parenting practices to enhance children's diet and physical activity behaviours from early infancy (Campbell et al., 2008). It was conducted in Melbourne, Australia (2008–2010) with first-time parents of children who were approximately four months of age at baseline. The details of the methodology used in the Melbourne InFANT Program have been reported previously (Campbell et al., 2008, 2013a). The Melbourne InFANT Program was conducted through the existing social setting of first-time parents’ groups, which were established by the universal Maternal and Child Health Service in Victoria, Australia. 96

Predictor

nine

Home availability of foods

97

Maternal fruit consumption

Mother coping with life

Mother offers the child different foods

Breastfeeding status

Child age of introduction to solid foods Maternal self-efficacy for promoting healthy eating

Maternal self-efficacy for limiting unhealthy foods

Parental eating/dieting

Family structure, stress and schedules

Parenting: styles, practices and rules

Parenting: styles, practices and rules

Parenting: styles, practices and rules In addition to the Modelc

In addition to the Modelc

Parental eating/dieting

nine

nine

nine

nine

nine

nine

nine

nine

nine

Parental eating/dieting

Mother eats meals at same time as child (modelling) Maternal vegetable consumption

four

Domain: Socio-Cultural Environments Education/Nutrition Maternal nutrition knowledge knowledge

Availability of fruits

nine

four

Child age when assessed (months)

Domain: Built and Natural Environments Home availability of foods Availability of vegetables

Domain: Political and Economic Environments Family socioeconomic Family socioeconomic status status

Micro-level component

How old was your baby when you started giving him/her solid foods? (ICC = 0.92)b Four items asked mothers about their level of confidence in getting their child to: 1. eat enough vegetables (not including potato or potato chips), 2. drink plain water (with no flavours or juice added), 3. eat a good range of foods and 4. eat enough fruit (not including fruit juice) over the next year. (Kappa = 0.38–0.65a for individual items; Cronbach's α = 0.79 when items were grouped). Four items asked mothers about their level of confidence in declining their baby's demands/fussing for: 1. softdrinks, fruit juice, cordials and other sweetened drinks, 2.

How long did you breastfeed your baby for? (Kappa = 1.00)a

I offer my baby a different kind of food if I find they don't like what I am offering. (Kappa = 0.33)a

Overall, how do you think you are coping with life at present?

In the last 6 months, how many serves of fruit did you usually eat per day?

In the last 6 months, how many serves of vegetables, including potatoes, did you usually eat per day?

17 nutrition knowledge items assessed mothers' knowledge of the Dietary Guidelines for Australian Adults (National Health and Medical Research Council, 2003) (9 items), nutrient sources of foods (3 items), selecting everyday foods (2 items) and relationships between diet and diseases (3 items) I eat my meals at the same time as I feed my baby. (Kappa = 0.47)a

About how often are these foods available in your home? Fruit (Kappa = 0.64)a

About how often are these foods available in your home? Vegetables other than potato (Kappa = 0.55)a

What is the highest level of schooling you have completed?

Measure

Table 1 List of the Melbourne Infant Feeding, Activity and Nutrition Trial measures used as potential predictors.

Campbell et al. (Campbell et al., 2010)

Response options ranged from “extremely confident” to “not at all confident”. All items were averaged to give a composite score for unhealthy

(continued on next page)

Campbell et al. (Campbell et al., 2010)

Purpose-designed based on unpublished research (Campbell, 2004)

Adapted from Australian Temperament Project (Prior, Sanson, Smart, & Oberklaid, 1983)

Adapted from Cancer Council Victoria Dietary Questionnaire for Epidemiological Studies (version 3) (Cancer Council Victoria, 2018)

Adapted from Cancer Council Victoria Dietary Questionnaire for Epidemiological Studies (version 3) (Cancer Council Victoria, 2018)

Purpose-designed based on unpublished research (Campbell, 2004)

Items adapted from Parmenter and Wardle (Parmenter & Wardle, 1999) and validated for use in Australia by Hendrie et al. (Hendrie, Cox, & Coveney, 2008)

Adapted from Macfarlane et al. (Macfarlane, Abbott, Crawford, & Ball, 2010) and Campbell et al. (Campbell et al., 2013b) Adapted from Macfarlane et al. (Macfarlane et al., 2010) and Campbell et al. (Campbell et al., 2013b)

Source

Response options ranged from “extremely confident” to “not at all confident”. All items were averaged to give a composite score for healthy eating. Composite scores were grouped into two levels: not at all to very confident and extremely confident.

Correct responses were awarded one point and incorrect responses were awarded no points. Points for all 17 items were totalled to obtain an overall nutrition knowledge score, with higher scores indicating better nutrition knowledge (Collins, Lacy, Campbell, & McNaughton, 2016). Four response options ranging from “strongly agree” to “strongly disagree” grouped into two levels: agree and disagree Nine response options ranging from “I don't eat vegetables” to “7 serves or more/day” grouped into two levels and reported as: < 3 servings/day and ≥3 servings/day Eight response options ranging from “I don't eat fruit” to “6 serves or more/day” grouped into two levels and reported as: < 2 servings/day and ≥2 servings/day Five response options ranging from “not at all” to “extremely well” grouped into two levels: not at all/a little/fairly well and very well/extremely well Four response options ranging from “strongly agree” to “strongly disagree” grouped into two levels: agree and disagree Three response options ranging from “never” to “still breastfeeding this baby” grouped into two levels: never/stopped and still Open responses for months and weeks

Four response options ranging from “never” to “always” grouped into two levels: sometimes/ usually and always Four response options ranging from “never” to “always” grouped into two levels: sometimes/ usually and always

Seven categorical responses grouped into two levels: less than university degree and university degree or beyond

Response

K.E. Lacy, et al.

Appetite 139 (2019) 95–104

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

2.3. Predictor measures Elements within the micro-level components of the model proposed by Rosenkranz and Dzewaltowski (Rosenkranz & Dzewaltowski, 2008), with the additional elements of maternal self-efficacy for promoting healthy eating and maternal self-efficacy for limiting unhealthy foods, were considered as predictors. The present study focused on mothers, as females tend to be the household food gatekeepers (main household food shopper and food preparer) in Australia (Burton, Reid, Worsley, & Mavondo, 2017). Data on potential predictor measures were collected via paper-based questionnaires completed by parents when children were approximately four and nine months of age. Table 1 presents details about the predictor measures, including the age of the child when they were assessed and the sources and reliability of the measures (reliability was examined via a two-week test-retest study in a separate (unrelated) sample of 60 mothers of nine-month-old infants). 2.4. Covariates Two covariates were assessed in the present study (child age and child energy intake) because of their potential associations with vegetable or fruit intakes (Kudlova & Schneidrova, 2012; Lorson, MelgarQuinonez, & Taylor, 2009). Child date of birth (used to calculate child age) was assessed via the questionnaire completed when children were around four months of age. Child average daily energy intake was calculated from the 24-h dietary recalls performed when children were 18 months of age. 2.5. Statistical analysis A total of 542 child-parent dyads were recruited to the Melbourne InFANT Program. Child-parent dyads were excluded from this analysis if parent participants were not first-time parents (n = 14; excluded due to the possibility that more experienced parents or those with multiple children may feed their children differently (Andrew & Harvey, 2011) or may have different dietary intakes themselves (Nasuti et al., 2014) compared with first-time parents), they did not participate in data collection when children were aged 18 months (n = 48), they did not complete at least two dietary recalls at child age 18 months (n = 82), fathers completed the data collection when children were four and/or nine months of age (n = 2) or they had incomplete data for predictor variables (n = 32) or covariates (n = 0). Children whose energy intakes were more than three standard deviations (SD) above (n = 3) or below (n = 0) the mean energy intake were considered outliers and were also excluded from analyses (Campbell et al., 2014; Lioret, McNaughton, Spence, Crawford, & Campbell, 2013; Spence et al., 2013). The analytical sample for the current study was 361 (n = 182 (50%) control) children. For analyses pertaining to vegetable intakes, children whose vegetable intakes were more than three SD above (n = 5) or below (n = 0) the mean vegetable intake were excluded, resulting in a sample size of 356. For analyses pertaining to fruit intakes, children whose fruit intakes were more than three SD above (n = 2) or below (n = 0) the mean fruit intake were excluded, resulting in a sample size of 359. These exclusions were made because the intakes were considered improbable. There were no significant mean differences in children's vegetable or fruit intakes at 18 months between the intervention and control arms (Campbell et al., 2013a) so all analyses were conducted with both groups combined but controlling for treatment arm and clustering by parent group. Descriptive statistics were performed for outcome and predictor

c

eating. Composite scores were grouped into two levels: not at all to very confident and extremely confident. potato chips/twisties/cheezels and similar foods, 3. sweet snacks, confectionary, lollies and/or ice-cream and 4. foods like lollies, chocolate or biscuits over the next year. (Kappa = 0.28–0.54a for individual items and Cronbach's α = 0.90 when items were grouped).

Two-week test-retest assessment of question reliability conducted in an unrelated sample of mothers (n = 60) of nine-month-old infants. ICC = intraclass correlation coefficient. Considered as part of the socio-cultural environment in addition to the components of the Model of the home food environment pertaining to childhood obesity (Rosenkranz & Dzewaltowski, 2008). b

a

Response Micro-level component

Table 1 (continued)

Predictor

Child age when assessed (months)

Measure

Source

vegetables consumed (Campbell et al., 2013a; O'Connor, Walton, & Flynn, 2016). The calculation for the amount of fruit consumed included all raw, cooked, tinned and dried fruits, but excluded juice and fruit in yoghurts, jams, smoothies, cake, breakfast cereals, pies and crumbles.

98

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

variables and covariates. Means and SD were calculated for normally distributed outcome and predictor variables and covariates. Frequencies were calculated for categorical predictor variables and sex of the child. Medians and interquartile ranges (IQR) were calculated for the non-normally distributed predictor variables maternal nutrition knowledge and self-efficacy for promoting healthy eating and limiting unhealthy foods. A new dichotomised variable was created for the predictor variable maternal nutrition knowledge based on the median for this score (Collins et al., 2016). New dichotomised variables were also created for categorical predictor variables (maternal education, home availability of vegetables, home availability of fruits, mother eating meals with her child, maternal fruit consumption, mother coping with life, mother offering child different foods, breastfeeding status, child's age of introduction to solid foods, maternal self-efficacy for promoting healthy eating and limiting unhealthy foods) using theoretically meaningful cut-points. For example, maternal fruit consumption was compared to the Dietary Guidelines for Australian Adults (National Health and Medical Research Council, 2003). Fruit intake was dichotomised at the guideline of 2 servings/day. Given so few mothers (< 10%) met the vegetable guideline of 5 servings/day (National Health and Medical Research Council, 2003), this variable was dichotomised at 3 servings/ day based on the distribution of mothers' responses (Inglis, Ball, & Crawford, 2008). Cut-points for dichotomised variables are presented in the Supplementary Table. Bivariate linear regression (accounting for treatment arm and clustering by parent group) was performed to assess the relationship between children's average daily vegetable or fruit intake and each predictor. Several multivariable linear regression models were used to test associations between children's average daily vegetable or fruit intake and the predictors while including treatment arm, clustering by parent group, child age, child energy intake and potential confounders (Campbell et al., 2013b; Cameron et al., 2015) identified through the DAGitty (Textor, Hardt, & Knuppel, 2011; Williamson et al., 2014) online program (http://dagitty.net/dags.html). For example, when examining maternal nutrition knowledge, mother eating meals with her child, mother coping with life, breastfeeding status and child's age of introduction to solid foods as predictors, maternal education was considered a potential confounder as socioeconomic position may be related to these predictors and child diet. Similarly, when examining home availability of vegetables and home availability of fruits as predictors, maternal education and maternal nutrition knowledge were considered potential confounders. When examining maternal vegetable consumption and maternal fruit consumption as predictors, maternal education, maternal nutrition knowledge and home availability of vegetables or fruits were considered potential confounders. Results of regression analyses were considered statistically significant at P < 0.005 (using Bonferroni adjustment 0.05/11 predictors = 0.0045–0.005) (Howell, 2008). Regression model multicollinearlity diagnostics were calculated using variance inflation factors for predictors and covariates. Stata (Version 13, 2015, College Station, Texas, USA) was used for all statistical analyses.

Table 2 Participant characteristics (n = 361). Political and Economic Environments predictors Maternal educationa, n (%) Less than university degree 151 (42%) University degree or beyond 210 (58%) Built and Natural Environments predictors Availability of vegetables other than potatob, n (%) Never Sometimes Usually Always Availability of fruitsb, n (%) Never Sometimes Usually Always

0 2 (1%) 36 (10%) 323 (89%) 0 7 (2%) 52 (14%) 302 (84%)

Socio-Cultural Environments predictors Maternal nutrition knowledgea, Median (IQR) 14 (13, 15) Mother eats meals at same time as childb, n (%) Strongly disagree 22 (6%) Disagree 148 (41%) Agree 158 (44%) Strongly agree 33 (9%) Servings of vegetables consumed daily by mother in previous six monthsb, n (%) 1 or fewer servings per day 65 (18%) 2 servings per day 100 (28%) 3 servings per day 113 (31%) 4 servings per day 48 (13%) 5 or more servings per day 35 (10%) Servings of fruit consumed daily by mother in previous six monthsb, n (%) I don't eat fruit 3 (1%) 1 or fewer servings per day 159 (44%) 2 servings per day 131 (36%) 3 servings per day 53 (15%) 4 or more servings per day 15 (4%) b Coping with life , n (%) Not at all 0 A little 12 (3%) Fairly well 99 (27%) Very well 173 (48%) Extremely well 77 (21%) b Offering different foods , n (%) Strongly agree 29 (8%) Agree 186 (52%) Disagree 125 (35%) Strongly disagree 21 (6%) Breastfeeding statusb, n (%) Never 8 (2%) Stopped 168 (47%) Still 185 (51%) b Age of introduction to solid foods (months) , Mean (SD) 5.3 (0.8) Maternal self-efficacyb, c, Median (IQR) Promoting healthy eating 3.5 (3.3, 4.0) Limiting unhealthy foods 3.3 (3.0, 4.0)

Total proportions may not equal 100 due to rounding. a Assessed at child age four months. b Assessed at child age nine months. c Maternal self-efficacy possible score range: 1–4, with higher scores indicating greater confidence.

3. Results The average ( ± SD) age of children was 18.0 ( ± 1.5) months with 54% boys (n = 196). More than half (58%) of the mothers were tertiary educated (Table 2). The average daily vegetable intake for children was 81 ( ± 51; minimum 0; maximum 249) grams and the average daily fruit intake was 158 ( ± 79; minimum 0; maximum 415) grams. The average daily energy intake was 4482 ( ± 834) kJ. Vegetables and fruits were reported to be always available in most homes. About half (53%) of mothers reported eating meals with their child. Just over half (54%) of mothers reported consuming at least three servings of vegetables per day and a similar proportion (56%) reported consuming at least two servings of fruit each day. More than two thirds (69%) of

mothers reported coping at least “very well” with life. More than half (60%) of mothers reported offering different foods if their child did not like the food on offer, and about half (51%) of mothers were still breastfeeding when the child was nine months of age. The Supplementary Table reports children's average vegetable and fruit intakes according to each of the predictors. Children's average daily vegetable intake at 18 months of age was positively associated with the availability of vegetables in their home when they were nine months of age (Table 3; β = 23.99, 95% CI: 10.39, 37.58, p = 0.001). This association remained significant after adjusting 99

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

Table 3 Relationships between average vegetable and fruit intakes at age 18 months and covariates and predictors. Average vegetable intake (n = 356) Bivariate Regressiona

Multivariable Regression

β

95% CI

p-value

Β

95% CI

p-value

Predictors Political and Economic Environments predictors Maternal educationb, c

5.28

−4.48, 15.04

0.284

4.09

−5.33, 13.51

0.388

Built and Natural Environments predictors Availability of vegetables other than potatod, e

23.99

10.39, 37.58

0.001

20.19

7.23, 33.15

0.003

3.24 −0.04 10.21 4.44 −2.45 7.38 −3.12

−7.72, 14.20 −12.27, 12.20 −0.82, 21.25 −7.95, 16.84 −13.13, 8.23 −2.78, 17.55 −14.16, 7.91

0.557 0.995 0.069 0.476 0.648 0.151 0.573

0.22 −0.38 6.73 4.21 −1.01 5.82 1.49

−10.83, 11.27 −11.93, 11.17 −3.36, 16.82 −8.51, 16.93 −12.01, 10.00 −5.20, 16.84 −10.56, 13.53

0.968 0.948 0.187 0.511 0.856 0.295 0.806

−0.02 3.91

−8.47, 8.43 −7.97, 15.78

0.997 0.513

−1.91 3.03

−10.67, 6.85 −8.18, 14.25

0.664 0.590

Average fruit intake (n = 359) Bivariate Regressiona β 95% CI

p-value

Multivariable Regressionb Β 95% CI

p-value

Predictors Political and Economic Environments predictors Maternal educationb, c

31.79

13.63, 49.95

0.001

30.83

12.17, 49.48

0.002

Built and Natural Environments predictors Availability of fruitd, e

41.10

17.98, 64.22

0.001

32.41

9.64, 55.19

0.006

22.77 15.02 8.90 −2.80 0.34 23.26 −12.08

5.10, 40.43 −2.58, 32.61 −8.33, 26.12 −21.07, 15.46 −16.14, 16.82 5.60, 40.92 −30.90, 6.75

0.012 0.093 0.306 0.760 0.967 0.011 0.204

13.61 14.06 −8.33 −3.99 2.08 16.43 −6.49

−3.02, 30.23 −2.86, 30.98 −26.72, 10.07 −21.46, 13.48 −14.57, 18.74 −0.78, 33.63 −24.91, 11.93

0.107 0.102 0.369 0.649 0.803 0.061 0.484

4.91 7.02

−10.34, 20.17 −7.66, 21.69

0.522 0.343

2.12 5.07

−13.14, 17.39 −9.15, 19.29

0.782 0.478

Socio-Cultural Environments predictors Maternal nutrition knowledgeb, f Mother eats meals at same time as childd, f Servings of vegetables consumed daily by mother in previous six monthsd, Coping with lifed, f Offering different foodsc, d Breastfeeding statusd, f Age of introduction to solid foodsd, f Maternal self-efficacyc, d Promoting healthy eating Limiting unhealthy foods

Socio-Cultural Environments predictors Maternal nutrition knowledgeb, f Mother eats meals at same time as childd, f Servings of fruit consumed daily by mother in previous six monthsd, Coping with lifed, f Offering different foodsc, d Breastfeeding statusd, f Age of introduction to solid foodsd, f Maternal self-efficacyc, d Promoting healthy eating Limiting unhealthy foods

g

g

a

Bivariate regression models accounted for treatment arm and clustering by parent group. Assessed at child age four months. c Multivariable regression model accounted for treatment arm, clustering by parent group, child age and energy intake. d Assessed at child age nine months. e Multivariable regression model accounted for treatment arm, clustering by parent group, child age, child energy intake, maternal education and maternal nutrition knowledge. f Multivariable regression model accounted for treatment arm, clustering by parent group, child age, child energy intake and maternal education. g Multivariable regression model accounted for treatment arm, clustering by parent group, child age, child energy intake, maternal education, maternal nutrition knowledge and home availability of vegetables or fruit. b

multivariable regression model (β = 30.83, 95% CI:12.17, 49.48, p = 0.002). However, the availability of fruits in the home when the child was nine months of age was not considered significantly associated (P = 0.006 compared with Bonferroni adjustment of P < 0.005) with children's fruit intake at 18 months of age after adjusting for treatment arm, clustering by parent group, child age, child energy intake and potential confounders maternal education and maternal nutrition knowledge in the multivariable regression model. There were no significant associations between children's fruit intake at 18 months and the other predictors examined.

for treatment arm, clustering by parent group, child age, child energy intake and the potential confounders, maternal education and maternal nutrition knowledge, in the multivariable regression model (β = 20.19, 95% CI:7.23, 33.15, p = 0.003). There were no significant associations between children's vegetable intake at 18 months and the other predictors examined. Children's average daily fruit intake at 18 months of age was positively associated with maternal education when the child was four months of age (β = 31.79, 95% CI: 13.63, 49.95, p = 0.001) and the availability of fruits in their home when the child was nine months of age (β = 41.10, 95% CI: 17.98, 64.22, p = 0.001). The association between maternal education at four months and children's fruit intake at 18 months remained significant after adjusting for treatment arm, clustering by parent group, child age and child energy intake in the

4. Discussion The present study is the first to examine early childhood predictors 100

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

be related to a lack of maternal modelling of vegetable and fruit consumption. A recent study comparing the influence of parental role modelling with parental dietary intake on 2- to 5-year-old children's diet quality demonstrated that parental modelling of healthy eating was particularly important for influencing children's diet quality and may be even more important than the quality of parents' diets (Vaughn, Martin, & Ward, 2018). Another possible explanation for the lack of association between maternal and child vegetable and fruit consumption could be due to many mothers offering different kinds of food if their infant refused what was initially provided. This suggests the possibility that mothers might be offering alternative foods to their children when vegetables and fruit are being consumed by mothers in the presence of their child. What is not clear from the data though is whether the other foods offered are different vegetables or fruits that are known to be liked, foods from other food groups or less healthy alternatives. It is also possible that the different methods used to estimate mothers' (two questions asking about vegetable and fruit intake over the previous six months) and children's (multiple 24-h recalls with a median of 10 days between first and last recalls) vegetable and fruit intakes or the dichotomisation of the variables for maternal vegetable and fruit consumption could have influenced the results. Breastfeeding status when children were nine months of age was not predictive of children's vegetable or fruit intakes at 18 months of age. Exclusive (Burnier et al., 2011; Möller et al., 2013; Perrine et al., 2014) and any breastfeeding duration (Perrine et al., 2014) have previously been associated with vegetable and/or fruit intakes of older children (2–6 years). It is important to note, however, that in the present study the breastfeeding measure had limited response options (never, stopped and still). Additionally, only 2% of mothers reported never breastfeeding and their data were grouped with those of mothers who had stopped breastfeeding by the time the child was nine months of age, thus reducing the sensitivity of this measure. It is thought that breastfeeding promotes the development of healthy food preferences by exposing the infant to the flavours of healthy foods so that they become accepted by the infant (Mennella, 1995; Mennella, Jagnow, & Beauchamp, 2001), but this requires that the mother eats these foods regularly (Forestell & Mennella, 2007). Many women in the present study had suboptimal intakes of vegetables and/or fruits and this could have affected their infant's exposure to healthy food flavours while breastfeeding. It is also possible that the effects of exposure to flavours in breastmilk on children's initial acceptance of vegetables and fruit may wash out with children's repeated exposure to vegetables and fruits as solid foods. Maternal self-efficacy for promoting healthy eating and limiting unhealthy foods were not predictive of children's vegetable or fruit intakes at 18 months of age. Research examining maternal feeding selfefficacy and vegetable and fruit intakes of children under two years of age is fairly limited (Campbell et al., 2010; Koh et al., 2014). One study reported positive associations between maternal self-efficacy to promote healthy eating and vegetable intakes by children aged 1 year but found no association with fruit intake (Campbell et al., 2010). Maternal feeding self-efficacy has also been associated with 6- to 7-month-old children's vegetable variety, with no association found with fruit variety (Koh et al., 2014). Additionally, a study by Spence et al. found cross-sectional associations between maternal feeding self-efficacy and a measure of child diet quality, which included vegetable and fruit intakes, among children aged ∼18 months using the same dataset as the present study (Spence et al., 2014). To our knowledge, the present study is the first to assess longitudinal associations between maternal self-efficacy for promoting healthy eating/limiting unhealthy foods and toddlers' vegetable and fruit intakes. The lack of associations reported here could be related to the ceiling effects in the levels of self-efficacy reported at nine months (the majority of mothers reported being at least “very confident”). It is also possible that maternal feeding self-efficacy at an early stage in a child's life is not predictive of a child's food intakes later when the child has gained greater autonomy with age.

of vegetable and fruit intakes for Australian toddlers aged 18 months. Using a comprehensive model to identify potential predictors of vegetable and fruit intakes, only availability of vegetables in the home was found to be predictive of vegetable intake at 18 months of age, and maternal education was predictive of fruit intake at the same age. Home availability of fruit was predictive of children's fruit intake in the bivariate regression analysis but not the multivariable regression analysis. No other predictors examined were identified as predictive of vegetable or fruit intakes at 18 months of age. The findings relating to home availability of vegetables extend existing evidence from systematic reviews examining correlates of older children's (4–12 years (van der Horst et al., 2007) and 6–11 years (Pearson et al., 2009)) vegetable intakes. Those systematic reviews have concluded from largely cross-sectional studies that children's vegetable and fruit intakes are positively related to home availability of these foods (Pearson et al., 2009; van der Horst et al., 2007). A study of children in Australia who were closer in age (3–5 years) to the children in the present study reported similar findings (Wyse, Campbell, Nathan, & Wolfenden, 2011). Although fruit availability was not considered statistically significantly associated with children's fruit intake in the multivariable regression model due to the use of a Bonferonni adjustment to the p-value in the present study, the result was in the anticipated direction. Maternal education, used as a measure of socioeconomic position, was predictive of children's fruit intake, but not vegetable intake. Two systematic reviews of literature in older children (4–12 years (van der Horst et al., 2007) and 6–11 years (Pearson et al., 2009)) previously reported limited and mixed findings on the relationships between indicators of socioeconomic position (household education, household income, socioeconomic status/parental occupational class) and children's vegetable and fruit intakes. However, a broader review of the literature concluded that low socioeconomic position is consistently associated with less frequent or low intake of fruit and vegetables among children and adolescents (6–18 years), and that four of eight papers specifically examining maternal education found higher or more frequent intake of vegetables and/or fruit with increasing educational level of the mother (Rasmussen et al., 2006). A study in Italy among toddlers of the same age as those in the present study reported associations between maternal education and both vegetable and fruit intakes (Carletti et al., 2013). Findings from the present study appear to align with those from a nationally representative sample of 2- to 8-yearold children in Australia in which fruit intake was related to the education level of the primary carer for boys and to the education level of the secondary carer for girls but vegetable intake was not related to the education levels of primary or secondary carers for boys or girls (Cameron et al., 2012). It may be the case that vegetable intakes are consistently low among Australian children, regardless of socioeconomic position, or that relationships between socioeconomic position and children's vegetable and/or fruit intakes could vary across different measures of socioeconomic position (Cameron et al., 2012; Rasmussen et al., 2006) or be mediated by other factors (Vereecken, Keukelier, & Maes, 2004). Maternal vegetable and fruit consumption when children were nine months of age were not predictive of children's vegetable and fruit consumption when children were 18 months of age. Several studies have reported cross-sectional relationships between maternal or parental vegetable and fruit consumption and children's consumption of these foods among older age groups (Cooke et al., 2004; Fisher, Mitchell, Smiciklas-Wright, & Birch, 2002; McGowan, Croker, Wardle, & Cooke, 2012; Wolnicka, Taraszewska, Jaczewska-Schuetz, & Jarosz, 2015), although other studies have found relationships between parental and older children's intakes of fruit only (Gibson, Wardle, & Watts, 1998; Longbottom, Wrieden, & Pine, 2002). As a large proportion of mothers reported not eating meals at the same time as their infant, it is possible that the absence of associations between maternal vegetable and fruit consumption and children's intakes of these foods could also 101

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

Findings from the present study have implications for informing strategies to support healthy vegetable and fruit intakes among young children. Firstly, they indicate that always having vegetables (and fruits) available in the home is important. Vegetables and fruit are often introduced as first foods, so it may be the case that toddlers have had many exposures to, grown to accept and even developed preferences for these foods by the time they reach 18 months of age. Parents may also stock these foods (or pre-prepared versions of them, such as pouches) in their home because they are generally quick and easy to prepare for young children, and frequent provision of these foods throughout the day may increase children's intakes (Wyse, Wolfenden, & Bisquera, 2015). While it appears to be important to encourage the availability of vegetables and fruits in the home from early in life, it is also possible that always having vegetables or fruit available in the home is a proxy for other potentially important factors such as having a generally healthy home food environment with few competing unhealthy foods or a consistent home food environment or approach to child feeding. Consideration of the facilitators (e.g. convenience) and barriers (e.g. perceived cost, potential for food waste) to making vegetables and fruits available in the homes of young children and greater understanding of the mechanisms underlying the relationship between availability and consumption are needed to better inform strategies to support healthy vegetable and fruit intakes among young children. Secondly, the finding that maternal education level is predictive of toddlers' fruit intakes suggests that it may be necessary for messages encouraging fruit consumption to be tailored to mothers with lower levels of education as their children could be particularly vulnerable to lower fruit intakes (Supplementary Table). The present study has several strengths, including the use of multiple 24-h dietary recalls to collect children's dietary data and the calculation of vegetable intakes including content from the disaggregation of mixed dishes. These dietary data are unique (Smithers, Golley, Brazionis, & Lynch, 2011), as, to the authors' knowledge, the Melbourne InFANT Program is the only contemporary Australian study with data from multiple 24-h dietary recalls available for children under two years of age (Lioret et al., 2013). Another strength is the use of an established model (Rosenkranz & Dzewaltowski, 2008) to guide the selection of potential predictors. Additionally, most of the measures that were utilised as potential predictors were used in previous research or were shown to be reliable through test-retest assessments conducted as part of the Melbourne InFANT Program. Limitations of the present study include the use of largely self-/ parent-reported measures, which may increase the likelihood of bias, including recall bias and social desirability bias. There was a relative lack of variation in the sample for some predictor measures and the generalisability of the study may be limited due to the high proportion of highly educated mothers as well as the possibility that mothers who regularly attend first-time parent groups may parent their children or perform health behaviours differently to those who do not attend these groups. Studies conducted among a more diverse group of mothers and studies investigating predictors related to fathers and other caregivers are needed to confirm and extend the findings of the present study. It is also worth noting that two of the potential predictors from child age nine months (mother eating meals with her child, mother offering her child different foods), which were not shown to be predictive of children's vegetable and fruit intakes at child age 18 months, could likely have changed during this 9-month period. Such changes may be due to several factors. For example, the child being included in more family meals as they age or the mother returning to work and the child eating meals while in the care of others could affect whether and how often the mother eats meals with her child. Age-related factors such as increased child autonomy in eating or the onset of food fussiness in the child could lead to the mother offering her child different foods when what is originally offered is not accepted. Conversely, a history of foods being rejected by her child could lead a mother to become reluctant to offer new foods and to offer only previously accepted foods.

Although the present study examined multiple potential predictors, it is likely that other predictors related to the micro- and macro-level components of a child's home food environment that were not included are relevant to toddlers' vegetable and fruit intakes and warrant further investigation. One example at the micro level is parenting practices around the introduction of vegetables into an infant's diet. In particular, repeatedly offering a variety of vegetables as first foods (before offering sweeter foods like fruit) during the start of complimentary feeding is likely to influence young children's vegetable acceptance and consumption (Chambers, 2016; Chambers et al., 2016). Another example is parental food preparation skills, which includes food literacy around planning for providing vegetables and fruit and procuring and preparing these foods. Pertinent macro-level examples include vegetable and fruit availability and accessibility in the community. Further research into these and a variety of other predictors is needed. 5. Conclusions The present study is the first to examine early childhood predictors of vegetable and fruit intakes for Australian toddlers aged 18 months. Home availability of vegetables was found to be predictive of toddlers' vegetable intakes and maternal education was found to be predictive of toddlers’ fruit intakes. Strategies to support adequate vegetable and fruit intakes among young children should address known barriers to the availability of vegetables and fruits in the home from early in life. Additionally, messages encouraging vegetable consumption should target all mothers of infants and messages encouraging fruit consumption may need to be tailored to mothers with lower levels of education. Declarations Ethics approval and consent to participate The Melbourne InFANT Program was approved by the Deakin University Human Research Ethics Committee (ID number: EC 175–2007) and by the Victorian Office for Children (Ref: CDF/07/ 1138). Families provided informed consent to participate in the Melbourne InFANT Program. Availability of data and material The datasets analysed for the current study are not publicly available due to ethical restrictions related to the consent given by participants at the time of study commencement. An ethically compliant dataset may be made available by the corresponding author on reasonable request and upon approval by the Deakin University Human Research Ethics Committee. Competing interests The authors declare that they have no competing interests. Funding The Melbourne InFANT Program was supported by the National Health and Medical Research Council (grant #425801). Additional funds were supplied by the Heart Foundation Victoria and Deakin University. Authors' contributions KEL undertook and managed the majority of data analyses, wrote the initial manuscript, and revised the final manuscript. ACS participated in day-to-day management of the Melbourne InFANT Program including recruitment, data collection and co-ordination of the dietary 102

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

recall coding, contributed to the design of the data analysis and critically revised the manuscript for important intellectual content. SAM (associate investigator for the Melbourne InFANT Program) designed and managed all dietary data collection for the Melbourne InFANT Program and contributed to writing the manuscript. All authors read and approved the final manuscript. DAC (chief investigator for the Melbourne InFANT Program) was involved in the conceptualization, design, and management of all aspects of the Melbourne InFANT Program and critically revised the manuscript for important intellectual content. RJW and LW critically revised the manuscript for important intellectual content. KJC (lead chief investigator for the Melbourne InFANT Program) conceptualized, designed, and managed/co-managed all aspects of the Melbourne InFANT Program and critically revised the manuscript for important intellectual content.

(2015). A review of the relationship between socioeconomic position and the earlylife predictors of obesity. Current Obesity Report, 4(3), 350–362. Campbell, K. J. (2004). Family food environments as determinants of children's eating: Implications for obesity prevention. Deakin University. Campbell, K. J., Abbott, G., Spence, A. C., Crawford, D. A., McNaughton, S. A., & Ball, K. (2013b). Home food availability mediates associations between mothers' nutrition knowledge and child diet. Appetite, 71, 1–6. Campbell, K. J., Hendrie, G., Nowson, C., Grimes, C. A., Riley, M., Lioret, S., et al. (2014). Sources and correlates of sodium consumption in the first 2 years of life. Journal of the Academy of Nutrition and Dietetics, 114(10), 1525–1532 e2. Campbell, K., Hesketh, K., Crawford, D., Salmon, J., Ball, K., & McCallum, Z. (2008). The Infant Feeding Activity and Nutrition Trial (INFANT) an early intervention to prevent childhood obesity: Cluster-randomised controlled trial. BMC Public Health, 8, 103. Campbell, K., Hesketh, K., Silverii, A., & Abbott, G. (2010). Maternal self-efficacy regarding children's eating and sedentary behaviours in the early years: Associations with children's food intake and sedentary behaviours. International Journal of Pediatric Obesity, 5(6), 501–508. Campbell, K. J., Lioret, S., McNaughton, S. A., Crawford, D. A., Salmon, J., Ball, K., et al. (2013a). A parent-focused intervention to reduce infant obesity risk behaviors: A randomized trial. Pediatrics, 131(4), 652–660. Cancer Council Victoria (2018). Dietary questionnaires. Available from: https://www. cancervic.org.au/research/epidemiology/nutritional_assessment_services. Carletti, C., Macaluso, A., Pani, P., Caroli, M., Giacchi, M., Montico, M., et al. (2013). Diet and physical activity in pre-school children: A pilot project for surveillance in three regions of Italy. Public Health Nutrition, 16(4), 616–624. Chambers, L. (2016). Complimentary feeing: Vegetables first, frequently and in variety. Nutrition Bulletin, 41(2), 142–146. Chambers, L., Hetherington, M., Cooke, L., Coulthard, H., Fewtrell, M., Emmett, P., et al. (2016). Reaching consensus on a ‘vegetables first’ approach to complementary feeding. Nutrition Bulletin, 41, 270–276. Collins, L. J., Lacy, K. E., Campbell, K. J., & McNaughton, S. A. (2016). The predictors of diet quality among Australian children aged 3.5 years. Journal of the Academy of Nutrition and Dietetics, 116(7), 1114–1126 e2. Conway, J. M., Ingwersen, L. A., & Moshfegh, A. J. (2004). Accuracy of dietary recall using the USDA five-step multiple-pass method in men: An observational validation study. Journal of the American Dietetic Association, 104(4), 595–603. Cooke, L. J., Wardle, J., Gibson, E. L., Sapochnik, M., Sheiham, A., & Lawson, M. (2004). Demographic, familial and trait predictors of fruit and vegetable consumption by preschool children. Public Health Nutrition, 7(2), 295–302. Craigie, A. M., Lake, A. A., Kelly, S. A., Adamson, A. J., & Mathers, J. C. (2011). Tracking of obesity-related behaviours from childhood to adulthood: A systematic review. Maturitas, 70(3), 266–284. Department of Health and Ageing (2008). 2007 Australian National Children's Nutrition and Physical Activity user guide. Canberra, Australia: AGPS. Devaney, B., Kalb, L., Briefel, R., Zavitsky-Novak, T., Clusen, N., & Ziegler, P. (2004). Feeding infants and toddlers study: Overview of the study design. Journal of the American Dietetic Association, 104(1 Suppl 1), s8–13. Fisher, J. O., Mitchell, D. C., Smiciklas-Wright, H., & Birch, L. L. (2002). Parental influences on young girls' fruit and vegetable, micronutrient, and fat intakes. Journal of the American Dietetic Association, 102(1), 58–64. Fitt, E., Mak, T. N., Stephen, A. M., Prynne, C., Roberts, C., Swan, G., et al. (2010). Disaggregating composite food codes in the UK National Diet and Nutrition Survey food composition databank. European Journal of Clinical Nutrition, 64(Suppl 3), S32–S36. Food Standards Australia New Zealand (2018). AUSNUT 2007. [cited 2018 Sept]. Available from: http://www.foodstandards.gov.au/science/monitoringnutrients/ ausnut/Pages/ausnut2007.aspx. Forestell, C. A., & Mennella, J. A. (2007). Early determinants of fruit and vegetable acceptance. Pediatrics, 120(6), 1247–1254. Foster, E., & Bradley, J. (2018). Methodological considerations and future insights for 24hour dietary recall assessment in children. Nutrition Research, 51, 1–11. Gibson, E. L., Wardle, J., & Watts, C. J. (1998). Fruit and vegetable consumption, nutritional knowledge and beliefs in mothers and children. Appetite, 31(2), 205–228. Harrison, K., Bost, K. K., McBride, B. A., Donovan, S. M., Grigsby-Toussaint, D. S., Kim, J., et al. (2011). Toward a developmental conceptualization of contributors to overweight and obesity in childhood: The six-Cs model. Child Development Perspectives, 5(1), 50–58. Hendrie, G. A., Cox, D. N., & Coveney, J. (2008). Validation of the general nutrition knowledge questionnaire in an Australian community sample. Nutrition and Dietetics, 65(1), 72–77. Hodder, R. K., Stacey, F. G., Wyse, R. J., O'Brien, K. M., Clinton-McHarg, T., Tzelepis, F., et al. (2017). Interventions for increasing fruit and vegetable consumption in children aged five years and under. Cochrane Database of Systematic Reviews, 9 CD008552. van der Horst, K., Oenema, A., Ferreira, I., Wendel-Vos, W., Giskes, K., van Lenthe, F., et al. (2007). A systematic review of environmental correlates of obesity-related dietary behaviors in youth. Health Education Research, 22(2), 203–226. Howell, D. C. (2008). Fundamental statistics for the behavioral sciences (6th ed.). United States of America: Thomson Wadsworth, a part of the Thomas Corporation. Inglis, V., Ball, K., & Crawford, D. (2008). Socioeconomic variations in women's diets: What is the role of perceptions of the local food environment? Journal of Epidemiology & Community Health, 62(3), 191–197. Koh, G. A., Scott, J. A., Woodman, R. J., Kim, S. W., Daniels, L. A., & Magarey, A. M. (2014). Maternal feeding self-efficacy and fruit and vegetable intakes in infants. Results from the SAIDI study. Appetite, 81, 44–51. Krebs-Smith, S. M., Guenther, P. M., Subar, A. F., Kirkpatrick, S. I., & Dodd, K. W. (2010). Americans do not meet federal dietary recommendations. Journal of Nutrition,

Acknowledgements We thank all participating local government regions for their encouragement, advice, and access to Maternal and Child Health nurses. We are grateful to the participating families and nurses for their enthusiastic commitment to the study and in particular to the Melbourne InFANT Program research team and student volunteers. LW is supported by a National Health and Medical Research Council Career Development Fellowship, and a Heart Foundation Future Leaders Fellowship. RW is supported by a National Health and Medical Research Council Translating Research Into Practice (TRIP) Fellowship. SAM is supported by an National Health and Medical Research Council Career Development Fellowship Level 2, ID1104636 and was previously supported by an Australian Research Council Future Fellowship (2011–2015, FT100100581). Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.appet.2019.04.009. References Anater, A. S., Catellier, D. J., Levine, B. A., Krotki, K. P., Jacquier, E. F., Eldridge, A. L., et al. (2018). The Feeding Infants and Toddlers Study (FITS) 2016: Study design and methods. Journal of Nutrition, 148(9S), 1516S–1524S. Andrew, N., & Harvey, K. (2011). Infant feeding choices: Experience, self-identity and lifestyle. Maternal and Child Nutrition, 7(1), 48–60. Australian Bureau of Statistics (2014). Australian Health Survey: Users' guide, 2011-13. 24Hour dietary recall. ABS catalogue No. 4363.0.55.001 Canberra, Australia: Commonwealth of Australia. [cited 2018 Sept]. Available from: http://www.abs. gov.au/ausstats/[email protected]/Lookup/4363.0.55.001Chapter65022011-13. Australian Bureau of Statistics (2016). Australian Health Survey: Consumption of food groups from the Austrlaian Dietary Guidelines. ABS catalogue No. 4364.0.55.012 Canberra, Australia: Commonwealth of Australia. Available from: http://www.abs.gov.au/ AUSSTATS/[email protected]/DetailsPage/4364.0.55.0122011-12?OpenDocument. Banfield, E. C., Liu, Y., Davis, J. S., Chang, S., & Frazier-Wood, A. C. (2016). Poor adherence to US dietary guidelines for children and adolescents in the national health and nutrition examination survey population. Journal of the Academy of Nutrition and Dietetics, 116(1), 21–27. Birch, L. L. (1999). Development of food preferences. Annual Review of Nutrition, 19, 41–62. Birch, L. L., & Fisher, J. O. (1998). Development of eating behaviors among children and adolescents. Pediatrics, 101(3 Pt 2), 539–549. Black, J. L., & Billette, J. M. (2013). Do Canadians meet Canada's Food Guide's recommendations for fruits and vegetables? Applied Physiology Nutrition and Metabolism, 38(3), 234–242. Burnier, D., Dubois, L., & Girard, M. (2011). Exclusive breastfeeding duration and later intake of vegetables in preschool children. European Journal of Clinical Nutrition, 65(2), 196–202. Burrows, T. L., Martin, R. J., & Collins, C. E. (2010). A systematic review of the validity of dietary assessment methods in children when compared with the method of doubly labeled water. Journal of the American Dietetic Association, 110(10), 1501–1510. Burton, M., Reid, M., Worsley, A., & Mavondo, F. (2017). Food skills confidence and household gatekeepers' dietary practices. Appetite, 108, 183–190. Cameron, A. J., Ball, K., Pearson, N., Lioret, S., Crawford, D. A., Campbell, K., et al. (2012). Socioeconomic variation in diet and activity-related behaviours of Australian children and adolescents aged 2-16 years. Pediatric Obesity, 7(4), 329–342. Cameron, A. J., Spence, A. C., Laws, R., Hesketh, K. D., Lioret, S., & Campbell, K. J.

103

Appetite 139 (2019) 95–104

K.E. Lacy, et al.

https://aifs.gov.au/publications/pathways-infancy-adolescence. Public Health England and Food Standards Agency (2014). National Diet and Nutrition Survey results from years 1, 2, 3 and 4 (combined) of the rolling programme (2008/2009 2011/2012). London, UK: Public Health England. [cited 2016 Sept]. Available from: https://www.gov.uk/government/statistics/national-diet-and-nutrition-surveyresults-from-years-1-to-4-combined-of-the-rolling-programme-for-2008-and-2009-to2011-and-2012. Rasmussen, M., Krolner, R., Klepp, K. I., Lytle, L., Brug, J., Bere, E., et al. (2006). Determinants of fruit and vegetable consumption among children and adolescents: A review of the literature. Part I: Quantitative studies. International Journal of Behavioral Nutrition and Physical Activity, 3, 22. Rosenkranz, R. R., & Dzewaltowski, D. A. (2008). Model of the home food environment pertaining to childhood obesity. Nutrition Reviews, 66(3), 123–140. Saavedra, J. M., Deming, D., Dattilo, A., & Reidy, K. (2013). Lessons from the feeding infants and toddlers study in north America: What children eat, and implications for obesity prevention. Annals of Nutrition & Metabolism, 62(Suppl 3), 27–36. Smithers, L. G., Golley, R. K., Brazionis, L., & Lynch, J. W. (2011). Characterizing whole diets of young children from developed countries and the association between diet and health: A systematic review. Nutrition Reviews, 69(8), 449–467. Spence, A. (2012). Mediators of dietary patterns in an intervention to improve early childhood nutrition. Deakin University. Spence, A. C., Campbell, K. J., Crawford, D. A., McNaughton, S. A., & Hesketh, K. D. (2014). Mediators of improved child diet quality following a health promotion intervention: The Melbourne InFANT program. International Journal of Behavioral Nutrition and Physical Activity, 11, 137. Spence, A. C., Campbell, K. J., Lioret, S., & McNaughton, S. A. (2018). Early childhood vegetable, fruit, and discretionary food intakes do not meet dietary guidelines, but do show socioeconomic differences and tracking over time. Journal of the Academy of Nutrition and Dietetics, 118(9), 1634–1643.e1. Spence, A. C., McNaughton, S. A., Lioret, S., Hesketh, K. D., Crawford, D. A., & Campbell, K. J. (2013). A health promotion intervention can affect diet quality in early childhood. Journal of Nutrition, 143(10), 1672–1678. Textor, J., Hardt, J., & Knuppel, S. (2011). DAGitty: A graphical tool for analyzing causal diagrams. Epidemiology, 22(5), 745. Vaughn, A. E., Martin, C. L., & Ward, D. S. (2018). What matters most - what parents model or what parents eat? Appetite, 126, 102–107. Vereecken, C. A., Keukelier, E., & Maes, L. (2004). Influence of mother's educational level on food parenting practices and food habits of young children. Appetite, 43(1), 93–103. Webb, K., Rutishauser, I., & Knezevic, N. (2008). Foods, nutrients and portions consumed by a sample of Australian children aged 16–24 months. Nutrition and Dietetics, 65(1), 56–65. Williamson, E. J., Aitken, Z., Lawrie, J., Dharmage, S. C., Burgess, J. A., & Forbes, A. B. (2014). Introduction to causal diagrams for confounder selection. Respirology, 19(3), 303–311. Wolnicka, K., Taraszewska, A. M., Jaczewska-Schuetz, J., & Jarosz, M. (2015). Factors within the family environment such as parents' dietary habits and fruit and vegetable availability have the greatest influence on fruit and vegetable consumption by Polish children. Public Health Nutrition, 18(15), 2705–2711. Wyse, R., Campbell, E., Nathan, N., & Wolfenden, L. (2011). Associations between characteristics of the home food environment and fruit and vegetable intake in preschool children: A cross-sectional study. BMC Public Health, 11, 938. Wyse, R., Wolfenden, L., & Bisquera, A. (2015). Characteristics of the home food environment that mediate immediate and sustained increases in child fruit and vegetable consumption: Mediation analysis from the healthy habits cluster randomised controlled trial. International Journal of Behavioral Nutrition and Physical Activity, 12, 118. Wyse, R., Wolfenden, L., Campbell, E., Campbell, K. J., Wiggers, J., Brennan, L., et al. (2012). A cluster randomized controlled trial of a telephone-based parent intervention to increase preschoolers' fruit and vegetable consumption. American Journal of Clinical Nutrition, 96(1), 102–110.

140(10), 1832–1838. Kudlova, E., & Schneidrova, D. (2012). Dietary patterns and their changes in early childhood. Central European Journal of Public Health, 20(2), 126–134. Lioret, S., McNaughton, S. A., Spence, A. C., Crawford, D., & Campbell, K. J. (2013). Tracking of dietary intakes in early childhood: The Melbourne InFANT program. European Journal of Clinical Nutrition, 67, 275–281. Longbottom, P. J., Wrieden, W. L., & Pine, C. M. (2002). Is there a relationship between the food intakes of Scottish 5(1/2)-8(1/2)-year-olds and those of their mothers? Journal of Human Nutrition and Dietetics, 15(4), 271–279. Lorson, B. A., Melgar-Quinonez, H. R., & Taylor, C. A. (2009). Correlates of fruit and vegetable intakes in US children. Journal of the American Dietetic Association, 109(3), 474–478. Macfarlane, A., Abbott, G., Crawford, D., & Ball, K. (2010). Personal, social and environmental correlates of healthy weight status amongst mothers from socioeconomically disadvantaged neighborhoods: Findings from the READI study. International Journal of Behavioral Nutrition and Physical Activity, 7, 23. Madruga, S. W., Araujo, C. L., Bertoldi, A. D., & Neutzling, M. B. (2012). Tracking of dietary patterns from childhood to adolescence. Revista de Saúde Pública, 46(2), 376–386. Mameli, C., Mazzantini, S., & Zuccotti, G. V. (2016). Nutrition in the first 1000 Days: The origin of childhood obesity. International Journal of Environmental Research and Public Health, 13(9). McGowan, L., Croker, H., Wardle, J., & Cooke, L. J. (2012). Environmental and individual determinants of core and non-core food and drink intake in preschool-aged children in the United Kingdom. European Journal of Clinical Nutrition, 66(3), 322–328. Mennella, J. A. (1995). Mother's milk: A medium for early flavor experiences. Journal of Human Lactation, 11(1), 39–45. Mennella, J. A., Jagnow, C. P., & Beauchamp, G. K. (2001). Prenatal and postnatal flavor learning by human infants. Pediatrics, 107(6) E88. Miles, G., & Siega-Riz, A. M. (2017). Trends in food and beverage consumption among infants and toddlers: 2005-2012. Pediatrics, 139(6), e20163290. Möller, L. M., de Hoog, M. L. A., van Eijsden, M., Gemke, R. J. B. J., & Vrijkotte, T. G. M. (2013). Infant nutrition in relation to eating behaviour and fruit and vegetable intake at age 5 years. British Journal of Nutrition, 109(03), 564–571. Moshfegh, A. J., Vinyard, B. T., Ingwersen, L. A., & Conway, J. M. (2003). Effectiveness of the US Department of Agriculture 5-step multiple-pass method in assessing food intake in obese and nonobese women. American Journal of Clinical Nutrition, 77(5), 1171–1178. Nasuti, G., Blanchard, C., Naylor, P. J., Levy-Milne, R., Warburton, D. E., Benoit, C., et al. (2014). Comparison of the dietary intakes of new parents, second-time parents, and nonparents: A longitudinal cohort study. Journal of the Academy of Nutrition and Dietetics, 114(3), 450–456. National Center for Health Statistics (2018). National Health and Nutrition Examination Survey. Available from: https://www.cdc.gov/nchs/nhanes/index.htm. National Health and Medical Research Council (2003). Dietary Guidelines for Australian Adults. Canberra, Australia: Commonwealth of Australia. National Health and Medical Research Council (2013). Australian Dietary Guidelines. Canberra, Australia: National Health and Medical Research Council. O'Connor, L., Walton, J., & Flynn, A. (2016). Fruit and vegetable intakes, sources and contribution to total diet in very young children (1-4 years): The Irish national preschool nutrition survey. British Journal of Nutrition, 115(12), 2196–2202. Parmenter, K., & Wardle, J. (1999). Development of a general nutrition knowledge questionnaire for adults. European Journal of Clinical Nutrition, 53(4), 298–308. Pearson, N., Biddle, S. J., & Gorely, T. (2009). Family correlates of fruit and vegetable consumption in children and adolescents: A systematic review. Public Health Nutrition, 12(2), 267–283. Perrine, C. G., Galuska, D. A., Thompson, F. E., & Scanlon, K. S. (2014). Breastfeeding duration is associated with child diet at 6 years. Pediatrics, 134(Suppl 1), S50–S55. Prior, M., Sanson, A., Smart, D., & Oberklaid, F. (2000). Pathways from infancy to adolescence; Australian temperament project 1983-2000 (Research Report #4) Melbourne, Australia: Australian Institute of Family Studies. [cited 2018 Sept]. Available from:

104