Associations Between Swedish Mothers' and 3- and 5-Year-Old Children's Food Intake

Research Article Associations Between Swedish Mothers’ and 3- and 5-Year-Old Children’s Food Intake

Lena M. Hansson, PhD1; Berit L. Heitmann, PhD2,3,4; Christel Larsson, PhD5,6; Per Tynelius, MsC1; Mikaela Willmer, PhD1; Finn Rasmussen, PhD1 ABSTRACT

Objective: To investigate associations between mothers’ and children’s food intake. Design: Cross-sectional study. Background variables collected through self-reports and from the register of the total population. Mothers recorded their own and their children’s food intake in a diary during 2 4-day periods. Setting: Eight counties in mid Sweden. Participants: Three- and 5-year-old children and their mothers were randomly selected from the register of the total population. A total of 2,045 families were invited, 355 of whom accepted. Mothers who accepted were older and to a larger extent born in Sweden. The final sample of mother–child pairs with complete food records was 189. Main Outcome Measures: Mothers’ and children’s food intake (16 food items). Analysis: Spearman rank-order correlation with 95% confidence intervals (2-sided). Moderation was investigated using generalized estimation equations with robust variance. Results: The strongest correlations between mothers’ and children’s food intake were found for pizza and oily fish (r ¼ .70–.80). The weakest correlations were found for sugared drinks and fruit and berries (r ¼ .24–.26). Children’s age moderated the relationship between mothers’ and children’s intake of savoury snacks, as did place of residence for pizza intake. Conclusions and Implications: There were substantial correlations between children’s and mothers’ intake of various foods. Modeling of mothers’ intake might be more effective in influencing young children’s intake of certain foods, whereas other strategies, such as encouraging parents to influence food availability (eg, gatekeeping), might be more useful for some foods. Key Words: children, nutrition, parent–child relations, dietary intake (J Nutr Educ Behav. 2016;-:1-10.) Accepted May 28, 2016.

INTRODUCTION Despite the recognition that lifestyle habits are formed early in life and that

interventions need to target preschool children to prevent the development of childhood obesity,1 behaviors associated with healthy or unhealthy eating

1

Child and Adolescent Public Health Epidemiology Group, Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden 2 Research Unit for Dietary Studies, Institute of Preventive Medicine, Copenhagen University Hospital, Copenhagen, Denmark 3 Boden Institute of Obesity, Nutrition, Exercise, and Eating Disorders, University of Sydney, Sydney, Australia 4 National Institute of Public Health, University of Southern Denmark, Odense, Denmark 5 Department of Food and Nutrition, Ume
a University, Ume
a, Sweden 6 Department of Food and Nutrition, and Sport Science, University of Gothenburg, Gothenburg, Sweden Conflict of Interest Disclosure: The authors’ conflict of interest disclosures can be found online with this article on www.jneb.org. Address for correspondence: Lena M. Hansson, PhD, Child and Adolescent Public Health Epidemiology Group, Department of Public Health Sciences, Karolinska Institutet, Widerstr€ omska Huset, S-171 77 Stockholm, Sweden; Phone: þ46736791063; Fax: þ4686583350; E-mail: [email protected] Ó2016 Society for Nutrition Education and Behavior. Published by Elsevier, Inc. All rights reserved. http://dx.doi.org/10.1016/j.jneb.2016.05.015

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016

in young children have not been reviewed systematically.2 To a high extent, health behaviors, including eating habits, are learned within families.3 Young children are highly dependent on parents or caregivers with regard to the food that is served. Research has found links between both parenting styles4 and child feeding behaviors5 and children's eating habits. Parenting style can be categorized in 4 dimensions based on levels of responsiveness and demandingness: authoritative (high demandingness/high responsiveness), authoritarian (high demandingness/low responsiveness), permissive (low demandingness/high responsiveness), and uninvolved (low demandingness/low responsiveness).6 Parents with an authoritative parenting style might use specific child feeding behaviors such as role modeling and/ or restriction, and these specific child feeding behaviors may have different types of impact on children's food

1

2 Hansson et al intake. A recent review showed that few high-quality studies have investigated whether parenting style actually predicts a specific child feeding behavior.7 There were some indications that an authoritative parenting style was related to parental monitoring, whereas an authoritarian parenting style was associated with pressuring the child to eat. Conclusions from other reviews suggested that an authoritative parenting style accompanied by specific feeding behaviors such as role modeling, making healthy foods available, and covertly restricting unhealthy foods are associated with the healthiest eating habits among children.4,8 Counseling parents about their own health behaviors,9 child feeding behaviors,10 and parenting style11 may thus result in the development of more healthy eating habits among their young children. Previous studies among young children seemed to have focused more on negative child feeding behaviors such as parental pressure or restriction than on positive food parenting strategies, ie, role modeling and making healthy food available.6 With regard to young children's sugar-sweetened beverage intake, a positive relationship was found for parental consumption of sugar-sweetened beverages and a negative relationship for positive parental role modeling, whereas the association was inconclusive for food availability and restriction.12 A review of qualitative studies exploring the barriers and facilitators for obesogenic dietary intake in children aged 0–6 years showed that modeling is an important aspect.13 A core construct for changing behavior according to Social Cognitive Theory is observational learning, as well as social experience and reciprocal determinism.14,15 According to Social Cognitive Theory, children's consumption of certain foods may thus increase if parents model the consumption of these foods, provide opportunities to consume the desired food, and reward the child for the eating behavior. If this is true, parents and children are anticipated to show similarities in their food intake and eating patterns. Several systematic reviews investigated determinants of mainly schoolchildren's food intake.16-21 Although there seemed to be positive associations between parents' and children's eating habits, a recent systematic review

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016 suggested that the agreement was weak.22 It was found that the relationship between children's and parents' food intake varied across foods, nutrients, dietary assessment methods, study quality, and countries studied. It was also found that the relationship between parents' and children's dietary intake was stronger among children aged < 10 years, possibly because children are also influenced by other environmental factors such as peer pressure, and as children grow older they also develop independence.23 For preschool children, in addition to day care, parents are presumably the strongest influence for forming children's food habits.24 A handful of studies investigated the relationship between parents' and young children's intake of certain foods.25-30 These studies showed there were moderate to strong relationships between parents' and children's intake of fruit and vegetables, which were the foods that had been studied most extensively. However, the studies had some limitations, such as investigating selected groups of children or not using methodologically strong dietary measures for both the parent and the child. For young children it therefore still remains to be explored whether certain parental eating habits, rather than others, are more easily transmitted to their children.31 Parental modeling seems to be an important factor for young schoolchildren's eating habits, but it seems to decrease with children's increasing age. However, it is unknown whether the influence of parental modeling may change as early as during the preschool years. The food environment and social norms concerning food and eating are potential determinants of food intake and eating habits.31 The influence of demographic characteristics on the relationship between parents' and young children's eating habits has not been investigated. Some research demonstrated that healthier foods are less available in rural than urban areas, whereas other studies did not show a difference.32 However, rural areas often also have lower availability of supermarkets, and these stores generally have a greater availability of healthier products than do smaller grocery stores.33 In qualitative studies, parents and young children reported that they were influenced by advertising and the

high availability of unhealthy foods in public places.13 Social norms regarding food, eating and body size, and exposure to health campaigns may also differ between rural and urban areas. In Sweden, for instance, the level of adult obesity is higher in more rural areas.34 A recent report showed that the risk of overweight among 12-yearold children in main cities (including the Stockholm metropolitan area) in 2011 was approximately 30% and 45% lower than that for 12-year-olds living in large and small cities, respectively, in Sweden.35 The current study investigated the associations between 3- and 5-year-old children's and their mothers' intakes of various foods. The primary hypothesis was that mothers and children would show a similarity in their intakes of certain healthy and unhealthy foods. Furthermore, the present study explored whether the associations were moderated by children's age and place of residence.

METHODS Participants In 2008, parents of children aged 3 or 5 years were invited by a mailed letter to take part in the study. The children were randomly selected from Statistics Sweden's register of the total population. The families were recruited from 8 counties in the mid-parts of Sweden, including the Stockholm metropolitan area. The included counties well reflect the variation in Sweden's demographics. A total of 2,400 were invited (Figure). Approximately half of the families did not respond to the invitation. The majority of families that declined participation claimed that the reason for nonparticipation was time constraints. In total, 514 families (21%) accepted taking part in the study. Families were sent 1 reminder for each of 3 assignments, 2 4-day food diaries (FD), and a food-frequency questionnaire (FFQ) with background questions. Of the families that accepted participation, 355 (69%) mailed back only the first 4-day FD, whereas 287 (56%) returned both 4-day FDs for a parent, a child, or both. The proportion of fathers who recorded their food intake was low, and therefore only mother–child pairs were included. Furthermore, 8-day recordings for both the mother and the

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016 Families invited to participate n = 2400 None-responders n = 1158 Declined n = 626 Invalid contact information n = 18 Families accepting to participate n = 514 Did not return 1st 4-day food record

Hansson et al 3 child had to be complete for them to be included in the analysis. Three observations were excluded because mothers and fathers had answered different parts. Additional mother–child pairs were excluded because of missing data for certain meals (see Data Processing). This resulted in a final sample of 189 mother–child pairs (37% of those who accepted participation). Parts of this sample were used previously to validate the FFQ against the FD.1 In this study only data from the FD were used. Ethical approval for the study (registered number 5106/2005) was granted by the Regional Ethics Committee at Karolinska Institutet, Stockholm, Sweden. Written informed consent was obtained from all subjects.

n = 159 Families asked to record the

Food Record

2nd 4-day food record n = 355 Did not return record n = 68 Families with both food records of either the parent or child or both n = 287 Fathers excluded n = 35 Mothers or children with both food records n = 252 Incomplete 8-day food records n = 47 Fathers and mothers answering different parts n=3 Mother-child pairs with complete food records n = 202 Children that had not had breakfast and/or dinner for 2 days or more of the 8 days n = 13 Final analytic sample n = 189

Figure. Flowchart of included and excluded participants.

Parents were instructed to record their entire food intake (including beverages). The first 4 days corresponded to Wednesday through Saturday, and the second period, Sunday through Wednesday. The 2 FDs were sent out approximately 2–3 weeks apart. Parents were asked to record the amount of the food, dish, or product they consumed, the time and place of consumption, the cooking method, and what label the food had. A food template with food pictures was provided to estimate portion sizes. Parents were also instructed to use measuring cups or weigh the food. Children's food intake was recorded by their parents, which meant that parents knew the amount and types of foods eaten by their child. Therefore, meals eaten away from home without the parent, for instance, at public day care, were not recorded. Besides what the child ate and drank, parents recorded whether the child had attended day care, had been away from home for other reasons, or had been at home. They were also instructed to note which meals the child had had away from home that could not be recorded. The Swedish National Food Agency previously analyzed which kind of foods that can be used to indicate diet quality with regard to intake of saturated fat, sugar, and dietary fiber.36 Foods that were investigated in the current study were intended to reflect the quality of the diet. The foods in this study were fruit and berries; vegetables; fish (both oily and

4 Hansson et al lean); sausage (both in dishes and in sandwiches); hamburgers; french fries; pizza; ice cream; buns and cookies; sweets, including chocolate; salty snacks; and cakes and pastries. Sugared drinks included carbonated soft drinks, sweetened or flavored milk, and fruit drinks, such as lemonade. The intake of carbonated soft drinks was also investigated as its own category. Fruit juice was a separate category, but it was not possible to distinguish between natural fruit juice and fruit juice with added sugar.

Background Variables The FD was accompanied by a questionnaire on different background variables, such as parental educational level, country of birth, parental weight and height, and child weight and height. Body mass index (BMI) was calculated as weight (kilograms) divided by squared height (m2). Parental overweight or obesity was categorized as a BMI of $ 25 according to the World Health Organization's criteria for adults.37 For children, overweight and obesity were classified according to the cutoffs of the International Obesity Task Force criteria.38 Other health behaviors were also collected but were not used in this analysis. Both the FD and the questionnaire were in Swedish. Parents' and children's age and place of residence were collected from the register of the total population and their ages at the time of sampling were used in the analysis. Place of residence was categorized into the Stockholm metropolitan area and outside Stockholm, and was intended to reflect urban and rural areas.

Data Processing The 2 4-day FDs were entered and analysed using a computerized nutrition program, Dietist XP (Kost och N€aringsdata, Stockholm, Sweden, 2009). For instance, Dietist XP was used to transform measuring cups, portion sizes, or whole dishes into grams of each food item. If parents had not specified the amount of a certain food item or dish, it was assumed to be a standard portion. The average intake in grams per day of different foods, dishes, and products, and their contribution to energy and nutrient intake were calculated. The Dietist XP program in-

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016 cludes approximately 2,000 food items from the Swedish National Food Agency's nutritional database and nutritional data of 14,000 food items from the food industry and other companies specializing in nutrition. Furthermore, the researchers worked extensively to enter information on foods not yet available in the Dietist XP. Nutritional information was collected through producers' Web pages or by contacting them by telephone if foods that were recorded by the parent were not identified in the database. The information was then added to the Dietist XP program. Under the supervision of a registered dietitian, 2 research assistants with bachelors degrees in public health entered the reported food intakes in Dietist XP. The dietitian also made spot checks for entering errors and ascertained quality in the data process. Two nutritionists and 1 dietitian classified the diary data into the food categories. Foods in mixed dishes were also categorized. If the parents reported ready-made dishes, information on the specific food content was collected from the producer. In Sweden, approximately 95% of children aged 1–5 years attend day care for a mean of 31 h/wk.39 Most children eat a cooked lunch at day care and some also have 1 or possibly 2 snacks during the day. The majority of children in this sample went to day care, and therefore lunch during weekdays often was not recorded. To be able to relate the total recorded food intake of mothers with the infrequently recorded food intake of their children, children's and mothers' intake needed to be harmonized. This meant that all recorded weekday lunches for children and mothers were excluded. When parents were able to record their children's food intake for a whole day (that is, weekend days), most children (98%) had breakfast, lunch, dinner, and 1–2 snacks. Children who had not recorded breakfast and/or dinner for $ 2 days of the 8 days were therefore excluded from the analysis (n ¼ 13). The final sample consisted of 189 mother–child pairs (Figure).

Statistical Analysis Data from the questionnaire and the food data from Dietist XP were analyzed using SAS statistical software (version 9.3, SAS Institute, Cary, NC). First, the

mean and SD as well as the median intake in grams per day, grams per week, or grams per month for the different foods for both children and mothers were calculated. Relationships between children's and mothers' intake of each food group then were investigated by Spearman rank-order correlations with 95% confidence intervals (2-sided). Investigation of possible moderation by children's age and place of residence on the association between mothers' and children's food intake was examined by using multiple linear regression models. Maternal age, maternal educational level, maternal country of birth, maternal BMI, place of residence, and children's age and gender were adjusted for in all analyses, unless the variable was investigated as a moderator. Regression models were assessed graphically and residuals showed overall good fit. However, for some outcomes the residual distribution showed slight deviations from a normal distribution, and therefore robust variance estimated by generalized estimation equations when calculating P and 95% confidence intervals (CIs) was used. This resulted in somewhat higher P and broader CIs.

RESULTS Participant Characteristics Mothers in the families that agreed to participate (n ¼ 355) were slightly older than were mothers in families that did not participate (n ¼ 2,045): 32.5 years (SD, 4.5) vs 31.4 years (SD, 5.1) (P < .001). Among participating mothers, 92% were born in Sweden, compared with 77% of nonparticipants (P < .001). Seventy-three percent of mothers who participated and 68% of those who did not came from the Stockholm area (P ¼ .06). Children's gender and age did not differ between groups. Thus, participating mothers differed somewhat from the population from which they were drawn. Mother–child pairs that had complete food records (final analytical sample) differed from mother–child pairs with incomplete food records with regard to the following background variables: mother's and child's age, and mother's weight status (Table 1). There were some missing values for background variables for both the excluded and the analytic

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016

Hansson et al 5

Table 1. Characteristics of Mother–Child Pairs With Complete and Incomplete Food Records

pizza intake showed the strongest. Correlations for cakes and pastries, french fries, and fish, both lean and oily, were in the midrange. However, for the majority of foods correlations between mothers and children were below 0.50. For instance, the correlation between mothers' and children's intake of sugared drinks was 0.26 (95% CI, 0.12–0.39).

Complete Food Records (n ¼ 189)

Incomplete Food Records (n ¼ 166)

Background Characteristics Mother had $ 3 y postgraduate education

% 63

n 119

% 66a

Mother born in Sweden

93b

174

Living in Stockholm metropolitan area

76

144

Child female gender

52

99

53

n 69

P for Difference .64

91

150

.58

69

114

.12

88

.90

Maternal overweight (World Health Organization cutoff)

18

c

d

34

32

33

.01

Child overweight (International Obesity Task Force cutoff)

12e

20

15f

16

.31

Child age 3 y

58

109

43

72

.01

Maternal age, y (mean [SD])

33.0

4.3

32.0

4.6

.02

Maternal BMI, kg/m2 (mean [SD])

23.2c

3.8

24.1d

4.1

.07

Child age, years (mean [SD])

3.7

1.0

4.0

1.0

.01

Child BMI, kg/m2 (mean [SD])

15.9e

1.6

15.9f

1.8

.97

Association Between Mothers’ and Children’s Food Intake: Moderation of Child’s Age

BMI indicates body mass index. a n ¼ 105; bn ¼ 188; cn ¼ 187; dn ¼ 104; en ¼ 171; fn ¼ 93. sample; this was especially salient for children's height and weight (Table 1).

Mean Food Intake Among Children and Mothers Table 2 lists mean and median food intakes. Children consumed approximately 135 g fruit, berries, and vegetables/d and 50 g fish/wk. Corresponding figures for mothers were 300 and 150 g, respectively. Children drank on average 550 g sugared drinks/wk, 105 g of which was carbonated soft drinks. Mothers consumed higher amounts of carbonated soft drinks than did the children, but their total consumption of sugared drinks was lower than their children's consumption. In general, mothers had higher intakes of energy-dense foods than those of their children, except for ice cream. Children ate approximately 70 g ice cream/wk whereas mothers ate 45 g/wk.

Correlations Between Mothers’ and Children’s Food Intake Correlations between mothers' and children's mean intake of the different foods ranged from 0.24 to 0.80 (Table 3). All correlations were statistically signifi-

cant. The correlation between mothers' and children's intake of fruit and berries showed the weakest correlation whereas

Associations between mothers' and children's intake of buns and cookies as well as salty snacks, adjusted for maternal age, maternal BMI, maternal educational level, maternal national background, place of residence, and child gender, were statistically significantly moderated by child age (Table 4). The intake of salty snacks among 3-yearolds was not associated with mother's salty snack intake whereas 5-year-olds’ intake was significantly (P < .05) related to their mothers' intake. Carbonated soft drinks and sugared drinks were the only foods for which b estimates were attenuated after adjusting for the confounding factors. That

Table 2. Intakes of Different Foods of Mother–Child Pairs Children’s Intake Foods

n

Mean

SD

189

85.2

77.1

70.0

Cakes and pastries, g/wk 189

26.7

52.9

0.0

189 105.1 215.3

0.0

Buns and cookies, g/wk Carbonated soft drinks, g/wk

Mothers’ Intake

Median Mean

SD

165.6 140.4 54.5

Median 133.9

88.8

0.0

163.0 306.4

0.0

French fries, g/mo

189

60.6 127.3

0.0

129.4 264.8

0.0

Fruit and berries, g/d

189

92.0

59.6

81.9

146.4 106.3

126.3

Fruit juice, g/d

189

36.3

48.0

18.8

Hamburgers, g/mo

189

44.3 110.0

0.0

63.2

25.0

78.2 203.0

50.8

0.0

Ice cream, g/wk

189

66.5

73.1

Lean fish, g/wk

189

26.6

56.7

52.5

45.9

67.9

0.0

0.0

50.9

87.1

0.0

Oily fish, g/wk

189

27.5

44.1

Pizza, g/mo

189 121.3 268.7

0.0

100.1 135.5

52.5

0.0

364.8 664.2

0.0

Salty snacks, g/wk

189

14.0

22.9

5.3

61.3

26.3

Sausage, g/wk

189

96.3

83.1

87.5

136.3 134.4

109.4

Sugared drinks, g/wk

189 548.5 538.4

400.0

421.9 553.3

218.8

Sweets including chocolate, g/wk

189

64.0

79.5

43.8

89.7

88.9

65.6

Vegetables, g/d

189

42.5

26.8

39.1

150.8

73.0

140.0

49.2

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016

6 Hansson et al

Table 3. Spearman Correlations Between Mothers’ and Children’s Food Intake Foods Pizza

na 189

Spearman r .80

95% Confidence Interval .74–.84

Oily fish

189

.70

.62–.76

Lean fish

189

.68

.59–.75

Cakes and pastries

189

.67

.58–.74

French fries

189

.67

.58–.74

Sausage

189

.48

.37–.59

Vegetables

189

.43

.30–.54

Buns and cookies

189

.42

.29–.53

Fruit juice

189

.42

.30–.53

Ice cream

189

.42

.30–.53

Salty snacks

189

.36

.23–.48

Hamburgers

189

.35

.22–.47

Carbonated soft drinks

189

.29

.16–.42

Sweets and chocolate

189

.27

.13–.39

Sugared drinks

189

.26

.12–.39

Fruit and berries

189

.24

.13–.40

a

Mother–child pairs. Note: All correlations are P < .001. is, the adjustment resulted in a weaker association between mothers' and children's intakes of sugared drinks, for both the 3- and 5-year-olds, and for 5-year-olds regarding carbonated soft drinks. In the analyses, children's weight status was not accounted for, because only 171 children had data on height and weight. Running the multiple regressions analyses including child weight status resulted in an additional attenuation of associations for sugared drinks for 3-year-olds and carbonated soft drinks for 5-year-olds (data not shown).

Association Between Mothers’ and Children’s Food Intake: Moderation of Place of Residence The only association that was moderated by place of residence was intake of pizza (P < .001) (Table 5). For each 100 g of pizza that a parent in the Stockholm area ate, the child ate 58 g (95% CI, 39–76), whereas the association was much weaker for children outside Stockholm: 23 g (95% CI, 18–28). Mothers in both areas had approximately the same average intake (360 g), whereas children in the Stockholm area ate on average 100 g less

than did children outside Stockholm (94 g vs 207 g). Looking at the overall pattern, there was a tendency for stronger associations (not significant) between mothers' and children's intake in the Stockholm area compared with outside Stockholm (Table 5). All of the analyses were adjusted for maternal age, maternal BMI, maternal educational level, maternal national background, and children's age and gender. Adjustment for children's weight status changed the associations only marginally (data not shown). Average intakes of french fries were higher and carbonated soft drinks lower among children and mothers in the Stockholm area than for those outside Stockholm. For the rest of the foodstuffs average intakes were similar.

DISCUSSION This study showed that children's and mothers' intakes of different foods in general were related. However, the strength of the associations differed depending on the foods investigated. The foods that showed the strongest correlations were pizza, oily and lean fish, cakes and pastries, and french fries. For vegetables, fruit and berries,

and sugared drinks, the correlations were much weaker. The differences in correlation strength are difficult to interpret. One possible explanation for the high correlation between mothers' and children's intakes of pizza and french fries is that mothers act not only as role models but also as gatekeepers for these foods. That is, these foods are more likely to be eaten within a certain context, often outside the home. Shared environmental factors, such as food availability and food modeling, are suggested to be the most important factors for young children's eating habits,31 even though one cannot rule out that genetics may also have a role in children's taste preferences, for example. Parental food modeling is just one suggested mechanism by which children form their eating behaviors. Therefore, the low correlations seen between mothers' and children's intakes for some of the foods may reflect other underlying mechanisms. Other child feeding behaviors thus may be of more importance when it comes to vegetables, sugared drinks, and fruits. A study among 2- to 5-year-olds showed that besides mothers' own intake of fruits and vegetables, parental monitoring and encouragement as well as child liking were associated with children's fruit and vegetable intakes.30 Previous studies among younger children (aged < 8 years) showed that the correlation between parents' and children's intakes of fruit varied between 0.38 and 0.74.27,28,30,40 One of the previous studies with the highest correlation used prospective weighed recordings, whereas the other studies used FFQs with unclear validity. Two other studies41,42 investigated vegetable and fruit intake combined, and both reported a correlation of 0.23. In these studies parental and child intakes were measured by different methods, namely FFQs and 3-day recordings, weighed or by a 24-hour recall, respectively. There are likely several reasons other than methodology for these inconsistent results, such as differences in the distribution of cultural, socioeconomic, gender, and weight status. For instance, children's mean age in those previous studies differed, and few included 3-year-olds. However, the current study found no statistically significant difference in the association

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016 Table 4. Multiple Regression Analysis of Children’s Food Intake and Their Association With Mothers’ Food Intake (n ¼ 189 Pairs), Moderated by Age and Adjusted for Background Variables

Foods

P for Adjusted b Age Unadjusted (95% Confidence Interaction by Age Group b Interval)a

Buns and cookies

3 5

.32 .16

.32 (.23–.42) .16 (.06–.27)

.02

Cakes and pastries

3 5

.30 .45

.30 (.20–.40) .46 (.29–.62)

.11

Carbonated soft drinks

3 5

.29 .23

.29 (.12–.46) .16 (–.24 to .56)

.55

French fries

3 5

.30 .30

.30 (.18–.42) .27 (.07–.47)

.79

Fruit and berries

3 5

.16 .12

.17 (.07–.28) .14 (–.03 to .31)

.74

Fruit juice

3 5

.31 .32

.29 (.18–.47) .30 (.07–.53)

.95

Hamburgers

3 5

.21 .13

.22 (–.01 to .44) .14 (–.04 to .33)

.55

Ice cream

3 5

0.36 0.38

.38 (.15–.60) .38 (.18–.58)

.97

Lean fish

3 5

0.46 0.42

.46 (.26–.67) .43 (.28–.59)

.86

Oily fish

3 5

0.21 0.15

.20 (.13–.28) .13 (–.06 to .38)

.63

Pizza

3 5

0.28 0.36

.28 (.15–.44) .35 (.23–.46)

.51

Salty snacks

3 5

0.02 0.17

.02 (–.02 to .06) .17 (.04–.31)

.03

Sausage

3 5

0.32 0.41

.32 (.17–.47) .42 (.21–.62)

.45

Sugared drinks

3 5

0.33 0.09

Sweets and chocolate

3 5

0.18 0.44

.19 (.06–.31) .46 (–.10 to 1.01)

.35

Vegetables

3 5

0.24 0.25

.24 (.14–.34) .25 (.12–.34)

.87

.25 (.05–.45) –.01 (–.22 to .20)

.07

a

Adjusted for maternal age, maternal body mass index, maternal education, place of residence, national background, and child gender. between mothers' and children's intake of fruits and vegetables in 3-year-olds compared with 5-year-olds. Nevertheless, children's age moderated the association between mothers' and children's intake of salty snacks and buns and cookies. The association between mothers' and children's salty snack intake was stronger for 5-year-olds than for 3-year-olds. This may indicate that mothers of younger children restrict the intake of these food products. For instance, parents are often warned about giving peanuts

to younger children because of the risk of allergy and suffocation. The association between mothers' and children's intake of sugared drinks seemed to be moderated by children's age, although it was not statistically significant (P ¼ .07). The 5-year-olds had a higher intake of sugared drinks than did the 3-year-olds in the current sample, and their intake had no relationship to their mothers' intake. There is only 1 previous study among this age group that assessed sugared drinks; however, it excluded

Hansson et al 7 soft drinks.29 Children's sweetened beverage intake was significantly correlated (r ¼ .21) with mothers' intake when the child was aged 2 years, but not at 3 years (r ¼ .17). A correlation of 0.11 was documented in a study investigating soft drink intake between mothers and children aged 5–8 years40 whereas correlations of 0.25–0.52 have been documented for soft drink intake between mothers and children aged 2.5–7 years.27,29,30 This may indicate that as children grow older, parents may have more difficulties in controlling their children's intake of sugared drinks. That is, to act as a role model is not enough to influence children's intake. It is interesting that place of residence, in this case living inside or outside Stockholm, seemed to influence associations of mothers' and children's pizza intake. Mothers' mean pizza intake did not differ between the 2 areas, whereas children's mean intake was higher outside Stockholm. This may reflect that some children outside Stockholm are given pizza as an alternative for families' normally served meals. The underlying reason for this could be that the high availability of fast-food restaurants or convenience stores in more rural areas influences parents' choice of children's meals. Strengths of this study, compared with previous studies, are the relatively large sample size and the prospective recording of food intake. Another strength is the use of the same data collection method for mothers and children. However, recording of food intake through food diaries is subject to bias, as are all other methods of dietary data collection. There is a risk of altering or misreporting food intake, and attenuation in reporting variety and amount, especially when the data collection period takes place over several days. Some of the misreporting in the current study regards omission of the amount of a certain food item eaten. When the amount was missing, a standard portion was assumed. However, this approach was used in few cases and for few food items and meals, and it was estimated that this missing information was random and had only a minor impact on the results. Furthermore, foods eaten less frequently, such as fast food, may have been better captured by using

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016

8 Hansson et al

Table 5. Multiple Regression Analysis of Children’s Food Intake and Their Association With Mothers’ Food Intake (n ¼ 189 Pairs), Moderated by Place of Residence and Adjusted for Background Variables

Foods

Place of Unadjusted Residence b

P for Adjusted b (95% Interaction Confidence by Age Interval)a

Buns and cookies

S O

.21 .27

.24 (.08–.40) .26 (.18–.34)

.82

Cakes and pastries

S O

.30 .36

.30 (.19–.41) .36 (.23–.49)

.45

Carbonated soft drinks

S O

.26 .21

.24 (–.13 to .61) .21 (.04–.38)

.89

French fries

S O

.36 .28

.35 (.16–.54) .28 (.16–.39)

.49

Fruit and berries

S O

.12 .16

.15 (–.05 to .36) .17 (.06–.26)

.91

Fruit juice

S O

.44 .27

.44 (.11–.77) .25 (.14–.37)

.30

Hamburgers

S O

.15 .20

.18 (.3–.34) .20 (–.00 to .39)

.91

Ice cream

S O

.47 .35

.50 (.21–.79) .36 (.20–.52)

.39

Lean fish

S O

.54 .42

.54 (.31–.77) .43 (.27–.59)

.41

Oily fish

S O

.26 .17

.24 (.12–.36) .17 (.05–.29)

.41

Pizza

S O

.57 .24

.58 (.39–.76) .23 (.18–.28)

< .001

Salty snacks

S O

.13 .11

.12 (.02–.22) .11 (–.01 to .22)

.89

Sausage

S O

.44 .34

.47 (.27–.68) .34 (.19–.49)

.29

Sugared drinks

S O

.12 .23

.04 (–.16 to .24) .15 (–.06 to .36)

.45

Sweets and chocolate

S O

.44 .25

.41 (.20–.62) .28 (–.07 to .62)

.55

Vegetables

S O

.27 .24

.27 (.15–.39) .22 (.13–.32)

.54

O indicates outside Stockholm; S, Stockholm. a Adjusted for maternal age, maternal body mass index, maternal education, national background, child age and gender. an FFQ, which would estimate usual intake and the probability of consumption. Children's total food intake was not collected, only the intake when parents were able to observe the child. The children's infrequent intake, the omitted meals, and the assumptions that were made in the current study might have introduced bias into the results, leading to either a weakening or strengthening of true associations between mothers' and children's food intake. One reason for

the low correlation between mothers' and children's intake of fruit and berries observed in this study could be that the majority of children's fruit and berry intake occurred at their day care centers, a period not captured in the study. The data in the current study are cross-sectional, and therefore causality cannot be claimed. Longitudinal data are needed, preferably studies with long-term outcomes. The proportion of families that were willing to partici-

pate in the current study was low (12%). The families that participated differed somewhat from the population from which they were drawn: that is, the mothers were more likely to be born in Sweden and were older. The findings may therefore not be generalizable to the target population. Among the families that agreed to participate, quite a few returned the first (69%) or both recordings (56 %). However, because of missing data and data processing, only 37% of families could be evaluated. As in previous studies among young children, mothers as the sole information source for children's food intake was used. Mothers are likely to be the strongest influential source because they spend more time with their children compared with fathers during the preschool years. However, increasing numbers of fathers are taking parental leave in Sweden, and therefore, it would have been interesting to study the association between children's intakes with those of both parents. A recent study also highlighted the general scarcity of studies investigating similarities in father–child food intakes.43 Despite encouraging both parents to record their food intake, only 1 family sent in recordings for both parents.

IMPLICATIONS FOR RESEARCH AND PRACTICE The current study showed that parents' and preschool children's intakes of a selection of healthy and unhealthy foods were associated. However, the strength of the associations varied depending on food type. This may indicate that parental modeling may be more effective for certain foods such as pizza, fish, cakes and pastries, and french fries. For other foods, eg, fruits and berries, sugared drinks, and sweets and chocolates, other parental strategies may be of greater importance, such as making certain foods more or less available. Future studies, preferably qualitative, should explore how parents implement different food strategies for specific food types. Musher-Eizenman and Keifner acknowledged the challenges of capturing parenting practices related to eating, and they suggested that

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016 studies include parenting style, child feeding behavior, and specific feeding behaviors for specific foods at the same time to be able to deepen an understanding of this complex area.6 Studies investigating longitudinal changes in food habits among young children are also needed. Finally, the complexity of food parenting suggests that different strategies may be combined in intervention programs designed to promote change in food intakes in families with preschool children. However, modeling of positive behaviors clearly has a role within such programs.

ACKNOWLEDGMENTS Finn Rasmussen is the principal investigator of the Primrose Project of which the current study was a part. It was funded by the Swedish Council for Working Life and Social Research (Nos. 2006-0226 and 2011-0413), the Swedish Research Council (Nos. K2006-27X-20069-01-3 and K201269X-22058-01-3), the Research and Development Committee, Stockholm County Council (No. 2006-0324), the Regional Research Council of the € Uppsala and Orebro Health Care Region (No. RFR-12404), Uppsala County € dermanland County CounCouncil, So cil, the Public Health Committee of Stockholm County Council (No. 0803-377), the V
ardal Foundation (No. B2007-006), AFA Insurance (No. H-06:05/070001), the Foundation of the Swedish Diabetes Society (No. TMA2006-004), and the Karolinska Health Care Sciences Postgraduate School (2008). The trial also received faculty funds from the Karolinska Institutet for PhD student grants (2012 and 2013). The authors are grateful to Marit Eriksson and Sanna Wallin, who contributed substantially to the execution of the study.

REFERENCES 1. Doring N, Hansson LM, Andersson ES, et al. Primary prevention of childhood obesity through counselling sessions at Swedish child health centres: design, methods and baseline sample characteristics of the PRIMROSE clusterrandomised trial. BMC Public Health. 2014;14:335.

2. Lakshman R, Mazarello Paes V, Hesketh K, et al. Protocol for systematic reviews of determinants/correlates of obesity-related dietary and physical activity behaviors in young children (preschool 0 to 6 years): evidence mapping and syntheses. Syst Rev. 2013;2:28. 3. Novilla ML, Barnes MD, De La Cruz NG, Williams PN, Rogers J. Public health perspectives on the family: an ecological approach to promoting health in the family and community. Fam Commun Health. 2006;29:28-42. 4. Ventura AK, Birch LL. Does parenting affect children’s eating and weight status? Int J Behav Nutr Phys Act. 2008;5:15. 5. Faith MS, Scanlon KS, Birch LL, Francis LA, Sherry B. Parent-child feeding strategies and their relationships to child eating and weight status. Obes Res. 2004;12:1711-1722. 6. Musher-Eizenman DR, Kiefner A. Food parenting: a selective review of current measurement and an empirical examination to inform future measurement. Child Obes. 2013;9(suppl):S32-S39. 7. Collins C, Duncanson K, Burrows T. A systematic review investigating associations between parenting style and child feeding behaviours. J Hum Nutr Diet. 2014;27:557-568. 8. Blissett J. Relationships between parenting style, feeding style and feeding practices and fruit and vegetable consumption in early childhood. Appetite. 2011;57:826-831. 9. Burrows T, Morgan PJ, Lubans DR, et al. Dietary outcomes of the healthy dads healthy kids randomised controlled trial. J Pediatr Gastroenterol Nutr. 2012;55: 408-411. 10. Daniels LA, Mallan KM, Nicholson JM, Battistutta D, Magarey A. Outcomes of an early feeding practices intervention to prevent childhood obesity. Pediatrics. 2013;132:e109-e118. 11. Xu H, Wen LM, Rissel C, Flood VM, Baur LA. Parenting style and dietary behaviour of young children: Findings from the Healthy Beginnings Trial. Appetite. 2013;71:171-177. 12. Mazarello Paes V, Hesketh K, O’Malley C, et al. Determinants of sugar-sweetened beverage consumption in young children: a systematic review. Obes Rev. 2015;16:903-913. 13. Mazarello Paes V, Ong KK, Lakshman R. Factors influencing obesogenic dietary intake in young children (0-6 years): systematic review of

Hansson et al 9

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

qualitative evidence. BMJ Open. 2015; 5:e007396. Bandura A. Health promotion by social cognitive means. Health Educ Behav. 2004;31:143-164. Bandura A. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice-Hall, Inc; 1986. McClain AD, Chappuis C, NguyenRodriguez ST, Yaroch AL, SpruijtMetz D. Psychosocial correlates of eating behavior in children and adolescents: a review. Int J Behav Nutr Phys Act. 2009;6:54. van der Horst K, Oenema A, Ferreira I, et al. A systematic review of environmental correlates of obesity-related dietary behaviors in youth. Health Educ Res. 2007;22:203-226. Krolner R, Rasmussen M, Brug J, Klepp KI, Wind M, Due P. Determinants of fruit and vegetable consumption among children and adolescents: a review of the literature. Part II: qualitative studies. Int J Behav Nutr Phys Act. 2011;8:112. Rasmussen M, Krolner R, Klepp KI, et al. Determinants of fruit and vegetable consumption among children and adolescents: a review of the literature. Part I: quantitative studies. Int J Behav Nutr Phys Act. 2006;3:22. Blanchette L, Brug J. Determinants of fruit and vegetable consumption among 6-12-year-old children and effective interventions to increase consumption. J Hum Nutr Diet. 2005;18:431-443. Pearson N, Biddle SJ, Gorely T. Family correlates of fruit and vegetable consumption in children and adolescents: a systematic review. Public Health Nutr. 2009;12:267-283. Wang Y, Beydoun MA, Li J, Liu Y, Moreno LA. Do children and their parents eat a similar diet? Resemblance in child and parental dietary intake: systematic review and meta-analysis. J Epidemiol Community Health. 2011;65:177-189. Fletcher A, Bonell C, Sorhaindo A. You are what your friends eat: systematic review of social network analyses of young people’s eating behaviours and bodyweight. J Epidemiol Community Health. 2011;65:548-555. Mikkelsen MV, Husby S, Skov LR, Perez-Cueto FJ. A systematic review of types of healthy eating interventions in preschools. Nutr J. 2014;13:56. Papas MA, Hurley KM, Quigg AM, Oberlander SE, Black MM. Low-income, African American adolescent mothers and

10 Hansson et al

26.

27.

28.

29.

30.

31.

their toddlers exhibit similar dietary variety patterns. J Nutr Educ Behav. 2009;41: 87-94. Sonneville KR, Rifas-Shiman SL, Kleinman KP, et al. Associations of obesogenic behaviors in mothers and obese children participating in a randomized trial. Obesity (Silver Spring). 2012;20:1449-1454. Vereecken CA, Keukelier E, Maes L. Influence of mother’s educational level on food parenting practices and food habits of young children. Appetite. 2004;43:93-103. Cooke LJ, Wardle J, Gibson EL, Sapochnik M, Sheiham A, Lawson M. Demographic, familial and trait predictors of fruit and vegetable consumption by pre-school children. Public Health Nutr. 2004;7:295-302. Hoerr SL, Lee SY, Schiffman RF, Horodynski MO, McKelvey L. Beverage consumption of mother-toddler dyads in families with limited incomes. J Pediatr Nurs. 2006;21:403-411. McGowan L, Croker H, Wardle J, Cooke LJ. Environmental and individual determinants of core and non-core food and drink intake in preschoolaged children in the United Kingdom. Eur J Clin Nutr. 2012;66:322-328. Larsen JK, Hermans RC, Sleddens EF, Engels RC, Fisher JO, Kremers SP. How parental dietary behavior and

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016

32.

33.

34.

35.

36.

37.

food parenting practices affect children’s dietary behavior: interacting sources of influence? Appetite. 2015;89:246-257. Gustafson A, Hankins S, Jilcott S. Measures of the consumer food store environment: a systematic review of the evidence 2000-2011. J Community Health. 2012;37:897-911. Liese AD, Weis KE, Pluto D, et al. Food store types, availability, and cost of foods in a rural environment. J Am Diet Assoc. 2007;107:1916-1923. Neovius M, Rasmussen F. Place of residence and obesity in 1,578,694 young Swedish men between 1969 and 2005. Obesity (Silver Spring). 2008;16:671-676. de Munter J, Friedl A, Lind S, et al. (2016) Stability in the prevalence of Swedish children who were overweight or obese in 2003 and 2011 [published online ahead of print February 2, 2016]. Acta Paediatr. http://dx.doi.org/ 10.1111/apa.13351. Sepp H, Ekelund U, Becker W. Enk€atfr
agor om kost och fysisk aktivitet bland vuxna—underlag till urval av fr
agor i befolkningsinriktade enk€ater. Uppsala, Sweden: Livsmedelsverket (National Food Agency); 2004. [In Swedish]. World Health Organization. Body mass index. http://www.euro.who.int/en/ health-topics/disease-prevention/nutrition /a-healthy-lifestyle/body-mass-index-bmi. Accessed March 24, 2016.

38. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ. 2000;320:1240-1243. 39. Skolverket. F€or€aldrars val och inst€allning till f€orskola och fritidshem. Resultat fr
an f€or€aldraunders€okningen 2012. Stockholm, Sweden: Skolverket. Rapport 392; 2013. [In Swedish]. 40. Longbottom PJ, Wrieden WL, Pine CM. Is there a relationship between the food intakes of Scottish 5(1/2)-8(1/2)-yearolds and those of their mothers? J Hum Nutr Diet. 2002;15:271-279. 41. Thorsdottir I, Gunnarsdottir S, Ingolfsdottir E, et al. Fruit and vegetable intake: vitamin C and B-carotene intake and serum concentrations in sixyear-old children and their parents. Scand J Food Nutr. 2006;50:71-76. 42. Fisher JO, Mitchell DC, SmiciklasWright H, et al. Parental influences on young girls’ fruit and vegetable, micronutrient, and fat intakes. J Am Diet Assoc. 2002;102:58-64. 43. Robinson LN, Rollo ME, Watson J, Burrows TL, Collins CE. Relationships between dietary intakes of children and their parents: a crosssectional, secondary analysis of families participating in the Family Diet Quality Study. J Hum Nutr Diet. 2014;8:443-451.

Journal of Nutrition Education and Behavior  Volume -, Number -, 2016

CONFLICT OF INTEREST The authors have not stated any conflicts of interest.

Hansson et al 10.e1