INTRODUCTION

The preschool stage is characterized by rapid growth and development. Early childhood is characterized by rapid growth and development including the establishment of eating habits, food preferences, and activity behaviors that may influence children's health in later childhood and on into adulthood. Hence it is a critical period for teaching and modeling healthy habits to positively influence children's health and wellbeing throughout their lifetime (¹).

Children's eating and activity behaviors are influenced by intrinsic (genetics, age, gender) and environmental factors (fa­mily, peers, community, and society) (²). The family environment is one of the most important determinants for the development of healthy eating behaviors in childhood (³). Socioeconomic status and educational level are the main factors associated with parental knowledge about healthy behaviors, which are crucial in the development and consolidation of healthy attitudes and habits (⁴,⁵). However, few studies have assessed parental knowledge about the Spanish Society for Community Nutrition (SENC) dietary guidelines, which include recommendations on portion sizes and consumption frequencies of both healthy and unhealthy foods groups using a food pyramid as an icon (⁶,⁷).

In this regard, it is known that children whose mothers have a high educational level consume more fruit and vegetables, and are more likely to have daily breakfast than those children whose mothers have a low educational level (³).

Parents play an essential role as their offspring's health promoters and are responsible for shaping healthy eating behaviors in their children (⁴). Thus, children's eating habits and food consumption are likely to be influenced by their parent's nutritional knowledge and dietary attitudes (⁸).

Attitudes are defined as emotional, motivational, perceptive, and cognitive beliefs that influence in a positive or negative way the behavior of an individual (⁹). Attitudes may explain why an indivi­dual adopts one practice and influences future behaviors regardless of that individual's knowledge. Measuring dietary attitudes is challenging, and very few of the questionnaires or instruments used nowadays have been validated. Therefore, the evidence available in this regard is difficult to extrapolate or re­plicate (¹⁰,¹¹).

Dietary intake is also difficult to measure, especially in children. The Food Frequency Questionnaire (FFQ) is the most efficient instrument to examine dietary intake, and is one of the most widely used dietary assessment tools in epidemiological studies. FFQs are useful to assess diet quality by calculating food intake and defining dietary patterns that can be compared with dietary guidelines (¹⁰,¹²).

Evidence of an association of parental attitudes towards their offspring's diet with children's actual eating habits is scarce. The main aim of this study was to determine if parental attitudes towards their offspring's dietary habits were associated with children's actual dietary habits in a sample of Spanish preschoolers included in the Follow-up of Children for Optimal Development (SENDO) project. Additionally, we investigated the association between parental attitudes towards their offspring's diet and children's breakfast habits.

METHODS

EXPOSURE ASSESSMENT

A baseline questionnaire was not developed specifically for this study; baseline information of participants was taken from the SENDO project, and included medical history, anthropome­tric measures, eating habits, dietary intake, and physical activity.

Body mass index (BMI) was calculated as weight (in kilograms) divided by squared height (in meters) to classify participants according to their nutritional status using the International Obesity Task Force (IOTF) standard of reference (¹⁴), which includes sex- and age-specific BMI cut-off points for normal weight, overweight, and obesity.

Physical activity was collected with a questionnaire that inclu­ded 14 activities, including sports, and 9 categories of response from “never” to “more than 11 hours/week”. The METs-h/week for each activity were calculated by multiplying the number of Metabolic Equivalent of Task (MET) of each activity by weekly participation, weighted according to the months dedicated to that activity. Total physical activity was quantified by summing the METs-h/week dedicated to all activities performed during leisure time.

The baseline questionnaire included 16 questions to assess parental knowledge (based on the SENC children's dietary guidelines) (⁷) (Table I). These questions asked about consumption frequencies by food groups in children. We assigned 1 point if consumption frequencies were according to the SENC children's dietary guidelines published in 2018, and 0 points if otherwise. This parental knowledge score ranged from 0 to 16 points.

Table I.  Criteria used to calculate the parental nutrition knowledge score (based on the children's dietary guidelines by the SENC) (⁷) 

The baseline questionnaire of the SENDO project included also 8 closed (yes/no) questions about certain parental attitudes towards their children's eating habits (Table II). With them, we developed an index to capture the aggregated information from the 8 questions: we assigned 1 point if the attitude was presumed to be healthy and 0 if it was presumed to be unhealthy. Thus, participants received 1 point for each “Yes” answer if they made sure their children ate more fruit, more vegetables, more fiber, more fish, less fat, less meat, and less sweets and pastries, and if they tried to avoid the consumption of butter, and 0 otherwise.

Table II.  Criteria used to calculate the parental attitudes towards their children's dietary habits score 

When these 8 items were summed, the parent attitudes score could potentially range from 0 to 8 points, with higher scores meaning healthier attitudes towards their children's diet. The final score was expressed as a percentage of the maximum possible score (8 points). Finally, parents were classified into three groups: low (< 40 %), medium (40-70 %) and high (> 70 %) index scores. The lowest category was used as the reference group.

OUTCOME ASSESSMENT

The children's food consumption was evaluated with a validated semi-quantitative parent-reported FFQ that included 140 food items and 9 possible responses that ranged from “never or almost never” to “more than 6 times per day” (¹⁵). A trained team of dietitians derived nutrient content of each food item, calculated by multiplying intake frequency by the edible portion and nutrient composition of the specified portion size. We used data from updated Spanish food composition tables (¹⁶), and from online databases (¹⁷,¹⁸) to calculate total energy and nutrient intakes.

We also collected information on children's eating habits through 19 questions including habits and behaviors at mealtimes (breakfast, lunch, and dinner). Each question had 6 possible answers from “never or almost never” to “everyday”. If the answer complied or not with dietary recommendations, it was scored as 1 or 0, respectively (Table III); thus, the final score ranged from 0 to 19 points. This children's eating habits score was not previously validated. Participants were classified according to their score as having unhealthy (≤ 14 points) or healthy (≥ 15 points) eating habits.

Table III.  Criteria used to calculate the children's dietary habits score 

For further analysis we also developed a breakfast quality index based on the questions regarding children's habits and behaviors at breakfast (5 questions). The breakfast quality index ranged from 0 to 5 points (Table IV).

Table IV.  Criteria used to calculate the breakfast quality score 

RESULTS

A total of 423 participants from the SENDO project were inclu­ded in the present analyses (Fig. 1). Children's and pa­rents’ baseline characteristics according to the children's dietary ha­bits score are presented in table V. The mean age of participants was 5.3 years (SD: 0.9), 52.3 % were boys, and 49.1 % reported having healthy eating habits. Children with healthier eating habits scores belonged to larger families and were more likely to have normal weight. No differences were observed between groups for anthropometric measures or physical activity.

Figure 1.  Flowchart of participants recruited into the SENDO project, 2015-2019. 

Table V.  Baseline characteristics of the 423 participants from the SENDO project and their parents, according to the children's score in the dietary habits index. Numbers are means (standard deviations) or percentages 

Participants’ mean food consumption and energy and macronutrient intake according to their score in the eating habits index is described in table VI. Healthy eating habits were significantly associated with higher consumption of fruits and lower consumption of processed meat and candies. Regarding macronutrients, higher carbohydrate intake and lower fat intake were observed in children with healthy eating habits scores.

Table VI.  Mean food and macronutrients intake of the 423 participants of the SENDO project according to children's score in the dietary habits index. Numbers are means (standard deviations) 

*p < 0.05; ^†p < 0.01; ^‡p < 0.001.

Regarding children's eating habits score, we found that more than 80 % of the participants reported being accompanied by an adult for lunch, having dinner at home, and taking mid-morning or mid-afternoon snacks (Fig. 2). Fried foods were often consumed both at home (41.4 %) and outside (20.6 %). Moreover, 39 % of the participants reported watching television during dinner and 20.6 % during lunch.

Figure 2.  Percentage of positive answers to the children's eating habits index. 

Compared to the lowest category, a higher score in the parental attitudes index was associated with higher odds of healthy dietary habits in their offspring (OR: 3.05; 95 % CI: 1.40-6.65) in the crude model. The association was slightly attenuated but remained significant in the fully adjusted model (Table VII). Hence, children whose parents were in the highest category of the parental attitudes index showed an almost 3-fold increase in the odds of having healthy dietary habits (p for trend = 0.004).

Table VII.  Crude and multivariable adjusted odds ratio (OR) and 95% Confidence Intervals (CI) for children's healthy-eating habits associated with the parental score in the parental attitudes towards their child's dietary habits index 

Ref.: reference; Multivariable 1: adjusted for age (continuous) and sex (male and female); Multivariable 2: additionally adjusted for BMI (continuous); Multivariable 3: additionally adjusted for energy intake (continuous) and parental university education (no studies, graduate, high school, college degree, master or doctorate); Multivariable 4: additionally adjusted for nutrition knowledge (low, medium and high). *p < 0.05.

Figure 3 shows the proportion of participants that scored po­sitively in each of the questions included in the breakfast quality index by their score in the parental attitudes index. Significant differences (p < 0.05) were observed for the following items: 1) having breakfast regularly, 2) consuming some fruit, and 3) consuming some dairy at breakfast.

Figure 3.  Proportion of positive answers to the breakfast habits index according to parental attitudes towards their children's dietary habits. *p < 0.05. 

In further analyses of the association between parental attitudes and their offspring's breakfast habits we did not find any significant trend (Table VIII).

Table VIII.  Crude and multivariable-adjusted difference and 95 % confidence interval (CI) in the breakfast quality index associated with parental score in the parental attitudes towards their children's dietary habits index 

Ref.: reference; Multivariable 1: adjusted for age (continuous) and sex (male and female); Multivariable 2: additionally adjusted for BMI (continuous); Multivariable 3: additionally adjusted for energy intake (continuous) and parental university education (no studies, graduate, high school, college degree, master or doctorate); Multivariable 4: additionally adjusted for nutrition knowledge (low, medium and high).

DISCUSSION

In this study with a sample of Spanish children from the SENDO project we found that children whose parents reported he­althier eating attitudes had higher odds of adhering to healthy eating ha­bits. These results reinforce the previously published idea whereby parental attitudes might be more important that parental nutritional knowledge alone to foster healthy eating habits in their offspring (¹⁹).

To the best of our knowledge, this is the first study in Spain, aimed at specifically investigating the influence of parental knowledge about national children's dietary guidelines and parental attitudes about their preschool children's eating ha­bits on child dietary intake. Overall, parents have an impact on what, how, how much, how often, where, and with whom their children eat, and on the eating habits that they will develop in the future (²⁰,²¹). In Spain, data from a nationally representative survey of households with children ages 0-10 years, found that 54 % of children aged 5-10 years had excess weight (²²). Several studies reported that unhealthy eating habits such as skipping breakfast, non-parti­cipating in family meals, unestablished meals schedule, avoiding shopping list, overeating, and eating unhealthy food are asso­ciated with children's overweight and obesity (²³-²⁶).

The recommended calorie intake for boys and girls aged 4-7 years with moderate physical activity is 1400-1600 kcal/day (²⁷). In the SENDO project, children with unhealthy eating habits reported higher energy intakes (2047 kcal/day), which exceed the recommendations (⁷). Similar findings have been reported in previous studies with Spanish children (²⁸,²⁹). Those results could reflect an excess of food consumption, but also a parental overestimation of their offspring's food consumption derived to the FFQ, which is commonly used in nutritional epidemiology to estimate the distribution of usual intake (⁴,³⁰-³²).

For children aged 4-8 years old, the recommended energy intake from carbohydrates, protein, and fats is 45-65 % of total energy intake, 10-35 %, and 20-35 %, respectively (²⁷). In our study, the percentage of total energy intake from carbohydrates, protein, and fats was 45 %, 18 % and 36-37 %. Our results agree with those reported by the ENALIA project (²⁹).

Suggs et al. concluded that children's food choice and eating habits may be influenced by the presence of others and by the place where they eat (³³). More specifically, they reported that eating at home and accompanied by a family member was associated with smaller amounts of food consumed and better eating habits. In this regard, we observed that the percentage of children that had lunch or dinner every day was 82.5 % and 92 %, respectively. In our study, only a 24.1 % of the participants reported having lunch at home, whereas a 93.9 % reported having dinner at home in weekdays. The percentage of Spanish children having lunch at home varies across studies. A previous study found that 67 % of children had lunch at home, whereas anther study with children from Madrid (Spain) found that 63 % of participants had lunch at school (³⁴,³⁵).

Evidence suggests that breakfast is a key meal of the daily intake of children.

Although the content depends on cultural reasons, there is quite a lot of agreement that breakfast must be varied and in sufficient quantity to provide the 20-25 % of daily total energy intake (³⁶,³⁷). The Study of Growth, Food, Physical Activity, Child Development and Obesity Surveillance in Spain (ALADINO) 2019 (³⁸) analyzed the breakfast of Spanish children aged 6 to 9 years and found that most commonly reported breakfast meals consisted of a combination of dairy and pastries/cookies (30.3 % study population), and that only 2.2 % of the Spanish children had a full and healthy breakfast, including at least three food groups. The authors of a previous study in Spanish children defined a breakfast quality index that included the consumption of dairy products, cereals and fruits. The authors found that less than 10 % of the participants were classified as having a high-quality breakfast and that fruit was missing for 70 % of the participants.

In our study we found than 90 % of the participants had breakfast regularly, but also that less than 50 % consumed a fruit at breakfast. The proportion of children having breakfast regularly was in line with a study conducted in Madrid with children aged 3 to 12 years old (³⁴), and with another study in Switzerland with children aged 4-6 years (³⁵).

Our results showed a direct trend between the children's score in the breakfast quality index and parental score in the parental attitudes towards children's dietary habits index, but changes across categories were non-significant. Van Ansem et al. (²⁰) found that families with higher socioeconomic status reported having breakfast more often than those with lower socioeconomic status. Our analyses are not adjusted for socioeconomic status, but we adjusted for parental educational level and nutritional knowledge, which may be considered as proxies of the socioeconomic status. To our knowledge, no previous studies have evaluated that association in Spanish families. Therefore, we consider that further studies with larger sample sizes are needed to elucidate whether parental attitudes towards their children's dietary habits are associated with children's breakfast habits independently of socioeconomic status.

Despite our findings, we acknowledge several limitations. First, our sample may be not fully representative of the Spanish po­pulation. However, cohort studies are not based on representative samples of the study population and as far as the variability in the sample allows a comparison of the hypotheses investigated, the representativeness is not necessary (³⁹). Second, most of the participants in the SENDO project are white and come from high-educated families. Although we controlled for parental education, we cannot totally deny the possibility of residual confoun­ding by unmeasured factors. Third, since information was reported by parents, a misclassification bias is possible (⁴⁰). Nevertheless, since it is unlikely that misreporting was associated with parental attitudes, in case of error the estimate would be biased toward the null, not affecting our main results. Some of the information reported by parents in the SENDO project has been validated (⁴¹), but neither parental attitudes index, nor children's eating habits index, nor the breakfast quality index have been validated yet. Nevertheless, a study using the parental attitudes index in the SENDO project has been previously published (¹⁹). Although this index has not been previously validated, we considered that it might be useful to identify parents more concerned about their children's diet. Besides, this score had been used in the SUN cohort to assess change (⁴²) and its association with the risk of weight gain (⁴³) or cardiovascular disease (⁴⁴). Finally, in this study we did not exclude children who had some illness or food-related conditions that may have influenced their eating attitudes and eating habits. However, we think that parent attitudes towards children's diet and children's eating habits had changed in consonance and the associations would not be affected. Lastly, studies that evaluated the influence of parental attitudes on their offspring's eating habits are very scarce, particularly in Spain, therefore comparing our results was challenging.

CONCLUSION

In this cohort there is an association between parental attitudes towards their children's diet and children's actual eating habits, supporting that nutritional education programs should focus on fostering healthy eating attitudes beyond nutritional knowledge. Moreover, our findings suggest that public health interventions aimed at promoting healthy eating habits among children should shift from an individual perspective to a family-based one. Finally, our results will serve as a basis for future research on the impact of parental attitudes on their offspring children's dietary behaviors.