Eating behavior of preschool-age children during social distancing period due to COVID-19

The aim of the study was to investigate the eating behavior of preschool children during the social isolation period of the first wave of COVID-19 in Fortaleza, Ceará. This is a cross-sectional observational study, carried out with caregivers of preschool children aged from 2 to 6 years. The PENSE questionnaire (IBGE) was used to collect socio-economic data and the data referring to eating behavior was used the CEBQ. 286 parents (mother and father) and their children participated in the research, of which 53.8% were female. The preschoolers in the study constituted 5.2% of infant I, 9.8% infant II, 14.7% infant III, 20.6% infant IV, 23.1% infant V, 15.4% 1-year elementary school and 11.2% did not attend school. As for the education of most mothers, 83.9% completed higher education. Regarding the number of people who lived in the household, 44.4% lived with 4 people and 73.1% had 1 child. The highest mean values ​​are present in the dimension "Pleasure to eat" (EF) and "Response to satiety" (SR), while the lowest refer to the dimensions "About Emotional Intake" (EOE) and "Response to food" (FR). The statistical results of the “Food Fussiness” (FF) subscale point out important aspects about the selectivity in the sample of the present study, demonstrating a good tendency to picky eating. The dimension with the lowest mean was “Emotional Overeating”, demonstrating that when confined due to the COVID-19 pandemic, preschoolers still have EOE lower than all other subscales. It follows that children in preschool stage have presented a higher prevalence in subscales that are linked to food avoidance behavior.


Introduction
The disease caused by the new Coronavirus -2019, popularly known as COVID-19, which was firstly mentioned in Wuhan, China, in December 2019, and quickly spread around the world, has resulted in a global health crisis. Children represent one to five percent of the diagnosed cases of COVID-19, however some children who may be infected, may be asymptomatic and, therefore, not diagnosed. (Ludvigsson, 2020) In this scenario of COVID-19 pandemic, public health recommendations and government measures have enforced lockdown, restricting movement and physical contact between people and creating social isolation. Although restrictions have helped to reduce infection rates, these limitations have also caused negative effects, since they have restricted daily physical exercises and leisure activities. (Hossain;Sultana;Purohit, 2020). This new reality of social and physical distancing has caused some impacts due to the fact that children had to stop going to school and, at the same time, some parents lost their jobs or started working from home. The economic consequences of the COVID-19 pandemic as well as the control measures put in place have led to increased levels of psychological distress and uncertainties which may affect parents' behaviors, leading to parental behaviors that are more authoritarian, strict, and not affective, ultimately impacting the quality of family life. (Prime; Wade; Browne, 2020) Children's eating behavior must be seen as a systemic mode, because it is related to parental, cultural and socioeconomic environment. Behaviors and familiar decisions that are authoritarian and uncontrolled may trigger eating disorders, causing obesity or malnutrition resulting from impulsiveness, responsiveness, and autonomy. Therefore, environmental changes that modify lifestyle habits may be positive or negative, so, it is necessary to pay attention to the relationship established with kids since their first years of life, mainly after they start preschool, as it is when children become more active (Scaglioni et al., 2018;Taylor & Emmet, 2018) Changes in the family environment resulting from the lockdown have resulted in disruptions to children's daily routines and activities. Emerging research indicates that changes in the family environment in the context of the COVID-19 pandemic has adversely impacted the eating and physical activity behaviors of children. Changes in children's eating and dietary patterns, and physical activity and sedentary levels and behaviors have resulted in increased increased risk of excess weight gain (Carrol et al., 2020;Sidor & Rzymski, 2020). For example, a study by LIPPI and colleagues found that being in closed spaces resulted in children's increased screentime (e.g., online games use of social media), which may adversely increase risk of psychological disorders, musculoskeletal disease, and neurological complications. (Lippi et al., 2020).
Previous studies have reported that young individuals have difficulty with weight-control lifestyle programs while at home compared to when they are in their usual school curriculum (Rudle et al., 2020). Furthermore, recent observations show that during the lockdown, many households consumed more ultra-processed and calorie-dense foods than usual (Creswell,2021).
Parents play a central role in creating and maintaining a family environment that impacts their children's eating and physical activity behaviors. Parents create family eating environments that shape their children's access and experiences with foods and eating (Lindsay et al., 2006). The social isolation and economic uncertainties brought by the COVID-19 pandemic may negatively impact parents' mental health and parenting practices, ultimately affecting the family eating environment and children's daily routines and eating behaviors. Therefore, the aim of this study was to assess eating behaviors duringthe social distancing and lockdown resulting from the first wave of the COVID-19 pandemic among preschool-age children in Fortaleza, Ceará.

Study Design
This is a cross-sectional study conducted from August to September in 2020 in Fortaleza, the capital of the state of Ceará, Northeast Brazil. In 2020, Fortaleza had approximately 2.669.342 inhabitants with 274.142 children from 0 to 5 years old, which represented 10,27% of the population (IBGE, 2020).

Study Sample
A convenience sample of parents of preschool-age children (2 to 6 years) were recruited to participate in the study.
Sample size was calculated based on sample calculator. In order to define the sample, it was used a calculation/ equation that was suitable to develop the calculation of the finite population size (Hulley et al., 2014). During the sample calculation, the following equation parameters were used to estimate the sample: study confidence coefficient of 99% (Z=2,575); sampling error of 3% (e=0,03); and children population size of 274.142 (IBGE, 2019). In terms of the event prevalence, it was regarded the value of 50% (p=0,5) as an average prevalence to different indicators, because of the lack of knowledge about the event proportion. The equation has shown that it is necessary to have a sample of 229 participants considering the parameters established.

Data Collection
Parents were recruited to participate in this study using the "snowball" model when a family indicated another family that had children in school.Data was collected using an on-line questionnaire in Google forms and sent to students' parents (public and private school) in Whatsapp groups.
The questions related to socioeconomic information were based on National School Health Survey (PeNSE), which was applied in 2015 by IBGE. The parents of the children were asked about their level of education, the number of children living in the house who were at school, the number of bathrooms in the house, property holding and if domestic service was provided in their houses. To evaluate the eating behavior, it was used Children's Eating Behavior Questionnaire (CEBQ), which was created considering the current knowledge about nutritional causes of obesity, focusing on behavioral determinants. This questionnaire investigates CEBQ, eating behavior and it can also be used to plan behavioral strategies to interfere in obesity.
The CEBQ questionnaire, validated to Portuguese (Viana & Sinde, 2008) The answers were evaluated in scale like Likert, which has five points that vary from 1 (not often) to 5 (very frequent). The items 3, 4, 10, 16 and 32 are in a reverse scale. The scores of the questions that belong to the same dimension are added up. In each dimension analyzed will have an average value and standard deviation (Viana& Sinde, 2008).
The questions were gathered according to the analyzed dimension: question 12,14,19,28 and 34 were about Food Responsiveness; 1, 3, 20 and 22 were related to enjoyment of food; Satiety Responsiveness -3, 17, 21, 26 and 30; Slowness in eating -4, 8, 18 and 35; Food Fussiness -7, 10, 16, 24, 32 and 33; Emotional over-eating -2, 13, 15 and 27; Emotional undereating -9, 11, 23 and 25; and Desire for drinks -6, 29 and 31. Besides the subscales that were described, it was considered two groups that are associated to the subscales. The first group is about the behavior that shows attraction for food, it includes the following subscales: EF, RF, EOE and DD. The second group is related to the avoidance of food, which is consisting by SR, SE, FF and EUE. The result of these dimensions showed the way someone interacts with food, thus, this reflects a pattern or eating style that influences nutritional states.
In Google forms questionnaire all the questions were considered mandatory, thus the answers could only be submitted if the participant had answered all of them. The project was approved by the ethics committee of the Federal University of Ceara No. 36223020.3.0000.5054. All participants signed the Informed Consent Form, sent via the link to the questionnaire applied by Google Forms.

Data Analysis (Missing information)
Data entered in an Excel sheet and analyzed in Stata SE 14.0. The descriptive statistics was used to define the proportion of responses to each question and the total of scores that was distributed in each questionnaire.

Results and Discussion
286 parents and school-age children were interviewed in Fortaleza, Northeast Brazil. 53,8% of them were female. 83,9% had the higher education level of education and 3,5% did not have a school diploma. Concerning the number of people who lives in the house, 44,4% of them live with 4 people; 23,3% live with 3 people and 18,2% live with 5 people. The number of children each mother has was about 1 (73,1%) and 2 (25,5%).
The children who participated in this study were separated in the following level of schooling: 15 (5,2%) preschool education (level 1); 28 (9,8%) preschool education (level 2); 42 (14,7%) preschool education (level 3); 59 (20,6%) preschool Research, Society and Development, v. 10, n. 17, e07101722435, 2021 (CC BY 4. Source: Prepared by the authors. The highest average scores are in the dimensions "Enjoyment of food" (EF) and "Satiety responsiveness" (SR). On the other hand, the lowest average scores are in the dimension "Emotional over-eating" (EOE) and "Food Responsiveness" (FR).  In accordance with Pearson correlation analysis, it is possible to determine that there is no meaningful difference between the studied eating behavior's dimensions and children's gender.
When the group of scales that constitutes CEBQ was investigated separately, the authors have decided to separate the dimensions in different groups. There is the group that evaluates "Food attraction", which includes the dimension EF (Enjoyment of food), FR (Food Responsiveness), EOE (Emotional over-eating) and DD (Desire for drinks). The other group studies the behavior related to "food avoidance", and it is constitutes by the following dimensions: SR (Satiety responsiveness), SE (Slowness in eating), FF (Food Fussiness) and EUE (Emotional under-eating). The data of each group are presented in Table 3. According to the datas analysis, it was observed that preschoolers have a tendency to avoid food, since they present high levels of satiety responsiveness, which may also highlights the presence of food selectivity.
Preschoolers lifestyle is strongly influenced by parents, however, these interferences occur not only because of family, they are also associated to children's interaction with media, and the information it communicates. Nevertheless, the majority of participation, control and food intervention come from parents' imposition. Thus, CEBQ is a tool that enables to describe children's eating behavior from parents perspective, identifying the factors that cause more impact in their subscales (Wardle et al., 2001;Vitorino, 2018).
This study aimed to verify which CEBQ's dimensions have more influence in children who are in distancing isolation, considering that an unusual situation. This context may affect not only kids' nutrition, but it can also be linked to other kinds of negative conditions, such as disease susceptibility, especially in the pandemic scenario.
The restrictive environment created by COVID-19 has aggravated parents and kids' mental health and well-being.
Behavioral and mood changes tend to be negative, because they are caused by psychological disorders that are originated from social distancing. (Courtney et al., 2020;Gassman_pines et al., 2020). Blisset et al., (2019) evaluated that "Enjoyment of food" is higher when children has a neutral mood or when they are induced to bad mood, the same happens with the population that was investigated in this research, it is also relevant to emphasized that this dimension had the highest average in the sample.
Besides that, Blisset et al., (2019) realized that a higher number of calorie intake is related to EF subscale in all analyzed categories, according to the food that was offered in their research. This relation is also found in the study of Carnell et al. (2016) with preschoolers, showing a strong correlation.
"Satiety responsiveness" was the second highest average in the sample. However, according to Blisset et al., (2019), this result is not associated to bad mood, but it is related to Emotional under-eating. Carnell et al. (2016), in their study, observed that there is a high level of SR linked to low food intake, when it was served five different kinds of meals to preschoolers. In addition, there is a huge number of studies that shows a negative relation between SR and overweight and body measure, which supports the ideia that a decrease in this scale avoids the control of food intake and this fact contributes to gain weight. (Passos et al., 2015;Derks et al., 2018;Dalrymple, 2019;Spahic, 2019) Fussy eating is connected to a lot of experiences and emotions that kids experience, however, factor genic also interferes in eating decisions (Wolstenholme et al., 2020). Cole et al., (2017), through a systematic review, discovered an increase of 0,06 units in Food Fussiness (FF) in each month of life during one year in 30 month-old children, their study did not associate this result to gender and weight. Besides that, there is an inverted correlation between FF and the quantity of food intake, being vegetables, in particular, the food category more undesirable according to FF (the higher the FF is, the more undesirable this food category is (Cole et al., 2017;Fernandez el at., 2018;Sandvik et al., 2019). The statistic results of FF subscale indicate important aspects about fussiness in this study, because it had a high average and it also had the lowest score variation. This fact demonstrates that there is a tendency to fussy eating in the group analyzed.
Emotional overeating dimension presented the lowest average in a several studies with preschoolers, and it had no relation to their nutritional state. When these children are in social isolation, due to COVID-19 pandemic, the EOE is lower than the other scales. To compensate this, EF and SR are the highest subscales (Konings et al., 2018;Blissett et al., 2019;Sandvik et al., 2019;Kininmonth et al., 2020;Joseph Louis, & Tyug, 2021). This reveals that, at this age, children did not create a lot of emotional bonds with food to overcompensate, on the other hand, they demonstrate a great pleasure in food intake. In the same way, children have already presented good signs of Satiety responsiveness or food rejection, since they do not understand their needs, which is why they avoid some kinds of meals.
It was not possible to meet and monitor the participants face-to face during pandemic period, thus, the data related to weight and height were not obtained. Nonetheless, Pedrosa (2017) has shown, through a big sample, that the decrease in malnutrition and overweight/ obesity in children are connected to mother's schooling level, so the higher the mother's school level is, the higher children's BMI will be. The high percentage of mothers with university degree (83,9%), in this study, may be an indication that this research has a huge number of children with BMI above eutrophic intervals, considering the restrictive population. According to IBGE (2019), more than half of the adult population in Brazil do not complete high school, and this number is higher in northeast region.
Many children suffer pressure to eat some kinds of food that are considered healthy (fruit, dairy products, vegetables) by their parents. A consequence of this pressure is the tendency to create negative behaviors and the decrease of children's ponderal state. Besides that, the restriction of some food or group of food might cause gain weight, in other words, it has an adverse effect (Coelho et al., 2017).
Contrary to this hypothesis, it was observed that the group "Food avoidance" dimension exceeded "Food attraction".
In addition, the subscale EOE was the best, which was not what had been expected, since many studies associate this subscale to overweigh and obesity. According to Spahić and Pranjić (2018) and Passos et al. (2015), children with higher scales in the group "Food Attraction" have higher Body Mass Index. Konings et al. (2018)

Conclusion
Children in preschool stage have presented a higher prevalence in subscales that are linked to food avoidance behavior (Satiety responsiveness -SR, Slowness in eating -SE, Fussiness -FF and Emotional under-eating -EUE.
These children also had a tendency to avoid food, showing high levels in Satiety responsiveness, which may demonstrate that there is food selectivity.
This study has some limitations. This study was carried out in the period right after the 1st lockdown decreed by the state government, due to the high prevalence of covid-19 cases. The study was conducted remotely/online and does not represent children from Fortaleza. The fact that the questionnaire was answered by women with a high level of education presents a bias, since women living in neighborhoods with a lower HDI were not questioned.
Researches about nutritional state, especially the ones associated to deficit in children's growth, might be stimulated, due to the fact that a long time of malnutrition may influence children's development, especially at this time of pandemic by covid-19.
Eating behavior is considered important in education and good eating habits help to prevent health problems. It is also important to consider that eating behavior has a multifactorial influence, since it has a relation of dependence with parental, cultural and socioeconomic environment.