key: cord-0981925-1gq0ubyq authors: Larissa Costa, Marcela; George Oliveira Costa, Maycon; Ferreira Cândido de Souza, Márcia; Góes da Silva, Danielle; Aliete dos Santos Vieira, Diva; Simões Mendes-Netto, Raquel title: Cognitive restraint, emotional eating and uncontrolled eating: Exploring factors associated with the cycle of behaviors during the COVID-19 pandemic date: 2022-03-09 journal: Food Qual Prefer DOI: 10.1016/j.foodqual.2022.104579 sha: 68fa785aa602c4c846b3590d5d41f8ba89b0f4ab doc_id: 981925 cord_uid: 1gq0ubyq The objective of this study was to analyze the factors associated with Cognitive Restraint, Emotional Eating, and Uncontrolled Eating behaviors with perceptions of life habits, body image, eating habits, and food consumption during the COVID-19 pandemic. Sociodemographic data, eating behavior, perceptions about life and eating habits, body image, and food consumption were collected. Factors associated with each of the three eating behaviors were accessed through three multiple linear regressions controlled for age, sex, BMI, education level, and monthly income. Cognitive restraint was associated with being active during the pandemic (β=5.85), attempted weight loss (β=11.89), perception of overweight (β=9.10), better eating habits in the pandemic (β=6.55), and decreased consumption of refined cereals (β=5.60) and fast foods (β=8.23). Emotional eating was associated with increased stress during the pandemic (β=7.10), worse sleep (β=4.74), body dissatisfaction (β=4.85), perception of overweight (β=7.45), attempted weight loss (β=4.19), increase in the amount of food consumed (β=9.34), increased food delivery purchase (β=5.13), increase in consumption of sweets and desserts (β=7.17) and reduction in consumption of vegetables (β=- 5.26). Uncontrolled eating was associated with working more than 8 hours/day (β=4.04), increased stress during the pandemic (β=3.65), body dissatisfaction (β=6.59), worse eating habits in the pandemic (β=5.21), and an increase in the amount of food consumed (β=8.73). Cognitive restraint behavior was associated with healthy life habits and negative body image variables. In contrast, emotional eating and uncontrolled eating were associated with negative body image and unhealthy eating habits. During the COVID-19 pandemic, one of the measures adopted to interrupt the spread of the virus was social distancing, which has become part of the new daily life of the population. However, the need for social distancing implies a radical change in lifestyle, which affects other dimensions of health, including mental health (Jiménez-Pavón et al., 2020) . Stress and anxiety are the main consequences of social distancing. They are directly related to health behaviors, such as physical inactivity, lower sleep quality, body dissatisfaction, changes in alcohol and food consumption (Malta et al., 2020; Swami et al., 2021) . From a psychosocial point of view, changes in food intake can be explained by different changes in different dimensions of eating behavior, including Cognitive Restraint (CR), Uncontrolled Eating (UE), and Emotional Eating (EE) (Natacci & Ferreira Júnior, 2011) . The CR behavior refers to the adoption of a mental position toward food to reduce energy intake to control body weight or improve body image perception. CR theory proposes that attempts to control weight and body shape lead individuals to organize their eating behavior around their fear of failure. This state of inhibition is often interspersed with losses of control that cause episodes of EE or UE (Natacci & Ferreira Júnior, 2011) . The three domains interact with each other in such a way that cognitive restraint individuals who have negative emotions or stressful situations can consume certain foods because of their emotional state or increase food intake as a result of this altered emotional state (Apfeldorfer & Zermati, 2001) . During the pandemic, there was an increase in the consumption of sweets, snacks, and ultra-processed foods, rich in sugar and saturated fat (Malta et al., 2020) . However, results also indicate an increase in the consumption of vegetables, fruits, and beans (Sánchez-Sánchez et al., 2020) . Thus, it is also essential to know what eating behaviors are related to these different food consumption patterns and their associated factors during However, most of the studies during the pandemic have focused on the EE dimension and associated factors (Al-Musharaf, 2020; Coulthard et al., 2021) . Only one study proposed to analyze eating behavior from the perspective of these three dimensions together as an outcome, demonstrating that the social isolation process has led to an increase in the dimensions of EE and UE; however, it did not increase the dimension of CR when compared to the pre-COVID-19 period (Elmacioğlu et al., 2020) . Nonetheless, no study has identified the factors associated with these behaviors together as an outcome and from the perspective of the interaction of CR, EE, and UE behaviors. Considering that the pandemic has the potential to change eating behaviors, it is hypothesized that CR will be associated with a healthy lifestyle and eating habits. In contrast, EE and UE will be associated with unhealthy lifestyles and eating habits. Thus, the present study aimed to analyze the association between the three dimensions of eating behavior (CR, EE, and UE) and perceptions of life habits, body image, eating habits, and food consumption during the COVID-19 pandemic. A cross-sectional web-based online survey was carried out between December 2020 and January 2021. The survey link was distributed via academic/department e-mails of a public university in the Brazilian northeast and social media (Whatsapp and Instagram) using a snowball technique (Naderifar et al., 2017) . The questionnaire consisted of 73 objective questions in Portuguese, divided into six parts. The authors designed the first 52 questions and pre-tested them before the final application; the remaining 21 questions were from the previously validated eating behavior questionnaire (Natacci & Ferreira Júnior, 2011) . It required approximately 20 minutes to complete. Data from completed forms were imported into a Microsoft Excel spreadsheet. The survey obtained 643 responses, 40 of which were excluded due to duplicate responses, and five were excluded due to invalid answers. The study was performed in compliance with the Helsinki Declaration Guidelines and approved under number: 4,380,553 in the Ethical Committee of the Federal University of Sergipe. The study participants were adults (18 to 59 years) and residents in the Brazilian territory. Pregnant women were not eligible to participate. The sample size was estimated with the G power software program for multiple linear regression with 26 potential predictor variables using medium effect size, p-value <0.05, and power=0.95, which indicated that a minimum of 245 participants were required. Participants were required to respond about their age, sex, education level, marital status, and state and city of Brazilian territory to assess sociodemographic characteristics. State and city information were used to describe the sample. The states collected were categorized into Brazilian regions (northeast, southeast, or other); the cities were used to describe if the individual lived in the capital or other cities. For the purpose of analysis, the educational level was categorized into Higher education (post-graduate and graduate degree) and Lower education (high school and below). Work or study modality were also assessed: home office, working or studying out of home, or currently not working or studying, and how many hours of work or study per day: <4 hours, 4-6 hours, 6-8 hours,> 8 hours/day were also assessed. The number of hours worked or studied was categorized into: Up to 8 hours/day and more than 8 hours/day. The Body Mass Index (BMI) was calculated using self-reported weight and height. BMI was used in the analysis as a continuous variable. Participants were asked about their practice of physical activity during the pandemic (frequency and total time) and classified using the recommendation of 150 min/week (World Health Organization, 2020). Next, the participants were asked: "How do you rate your level of social isolation during the pandemic?" and the responses were categorized into: High level of social isolation (Not leaving home and Leaving 1-2 time/week) and Low level of social isolation (Leaving > 3 times/week). Another three questions were asked to assess the perception of stress level, changes in sleep, and consumption of alcoholic beverages during the pandemic. All questions were answered using a 5-point Likert scale. The answers were dichotomized for analysis purposes. In this section, three questions were made about perception on body satisfaction, perception of body weight and attempt to lose weight in the last six months. The two initials questions were: "Do you currently feel satisfied with your body?" and "How do you feel about your current weight?" The questions were answered using a 5-point likert scale. Moreover, a final question was asked: "Have you attempted to lose weight in the last 6 months?" with the option of answering "yes" or "no". Four questions evaluated participant perceptions on eating habits during the pandemic. The questions were: "In your opinion, your eating habits during the pandemic have been"; "Have you noticed a difference in the amount of food you are consuming during the pandemic?"; "Have you noticed a difference in the frequency of food preparation at home during the pandemic?" and "Have you noticed a difference in the frequency of purchasing ready-to-eat food via delivery during the pandemic?". All questions were answered using a 5point Likert scale. The answers were also dichotomized for analysis purposes. The three-factor eating questionnaire (TFEQ-R21) validated into Portuguese (Natacci & Ferreira Júnior, 2011 ) was used to assess eating behavior. The TFEQ-R21 assesses three dimensions of eating behavior through 21 questions: Cognitive Restraint (CR, six questions), Emotional Eating (EE, six questions), and Uncontrolled Eating (UE, nine questions). The scoring analysis of the questionnaire was performed according to the authors' guidelines. Item scores ranged from 1 to 4, and the subscales (CR, EE, UE) were added to calculate the individual mean of each subscale. The mean of each subscale was called the raw score. Raw scores were converted using formulas for standardization of results and comparison with other studies using this methodology (Elmacioğlu et al., 2020; Jáuregui-Lobera et al., 2014) . The final score was calculated using the following formula: Behavior score: 100*[(Raw mean of the subscale -Minimum raw score) / Maximum raw score]. The values of the final score range from 0-100, and higher scores indicate greater CR, EE, and UE. Finally, the participants answered questions about perception of changes in food consumption in the pandemic. The list of foods was adapted from the food frequency questionnaire (FFQ) used in ISACAMP-Nutri.(Universidade Estadual de Campinas, 2014). The food categories evaluated were: 1) Vegetables; 2) Fresh fruit; 3) Refined cereals; 4) Packaged snacks; 5) Ultra-processed juices; 6) Sweets and desserts; 7) Processed meats; 8) Fried foods; and 9) Fast Food. For the 9 food categories, participants were asked whether consumption (a) increased, (b) remained unchanged, or (c) decreased. The questionnaire developed by the authors underwent a pre-test before starting the data collection. The research group analyzed the initial questionnaire to help strengthen the available questions and answers. Next, a pilot questionnaire was applied to the target population to verify whether the elaborated questionnaire answered the following questions: Clarity and precision of the terms, quality of the data, and feedback. Thus, the questionnaire underwent subsequent corrections of the flaws and difficulties encountered by the participants. The invitation to respond to the pilot questionnaire was disseminated via the Google Forms link on social networks and in contact with national researchers in the area. The results were extracted to a Microsoft Excel table, and the data were interpreted and analyzed according to the research objective. A total of 340 responses were obtained, and after analyzing and discussing the questions and answers collected, there was a need to reformulate the "perception on life habits during the pandemic," "perception on body image and satisfaction," and "eating habits during the pandemic" sections to improve the order of responses, the flow of questions and standardize technical terms present in the questions. After the responses were reorganized, the questionnaire underwent a new analysis by the research group for the final application. Data were analyzed using Statistical Package for Social Sciences (SPSS) version 25.0. Continuous variables were subjected to the Kolmogorov-Smirnov test to verify the assumptions of normality. A descriptive analysis was performed using absolute numbers and percentages for categorical variables and mean, standard deviation and 95% confidence intervals for continuous variables. Three multivariate analyses by linear regression were performed using Cognitive Restraint, Emotional Eating, and Uncontrolled Eating as dependent variables. All variables which showed a p < 0.2 in the simple linear regression were included in the adjusted model. The stepwise method was used to simultaneously remove the weakest correlated variables and come up with a model that best explained the distribution to examine the relationship between several variables simultaneously and eliminate confounding factors. For the three multiple linear regressions analyses, the normality of the residuals was assessed, and the Durbin-Watson test verified the independence of the residuals. Multicollinearity was refuted using the variance inflation factor (VIF), and the significance of the models were assessed by ANOVA test. All analysis were controlled for sex, age, BMI, education level, and monthly income. A p-value of p < 0.05 was considered significant. Table 1 shows the descriptive analysis of the sample. In supplementary table 1 are presented the descriptive analysis of the perception variables on life habits, body image and satisfaction, eating habits, and food consumption during the pandemic. (Table S1 ). The mean of cognitive restraint in the total sample was 45.43 (22.06), constituting the highest mean of all three eating behaviors dimensions by 95% confidence interval observation (Table 1 ). In the final model, more active during the pandemic (β=5.29), weight loss attempt (β=11.99), perception of overweight (β=8.71), better eating habits during the pandemic (β=6.41) and decreased consumption of refined cereals (β=5.34) and fast foods (β=8.88) were associated with cognitive restraint. The model was controlled for age, sex, income, education level, and BMI. The regression formed a good fit for the data F(10,587) = 23.072, p<0.001, which accounted for 29.8% of the variance in increased cognitive restraint punctuation during the pandemic (R²=0.298). Table 2 shows the crude and adjusted models. The mean of emotional eating was 34.44 (25.63) . In the final model, increased stress during the pandemic (β=7.19), worse sleep (β=4.89), body dissatisfaction (β=4.88), perception of overweight (β=7.35), weight loss attempt (β=4.19), increased amount of food consumption (β=9.32), increased purchase of delivery (β=4.93), increased sweets and desserts (β=7.14) and decreased vegetables consumption (β=-5.21) were associated with emotional eating. The model was controlled for age, sex, income, education level, and BMI. The regression formed a good fit for the data F(13,584) = 21.984, p<0.001, which accounted for 32.9% of the variance in increased emotional eating punctuation during the pandemic (R²=0.329). Table 3 shows the crude and adjusted models. The mean of uncontrolled eating was 32.65 (19.15). In the final model, working>8 hours/day (β=3.85), increased stress during the pandemic (β=3.83), body dissatisfaction (β=6.61), worse eating habits during the pandemic (β=5.26), and increased the amount of food consumption (β=8.65) were associated with uncontrolled eating. The model was controlled for age, sex, income, education level, and BMI. The regression formed a good fit for the data F(9,589) = 20.663, p<0.001, which accounted for 25.3% of the variance in increased uncontrolled eating punctuation during the pandemic (R²=0.253). Table 4 shows the crude and adjusted models. This article sought to analyze the association between perceptions of health-related habits (life habits, body image, eating habits and food consumption) with CR, EE and UE behaviors during the COVID-19 pandemic. The dimension of eating behavior with the highest mean in the sample was cognitive restraint, and this dimension was associated with better eating habits, being more active, having attempted to lose weight in the last 6 months, having the perception of being overweight and decreased consumption of fast foods and refined cereals during the pandemic. The emotional eating dimension was associated with a greater number of variables such as an increase in the amount of food consumed during the pandemic, having the perception of being overweight, having attempted to lose weight in the last 6 months, increased stress, worse sleep, increased purchase of food delivery and decreased consumption of vegetables during the pandemic. The uncontrolled eating dimension was associated with an increase in the amount of food consumed, worse eating habits, increased stress, working or studying >8 hours/day and body dissatisfaction. Although developing healthier behaviors, as demonstrated in the present study, the CR state is often interspersed with episodes of disinhibition of food control (Natacci & Ferreira Júnior, 2011) . CR also debilitates the individual emotionally, causing individuals to eat for comfort (Apfeldorfer & Zermati, 2001) . These findings can also be justified by the emphasis that contemporary society places on the ideal of thinness by providing a sociocultural environment that justifies weight loss. This fact may have been aggravated in the pandemic by the increased time spent on social media that has the potential to aggravate body dissatisfaction and concern with appearance ( Ahuja et al., 2021) . These factors were associated with CR in the present study. This perspective becomes even more critical, especially in the most affected patients, those more sensitive to mood disturbances or those at risk of eating disorders, considering that having better eating and life habits are not necessarily related to healthy eating behaviors. As seen in the case of CR, a practice promoted by the media, accepted by the family and recommended by professionals nowadays (Alvarenga et al., 2015) , and may bring episodes of EE and UE, especially with increased stress during the pandemic. Emotional eating behavior is one in which emotions lead to food choices associated with the emotional context experienced by the individual. It was also possible to associate stress during the pandemic with higher EE scores in a study carried out in Saudi Arabia (Al-Musharaf, 2020); however, unlike the present study, they did not find an association with worse sleep quality. We believe that the association between worse sleep and EE in our study was caused by the ability of sleep disturbances to trigger EE (Palmer & Alfano, 2017) . One study pointed out that the EE score is higher when starting a diet for aesthetic reasons (Jáuregui-Lobera et al., 2014) . The result of the present study may imply that body dissatisfaction and negative perceptions related to the body, when added to greater stress during the pandemic, can lead individuals to have episodes of disinhibition, leading them to consume food as a result of their emotions. This hypothesis can be confirmed by a study carried out in the United Kingdom, where the stress related to the COVID-19 pandemic was associated with the greater drive for thinness in women. The authors suggest that the change in routine caused by the pandemic led women to feel dissatisfied with their bodies and increased restrictive eating and unhealthy weight management practices (Swami et al., 2021) . It is known that individuals who suppress cognitions, such as restrained eaters, tend to eat as a result of their emotions, and when this occurs, there is a search for foods with high energy content and rich in sugar and fat (Coulthard et al., 2021) . In the present study, EE was associated with an increase in the amount of food consumed and increased consumption of sweets and desserts. In addition to this, it is easy to acquire these foods via delivery apps, which may increase their consumption (Horta et al., 2021) . Therefore, we believe this facility justifies the association between EE and increased delivery purchases in our sample. Surprisingly, our study also found that decreased vegetable consumption was associated with a lower EE score. Therefore, a reduction in vegetable consumption for our sample would reduce the EE score during the pandemic. We hypothesized that this result occurred due the Compensatory Health Beliefs (Knäuper et al., 2004) ; according to this mechanism, individuals tend to engage in a healthy behavior (i.e., increasing consumption of vegetables) to compensate or counteracted an unhealthy behavior (i.e., stress, worse sleep, increased amount of food consumed or consumption of sweets and fast foods). These findings suggest that healthy eating can be part of a compensatory strategy. A recent study showed a significant increase in the UE score during the pandemic (Elmacioğlu et al., 2020 ). An association was found between UE and working more than 8 hours/day. It is known that the pandemic exposed workers to the home office and narrowed the boundaries between work and personal life (Shimazu et al., 2020) . This context has the potential to increase stress and consequently trigger UE episodes. The association between stress, worse eating habits, and increased food consumption with UE behavior can be explained by the tendency of stressed people to exhibit a lower level of eating competence, poorer quality of diet, and to overeat; moreover, increased food consumption was previously associated with higher stress and UE behavior (Alvarenga et al., 2015; Elmacioğlu et al., 2020) . As far as we know, this is the first study to assess factors associated with CR, EE and UE behaviors as an outcome and together during the pandemic. Previous studies have only focused on factors associated with EE; however, eating behavior can be understood as a cycle, where individuals can present a behavior over a period of time, and then for some specific reason trigger another type of eating behavior. Therefore, exploring the factors that influence dietary restraint and disinhibitory behaviors is important to assess the real impact of the pandemic on different dysfunctional eating behaviors. Although our results are consistent with studies carried out before the pandemic, we have some limitations. The cross-sectional characteristic of the study does not allow us to verify the precise interaction between changes from the restraint to the disinhibitory state, and therefore longitudinal studies are needed during the pandemic to identify these questions. Furthermore, the sample is not representative and should not be generalized to the entire population, as most of the sample has a high level of internet access. We also cite the use of self-reported variables on weight, height and perceptions of habits and behaviors during the pandemic as a limitation, however, this is a common limitation in behavioral studies during the pandemic and recent research has also used this method, which we consider to be viable and important for research conducted in times of uncertainty (Al-Musharaf, 2020; Coulthard et al., 2021) . Analyzing the associations between dysfunctional eating behaviors and factors related to the pandemic is essential for the management and guidelines to change behavioral patterns, especially during this period of intense changes in the population's routine. There was a significant and negative change in the quality of sleep, stress, physical activity, and food consumption during the pandemic (Malta et al., 2020) , in addition to the increased influence of social media, which has the power to increase body dissatisfaction. Thus, the data from this study can enable nutritionists and behavioral professionals to identify and work on nutritional counseling and a good relationship with food more assertively. Identifying factors associated with CR, EE, or UE behaviors during the pandemic can facilitate behavioral diagnosis during this period and assist in managing the dysfunctional eating behavior that the individual is expressing during the pandemic. Furthermore, these results can support behavioral intervention strategies such as intuitive eating and mindful eating and advanced therapies such as cognitive behavioral therapy and acceptance and commitment therapy (Alvarenga et al., 2015) . In addition, exploring the relationship between behaviors from a theoretical perspective can help professionals identify the cycle of restriction and disinhibition behaviors through lifestyle and eating habits. Our results showed that dysfunctional eating behaviors are present and associated with perceptions about lifestyle and eating habits during the pandemic, regardless of age, gender, BMI, education level, and income. However, restrictive behavior had the highest average in the sample, and this behavior was associated with variables negative body image; this has the potential to trigger emotional eating and uncontrolled eating behaviors when associated with greater stress during the pandemic, which in turn are associated with an intake of energy-dense foods. These disinhibitory behaviors are associated with worse habits and food consumption, which can increase the incidence of obesity and non-communicable chronic diseases, considered risk factors for aggravation and mortality from COVID-19. More longitudinal studies should reveal the relationship between behaviors to provide more guidance to the population in this and future pandemics. Weighty woes': Impact of fat talk and social influences on body dissatisfaction among Indian women during the pandemic Prevalence and Predictors of Emotional Eating among Healthy Young Saudi Women during the COVID-19 Pandemic La restriction cognitive face à l'obésité. Histoire des idées, description clinique [Cognitive restraint in obesity. History of ideas, clinical description Critério Brasil 2019. Associação Brasileira de Empresas de Pesquisa Eating in the lockdown during the Covid 19 pandemic; self-reported changes in eating behaviour, and associations with BMI, eating style, coping and health anxiety Evaluation of nutritional behavior related to Covid-19 Digital food environment of a Brazilian metropolis: food availability and marketing strategies used by delivery apps Psychometric Properties of Spanish Version of the Three-Factor Eating Questionnaire-R18 (Tfeq-Sp) and Its Relationship with Some Eating-and Body Image-Related Variables Physical exercise as therapy to fight against the mental and physical consequences of COVID-19 quarantine: Special focus in older people. Progress in Cardiovascular Diseases Compensatory health beliefs: scale development and psychometric properties A pandemia da COVID-19 e as mudanças no estilo de vida dos brasileiros adultos: um estudo transversal Snowball Sampling: A Purposeful Method of Sampling in Qualitative Research The three factor eating questionnaire -R21: tradução para o português e aplicação em mulheres brasileiras Sleep and emotion regulation: An organizing, integrative review Eating Habits and Physical Activity of the Spanish Population during the COVID-19 Pandemic Period Psychosocial impact of COVID-19 for general workers COVID-19-related stress and anxiety are associated with negative body image in adults from the United Kingdom ISACAMP-NUTRI 2014 -Manual do Entrevistador WHO guidelines on physical activity and sedentary behaviour: at a glance This study was financed in part by the