Age, skin color, self-rated health, and depression associated with co-occurrence of obesogenic behaviors in university students: a cross-sectional study

ABSTRACT BACKGROUND: The university context plays an important role in the health-disease process since students are potentially vulnerable to obesogenic behaviors that can influence long-term health. OBJECTIVE: To estimate the prevalence of and factors associated with the co-occurrence of obesogenic behaviors among university students. DESIGN AND SETTING This was a cross-sectional study at a Brazilian public university. METHODS: This study was conducted with all university students in the first and second semesters of 2019 at Universidade Federal de Ouro Preto, Minas Gerais, Brazil. Data were collected between April and September 2019, using a self-administered questionnaire. The outcome was the co-occurrence of obesogenic behaviors, measured as the sum of three risk behaviors: inadequate eating practices, leisure-time physical inactivity, and sedentary behavior. A Venn diagram was used to evaluate the simultaneous occurrence of risk behaviors. Pearson’s chi-square test and multivariate logistic regression were used for statistical analyses. RESULTS: A total of 351 students participated in the study. Inadequate eating practices constituted the most prevalent isolated risk behavior (80.6%), which was also the most prevalent when combined with sedentary behavior (23.6%). University students aged 20 years or younger, with non-white skin color, poor self-rated health, and symptoms of depression had increased chances of simultaneous occurrence of obesogenic behaviors. CONCLUSION: These findings highlight the importance of developing and implementing actions to reduce combined obesogenic behaviors in the university environment. Institutions should focus on creating an environment that promotes health-protective behaviors such as physical activity and healthy eating.

Understanding potentially obesogenic behavioral risk factors among university students is imperative for identifying more susceptible groups and recognizing the health effects of these factors, to facilitate the development of prevention and health promotion strategies targeted at the university environment. Additionally, this information can contribute to more effective public policies to reduce the rates of obesity-and overweight-related NCDs.

OBJECTIVE
This study therefore aimed to estimate the prevalence of cooccurrence of obesogenic behaviors and their associated factors in university students.

Study design and population
This cross-sectional study was integrated with a project on anxi- PADu is a longitudinal study conducted with university students entering undergraduate courses offered at the campi of Ouro Preto and Mariana of the UFOP. Data will be collected at three different time points (T0-in the first semester of the undergraduate course; T1-after attending two years; T2-after attending four years) to verify behavioral changes during academic life. For the present study, data from the baseline (T0) were used.
The study population included all university students entering the first and second semesters in the 2019 undergraduate courses in architecture and urbanism, performing arts, law, physical education, civil engineering, production engineering, geological engineering, pharmacy, history, journalism, mathematics, medicine, nutrition, and pedagogy. The students' lists were made available through the UFOP's teaching sections.
Students who met the following inclusion criteria participated in the research: regularly enrolled in the first period of the undergraduate courses evaluated in the study and aged 18 years or older.
The PADu sample comprised 355 university students. However, the final sample of this study consisted of 351 university students, since four participants did not answer all the questions related to the co-occurrence of obesogenic behaviors.

Data collection
Data were collected between April and September 2019 by project members who were previously trained to apply the instrument and clarify possible doubts of the students. A pilot study was conducted with students attending the eighth period of the nutrition course in the second semester of 2018 who would therefore not participate in the sample.
The questionnaires were administered during class hours, after taking prior appointments, and the teacher's presence in each selected course. The researchers oriented the university students about the study, risks, and benefits. They were also informed that their participation was voluntary and anonymous. Those who agreed to participate signed the informed consent form and answered a questionnaire consisting of socioeconomic characteristics, lifestyle habits, and health conditions.

Variables of the study
The outcome variable (co-occurrence of obesogenic behaviors) was obtained from the sum of three risk behaviors: inadequate eating practices, leisure-time physical inactivity, and SB. The responses were categorized as none to three obesogenic behaviors. These behaviors are justified because they are considered health risk factors and are associated with the most significant burden of NCDs and mortality. 17 The variable "inadequate eating practices" was obtained through a scale developed and validated by Gabe and Jaime for adults, which measures adherence to healthy eating practices based on the recommendations of the second edition of the Food Guide for the Brazilian Population. 18,19 For classification purposes, the cut-off points proposed by Gabe and Jaime were used, wherein eating practices were classified as "inadequate" when the sum of the individual scores assigned to the responses for each alternative resulted in a score of up to 31 points, at "risk" when the score was between 32 and 41 points, and "adequate" when it was greater than 41 points. 20,21 Subsequently, for the present study, eating practices were recategorized as "adequate" and "inadequate. " Leisure-time physical inactivity was assessed using the study Surveillance System for Risk and Protective Factors for Chronic Diseases by Telephone Survey (VIGITEL), with questions such as: "In the last three months, did you practice any physical exercise or sport? (Do not consider physical therapy). " 22 Participants who answered "no" were classified as "inactive in leisure time, " and those who answered "yes" were considered "active in leisure time. " SB was included in the study because a growing number of studies characterize it as a health risk factor, different from and independent of physical inactivity, and associated with the occurrence of adverse health effects, such as metabolic syndrome. 9, 23 SB was determined in the questionnaire using the following question: "In your free time, that is, when you are not studying or working, how much time (in hours) do you dedicate to using the cell phone, television, computer, or tablet?" This question was adapted from two questions from VIGITEL. 22 For each of the screen types evaluated, eight answers were possible: "I don't use, " "less than an hour, " "between one to two hours, " "between two to three hours, " "between four to five hours, " "between five to six hours, " and "more than six hours. " For analysis purposes, SB was analyzed as a continuous variable and responses were coded as 0, 0.5, 1.5, 2.5, 3.5, 4.5, 5.5, and 6.5 hours, respectively. Subsequently, the responses of the time spent on each type of screen were summed, and the classification of SB was established according to the median. University students with screen time ≤ 6 h were classified as "non-sedentary, " while those with screen time > 6 h were considered "sedentary. " The explanatory variables included in this study were grouped In the health condition domain, the following variables were evaluated: self-rated health, categorized as "good" (good and very good) and "bad" (regular, bad, and very bad); anthropometric profile (not overweight or overweight); use of medication for chronic diseases (no and yes); anxiety symptoms (no and yes); depression symptoms (no and yes); and stress symptoms (no and yes).
The anthropometric profile was evaluated by calculating the body mass index (BMI) through the anthropometric measurements of weight and height, self-reported by the participants. The classification was made according to the BMI reference values established by the World Health Organization for adults 24 and adolescents. 25 Individuals classified as underweight and eutrophic (BMI < 24.9 kg/m 2 ) were grouped in the "not overweight" category and those classified as overweight and obese (BMI ≥ 25 kg/m 2 ) in the "overweight" category. Additional details on the anthropometric profile classification methodology can be found in a previous publication. 26 The variables "anxiety symptoms, " "depression symptoms, " and "stress symptoms" were obtained through the reduced version of the Depression Anxiety Stress Scale-21 (DASS-21). 27 The scale is composed of a set of three subscales, designed to estimate in a self-reported way the symptoms of anxiety, depression, and stress in the week before data collection. The response scale to the items is a four-point Likert scale ranging from 0 (not applicable) to 3 (applicable most of the time), generating scores that allow the classification of anxiety, depression, and stress symptoms as "normal, " "mild, " "moderate, " "severe, " and "extremely severe. " In the present study, symptoms of mental disorders were re-classified as absence ("no"; normal and mild) and presence ("yes"; moderate, severe, and extremely severe).

Statistical analysis
The variables were descriptively analyzed using frequency distribution. A Venn diagram was used to represent the simultaneous occurrence of obesogenic behaviors among the evaluated university students. This representation method allows for the comparison and visualization of the overlap and differences among the datasets being analyzed based on the intersections of the graphical shapes. 28,29 Initially, the chi-square test was performed between the explanatory variables and the co-occurrence of obesogenic behaviors, and those with a P value < 0.20 in the bivariate analysis were included in the multivariate model. Multivariate logistic regression was used to verify the association between the co-occurrence of obesogenic behaviors and explanatory variables. In this analysis phase, three models were structured to represent the co-occurrence of the obesogenic behaviors evaluated: Models 1, 2, and 3 included no behavior versus one behavior, no behavior versus two behaviors, and no behavior versus three behaviors, respectively. For this, we used a reference category for university students with no obesogenic behavior versus the number of obesogenic behaviors (1, 2, or 3). To select sociodemographic and health condition variables, the backward method was adopted, and only the variables that presented a P value of < 0.05 remained in the multivariate model. All the models were adjusted for the variable "sex. " The odds ratio (OR) was used to measure the association with the respective 95% confidence intervals (95% CI). The level of statistical significance was 5%. The analyses were performed using Stata version 13.0 (Stata Corporation, College Station, Texas, United States).

RESULTS
Of the 351 university students included in this study, 57.6% were female and 65.8% were 20 years or younger, ranging from 18 to 31 years. Most participants self-reported their color or race as white (51.1%), single (95.4%), heterosexual (79.5%), living without family members (66.4%), and not employed (89.2%). Regarding family income, slightly more than half (56.7%) of the students reported a family income higher than or equal to three minimum wages. Regarding the distribution by area of knowledge, 41.0% were from life sciences courses, 34.5% from the humanities and social and applied sciences, and 24.5% from the exact sciences (Table 1).
Regarding health conditions, 41.0% of the university students selfrated their health as bad, 22.3% were overweight, and 13.7% reported using medications for chronic diseases. Anxiety, depression, and stress  Figure 1 shows the co-occurrence of obesogenic behavior.
The adoption of inadequate eating practices and SB (23.6%) was observed to be the most prevalent combination of risk behaviors among students, followed by inadequate eating practices, leisure-time physical inactivity, and SB (17.9%), and inadequate eating practices and leisure-time physical inactivity (15.7%). The absence of risk factors was observed in 9.7% of university students.
The prevalence distribution of obesogenic behaviors according to sociodemographic characteristics and health conditions is presented in Table 1. In the bivariate analysis, sex, age, skin color, selfrated health, anxiety, depression, and stress symptoms remained associated with the co-occurrence of obesogenic behaviors among university students.  with the co-occurrence of two or more obesogenic behaviors. With the transition from high school to higher education, university students face many changes, such as lack of time due to studies, overlapping activities, and new responsibilities, which may interfere with adopting healthy eating practices. 30 In addition, many factors, such as socioeconomic status, lack of ability to make healthy food choices, difficulty cooking, lack of healthy food in university cafeterias, and "environmental barriers, " such as opening hours of nearby food stores, influence the availability of food, and negatively affect students' eating behaviors. 30,31 These factors may favor new eating habits, reflected in unhealthy eating practices and health-related problems, including being overweight. [32][33][34] Thus, health promotion strategies, including promoting healthy eating in the university environment, are vital, as numerous health behaviors are developed and established during this period 30 and tend to continue into adulthood, increasing the risk of developing chronic diseases in subsequent years. 15 Exposure to health-risk behaviors has been described in studies with young populations. 9 Studies that evaluated the aggregation of inadequate eating practices and SB showed that these factors share contextual determinants and influence each other. 35 In the present study, we found that the most prevalent combination of risk behaviors among university students was the coexistence of inadequate eating practices and SB. In contrast, in a study of adults, SB, including the habit of watching television, using a computer, reading books, or magazines, remained associated with the consumption of healthy and unhealthy foods. 36 However, comparisons between the risk factors analyzed should be interpreted with caution, given the various methods used to assess food intake. It is noteworthy that the instrument used in the present study included other dimensions of adequate and healthy eating and food intake.
Scientific evidence shows that SB reduces energy expenditure and favors inadequate food consumption, including increased intake of foods rich in fat and sugars and decreased consumption of healthy foods such as fruits and vegetables. 35,36 Moreover, besides being risk factors for becoming overweight, this association between high screen time and inadequate eating habits may increase susceptibility to other health-risk behaviors, 35 resulting directly in series of unfavorable health outcomes. 37 The simultaneous occurrence of the three risk behaviors assessed, characterized by inadequate eating practices, leisure-time physical inactivity, and SB, was prevalent in 17.9% of university students. Few studies have investigated the clustering of health risk behaviors among university students. 13 In a study conducted with Brazilian university students, a high prevalence was observed for the simultaneous occurrence of the four primary behavioral risk factors for NCDs: physical inactivity, inadequate fruit and vegetable consumption, excessive alcohol consumption, and smoking. 38 The study did not include SB in its analyses since this risk factor has been less studied than other risk behaviors already established in the literature, such as food intake and physical activity.
It is worth noting the importance of investigating the aggregation of traditional and emerging risk behaviors among young people, especially university students, to provide information on which to base future actions. 13 Table 2. Odds ratio (OR) and 95% confidence interval (95% CI) for one or more obesogenic behaviors; multivariate model of sociodemographic characteristics and health conditions associated with the co-occurrence of obesogenic behaviors in university students entering the Universidade Federal de Ouro Preto in 2019. Ouro Preto, Minas Gerais, 2019 (n = 351) OR = odds ratio; CI = confidence interval.  39,40 In this study, we also observed that university students who self-rated their health as bad had a higher chance of one, two, or three obesogenic behaviors than those who self-rated their health as good. Studies show that individuals who perceive their health as bad tend to present more health risk behaviors, 41 such as inadequate intake of fruits and vegetables, physical inactivity, and SB. These behaviors are determinants of NCDs and are related to the negative subjective assessment of health. 42 Thus, these findings highlight the importance of considering how health is perceived by university students since perceptions of health can influence the adoption of healthy lifestyle behaviors. 43 The presence of depressive symptoms was associated with the co-occurrence of obesogenic behaviors among university students, corroborating the findings of Champion et al., 13 who evaluated 18-year-old Australian youth and observed a significant association between the clustering of multiple health-risk behaviors and mental health outcomes such as anxiety and depression.
One hypothesis to justify this association is that individuals may engage in unhealthy behaviors to help cope with mental health problems. 44 In addition, stress and mental health disorders may interfere with a person's choice to adopt healthy lifestyle behaviors such as physical activity, while also exposing themselves to health-risk behaviors. 45 There was a greater chance of exposure to multiple obesogenic behaviors among university students aged 20 years or younger.
Evidence shows that the prevalence of simultaneous exposure to health risk behaviors increases with age, 9,46 since young people acquire greater autonomy and economic independence with advancing age. 9 However, this association has not been well established in the literature. In a systematic review of the co-occurrence of multiple risk behaviors, older age groups were considered risk factors for aggregating multiple risk behaviors. 12 Although the findings of this study are consistent with those reported in the literature, some limitations should be considered when interpreting the results. Students from a single university were included, limiting comparisons with students from other higher education institutions. Another limitation of the study is the methodological design, which does not establish a cause-effect relationship between the variables and temporal relationships on the associations found. In addition, as the students were evaluated in their initial semester at the university, their recent entry into academic life may not have set their lifestyles.
It is important to highlight that this study was based on self-reported behaviors, which may have generated information bias, as young people tend to overestimate or underestimate their exposure to health risk behaviors. Despite these limitations, the findings obtained add essential evidence regarding the prevalence and factors associated with the co-occurrence of obesogenic behaviors among university students.

CONCLUSION
The findings from this study showed that a high proportion of university students with simultaneous obesogenic behaviors, especially among those who self-reported colored skin, rated their health as bad, and reported depressive symptoms. These findings contribute to a better understanding of the associations between various obesogenic behaviors, highlighting the need for interventions directed at university students. In addition, these results highlight the importance of health promotion in the university environment, with actions aimed at a healthy lifestyle.
Public policies that target risk behaviors in groups and stimulate a healthy food environment and physical activity in universities are essential for reducing the risk of major chronic diseases related to excess weight.