Prevalence of Emotional Eating and Its Associated Psychosocial Factors Among Urban Chinese Undergraduates In Hong Kong: A Cross-Sectional Study

there are limited data from this study to examine the prevalence and of A cross-sectional study was conducted on 424 (aged from two in 2019. Respondents were randomly invited to complete an anonymous online questionnaire that obtained background information, emotional eating subscale of Dutch Eating Behaviour Questionnaire (DEBQ), and Depression Anxiety and Stress Scales (DASS-21). Two-sample independent t-test and multiple regression analyses were conducted to test the association of study variables with emotional eating. adequate sample and the use of the validated questionnaire our results could be compared to It examined multiple psychosocial factors and, to our knowledge, the relationship between EE and self-reported health was not previously examined. The was that sensitive (i.e. were least by social desirability.

In studies conducted in western countries, various psychosocial parameters were associated with EE. In general, EE was associated with psychological factors (e.g. impulsiveness and the presence of restricted eating), biological factors (e.g. sex and current weight), and situational factors (e.g. level of stress and current mood) (12). Yet, the current evidence suggested a prominent gender difference in the determinants of EE. In particular, females were more prone to EE to numb their emotions and were more in uenced by depression, anxiety, and stress (13). Bennett et al. suggest that EE was triggered by stress in female college students but was triggered by anxiety in male students (14).
Life in universities can be stressful due to adaptation into a new environment, interpersonal di culties, academic pressures, and anxieties about one's future career; and university students reported a high level of substance abuse and alcohol consumption (15,16). However, there is a lack of similar research describing EE, including its prevalence and its associated psychosocial factors, in Chinese undergraduates. Most of the relevant studies were conducted in Western countries and previous Chinese research focused on much younger adolescents living at home and found that only 3% had EE (17). Other studies from China found that stress and mood problems were associated with EE in females and secondary school students (18,19). Existing studies also described that EE behavior as a negative coping strategy to relieve mental tension (14,20,21). Furthermore, in contrast to most urban areas of China, Hong Kong has a long tradition of preprocessed foods and Western fast-food chains such as McDonald's; and it was unclear if EE is prevalent in older, university-aged students in Hong Kong.
This study aimed to determine the prevalence of EE among Hong Kong university students in Hong Kong and examine the characteristics of emotional eaters. As the primary objective, we hypothesized that EE was prevalent among this population and had prevalence approaching those reported in previous Western studies (approximately 20%) (5). For secondary objectives, we hypothesized that EE was associated with greater dysphoric mood (depression, anxiety, stress), higher body mass index (BMI), worse life satisfaction, and worse perceived health.

Participants
This survey study recruited university students in public areas of the two major universities in Hong Kong (The University of Hong Kong and The Chinese University of Hong Kong) between April and June 2019.
Participants were included if they were (i) of 18-24 years of age and (ii) studying an undergraduate degree.
International and exchange students were excluded from this study. The study recruited 424 participants (based upon sample size calculations, All participants were invited to ll in a standardized questionnaire on the Google survey platform (details see below). Ethical approval was obtained from the Survey and Behavioural Research Ethics Committee of the university sponsoring the study prior to participant recruitment (Approval #043 − 19).
Assuming the prevalence of EE was estimated to be 20%, the precision of 4%, and the type I error at 5%, 385 participants were needed. 424 participants were recruited to account for possible missing data and drop out Page 4/17 at 10%.

Instruments
Dutch Eating Behavior Questionnaire (DEBQ) The 13-item Emotional Eating Subscale of the Dutch Eating Behaviour Questionnaire (DEBQ) was the primary outcome of the current study and was one of the most commonly used instruments to detect EE in research (5). Each statement in the DEBQ was rated with a 5-point Likert scale (ranging from 1 "never" to 5 "very often"). The scores from each statement were summated and averaged to obtain the nal score. DEBQ had high internal and factorial consistency, with a Cronbach's alpha of 0. 964 among Chinese adolescents (22). This study used a score of > 3.25 to classify respondents as having EE (5).
Higher scales scores for perceived health life satisfaction and study satisfaction re ect greater levels of health and satisfaction.
Covariates and other outcomes Demographic data including age, sex, place of birth, faculty, presence of romantic relationship, and exercise habit were collected. As the questionnaire was conducted in an open area in the universities and measuring instruments were not feasible, the weight and height of the participants were self-reported. Life satisfaction and perceived health were rated on a 5-point Likert Scale (ranging from "1": totally disagree to "5": totally agree) to the statements "You are satis ed with your life" and "You perceive yourself as healthy", respectively.

Statistical analysis
Demographic data and psychosocial factors among female and male students were described by mean and standard deviation (SD) and by percentage and absolute number for continuous and categorical variables respectively. Since EE was strongly linked to gender, we strati ed all data analyses by gender. Body mass index (BMI) was calculated by weight (in kilograms) over square of height (in meters) and obesity was de ned as body mass index of > 23 kg/m 2 , which was the international cut-off for Southeast Asians (24

Association of EE with bodyweight, perceived health, and perceived life satisfaction
Both male and female EE participants reported higher BMI, poorer self-perceived health, and life/study satisfaction than their non-emotional eating counterparts (Tables 3 and 4). Furthermore, obesity was more signi cantly more common in females with EE; obesity was found in 48.5% of EE females (vs 11.1% in non-EE females, p < 0.001) and 66.7% in EE males (vs 30.7% in non-EE males, p = 0.24) (non-tabulated).  Emotional eaters (Mean DEBQ score > 3.25) versus Normal emotional eating score (Mean DEBQ score ≤ 3.25) *p < .05, **p < .01. † Perceived health score possible range from 1-5; ‡ Life Satisfaction Scale score possible range from 1-5; § Study satisfaction score possible range from 1-5.
Higher scales scores for perceived health life satisfaction and study satisfaction re ect greater levels of health and satisfaction. Emotional eaters (Mean DEBQ score > 3.25) versus Normal emotional eating score (Mean DEBQ score ≤ 3.25) *p < .05, **p < .01. † Perceived health score possible range from 1-5; ‡ Life Satisfaction Scale score possible range from 1-5; § Study satisfaction score possible range from 1-5.
Higher scales scores for perceived health life satisfaction and study satisfaction re ect greater levels of health and satisfaction.

Discussion
This is one of the rst studies to describe the prevalence of EE in Chinese college students and found that EE was present in around 10% of the college students. A striking gender difference was found in terms of prevalence of EE (14.8% in female and 4.5% in male students) and its associated psychosocial factors. In male students, EE was only predicted by depression and stress scores, but EE was also predicted by the absence of a romantic relationship among females. Also, all other demographic data did not associate with EE. Furthermore, the current study showed that EE was associated with obesity, mood problem, poor self-rated health and poor life/study satisfaction.
The prevalence of EE in the current study was 10%, which was similar or lower to the prevalence reported in Western countries (2, 3, 25); however, this was substantially higher than the prevalence of EE in secondary school-age students in Hong Kong (17). The underlying reasons could not be concluded from the current study -it could be a genuine change so that EE was increasingly prevalent or that EE was more common in college students than in secondary school-age students. However, this may also be due to the use of different de nitions and instruments used to de ne EE in different studies (17). This could be investigated in longitudinal studies. Yet, in concordance with other studies, EE was more common in female than male and was associated with poorer health, mood, and life satisfaction (26,27).
Our study was one of the rst Chinese study to note that emotional eating behaviors and its relationship with psychosocial factors differ between genders. For instance, having a romantic partner was protective for emotional eating among female students only. Due to the cross-sectional nature of this study, it is unclear whether close interpersonal interactions from a romantic relationship may reduce reliance on emotional eating as a coping mechanism or if those who partake in maladaptive coping behaviors are less likely to form these relationships. Similar to previous studies (47,50), EE was associated with dysphoric mood. However, our study did not show that anxiety was an independent trigger for EE. It may re ect that EE is not a common coping mechanism for anxiety in Chinese adolescents or that the sample size was insu cient to detect an association.
The current research suggested that EE was common among university students, especially affecting female students. Therefore, clinicians may consider screening for EE, especially when seeing young adults with mood problems, and offer appropriate counseling and interventions. However, despite being associated with adverse physical and psychological consequences, there is a lack of guideline-based treatments for EE in Hong Kong and worldwide. As EE was conceptualized as a poor stress-coping strategy, treatments that enhance emotional coping skills may reduce EE (28). For instance, the latest meta-analysis suggested that mindfulness-based interventions could reduce EE and body weight; however, the current evidence was limited by unclear/high risk of bias and there was a lack of similar studies in the Chinese population (29). Further research could be conducted to delineate the prevalence of EE in other populations. Moreover, longitudinal studies will clarify the relationship between mood problems and EE; and similarly, high-quality randomized controlled trials will be needed to examine treatment modalities in the Chinese population.
The strength of the current study included adequate sample size and the use of the widely validated questionnaire (DEBQ) so that our results could be compared to international studies. It also examined multiple psychosocial factors and, to our knowledge, the relationship between EE and self-reported health was not previously examined. The questionnaire was completely anonymous so that sensitive answers (i.e. emotional eating behaviors) were least affected by social desirability.
Yet, several limitations could be discussed. First, as a common limitation to all cross-sectional studies, the causality between EE and other factors could not be concluded. For example, EE could have a bi-directional relationship with the self-reported level of health -people with EE may overeat or eat unhealthy food and led to poor health; yet, poor health could be a stressor that facilitated EE. The students from these two universities are more academic achieving and may be more likely to experience stress because of high expectations as compared to other tertiary institutes in Hong Kong. However, the age distribution and socio-economic backgrounds of students from the sample are not expected to be appreciably different from other universities in the region. Second, although our sample size was adequate for the primary outcome, only a few male participants had EE (n = 9) and this might limit the power to detect the association between EE and psychosocial factors, especially among male students. Moreover, many outcomes, including BMI, were selfreported and were prone to reporting bias. Similarly, as the underlying causes of emotional eating could be multifactorial, there are likely other confounding variables that were not covered in this study such as aspects of social support.

Conclusion
Emotional eating is common in Chinese university students, with female being the most prominent risk factor.
EE was associated with mood problem, obesity, poorer health, and poorer life/study satisfaction. More studies are needed to determine the best treatment strategies, especially in the Chinese population.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests.

Funding
Not applicable.
Authors' contributions KS designed the study and collected the data. KS and RC conducted the statistical analyses. KS, EL and JHK interpreted the results and wrote the manuscript. All authors were involved in the production of the manuscript and approved the nal version.