Health Behaviors, Life Skills, Mental Health, and Demographic Factors Associated with Mental Health among University Students in a Developing Country.

University students in developing countries may be at high risk for mental health disorders due to many nancial, political, and social stressors. In the present study, we intended to explore the association of mental health with health behaviors and life skills. Participants included 2789 university students in Lebanon. Bivariate and multivariate regression analysis were performed. Results suggested that the risk for mental health disorders is associated with the socio-demographic factors of being a rst year university student, being female, and living in rural areas. Also, consumption of high-calorie dietary pattern, low physical activity, and high level of risky behaviors associated with higher occurrence of mental health disorders. Life Skills are associated with less occurrence of mental health disorders in our data. Sociodemographic factors such as gender, university year, and area of residence (rural vs. urban), healthy behaviors, and life skills showed to be strongly associated with university students’ mental health. Life skills based interventions addressing the aforementioned mental health determinants would benet university students in developing countries.

variables in a critical context such as Lebanon. University students in Lebanon are living under a variety of stressors as a result of the demographic transition, nutrition transition, high unemployment rates, political con ict and economic crisis which have their social and nancial implications on youth leading to an increase demand for educational employment and employment opportunities to mitigate health and economic burdens (UNDP, 2016). Therefore, exploring factors that might enhance university students' mental health in such a context is crucial to be adopted in other developing countries facing similar circumstances. This study is essential to frame health promotion interventions that are evidence, theory and innovation-driven interventions. Therefore, this paper aims to study the association between life skills, health behaviors and mental health as health outcome among university students in Lebanon allowing us to ultimately design an appropriate intervention accordingly.

Participants
Two thousand seven hundred eighty-nine university students, with mean age 20.9 (± 6.14) enrolled in 12 public and private universities have participated in this study. We were able to get the larger universities participate and have the one public university in the study.

Procedures
An online survey was lled randomly by university students. Written consent was obtained from all participants; it was added to the rst page of the questionnaire. All students over 18 years of age were eligible to participate, and no other exclusion criterion. The sample size needed for su cient power was achieved where 2789 questionnaires were completed. All the incomplete questionnaires were discarded. The Institutional Review Boards of the participating universities approved the present study.

BMI
The corrected BMI designed and validated by  was adopted in this study. It aims to mitigate the social desirability among responses. The equations are: for males [corrected weight = (1.003* reported weight) and corrected height = (0.959* reported height) + 7.59 and for females corrected weight= (0.942* reported weight) + 3.14 and corrected height = (0.943* reported height) + 9.42. The corrected body mass index (BMI) was calculated as follows: corrected weight in kilograms /corrected (height) 2 (in meter). BMI values were classi ed into four categories for emerging adults (> 18 years old); underweight (BMI < 18.5 kg/m2), normal weight (BMI 18.5-24.9 kg/m2), overweight (BMI 25-29.9 kg/m2), and obese ( > = 30 kg/m2) ( WHO, 2000). each variable in relation to MH. All results were considered statistically signi cant with p < .05, CI = 95%. Multivariate models were used to decompose the covariance between demographics, health behaviors and life skills to better explore the predictors that are correlated with the health outcomes. Four models were introduced in the multivariate analysis; in rst model, the demographic factors(Rural/Urban residence, university year, age, sex)and sleeping hours were studied in correlation with Mental Health. In Model 2, Life Skills components were added while in model 3, we added BMI, and Model 4 studied the correlation between health behaviors (physical activity, risk taking actions, smoking, food intake).

Results
A total of 2789 university students participated in this study. Their mean age is 20.9 (SD = 6.14) years.
The majority of them sleep ve to eight hours during the weekends (65%) and the weekdays (77%). The percentage of students who sleep less than ve hours during weekdays is 17% and 8% during weekends.
While 59% of the participants have a low physical activity level, the BMI mean is 22.72 Kg/m 2 (SD = 3.61) with 68% of the university students having a normal weight status, 16% are overweight, 11% are underweight, and 5% are obese.
Concerning the food frequency intake, 60% have low calorie food consumption and 19% adopt high calorie food, 18% adopt processed food, while 3% consume hot beverages. Moreover, the results show that 20% of the participants are prone to anxiety and/or depression. According to the life skills, 89% of the participants have the overall life skills: 90% of the students have Relationship and Communication skills, 89% have the Self-Care skills, 86% of them have the Daily living skills, 84% have the skills for Goal Setting, 82% have skills in Work and Study Skills, 80% have the Career and Education Planning skills. The lowest percentage is 67% for Housing and money management skills.
Bivariate analysis was performed to examine the association between university students' mental health and demographic and behavioral factors as well as life skills (Table 1).
Results showed a strong association between mental health and participants' sex (p = 0.028); females are more prone to anxiety/depression (M = 35.02 ± 7.23 compared to 34.60 ± 6.25 for males). Also, the residence of the participants is a signi cant variable associated with their mental health (p = 0.026). Post-Hoc analysis showed a signi cant difference in the mental health status between Beirut and regions outside Beirut (p = 0.024). Students living outside the capital were more prone to anxiety and depression (36.42 ± 8.42).
Smoking is another variable that showed to be signi cantly associated with students' mental health (p = 0.040). Post-Hoc analysis showed that differences were detected among sub variables (cigarettes, water pipes, and mixed smokers). Risk taking behaviors are also signi cantly and positively associated with students' mental health with p < 0.001 and R = 0.10.
Moreover, a signi cant association was observed between mental health and physical activity (p = 0.047) while according to the Post-Hoc analysis, there is an absence in signi cant difference among the sub categories; low, moderate, and vigorous activities.
BMI is another signi cant factor associated with students' mental health (p < 0.001). According to the Post-Hoc analysis, the signi cant differences are among underweight and normal students (p < 0.001) and between underweight and overweight students (p = 0.045). Underweight students have the highest mental health mean (M = 36.62 ± 5.95) while the normal weight students have the lowest mental health mean score (M = 34.74 ± 6.91). As for eating patterns, only low calorie diet (factor 1), and high calorie diet (factor 2) are signi cantly associated with students' mental health where this latter is negatively associated with factor 1 (R = − 0.072) and positively associated with factor 2 (R = 0.132), the correlation with processed food and hot beverages is non-signi cant.
Analysis of the relation between life skills and mental health showed lower mental health mean scores among students who have the life skills compared to those who do not have them; For instance, students who have the Daily Living skills have a mental health mean score equal to 34.73 ± 7.09 compared to 36.88 ± 5.15. The difference is very remarkable among students who have the Self-Care skills where their mental health mean score is equal to M = 22.71 ± 4.62 compared to M = 37.39 ± 7.08 among those who do not have this skill. Linear regression models were conducted to adjust the effect of covariates on the students' mental health. Based on testing several models in multivariate linear regression, the R 2 was 0.76 for the nal model which means that this model explains 76% of the variance for mental health (Table 2). This model was a signi cant predictor of the outcome variable with p value < 0.001.
In Model 1, region (p < 0.001), gender (p = 0.018), and university year (p = 0.028) were signi cant variables associated with students' mental health. Female students are more prone to have anxiety and depression than male students (B = 0.045). Students who come from outside Beirut are more at a higher risk for depression and anxiety (B= -0.075). As for university year; the more advanced the student's year is, the less likely they are to show signs of mental health disorders (B= -0.043) When life skills were added in Model 2, the region (p = 0.03), university year (p = 0.019), and gender (p < Age, sleeping hours, physical activity, smoking, body mass index, hot beverages and processed food, life skills components of Daily living, Self-Care, Relationship and Communications, Career and Education Planning showed to have non-signi cant association with university students' mental health.    Strati cation by gender The data were strati ed by gender ( Table 3). The same models were conducted. The R 2 was 0.83 for both males and females for the nal model. These models were signi cant predictors of the outcome variable, mental health, with p value < 0.001.
For female students, the results of model 1 showed that university years and region are both signi cantly and negatively associated with female students' mental health. Female students who belonged to areas outside Beirut are at higher risk for anxiety or/and depression (p = < 0.001, B= -0.082), also students who are at senior year in university are less likely to suffer with mental disorders (p = 0.014, B= -0.067).
In Model        his or her lifetime, leading to more comorbidities and even death. However, very few seek treatment due to the stigma and social taboos in this context preventing them from living healthy and productive lives (Embrace, 2020).
The socio-demographic factors of university years, sex, and urban vs. rural residence, signi cantly related to mental health. While no signi cant association between mental health status and the region students belonged to was shown among male students, results showed that female rural dwellers are at higher risk for anxiety or/and depression than female urban dwellers. This can be explained by the fear and concerns of rural emerging adults about making decisions especially among girls who are living away from their parents' home to continue their education as the result of social expectations. The social norms and the parents put more restrictions on female students (San Antonio, 2016). Emerging adults who are originally from rural areas are suffering from a higher risk of mental health disorders due to the lack of resources and stigmatization while they used to rely on spiritual leaders, family and friends to solve their problems in their villages (Gsell, 2010). Rural dwellers might experience culture shock due to their exposure to new social norms and a number of environmental stressors in the rst two university years (Gsell, 2010). Studies showed that male were more successfully able to adapt than female students as the latter have a higher concern regarding public transportation use, dressing styles, reactions to greetings, separation from family,, lack of social support structure, and the need for time and money management (Smith, 2004).
Interestingly, we also show that contrary to university year, age was not a signi cant variable in our study.
A meta-analysis showed that rst year students had the highest rates of internalized mental health problems then gradually decreased at the nal year. This might be explained by an increase in students coping and adaptation skills gradually throughout the university year (Puthran et al., 2016). This is similar to the results of the multivariate analysis in this study that showed an increase in depression and anxiety in the rst university year, then it will start to decrease gradually by years. This can be explained by the adjustments that rst year university student, female in speci c needs to adopt, accompanied by the social restrictions and taboos that will be accountable for as a female living in a developing country. However, students start to acquire the coping skills needed gradually throughout the university years.

Health Behavior Variables Related to MH
The behavioral factors of physical activity, risky behaviors, diet, and life skills signi cantly related to mental health. In contrast, BMI, smoking, and hours of sleep, did not signi cantly relate to mental health of college students.
Our ndings showed that risky behaviors (excluding smoking) are signi cantly and positively related with college students' mental health among both males and females; risk taking actions are associated with higher risk of depression and anxiety.  have shown similar results where they found that mental distress was associated with higher adoption of risky behaviors among the Lebanese university students. Same results were also explored by other researchers (Bersamin et al.,2014;Tavolacci et al., 2013). This transition phase in emerging adults' lives is a self-discovery period where students are more prone to experience risky behaviors and are more susceptible to media and peer in uences. Moreover, college students do not consider themselves as vulnerable to dangers (e.g., drugs, binge drinking, unsafe sex practices) while they are more exposed to risky behaviors on a regular basis associated with lack of continuous parental monitoring ( . In our study, physical activity is signi cantly associated with only male students' mental health, this might be explained by the higher level of physical activity conducted by male than female students (Arroyo et al., 2000).
As for the diet patterns, both the bivariate and multivariate analysis showed a signi cant association between high/low calorie food and students' mental health. While high calorie food showed to be positively associated with males' and females' mental health status, low calorie food showed to be a signi cant factor that is negatively associated with only female students' mental health. The consumption of low food calorie is associated with lower mental health disturbances in the literature. In  Singla et al., 2019), and strengthen one's sense of coherence, which plays a role in promoting mental and physical health (Kase, et al., 2019). Although LST usually aims to improve all the life skills components it comprises comprehensively, it may be more effective to focus on speci c skills that showed a strong correlation with students' mental health in our study (Fagan & Mihalic, 2003). The nancial, political, and social complexities of the developing countries such as Lebanon add more burdens on students' choices of healthy behaviors leading to negative health outcomes. Developing countries that are similar to the Lebanese context are in need for life skills based intervention to enhance youth wellbeing to overcome the ecological barriers and challenges (IYF, 2013). This will provide them with a safe and healthy environment to be more productive and exible in their university and career life (Al-Nakeeb et. Al, 2015). Intervention can be selective in focusing on the components that are of importance to the lives of university students in a speci c context to increase their effectiveness and e ciency. Limitations: The nature of this study is cross-sectional, this design precludes the causality relationship between variables in this present study. A cohort design might lead to better understanding of the association between mental health and other variables studied in this present paper. Also, the participants belonged to 12 universities and not all the universities in Lebanon, therefore, the results might not represent the whole university students' community. Reporting bias is another limitation where self-reporting data were collected through the questionnaire. Due to the large sample size, conducting an accurate clinical psychological assessment of the participants rather than relying on an epidemiological tool was di cult.
The results of the study cannot be extrapolated on emerging adults who are out of universities. Despite all these limitations, exploring the effect of life skills and healthy habits is very essential to better identify the social determinants factors affecting university students' mental health in order to design a proper health promotion intervention to mitigate the burden of diseases among emerging adults.

Declarations
Consent for publication: All participants have received consent form prior to the questionnaire.
Availability of data and material: The datasets during and/or analysed during the current study available from the corresponding author on reasonable request.
Competing interests: No con ict of interest Ethics approval and consent to participate: All the participated universities have shared with us their IRB approval based on Belmont report.

Funding:
No funding Authors' contributions: Mrs. Diana Maddah has analyzed the data and has written the nal paper, Dr. Tamar Kabakian has contributed in paper writing, Dr. Rouba Zeidan has contributed in data analysis and interpretation, Ms.
Nathalie Al Saady has collected and entered the data, Dr. Nael Alami has contributed in the discussion and article writing, Dr. Pascale Salameh has edited and supervised the whole process of paper preparation and writing.