Association Between Social Support and Mental Health Status Among College Students: A Nationwide Survey

Background Social support could improve college students’ mental health status. However, mental health varies across different geographic areas, and most of previous studies only included small sample sizes, thus the nationwide survey are warranted. This study sought to estimate the association between social support quality (SSQ) and mental health among Chinese college students using data yielding from a nationwide survey. Methods A cross-sectional study with a web-based questionnaire survey was conducted in Wuhan, China, from October 2017 to February 2018. A total of 11093 college students were recruited from 18 colleges or universities through multi-stage random cluster sampling method. The Multidimensional Scale of Perceived Social Support scale and 12-items General Health Questionnaire were used to estimate students’ SSQ and mental health status, respectively. Logistic regression analysis was used to identify the association between SSQ and the mental health disorder in college students. Results A total of 10,676 (96.24%) college students were included in this study. Of them, 21.4% were identied as having a mental disorder. Students being a female, aged 18-22 years old, whose mother held a college degree and above, and drinking alcohol had more risks of being mental health problems (P <0.05). However, having general or higher family economic levels, working and resting regularly, and sleeping ≥ 7 hours could be preventive factors for mental health problem (P <0.05). Especially, with the improvement of SSQ, there was a decreasing trend in the risk of being mental disorder. important public challenge among college students in China. Improving SSQ could be the practical and valuable method to prevent college students’ mental problem.

from each college/university.) Ultimately, a total of 11093 questionnaires were collected on a computer terminal. After excluding those questionnaires completed in less than ve minutes, 10676 quali ed questionnaires were included in further statistical analyses, yielding a 96.24% quali cation rate.
This study was approved by the ethics committee of Tongji Medical College institutional review board, Huazhong University of Science and Technology, Wuhan, China. All participants signed informed consent before lling out the questionnaire.

Social support
Multidimensional Scale of Perceived Social Support (MSPSS) [12] consisted of 12 items with response options scoring from 1 (very strongly disagree) to 7 (very strongly agree). It estimated SSQs from three sources: family (including item 3, 4, 8, and 11), friends (including item 6, 7, 9, and 12) and signi cant others (including item 1, 2, 5, and 10) [13]. All item scores were added up and then divided by 12. The mean scores ranging from 1 to 2.99, 3 to 5 and 5.01 to 7 were classi ed as low, medium, and high perceived support levels, respectively [12]. MSPSS had a sound factorial validity (with Cronbach's alpha coe cients of 0.953), and internal consistencies for the full scale and subscales were both satisfactory [14]. The Chinese version has been suggested as a reliable tool for assessing SSQ [15].

Mental health status
The 12-items General Health Questionnaire (GHQ-12) [16] was one of the most widely used abridged versions measuring mental health disorder. It included 12 items corresponding to three dimensions: anxiety/depression (including item 1, 2, 7, and 10), social dysfunction (including item 3, 4, 5, 6, 8, and 9) and de ciency of con dence (including item 11 and 12) [17]. There were four answers ranged from "better/healthier than normal" to "much worse/more than usual". The GHQ scoring method (the four options were scored by 0-0-1-1, respectively) was adopted in our study [18]. Higher score corresponded to worse mental health status. A total score of 4 or more was classi ed as having notable mental problem [19]. GHQ-12 had satisfactory reliability (with Cronbach's alpha coe cients of 0.886) and extensive sensitivity, which had been previously validated.

Demographics
The questionnaire included the following demographic variables: age, gender, ethnicity, religious belief, residence area, single-parent/-child family or not, paternal/maternal education level, family economic status. Lifestyle variables referred to physical exercise, regular work-rest or not, sleep duration, smoking and alcohol drinking status.

Data Analysis
Data analyses were performed using the SPSS software (Version 22 for Windows, SPSS Inc, Chicago, IL, U.S.A.). Descriptive analyses included means (standard deviations [SDs]) for continuous variables and percentages for categorical data. We analyzed the demographic characteristics of the respondents, and compared the differences between SSQs and mental health statuses among various demographics by χ 2 tests. Potential in uencing factors of psychological disorder were identi ed via multi-class logistic regression analysis. Furthermore, we described the correlation between MSPSS and GHQ-12 by matrix analysis and estimated relationships between SSQs and mental problems under different adjustments using trend analysis. Signi cance level was accepted as P<0.05 (two-tailed) for all tests.

Results
Participants in this study had a mean age of 19.66 (SD=2.22), and 56.7% were female. 2,284 (21.4%) students were identi ed as having a mental disorder. The results of χ 2 tests suggested that SSQ was associated with gender, age, religious belief, residence area, from single-parent or -child family or not, paternal or maternal education level, and family economic status (P <0.05). Statistical signi cantly lower SSQs (score 1-2.9) were reported in males, in students aged 18-22 years old, and those having religious beliefs. Lower SSQs were also found in students living in rural areas, from single-parent family or non-single-child family, whose parents' (both paternal and maternal) education levels were primary or below, and those from families with lower economic statuses (P <0.05) ( Table 1).
For mental disorder, gender, age, ethnicity, religious belief, residence area, from single-parent or -child family or not, paternal or maternal education levels, and family economic status were considered as potential relevant factors according to the results of χ 2 tests (P <0.05). And females, students aged 18-21 years old, minority students, students having religious beliefs, living in rural areas, from single-parent or non-single-child families, whose parents' education levels were primary or below and with lower family economic status were more likely to have mental problems (P <0.05).
Regression analysis indicated that gender, age, maternal education level, family economic status, work-rest routine, sleep duration, alcohol drinking and SSQ had potential in uences on students' psychological status (P <0.05). Of them, female, aged 18-21 years old, maternal education level was college and above, drinking alcohol were independently associated with poorer mental health statuses (odds ratios (ORs) were between 1.245 and 1.362, all P <0.05). While middle and higher family economic status, regular work-rest routine, sleep duration ≥7 hours, having moderate and high social support quality were more likely to have better mental health conditions (ORs ranged from 0.245 to 0.709, all P <0.05) ( Table 2).
According to results of regression analysis, SSQ was demonstrated as one of the most signi cant in uencing factors for students' mental health status (moderate vs. low: OR=0.521, 95% CI=0.383-0.708; high vs. low: OR=0.245, 95% CI=0.180-0.333). In order to thoroughly analyze the relationship between college students' SSQ and psychological problem, we conducted further analyses. The correlation matrix presented in Table 3 revealed that SSQ was negatively associated with mental problem (r=-0.184). Among the three social support sources, family provided the strongest effect on mental disorder (r=-0.182), then followed by friends (r=-0.167) and signi cant others (r=-0.157). Within the MSPSS, the family, friends and signi cant others subscales were highly correlated with each other (r between 0.586 and 0.717) and with the overall scale (r between 0.787 and 0.835).
In addition, we explored the relationships between SSQ and students' mental disorder under different adjustments. Compared with low SSQ, both moderate and high SSQs reduced the risk of mental disorder (ORs ranged from 0.184 to 0.526, P <0.05). In different models of each subscale, there were signi cant differences in the effects of different SSQs on mental health statuses (P fordifference all <0.05). Furthermore, in all three models with different adjustments, there were signi cant positive trends in the associations between both the full scale and subscales of SSQ and mental disorder (P for trend all <0.001). Especially, in model 3, with the full adjustment, both higher and moderate SSQs had greater negative impacts on mental disorder than low social support (OR=0.249, 95% CI=0.183-0.338; OR=0.526, 95% CI=0.386-0.715, respectively). In subscales, family support had the strongest preventive effect on students' mental disorder (high/moderate SSQ=0.412/0.219), then followed by friends support (0.434/0.234) and signi cant others supports (0.524/0.285) ( Table 4).

Discussion
A total of 10676 college/university students were included in this study and the prevalence of mental disorder was 21.4%. Although this result was lower than the nding (28.0%) from a survey conducted in Finnish university students [20], it was a little higher than results of a cross-national mental health survey (20.3%) [21] and another one (19.0%) from Hungary [22], all above studies were conducted among college students using the GHQ-12. These differences might be attributed to racial, cultural diversities, and socioeconomic disparities [23,24]. With the increase of SSQ, the risk of being mental disorder among college students showed a signi cant decreasing trend. It suggested that improving the SSQ could be an effective and practical method to prevent and treat mental disorders of college students.
The high prevalence of psychiatric disorder in college/university might due to the fact that most of the students were just entering adulthood, it was a crucial time for personal identity development. In this period, they were generally sensitive to the shift of surroundings, such as changes in living environment and learning style, crisis of interpersonal relationships, which was easily to cause psychological uctuations and mental problems [25,26]. Another possible explanation was that entering college/university followed by a series of unexpected challenges for many students. Apart from considerable pressure related to academic load, they had to bear more adult-like responsibilities, even though they had not yet captured enough cognitive maturity and developed foundational skills required for adulthood [27]. For example, college students might face with stresses from living with housemates with different lifestyles or personalities, graduating and looking for work, and dealing with love affairs. All of these might induce psychological problems, which were harmful to their health, academic performance, and even individual development [28]. Therefore, free psychological counseling centers were suggested to set up in campus, so that more professional resources could be offered to help students prevent and resolve psychological problems[29].
Students' mental disorder was associated with various factors. Similar to previous studies [11], gender was one of important factors in uencing college students' mental health in our study, as the female generally lived with more delicate emotions and were prone to have negative views of their health than males, they were more likely to have mental problems. Compared with the older college students, the freshmen, mostly aged 18-21, faced confused lifetime planning, cash-strapped living, and less social experience, which made them anxious and stressful. Mother played a pivotal role in the development of children's mental health[30], but we found mother with college degree or above was a risk factor for students' mental health. Higher academic degree generally corresponded to longer work time and heavier work stress, which deprived the time and energy to accompany with children. In this case, mothers might ignore mental problems in their children and delay their psychological counseling. In addition, students who drank alcohol tended to have decreased sensitivity, intense emotions, and interpersonal con icts, and eventually increased the risk of bad mood [31,32]. Nevertheless, the function between alcohol drinking and mental problem could be bi-directional, college students also opted to cope with their depression, stress and other mental illnesses by drinking.
Higher family economic status could protect students from being mental problems, which aligned with previous evidences that poorer family economic status was associated with lower quality of life and less warmth from parents[33]. Coupled with the increasing self-esteem of college students, the gap between subjective demands and objective facts contributed to psychological imbalance and ultimately induced mental problem [34] . Students followed regular workrest schedule had signi cant lower risk of mental disorder than those who never/seldom work-rest regularly. This nding was partly supported by a study based on the UK Bio-bank that circadian disruption was reliably associated with various adverse mental health outcomes [35]. However, as its study sample aged 37-73 years old, the result might be not applicable for college students, further prospective studies and random controlled trials were needed to con rm their relationship. Correspondingly, sleep seven hours and longer would decline the risk of mental disorder, which was supported by a recent meta-analysis [36]. Adequate sleep could preserve the homeostasis of affective brain, and optimally prepare next-day emotional functioning, leading to a stable and healthy mental status[37].
Compared with low SSQ, moderate and high SSQs were protective factors for students' mental health. From the correlation coe cient matrix analysis, we found that SSQ and its subscales were negatively correlated with mental disorder, which were in consistent with previous studies [38]. Under different levels of adjustment, medium and high SSQs and their subscales were all negatively associated with mental disorder when compared to low SSQ. Adequate highquality social supports could give individuals comfortable mental consolation and the sense of security, which bene ted college students to keep a healthy psychological status [39]. In general, family was the most important source of social support, followed by friends and signi cant others. Because of the conservative family values and collectivist nature of Chinese society, family ties were deemed as the most important social relationships [40]. Generally, people a liated with mental disorder often create the feelings of stigma or shame. In this case, the family should play a key role in guiding family members with psychological problems to receive treatment, and making patients healthy through active interventions[41]. However, several studies listed friends as the most important source of supports, ahead of family and signi cant others [42]. The reason might be that most college students lived with friends in the campus rather than their family members, friends could timely nd psychological change of each other and social supports from friends could offer su cient mental assistance. Finally, multinomial logistic regression analyses suggested that SSQ could improve college students' mental health, regardless of from family, friends or signi cant others. This result informed that university administrators and teachers should strive to improve students' SSQs, especially the supports from family and friends, so as to prevent and treat mental problems in college students better.

Strengths and limitations
There were several strengths in this study. Firstly, we included 10,676 students in this study via the multi-stage random cluster sampling. It was a big sample size among relevant studies, guaranteeing the results were credible to some extent. Secondly, we obtained some important ndings. For example, there was a decreasing trend of being mental problem with the improvement of SSQ. This result further supplemented and determined the in uencing factor network of college students' mental health, and provided statistical data for international comparisons around this topic.
However, some limitations also should be acknowledged. Firstly, the study was cross-sectional design, the results only suggested the observational correlation rather than the exact causality between SSQ and mental health. Secondly, students' SSQs and mental health statuses were collected through self-reported, which might be a potential source of information bias. However, as the questionnaire was lled out in anonymous forms, obtained data might be objective.
Furthermore, previous studies also suggested that there was quali ed agreement between online self-reported and measured data [43], indicating the bias derived from this method were permissible. Thirdly, this study was conducted in Wuhan city, which might limit the generality to other regions. However, college students in Wuhan were from across the country, which could make up for the lack of sample representation to some extent.

Conclusion
In conclusion, the SSQ was an critical factors associated with mental health. Compared with low SSQ, moderate and high SSQs signi cantly prevented mental disorder. Given college students was an important group, their mental health signi cantly correlated with the quality of social development, our ndings might provide valuable and practical clues for the prevention of mental disorder in college students.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this manuscript.

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

Funding
This study was supported by the National Social Science Foundation of China (Grant No. 18ZDA085). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Author contributions CW designed of the questionnaire, participated in the survey and analyzed the data, wrote the draft of the paper. SY, YYG and YG designed of the questionnaire, analyzed the data and cross-checked the paper. HJ and HX checked the draft of the paper and gave advice on statistical methodology. CL contributed to the design of the paper. ZL provided advices and supervision of the work. All authors read and approved the nal version.
Note. SSQ, Social support quality.  All correlation coe cients were statistically signi cant at the 0.01 level. Table 4. Multinomial logistic regression analyses for the associations between social support quality and mental health status.