College Students’ Well-being during the COVID-19 Pandemic: A Systematic Review of the Literature

Studies have shown increased levels of distress during the coronavirus disease-19 (COVID-19) pandemic, and college students are becoming more recognized as a vulnerable population . This narrative systematic review aims to synthesize the current understanding of mental health, lifestyle, and socioeconomic impacts that the pandemic had on college students in the United States. A search was conducted on PubMed, PsycInfo, and Web of Science. A total of 34 observational studies were included which examined aspects of college students’ health and experiences during the COVID-19 pandemic in the United States. A great deal of students was shown to experience a moderate level of stress and subsyndromal depression and anxiety in the early months of the COVID-19 pandemic. Several risk and protective factors have been characterized. Students experienced various academic, financial, and housing disruptions. Studies have highlighted the need for institutional support to reduce the adverse psychological impact of the pandemic. There is a need for further large-scale research to assess the scope of COVID-19-related biopsychosocial impact, especially in vulnerable populations such as racial/ethnic and sexual/gender minorities.


Introduction
The coronavirus disease-2019 (COVID-19) pandemic abruptly changed people's daily routines and continues to affect how people live across the globe. Self-care includes various behaviors, such as healthy eating, exercise, relaxation, social connection, spirituality. Many of these aspects have been affected by restrictions and changes instated during the COVID-19 pandemic that may have prevented people from practicing their usual health-related behaviors or added a substantial amount of uncertainty.
Several studies have been conducted across the globe to study mental health in young adults (18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29) years old) and college students. The current understanding is that young adults are suffering from relatively high rates of perceived stress and that symptoms of depression and anxiety are common psychological reactions to the pandemic (Rajkumar, 2020;Xiong et al., 2020). At baseline, the college student population is particularly vulnerable to stress due to concerns related to academic performance, finances, interpersonal relationships, and post-graduation plans (Beiter et al., 2015). Psychosocial challenges are therefore compounded when faced with public health emergencies.
There is a growing volume of literature available on this topic. It is important to evaluate the current evidence and determine trends in results to inform future research and interventions. The purpose of the study is to identify manifestations of psychological distress among college students, common coping behaviors, and challenges to managing self-care. The current review was designed to summarize the existing literature addressing health concerns among college students during the COVID-19 pandemic.
Limitations to the existing literature should be addressed in future research. To our knowledge, this is the first review of the literature on this topic in this population to date.

Method
This article is a narrative review of the existing literature on mental health symptoms and behaviors during the COVID-19 pandemic.

Information Sources & Eligibility Criteria
Records identified from: PubMed (n = 43) APA PsychInfo (n = 19) Web Of Science (n = 72) Records removed before screening: Duplicate records removed (n = 32) Records screened (n = 101) Records manually excluded due to location outside of United States (n = 40) Reports sought for retrieval (n = 64) Reports not retrieved (n = 1) Reports assessed for eligibility (n = 63) Reports excluded: Incorrect article type (n = 17) Lack of relevance (n = 8) Sample not limited to college students (n = 4) Studies included in review (n = 34)

Sample Characteristics
The number of subjects in the studies included ranged from 49 to 5,547, with an average of 766. 25 studies took place at one institution alone. Many of these institutions are located in the South (n=11) and Northeast (n=9) regions of the country. Two studies were conducted across more than one institution (Browning et al., 2021;Eden et al., 2020). Six studies recruited nationwide samples via social media, such as Facebook and Instagram (Cohen et al., 2020;Conrad et al., 2021;Firkey et al., 2021;Gonzales et al., 2020;Hoyt et al., 2021;Melcher et al., 2021).
All studies recruited full-time students and were typically open to all programs of study and both undergraduate and graduate levels. One study included examined only social work students (Lawrence et al., 2021). Articles on students in other health professions, such as medicine, nursing, and pharmacy, were typically commentaries or took place outside of the United States and were therefore excluded.
Overall, two surveys included an interview component that was conducted virtually (Melcher et al., 2021;Son et al., 2020).
Most survey questionnaires (n=29) included at least one standard validated scale to assess for symptoms of depression or anxiety or experiences of stress. Other survey questions addressed other aspects of student experiences such as stress-reducing behaviors, socioeconomic and academic disruptions, and COVID-19-related concerns. Likert scales were used more often than open-ended free response answer formats.
Generalized Anxiety Disorder Scale. Anxiety symptoms were assessed by ten studies using the 7-item Generalized Anxiety Disorder Scale (GAD-7) (Biber et al., 2020;Conrad et al., 2021;Fruehwirth et al., 2021;Hoyt et al., 2021;Krendl, 2021;Lawrence et al., 2021;Lechner et al., 2020;Melcher et al., 2021;Wang et al., 2020;Zhang et al., 2020). This questionnaire is used in primary care settings to screen for generalized anxiety disorder and has demonstrated excellent internal consistency with a Cronbach alpha of 0.92 . Participants were asked to rate on a 4-point Likert scale how often they experienced anxiety symptoms over the past two weeks. Possible scores range from 0 to 21. Scores from 6 to 10, 11 to 15, and 16 to 21 indicate mild, moderate, and severe anxiety, respectively.
The PHQ-4 is a shorter 4-item screening tool that comprises of two screening questions from each of the previous scales. The PHQ-4 was used by three studies (Eden et al., 2020;Haliwa et al., 2021;Hoyt et al., 2021).
Perceived Stress Scale. Levels of stress were assessed by nine studies using the Perceived Stress Scale (PSS) (Eden et al., 2020;Hathaway et al., 2021;Hoyt et al., 2021;Kecojevic et al., 2020;Krendl, 2021;Melcher et al., 2021;Son et al., 2020;Trammell et al., 2021;Zhen et al., 2021). The 10-item PSS (PSS-10) has demonstrated good internal consistency with a Cronbach alpha of 0.84-0.85 in college student samples as well as good validity with a positive correlation with scores on life event scales (Cohen et al., 1983). Participants were asked to rate on a 5-point Likert scale how often they felt or thought a certain way in the past month. Items numbered 4, 5, 7, and 8 were reverse scored. Possible scores range from 0 to 40. PSS does not have official score cutoffs for classification of severity, but higher scores indicate a higher level of perceived stress.
Other scales were administered relatively frequently. The Multidimensional Scale of Perceived Social Support was used by four studies to assess for perceived support from family, friends, and significant others (Fruehwirth et al., 2021;Haliwa et al., 2021;Hathaway et al., 2021). The Depression Anxiety and Stress Scale is a widely used screening tool that was administered in three studies, only one of which supplemented it with PHQ and GAD (Haliwa et al., 2021;Kibbey et al., 2021;Melcher et al., 2021). The Brief Resilience Scale was used by three studies to measure resiliency (Eden et al., 2020;Fruehwirth et al., 2021;Hunt et al., 2021). The University of California, Los Angeles Loneliness Scale was used by three studies that characterizes how much a person feels socially disconnected from others (Conrad et al., 2021;Krendl, 2021;Melcher et al., 2021).

Thematic Analysis
Four broad themes were identified across the 34 publications and were used to organize the review: (a) mental health in the general student population, (b) mental health in specific subgroups, (c) health-related behaviors, and (d) socioeconomic implications.

Mental Health in the General Student Population
Review of titles, abstracts, and methods revealed that 14 studies investigated mental health as the primary research topic, with the objective to determine the severity of psychological distress and its manifestations as symptoms of depression or anxiety.
Almost all 14 publications showed that students overall were experiencing high levels of distress in the early stages of the pandemic. The frequency of students expressing increased levels of stress ranged from 14.6% (Browning et al., 2021) to as high as 71% Wang et al., 2020). Average PSS10 scores ranged from 18.8(4.9)   Protective factors for psychological health were investigated as well. Higher levels of mindfulness and social support were identified as protective factors for mental health in terms of reporting less perceived stress, greater happiness, and greater life satisfaction (Haliwa et al., 2021;Szkody et al., 2020;Zhen et al., 2021). Personality traits were also found to contribute to the ability to cope with major stressful events. For example, extraversion was negatively associated with mood indices but positively associated with engagement in healthy activities (Rettew et al., 2021).
Stress was associated with worse mental health outcomes during the pandemic. Symptoms of depression and anxiety were universally reported at high rates. Among all publications, PHQ scores

Mental Health in Specific Subgroups
Five publications focused on specific subgroups regarding demographics such as race/ethnicity and sexual/gender identity as the primary research topic ( One study found no statistically significant ethnic/racial differences in perceived stress, perceived COVID-19-related depression, health-related behaviors, or overall COVID-19 mental health impacts (Trammell et al., 2021). In this study, however, Hispanic/Latinx and Asian students perceived higher COVID-19-related threat and held more negative health beliefs than White students; Latinx students reported higher financial and resource impact scores than White students; and Asian students were significantly more likely than White students to report witnessing or experiencing discrimination.
One study found that relatively few participants, most of which identified as Asian, reported experiencing racial discrimination in April 2020 (Cohen et al., 2020). When surveying ethnically diverse students, the majority of these individuals personally experienced or witnessed discrimination individuals. These are notable findings, as one study found that Asian students experienced higher levels of psychological impact compared to White students (Browning et al., 2021). On the contrary, another study found that Asian students reported fewer anxiety symptoms than White students (Hoyt et al., 2021).
Female and sexual/gender minority individuals were shown to be at highest risk for increases in depression and anxiety across longitudinal studies (Fruehwirth et al., 2021). Two studies focused on members of the lesbian, gay, bisexual, transgender, queer, and/or questioning (LGBTQ) community (Gonzales et al., 2020;Hunt et al., 2021). Both studies found that nonbinary individuals reported higher psychological distress than binary individuals. Almost half of LGBTQ college students (45.7%) reported having families that do not support or know of their LGBTQ identity, and these students were more likely to experience frequent distress when compared with students who reported having supportive families (Gonzales et al., 2020). The majority of these respondents (58.5%) indicated that they would be able to receive mental health care if needed. However, 12.8% of respondents indicated that stay-at-home orders prevented their ability to receive mental health care services.

Health-related Behaviors
Behavior changes in response to the pandemic were found to be pervasive throughout the studies.
Students almost universally adopted public health recommendations of physical distancing, regular hand hygiene, and face protection (Cohen et al., 2020;Kecojevic et al., 2020;Szkody et al., 2020). In qualitative sections of cross-sectional surveys, students mentioned several coping strategies such as ignoring the news, using other media, turning to social support, drinking alcohol, meditating, and Although college students have recently shown higher rates of disordered eating behaviors as compared to the general population (Rodgers et al., 2020), change in eating patterns during the pandemic was not studied extensively in the included articles. Few studies highlighted examples such as inconsistent eating, increased or decreased appetite, and emotional eating Wang et al., 2020).
www.scholink.org/ojs/index.php/grhe Global Research in Higher Education Vol. 4, No. 4, 2021 Four studies examined screen time and use of media as a coping tool (Browning et al., 2021;Eden et al., 2020, Kecojevic et al., 2020Zhen et al., 2021). The intensity of anxiety was associated with spending more than one hour per day searching for information on COVID-19 in April 2020 (

Socioeconomic Implications
Several studies focused on the psychosocial impact of the pandemic on areas such as disruptions in income, housing, food security, and academic experiences. A considerable portion of students reported a reduction in household income (Cohen et al., 2020;Firkey et al., 2021;Hoyt et al., 2021;Jones et al., 2021;Owens et al., 2020;Son et al., 2020 Son et al., 2020). It is currently unclear to what extent these consequences will affect young adults entering the workforce.

Summary of the Evidence
The purpose of this review is to summarize the current understanding of well-being in college students during the COVID-19 pandemic. With school closures and stay-at-home orders, students have had to adapt to new environments, such as virtual classes and meetings and different living situations. National and regional studies across the United States have examined the mental health burden among students in higher education. Thematic analysis revealed several emerging themes from challenges that students face and their attempts to cope with hardships.
Students endured employment losses, housing changes, and difficulty meeting basic needs on top of the transition to online learning. Generally, a great deal of students was shown to experience a moderate level of stress as soon as American institutions started addressing the outbreak. individual, and those with clinically severe psychiatric conditions may benefit from seeking professional treatment. The increased need for mental health services suggests that universities should seek ways to increase access to these services for their students.
Several studies found increased severity of stress and psychiatric symptoms in females and LGTBQ individuals. However, females were overrepresented in several samples, and this imbalance may skew analyses. There is mixed evidence on racial/ethnic disparities in mental health burden, and researchers sought to examine these nuances in different ways. While some aspects of psychological responses to the pandemic may have been universal (Trammell et al., 2021), racial discrimination and financial impact was more likely to be reported among certain racial/ethnic groups. In particular, Asian, Black,  Wang et al., 2020) during the pandemic. Disturbances in sleep and eating patterns are also common symptoms of depression, but these may also be affected by external factors such as housing and food insecurity.
One disadvantage is that these results were typically collected in cross-sectional studies, which lack an objective baseline, but one study was able to record self-reports of substance use in real time via EMAs (Kleiman et al., 2020). Under-reporting is a concern for self-reporting of drug use (Bone et al., 2016).
There is not enough evidence to conclude long-term implications of these behavior changes, however and reduce rates of survey completion. However, free responses have provided a more complex and human narrative of student experiences. All studies relied on self-reported data and could be susceptible to recall bias. EMAs have the advantage of frequent sampling over time, which can minimize recall bias and maximize validity by capturing data in real time (Shiffman et al., 2008).
EMAs require smartphone compatibility, and caution must be taken to ensure data privacy.

Limitations
Limitations apply to this review. Inclusion criteria were relatively liberal to capture the wide perspective of available literature in the United States; specifically, a quality assessment tool was not used. Preliminary articles were not included, and the current study may not have captured latest research on data from the later stages of the outbreak. Substantial heterogeneity of methods (i.e., assessment tools, primary outcomes measured, data analysis) and data reporting was noted across the included studies. Authors of publications were not contacted for raw data or confirmation of results. All articles relied on self-reported data, thereby increasing the risk of recall bias, particularly in cross-sectional studies and assessments using large retrospective timeframes.
Owing to the breadth of the chosen research topic, many factors were not assessed in depth in this review, including somatic manifestations of stress (i.e., headaches, change in bowel habits), caretaking responsibilities, first-generation college student status, and domestic violence. These are important physiologic and socioeconomic characteristics to consider in future research. Other demographic features, such as city of residence and year in program of study, were not analyzed in the review.
In the current review, quantitative synthesis was not conducted, and study findings were summarized in a narrative fashion. As with observational studies generally, causality cannot be concluded from studies included in this review. Lastly, overrepresentation of female and White students was noted in several studies, introducing concern for sampling bias and indicating that presented findings may not be representative of the real collegiate population. Large population-based studies may suffer from ecological fallacy, and inferences about individuals cannot be made based on aggregate data.

Conclusions
College students in the United States have experienced a great deal of biopsychosocial stress during the era of COVID-19. There is concern for a -dual pandemic‖ pertaining to mental health and social isolation. The volume of literature on well-being in young adults during these times is expanding globally from various disciplines. University students are becoming increasingly recognized as a vulnerable subpopulation suffering from high rates of depression, anxiety, and substance use. These conditions are all risk factors for suicide and should be taken seriously.
The mental health burden of American collegiate is extensive, certain subgroups are vulnerable to greater impact, students have changed their behaviors in many ways, and the pandemic has impacted student's financial and living situations. Some demographic subgroups (e.g., sexual/gender and Studies included in the review took place in the spring and summer of 2020, with the latest data collection ending in August 2020. As COVID-19 vaccinations become more widely available, it will be helpful to assess how vaccination status affects mental health and behaviors as well. As the pandemic continues to evolve, ongoing psychological and economic effects of the pandemic should continue to be monitored. There is a need for more high-quality large-scale longitudinal studies that examine differences in mental health outcomes based on risk factors such as racial/ethnic and sexual/gender identity as well as socioeconomic status. More consistent survey measures, particularly highly validated quantitative assessments of psychosocial profiles, should be utilized. The long-term impact of COVID-19-related stress on physical health may take months to years to become fully apparent. Chronic distress has been associated with an increased risk of chronic diseases such as hypertension, diabetes, and cardiovascular disease. Early detection of these problems at the population level is essential to optimizing management, identifying disparities, and informing health promotion and disease prevention strategies. Health care systems and educational institutions should be prepared to address socioeconomic impacts beyond the immediate academic terms. There was a negative correlation between anxiety and optimism (r=-0.36), gratitude (r=-0.12), and perceived effectiveness of academic instruction (r=-0.11). Average GAD7 score of all respondents was 6.96(6.27).

Appendix A. Study Characteristics
Students with higher GAD7 scores perceived lower effectiveness of institutional response to COVID-19 when compared to mildly or moderately anxious respondents [F(2,1516) In open-ended responses, respondents reported symptoms of depression (27.3%) and anxiety (17.4%) and perceptions of stress (14.6%). Bivariate associations showed students who were female, were Asian, in fair/poor health, of below-average relative family income, or who knew someone infected with COVID-19 experienced higher levels of psychological impact. Students who were White, higher socioeconomic status, spent at least two hours outside, or less than eight hours on electronic screens were likely to experience lower levels of psychological impact. Multivariate modeling showed that being a female, having fair/poor general health status, spending 8 or more hours on screens daily, and knowing someone infected predicted higher levels of psychological impact when risk factors were considered simultaneously. Anxiety was associated with increased use of media (p<0.05). Stress was associated with more hedonic (p<0.05) and less eudemonic (p<0.05) media use.
Avoidant coping was associated with poorer mental health, whereas humor coping was associated with better mental health. Resilience positively moderated the effect of stress on problem-focused, avoidant, reframing, and humor coping. Prevalence of moderate to severe depression increased from 21.5% in Fall 2019 to 31.7% in Spring-Summer 2020 (p<0.001). Prevalence of moderate to severe anxiety increased from 18.1% to 25.3% (p<0.05). White, female, and sexual/gender minority students were at highest risk for increases in anxiety symptoms. Black, female and sexual/gender minority students were at highest risk of increases in depression. Loss of work and COVID-19 diagnosis or hospitalization of oneself, family members, or friends were not associated with increases in depression or anxiety symptoms. The majority of respondents reported experiencing depressive symptoms (60.4%) anxiety symptoms (65%) and frequent distress (61%). Compared to cisgender men, transgender students were much more likely to report frequent distress (aOR=3.41, CI 1.31, 8.86). 45.7% of LGBT college students have immediate families that do not support or know of their LGBT identity.

Hoyt et al., 2021
Levels of anxiety and perceived stress were higher in April than in July 2020. Average GAD7 scores decreased from 10.49(5.95) to 9.85(6.04). Average PSS10 scores decreased from 22.72(9.00) to 20.36(8.06). Black and mixed race/ethnicity individuals were the only groups to show increased anxiety from April to July. Women reported worse well-being compared with men. Gender diverse and sexual minority youths reported worse outcomes than their cisgender heterosexual peers at both time points.
Qualitative data illustrates educational, economic, and environmental stressors. Students reported several difficulties and high levels of perceived distress with a total average PSS10 score of 20.6(7.3). High levels of depression were associated with difficulties in focusing on academic work and employment losses. High levels of anxiety were more likely to be reported by upperclassmen and those who spent more than one hour per day searching for information on COVID-19. Difficulty focusing on academic work was associated with higher levels of somatization. Respondents with higher perceived stress were more likely to be female, unable to focus on academic work, and report difficulties obtaining medications and cleaning supplies. There was a statistically significant increase in depression and anxiety scores in May through July 2020 when compared to retrospective self-reporting of scores for February 2020. Average PHQ9 increased from 3.63 to 9.10 (p<0.0001). Average GAD7 increased from 4.03 to 9.11 (p<0.0001). There was a significant increase in screen time during the pandemic (p<0.0001). Alcohol consumption did not significantly increase during the pandemic (p=0.06) and moderate-to-vigorous physical activity did not significantly decrease during the pandemic (p=0.21). Bivariate analyses showed a positive association between depression and anxiety with screen time and the transition to virtual classes. 71% of respondents reported increased stress and anxiety during the pandemic.
Total average PSS10 score was 18.8(4.9). Common stressors include fear and worry about their health and of loved ones (91%), difficulty concentrating (89%), decreased social interaction due to physical distancing (86%), and increased concerns on academic performance (82%). Participants have adopted a variety of coping strategies. A vast majority of participants who indicated an increase in stress and anxiety (93%) reported that they had not used school counseling services during the pandemic.  Tasso et al., 2021 There was a positive relationship between academic frustrations and mental health symptoms. Worries about becoming infected with COVID-19 were positively correlated with mental health symptoms and negatively correlated with trust in the government. 49% and 53%of respondents reported an increase in PHQ9 and GAD7, respectively, over the course of the study. Greater late-night online activity volumes were positively correlated with increasing depression (r ranging between 0.32 and 0.75, p<0.04) and anxiety levels (r ranging between 0.39 and 0.74, p<0.006).