Connect with us

Health

A longitudinal investigation of COVID‐19 pandemic experiences and mental health among university students – Stamatis – – British Journal of Clinical Psychology

 


Background

On 13 March 2020, two days after the World Health Organization declared COVID-19 a pandemic (World Health Organization, 2020), the U.S. government imposed the first mandatory quarantine order since the 1960s (Kempe & Benenson, 1965). Since then, the world has undergone unprecedented societal changes surrounding the COVID-19 experience. The significant disruptions to daily life, high uncertainty about the future, and stress across multiple domains (e.g., social, economic, educational, health-related) appear to have impacted well-being (Vindegaard & Eriksen Benros, 2020). A recent meta-analysis of global changes in depression and anxiety symptoms between 2020 and 2021 estimated increases of 33.7% and 28.4% in the prevalence of major depressive disorder and anxiety disorders, respectively, in the United States due to the effects of the COVID-19 pandemic (Santomauro et al., 2021), with similar increases worldwide (depression: 27.6%; anxiety: 25.6%).

While previous studies have established a link between the COVID-19 pandemic and poorer mental health (Santomauro et al., 2021), less research has investigated which specific aspects of the COVID-19 experience confer vulnerability for psychiatric symptoms. In the United States, the COVID-19 pandemic has disproportionately impacted minoritized groups such as Black, Latinx, Native American, and low-income communities (Anyane-Yeboa et al., 2020), as well as Asian communities subjected to anti-Asian hate crimes (Wu, Qian, & Wilkes, 2021). However, evidence is mixed as to whether this impact has extended to mental health. In one study, Latinx and Asian students reported higher COVID-19-related threat and negative beliefs than White students, although there were no racial or ethnic differences in mental health (Trammell, Joseph, & Harriger, 2021). Similarly, another study indicated a higher prevalence of pandemic stressors among Black and Hispanic respondents, but no difference in mental health indicators across racial or ethnic groups (Goldmann et al., 2021), and a third found that identifying as Black was associated with better mental health during the pandemic (Kantor & Kantor, 2020).

Additional sociodemographic and historical factors may also be associated with differences in mental health risk. Living in a larger home and being male have both been linked with lower risk for anxiety and depressive symptoms (Kantor & Kantor, 2020). Previous history with pandemics has also been found to play a role in predicting mental health status during the pandemic; for example, Choi et al. (2020) reported that Hong Kong residents who had experienced the SARS outbreak in 2003 reported better mental health during COVID-19. Personality and personal history, including loneliness and a history of hospitalizations, are additional factors associated with higher risk of faring poorly during COVID-19 (Kantor & Kantor, 2020). This growing literature indicates that it will be critical for us to better identify potential factors that might increase risk or resilience in response to a pandemic, which could have implications for individual- and community-level responses to national crises of this nature.

An important caveat of past studies examining risk factors is that the vast majority have relied on cross-sectional designs. With several notable exceptions (e.g., Magson et al., 2021), few have examined risk factors for the development of psychological symptoms over time, making it difficult to parse the timeline over which symptoms may unfold in relation to potential risk factors. An additional consideration is that most studies have been conducted with the general population (Abir et al., 2021; Bu, Steptoe, & Fancourt, 2020; Kantor & Kantor, 2020), and findings may not apply to specific sub-populations with increased vulnerability for negative mental health ramifications. In particular, there is a need to consider the impact of COVID-19 among young adults and college students.

Despite being less at risk for serious complications from COVID-19, young adults appear to be especially vulnerable to worsened mental health during the pandemic. In a large Web-based survey of Chinese adults, younger people reported a higher prevalence of depression and anxiety symptoms than older adults at the time of the initial outbreak (Huang & Zhao, 2020). Another study indicated that older adults perceived fewer COVID-19-related stressors and lower levels of depression and anxiety than younger adults, even after controlling for pre-pandemic depression (Bruine de Bruin, 2020). The relationship between pandemic-related experiences and mental health may be particularly strong among university students, who experienced a significant change in day-to-day experiences when colleges abruptly transitioned to online learning. Indeed, a high proportion of college students have experienced clinically significant depression (34.2–48.1%), anxiety (24.0–38.5%), and stress (50.4%) during COVID-19 (Cao et al., 2020; Odriozola-González, Planchuelo-Gómez, Irurtia, & de Luis-García, 2020; Wang, Hegde, et al., 2020), and these symptoms have increased by 42.5–63.3% during the first lockdown of the pandemic (Kaparounaki et al., 2020). While one study pointed to psychological protective factors among Chinese university students, such as living in an urban area, having a stable family income, and living with parents (Cao et al., 2020), there is a need to further investigate which specific aspects of the pandemic experience relate to mental health problems over time among college students. Moreover, it is important to consider additional mental health outcomes for young adults, such as alcohol and substance use, since COVID-19-related worry has been linked with substance use coping motives (Rogers, Shepherd, Garey, & Zvolensky, 2020) and externalizing symptoms (Copeland et al., 2021).

The overarching objective of the present study was to characterize the experiences of U.S. college students during the initial months of the pandemic, and to examine associations between these experiences and mental health symptoms – including depression, anxiety, stress, alcohol use, and substance use. We used a longitudinal approach to understand how psychological symptoms may have changed over the course of a month, between the start of the national quarantine period and the end of the semester. Given the limited literature on specific predictors of mental health responding during a pandemic, we elected a novel analytic method that helps with variable selection when theoretical background is limited. Namely, we opted to test associations between COVID-related factors and mental health using elastic net regression. Elastic net combines two regularization parameters, lasso and ridge penalties, to reduce potentially spurious associations (Zou & Hastie, 2005); thus, it stands to identify COVID-related predictors of mental health that are more likely to generalize to other samples.

Our first aim was to characterize the psychological health and early pandemic-related experiences of U.S. university students. Both mental health symptoms and pandemic-related experiences were assessed at the start of quarantine (Time 1: late March/early April) and at the end of the semester (Time 2: late April/early May). Various pandemic-related experiences were examined, including COVID-19-related worry and interference, actual exposure, adherence to guidelines, perceived structural support, behavioural responses, and expectations of future exposure and ongoing disruption to daily life. We assessed changes in mental health symptoms and pandemic experiences from Time 1 to Time 2, and we also examined whether mental health symptoms varied according to sociodemographic characteristics, given the disproportionate impact of COVID-19 on Black and Hispanic groups (e.g., Anyane-Yeboa et al., 2020).

Our second aim was to investigate which pandemic-related experiences, as _ at the start of quarantine, predicted mental health symptoms at the end of the semester. Using elastic net regression, we first tested the association between Time 1 pandemic-related experiences and Time 2 mental health symptoms; sociodemographic variables were also included as predictors. Follow-up analyses controlled for Time 1 mental health symptoms, which allowed us to consider symptom course. Our third aim was to consider associations between end of the semester pandemic-related experiences and Time 2 mental health symptoms to gain a better understanding how relationships may have shifted across time as the pandemic became more protracted. We repeated the same procedure as in Aim 2, first testing a series of elastic net regression models predicting Time 2 mental health from Time 2 pandemic experiences; the models were re-run controlling for Time 1 mental health where possible.

We hypothesized that in line with prior studies (e.g., Odriozola-González et al., 2020), a substantial number of students would report affective symptoms above clinical cut-offs, and that these levels would increase at the end of the semester, in line with self-reported increases in Greek college students during a similar period (Kaparounaki et al., 2020). We further predicted that individuals belonging to U.S. minoritized groups would be more strongly impacted by the pandemic (Anyane-Yeboa et al., 2020). Due to the limited literature on specific pandemic-related experiences as predictors of mental health, hypotheses for Aims 2 and 3 were largely exploratory in nature; however, we generally expected that students reporting greater pandemic-related worry and interference would endorse heightened levels of mental health symptoms at Time 2. Additionally, in line with previous literature (Leigh & Stall, 1993; Scott-Sheldon et al., 2016), we hypothesized associations between alcohol/substance use symptoms and increased risk behaviours (e.g., reduced adherence to local and national guidelines) at Time 2.

Method

Participants and procedure

Participants (N = 176) were undergraduate students at a large, private university in the South Florida region of the United States. The sample was recruited through the Psychology Department’s undergraduate participant pool (n = 107), and from the larger university student community (n = 69). Participants enrolled and completed the Time 1 survey in late March/early April 2020, following the university’s decision to shift to a virtual format in mid-March. Participants were then invited to complete the Time 2 follow-up survey in late April/early May 2020, which coincided with the end of the semester.

Prior to analysis, data cleaning procedures were performed. We excluded any responses that were completed in less than 60% of the expected completion time (n = 14), as well as responses in which less than 80% of embedded attention checks were answered correctly (n = 2). The final sample included 165 participants at Time 1 and 98 participants at Time 2. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Participants were predominantly female (67.7%), and ages ranged from 17 to 31 (M = 19.57, SD = 1.70). Approximately 55% of the sample identified as White, 15.6% as Black, 15% as Asian, and 10.8% as multi-racial or other. With respect to ethnicity, 24.6% of participants identified as Latino/a. For analyses, participants who identified their race as Black, Asian, multi-racial, or other were included as part of a Black, Indigenous, People of Color (BIPOC) group (39.4%). Participants subjectively ranked their socioeconomic status (SES) relative to others in the United States using the MacArthur SES ladder (Adler, Epel, Castellazzo, & Ickovics, 2000), on a 10-point scale from 1 (worst off) to 10 (best off). On average, participants reported believing they were better off than others (M = 6.73, SD = 1.7), although ratings spanned the entire scale range. In terms of political affiliation, the majority of the sample identified as either Liberal (38.3%) or Moderate (35.3%).

Measures

Sociodemographic Information

Participants completed four items to report their basic demographic information (gender, age, race, and ethnicity), as well as one item on their political views (1 = Very conservative, 5 = Very liberal). The MacArthur SES Ladder was used to measure SES (Adler et al., 2000).

Mental health symptoms

The Patient Health Questionnaire-9 (PHQ-9; Kroenke, Spitzer, & Williams, )

The PHQ-9 is a nine-item measure of depression severity that assesses the extent to which participants are bothered by depression symptoms over the past two weeks (e.g., ‘Little interest or pleasure in doing things’; ‘Feeling down, depressed, or hopeless’). Responses are recorded on a four-point Likert scale (0 = Not at all, 3 = Nearly every day). Scores are summed to create a total score ranging from 0 to 27, with higher scores indicating greater symptoms of depression. The clinical cut-off for moderate depression is ≥10 (Kroenke et al., 2001). In this sample, the PHQ-9 demonstrated good reliability at both Time 1 (α = .92) and Time 2 (α = .92).

The Generalized Anxiety Disorder 7-Item Scale (GAD-7; Spitzer, Kroenke, Williams, & Löwe, )

The GAD-7 is a seven-item measure of anxiety severity that measures the frequency of anxiety-related symptoms over the past two weeks (e.g., ‘Feeling nervous, anxious or on edge’). Responses are recorded on a four-point Likert scale (0 = Not at all, 3 = Nearly every day). Scores are summed to create a total score ranging from 0 to 21, with higher scores indicating higher levels of anxiety. The clinical cut-off for moderate anxiety is >10 (Spitzer et al., 2006). In this sample, the GAD-7 demonstrated good reliability at both Time 1 (α = .93) and Time 2 (α = .94).

Impact of Events Scale-6 (IES-6; Thoresen et al., )

The IES-6 is a six-item measure of the extent to which participants have been bothered by stress-related difficulties over the past week (e.g., ‘I thought about it when I didn’t mean to’) following a traumatic event. The IES-6 used in this study was modified to specifically reflect stress responses due to the COVID-19 pandemic (i.e., the instructions read: ‘Below is a list of difficulties people sometimes have after stressful life events. Please answer the following questions with regard to the impact of COVID-19’). This tailoring of the instructions is similar to that used in other published reports (Di Crosta et al., 2020). As with the original IES-6, responses on the modified COVID version are recorded on a 5-point Likert scale (0 = Not at all, 4 = Extremely). Scores are summed and then averaged to create a total IES-6 score ranging from 0 to 4, with higher scores indicating higher levels of pandemic stress reactions. The clinical cut-off for the original IES-6 is ≥1.75 (Hosey et al., 2019). In this sample, the IES-6 demonstrated adequate reliability at both Time 1 (α = .87) and Time 2 (α = .85).

World Health Organization (WHO) Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST; WHO ASSIST Working Group, )

We used a modified version of the WHO’s ASSIST to measure alcohol and substance use disorder symptoms. Participants were asked to reflect on their alcohol and substance use over the past three months, answering four questions related to each. For alcohol use, participants first indicated how often they consumed 4 or more (for women) or 5 or more (for men) alcoholic drinks in a day. They then reported how frequently they (1) experienced a strong desire or urge to use; (2) had use-related health, social, legal, or financial problems; and (3) failed to do what was normally expected of them because of their use over the past three months. All four items were rated on a 0 to 4 scale (0 = Never, 1 = Once or Twice, 2 = Monthly, 3 = Weekly, 4 = Daily or Almost Daily) and were summed to create a total alcohol use disorder symptom score ranging from 0 to 16. For substance use, participants first indicated how often they used illegal substances over the past three months, followed by the same three questions pertaining to urges, use-related problems, and failure to fulfil expectations. The four substance use items were rated on the same 0 to 4 scale and were summed to create a total substance use disorder symptom score ranging from 0 to 16.

Pandemic-related experiences

Participants completed a battery of questions at Time 1 and Time 2 assessing experiences related to the COVID-19 pandemic since the pandemic began to impact life in the United States. See Supplement for more details about this measure.

Perceived burden

Participants completed nine items assessing their level of perceived burden related to the pandemic, including questions measuring the level of COVID Worry, School/Work Interference, Financial Stress, and Pandemic Disruption. Participants rated all items on a 5-point Likert scale (1 = Not at all, 5 = Extremely).

Actual risk

Participants completed two items assessing their level of actual risk of exposure to COVID-19, which were used to generate a single score labelled COVID Exposure. Participants were asked whether they had been diagnosed with or were exhibiting symptoms of COVID-19 (0 = No, 1 = Yes) and whether they had been in contact with someone with a confirmed case of COVID-19 (0 = No, 1 = Yes).

Pandemic behaviours

Participants completed six items assessing their pandemic-related behaviours, including adaptive behaviours (Social Distancing Compliance and Guideline Adherence) and behaviours that were maladaptive or potentially maladaptive (Supply Hoarding, News Checking, Social Media Use). Participants rated all items using a 5-point Likert scale.

Expectancies

Three items assessed two different types of pandemic-related expectancies. Exposure Likelihood was assessed using two items (‘How likely is it that you have been exposed to COVID-19?’ and ‘How likely is it that you will be exposed to COVID-19 in the future?’). Both items were rated on a 5-point Likert scale (1 = Not at all, 5 = Extremely). Pandemic Course Prediction was measured with one item (‘For how much longer do you expect that COVID-19 will continue to disrupt daily life?’). Participants rated this item on a five-point scale (1 = Less than 2weeks, 2 = 2 weeks to 1month, 3 = 1 to 3months, 4 = 3 to 6months, 5 = More than 6months).

Structural support

Participants completed two items measuring their perceived structural support from the university (University Support) and the federal government (Government Support). Both items were rated on a 5-point Likert scale (1 = Not at all, 5 = Extremely).

Data analysis plan

At Time 2, 69 participants (41%) were lost to follow-up, for a final sample of 98 participants. Analyses revealed that attrition at Time 2 was not conditional on any variables measured in this study, including demographics, symptom measures, or pandemic-related experiences. Thus, since the data appeared to be missing at random, we used listwise deletion for all analyses.

For the first aim, we used paired t-tests to assess changes in mental health symptoms and pandemic experiences from Time 1 to Time 2, and to test for differences in mental health symptoms across demographic groups at both Time 1 and Time 2.

For the second aim, we used a series of elastic net regression models to investigate whether aspects of the pandemic experience predicted mental health symptoms at Time 2. For non-zero predictors, we followed up with traditional regression models to determine whether the selected variables significantly predicted the outcome of interest (a priori alpha level of .05), as elastic net regression does not produce traditional p-values. Specifically, we entered each of the non-zero elastic net predictors for a given outcome into a multiple linear regression model predicting that outcome.

For all models, we included demographic factors as covariates (age, gender, SES, BIPOC, ethnicity; all measured at Time 1). For the models with depression, anxiety, and pandemic stress response as outcomes, we re-ran the analyses controlling for Time 1 depression, anxiety, and pandemic stress response. We were unable to test the alcohol and substance use models controlling for Time 1 alcohol and substance use symptoms because these symptoms were only assessed at Time 2.

Results

Aim 1. Characterizing the sample: Psychological health and pandemic-related experiences

Most participants (61.2%) reported needing to relocate at Time 1 because of the COVID-19 university shutdown. Participants’ geographical locations and case rates by location at time of assessment are depicted in Figure 1. At Time 1, only 2.4% of the sample was experiencing current COVID-19 symptoms and only 7.9% reported contact with a confirmed case of COVID-19. 15.2% of the sample reported a pre-existing health condition at Time 1. See Table 1 for the means and standard deviations for the mental health symptoms and pandemic-related experiences measures at Time 1 and Time 2. Zero-order correlations by timepoint are presented in Figure 2.

image

Participants’ Geographical Locations (black dots) and community-level COVID-19 Case Rates (heatmap) by Location at both Time 1 and Time 2.

Table 1.
Mental health symptoms and pandemic experiences across time points

Time 1

n = 165

Time 2

n = 98

M (SD) M (SD)
Clinical characteristics
PHQ-9 8.44 (6.98) 9.36 (7.30)
GAD-7 6.68 (6.17) 7.20 (6.62)
IES-6 1.82 (1.01)a 1.62 (0.99)a
Alcohol use 2.27 (2.43)
Substance use 1.05 (2.52)
Pandemic experiences
Perceived burden
COVID worry 3.63 (0.87)a 3.05 (1.02)a
School/Work interference 3.38 (1.13) 3.35 (1.30)
Financial stress 2.39 (1.15) 2.46 (1.22)
Pandemic disruption 3.23 (0.94)
Actual risk
COVID exposure 0.05 (0.16) 0.05 (0.17)
Pandemic behaviours
Social distancing compliance 4.15 (0.73) 4.31 (0.74)
Guideline adherence 4.37 (0.61)
Supply hoarding 3.16 (0.87)a 2.89 (1.05)a
News checking 2.98 (1.16)
Social media use 4.18 (0.90)
Expectancies
Exposure likelihood 2.49 (0.87) 2.57 (0.88)
Pandemic course prediction 3.35 (0.75)a 3.58 (0.91)a
Structural support
University support 3.08 (0.94) 2.99 (1.07)
Government support 2.48 (1.04) 2.40 (1.07)

Note

  • GAD-7 = General Anxiety Disorder-7; IES-6 = Impact of Event Scale-6; PHQ-9 = Patient Health Questionnaire-9.
  • All values are means and standard deviations.

  • a

    A significant difference from T1 to T2 (p < .05).

image

Intra-temporal zero-order correlations between pandemic-related experiences and mental health symptoms. Note. Pearson correlation coefficients are displayed for each relationship. Significant correlations (p < .05) are highlighted in blue or red, with colour varying depending on the strength and direction of the correlation. GAD-7 = General Anxiety Disorder-7; IES-6 = Impact of Event Scale-6; PHQ-9 = Patient Health Questionnaire-9.

Mental health symptoms

At Time 1 (n = 165), 34.5% (n = 57) scored above the clinical cut-off for moderate depression on the PHQ-9, 28.5% (n = 47) scored above the clinical cut-off for moderate anxiety on the GAD-7, and 50.3% (n = 83) scored above the clinical cut-off for pandemic stress response on the IES-6. At Time 2 (n = 98), 43.9% (n = 43) scored above the clinical cut-off for depression, 27.6% (n = 27) scored above the clinical cut-off for anxiety, and 42.9% (n = 42) scored above the clinical cut-off for pandemic stress response. There was a significant decrease in the level of pandemic stress response from Time 1 to Time 2, t(95) = 2.50, p = .01, and a trending significant increase in depression from Time 1 to Time 2, t(95) = 1.88, p = .06; however, these symptom changes do not appear to reflect clinically meaningful differences. There was no change in anxiety from Time 1 to Time 2, t(95) = −1.17, p = .24.

We next examined follow-up comparisons across core sociodemographic groups. There were no differences in clinical symptoms between individuals who identified as BIPOC and those who did not, nor were there any differences between individuals who identified as Hispanic and non-Hispanic (all p’s > .05). We also compared mental health symptoms across gender; similarly, there were no differences between male and female participants, with the exception of the IES-6. There was a significant difference in the level of pandemic stress response at Time 1 by gender, t(157) = −2.02, p = .045, such that women reported significantly higher levels of pandemic stress reactions than men at Time 1.

Pandemic-related experiences

Next, we investigated whether pandemic-related experiences changed over time using a series of paired t-tests. From Time 1 to Time 2, there was a significant decrease in COVID Worry, t(95) = 7.21, p < .001, a significant decrease in Supply Hoarding, t(95) = 3.49, p < .001, and a significant increase in the Pandemic Course Prediction, t(95) = −2.30, p = .02. There were no other significant changes in pandemic-related experiences from Time 1 to Time 2 (all p’s > .05).

Aim 2. Time 1 early pandemic experiences predicting time 2 mental health

Figure 3 contains the variable importance plots for each elastic net regression model. For each model, the blue bars indicate non-zero predictors from elastic net regression; these non-zero predictor variables were entered into multiple linear regression models in order to determine whether the associations were statistically significant.

image

Variable importance from elastic net regression of time 2 mental health on time 1 pandemic experiences. Note. Models include demographic factors and pandemic experience variables measured at Time 1. Dark blue bars indicate predictors selected by elastic net that were significant at the p < .05 level when entered into a traditional multiple regression model; light blue bars indicate predictors selected by elastic net that were not significant according to a multiple regression model. EtOH = alcohol use; GAD-7 = General Anxiety Disorder-7; IES6 = Impact of Event Scale-6; PHQ-9 = Patient Health Questionnaire-9; Sub = substance use.

Depression symptoms at Time 2 were significantly predicted by Time 1 COVID Worry (β = .42, 95% CI [0.22, 0.61], p < .001) and School/Work Interference (β = .23, 95% CI [0.04, 0.42], p = .018). Similarly, Time 2 anxiety symptoms were predicted by Time 1 COVID Worry (β = .40, 95% CI [0.21, 0.59], p < .001), School/Work Interference (β = .24, 95% CI [0.06, 0.43], p = .011), and having a pre-existing condition (β = .67, 95% CI [0.18, 1.15], p = .007). Time 2 stress response was predicted by Time 1 COVID Worry (β = .44, 95% CI [0.25, 0.64], p < .001) and School/Work Interference (β = .21, 95% CI [0.02, 0.40], p = .032). Finally, greater alcohol use symptoms at Time 2 were predicted by having a pre-existing condition at Time 1 (β = .96, 95% CI [0.44, 1.49], p < .001), reporting lower Social Distancing Compliance (β = −.40, 95% CI [−0.59, −0.21], p < .001), and greater School/Work Interference at Time 1 (β = .22, 95% CI [0.03, 0.42], p = .027). The only significant Time 1 predictor of substance use symptoms at Time 2 was School/Work Interference (β = .24, 95% CI [0.02, 0.46], p = .031).

For models where we had Time 1 measures of the dependent variables (i.e., depression, anxiety, and stress response), we repeated analyses controlling for Time 1 symptoms. In each case, Time 1 symptoms became the strongest predictor of Time 2 symptoms, with all other predictors becoming non-significant.

Aim 3. Time 2 end of semester pandemic experiences predicting time 2 mental health

Figure 4 contains the variable importance plots for each elastic net regression model. Time 2 COVID Worry (β = .21, 95% CI [0.00, 0.42], p = .046) and Pandemic Disruption (β = .42, 95% CI [0.14, 0.70], p = .004) were both significantly and positively associated with Time 2 depression symptoms. In contrast, those perceiving greater Government Support at Time 2 reported significantly lower depression symptoms (β = −.21, 95% CI [−0.39, −0.03], p = .02). When we included Time 1 depression symptoms in the model, it was a strong predictor of Time 2 symptoms (β = .63, 95% CI [0.48, 0.79], p < .001). In this model, Time 2 depressive symptoms were no longer linked with Time 2 COVID Worry (β = .04, 95% CI [−0.12, 0.20], p = .62), but they remained significantly associated with Time 2 Pandemic Disruption (β = .25, 95% CI [0.04, 0.45], p = .021) and Government Support (β = −.14, 95% CI [−0.27, −0.01], p = .038).

image

Variable importance from elastic net regression of time 2 mental health on time 2 pandemic experiences. Note. Models include demographic factors measured at Time 1 and pandemic experience variables measured at Time 2. Dark blue bars indicate predictors selected by elastic net that were significant at the p < .05 level when entered into a traditional multiple regression model; light blue bars indicate predictors selected by elastic net that were not significant according to a multiple regression model. EtOH = alcohol use; GAD-7 = General Anxiety Disorder-7; IES6 = Impact of Event Scale-6; PHQ-9 = Patient Health Questionnaire-9; Sub = substance use.

Greater COVID Worry at Time 2 (β = .28, 95% CI [0.08, 0.47], p = .006), and having a pre-existing condition (β = .54, 95% CI [0.05, 1.02], p = .032) were significant predictors of Time 2 anxiety symptoms. However, when we included Time 1 anxiety symptoms in the model, this variable was the strongest predictor of Time 2 anxiety (β = .62, 95% CI [0.46, 0.78], p < .001), and all other associations became non-significant.

Greater levels of Time 2 COVID Worry (β = .42, 95% CI [0.24, 0.61], p < .001), News Checking (β = .20, 95% CI [0.04, 0.36], p = .014), and Course Prediction (β = .19, 95% CI [0.03, 0.35], p = .019) were all significantly linked with greater pandemic stress response at Time 2. Notably, these associations remained significant after Time 1 pandemic stress response symptoms were included in the model (COVID Worry: β = .25, 95% CI [0.07, 0.42], p = .006; News Checking: β = .16, 95% CI [0.02, 0.29], p = .03; Course Prediction: β = .15, 95% CI [0.01, 0.29], p = .032), even though Time 1 pandemic stress response symptoms were a strong predictor of Time 2 pandemic stress symptoms (β = .42, 95% CI [0.26, 0.58], p < .001).

Greater alcohol use disorder symptoms at Time 2 were significantly associated with having a pre-existing condition (β = .76, 95% CI [0.23, 1.28], p = .005), and with Exposure Likelihood (β = .33, 95% CI [0.13, 0.53], p = .001). Conversely, Time 2 Guideline Adherence (β = −.22, 95% CI [−0.41, −0.03], p = .024) and News Checking (β = −.26, 95% CI [−0.46, −0.07], p = .009) were significantly linked with reduced alcohol use disorder symptoms. At Time 2, greater substance use disorder symptoms were associated with lower Time 2 Guideline Adherence scores (β = −.26, 95% CI [−0.45, −0.06], p = .010) and lower levels of Time 2 Social Media Use (β = −.22, 95% CI [−0.41, −0.02], p = .033).

Discussion

The overall aim of the current study was to capture the psychological health and pandemic-related experiences of U.S. college students at two timepoints during the early months of the COVID-19 pandemic. As expected, a significant number of students reported depression, anxiety, and pandemic-related stress symptoms above clinical cut-offs at both timepoints, consistent with prior studies (e.g., Odriozola-González et al., 2020). These figures suggest high levels of general distress experienced by the U.S. college student population during COVID-19. Students reporting elevated mental health symptoms at the beginning of the pandemic continued to endorse negative mental health at the end of the semester, similar to findings in a Chinese community sample (e.g., Wang, Pan, et al., 2020). Contrary to our expectations, there were no clinically significant changes in mental health symptoms from the start of the pandemic to the end of the semester. These findings are in line with more recent reports from across the globe that as the pandemic became protracted, rates of psychiatric symptoms either did not change or in some cases even declined (Belz et al., 2021; Brunoni et al., 2021; Fancourt, Steptoe, & Bu, 2021).

Contrary to hypotheses, we found few differences in mental health symptoms between sociodemographic groups. These findings are surprising in the context of the known disproportionate impact of the pandemic on U.S. racial and ethnic minority populations (Tai, Shah, Doubeni, Sia, & Wieland, 2020). Asian communities have experienced greater levels of discrimination throughout the pandemic (Wu et al., 2021), and both COVID-19-related discrimination and associated stigma have been linked with poorer mental health (Miconi et al., 2021). African American, Native American, and Latinx communities experienced significant disparities in COVID-19 exposure, susceptibility, and treatment access (Tai et al., 2020). Additional research indicates that marginalized populations also experienced disproportionately negative psychosocial repercussions during the pandemic (Ruprecht et al., 2021), which may interact with and compound existing health disparities (Gezici & Ozay, 2020). With respect to our findings, it may be that the relatively early assessment window (i.e., Spring 2020), within what ultimately became an extremely drawn-out pandemic, did not capture differences between racial and ethnic groups that may have emerged later in the year. Our findings may also stem from unique considerations relevant to our sample. College students at a private institution with an annual tuition above $70,000 may not be representative of the general U.S. population. Further, Latinx students enrolled in the current study may not be representative of the larger Latinx population in the United States, given the demographic makeup of South Florida and Miami in particular. As Mahler, Cogua-López, and Chaudhuri (2018) describe, ‘In this region [Miami metro area] Latin@s are not, on the whole, a disadvantaged minority but rather the fulcrum around which the rest of the population rotates’. There are also additional reports emerging that suggest that individuals identifying as Black may have decreased odds of endorsing clinically significant anxiety and depression in the context of the pandemic (Kantor & Kantor, 2020). One recent qualitative study of Black adolescents found that despite reports of the negative impact of COVID-19, participants also highlighted many positive and adaptive responses (Banks, 2021), all of which indicates that further research is warranted.

Pandemic-related experiences at the start of the pandemic and end of the semester appeared to have a distinct impact on internalizing symptoms at the end of the semester. While several early pandemic experiences, including COVID worry and school/work interference, significantly predicted depression, anxiety, and stress at Time 2, these associations became non-significant after controlling for mental health symptoms at Time 1. The relative stability of depression, anxiety, and stress aligns with prior research supporting their trait-like nature across 2-, 3-, and 8-year time periods (Lovibond, 1998). In contrast, even controlling for Time 1 symptoms, pandemic-related worry at Time 2 was linked with Time 2 stress symptoms, and pandemic-related disruption at Time 2 was associated with Time 2 depression symptoms. Our measure of disruption was an index score that included issues with motivation, concentration, and family conflict (see Supplement) – all constructs that have been previously linked with depression (e.g., Lam, Kennedy, McIntyre, & Khullar, 2014; Sheeber, Hops, Alpert, Davis, & Andrews, 1997). Many students in our sample (61.2%) reported changes in their living situation due to the pandemic, and 79.4% reported living with family. Thus, home-related stressors, such as reduced independence and increased family conflict, may have influenced depressive symptoms in this population.

We also found several potential associations of news and policy factors with mental health symptoms. For one, students perceiving an inadequate government response to the pandemic reported significantly higher depression symptoms later in the semester. Similar to a prior study reporting that over two-thirds of students lacked confidence in the government’s handling of COVID-19 (Copeland et al., 2021), our sample endorsed limited confidence that the government was taking appropriate steps to address the COVID-19 pandemic (M = 2.40; range = 1–5). This association of mistrust in the government’s handling of the pandemic and higher depression symptoms, even after controlling for Time 1 depression, highlights the important connection between policy decisions and mental health at the individual level. Additionally, similar to recent research (e.g., Boyraz, Legros, & Tigershtrom, 2020), we found that news checking behaviour was associated with Time 2 pandemic stress response, even controlling for Time 1 pandemic stress symptoms. Prior studies have linked COVID-related news exposure with increased COVID-related distress (e.g., Chao, Xue, Liu, Yang, & Hall, 2020; Mertens, Gerritsen, Duijndam, Salemink, & Engelhard, 2020), suggesting that media exposure may be an important factor influencing mental health during the pandemic. If replicated, these results suggest that reducing the amount of time spent checking the news or other media sources for information about the pandemic could be beneficial for students’ mental health. Research on problematic internet use during the pandemic suggests that self-monitoring and regulating one’s screen time, as well as reducing physical access to smartphones and other devices, are also important tools to prevent problematic internet use and improve mental well-being (e.g., Király et al., 2020).

The expected duration of the pandemic was another significant predictor of pandemic stress response at the end of the semester, even after controlling for Time 1 pandemic stress. Over time, students appeared to become more worried about the length of the pandemic’s disruption to their day-to-day lives: At the beginning of the pandemic, only 8.5% reported believing that the pandemic would continue longer than 6 months, whereas by the end of the semester, this percentage had increased to 19.4%. Negative expectancies, defined as persistent and negative biases about the self, others, and the world (American Psychological Association, 2013), are linked with post-traumatic stress symptoms (e.g., Gallagher, Long, & Phillips, 2020; Kimble, Sripad, Fowler, Sobolewski, & Fleming, 2018). Though limited by the fact that the impact of the COVID-19 pandemic does not qualify as a Criterion A event, our findings may nevertheless be informed by the associations in the trauma literature between expectancies (e.g., optimism/hopelessness) and symptom severity. In the context of this literature, our finding linking pandemic stress response symptoms and longer course prediction may reflect increased hopelessness about the future. As an additional consideration, even our ‘later’ time point (Time 2) was relatively early in the pandemic period. As it has become clear that the pandemic will continue well into 2022, we wonder how predictions of continued disruption to daily life may have intensified over time, with implications for stress symptoms.

While a range of alcohol and substance use disorder symptom severity was represented in our sample, the majority of our sample reported limited use and associated problems. Despite this limitation, several aspects of the pandemic experience were associated with increased alcohol and substance use disorder symptoms at the end of the semester. Pandemic-related interference early in the semester was a significant predictor of alcohol and substance use symptoms at the end of the semester, pointing to potential coping motives behind drug and alcohol use (Kuntsche, Knibbe, Gmel, & Engels, 2005). Additionally, we found that alcohol and substance use disorder symptoms at Time 2 were associated with lower compliance with social distancing at Time 1 and lower general CDC guideline adherence at Time 2. These findings suggest that alcohol and substance use disorder symptoms may represent relevant considerations in delivering COVID-19 guidelines to young adults. Notably, because we measured alcohol use, substance use, and guideline adherence only at Time 2, these regression models cannot speak to the directionality of associations. However, as has been demonstrated in previous studies of alcohol/substance use and health-related decision-making (Leigh & Stall, 1993), one possible interpretation of these findings is that the use of alcohol or drugs could contribute to greater disinhibition and willingness to engage in behaviours that may have potentially increased risk of contracting COVID-19 (Scott-Sheldon et al., 2016).

There were several limitations of the current study. The length of time between Time 1 and Time 2 was approximately one month, and given that the pandemic is still continuing over a year later, our timeframe may have been too short to capture meaningful changes in mental health and behaviour. A second limitation is that we only measured alcohol and substance use at Time 2. As a result, we could not control for Time 1 alcohol and substance use in the respective elastic net regression models, and thus do not know whether pandemic experiences predict alcohol and substance use above baseline symptom levels. Another limitation is the use of the modified IES-6 to assess pandemic stress response symptoms. While other studies have used a version of the IES to measure reactions to the COVID-19 pandemic (Di Crosta et al., 2020), the IES-6 was initially developed to measure symptoms of PTSD after a Criterion A event, and asking participants to report on stress symptoms related to ‘the impact of COVID-19’ is not equivalent to assessing for a Criterion A stressor. A fourth limitation is that our measures of alcohol and substance use disorder symptoms were positively skewed, despite participants reporting a range of alcohol and substance use disorder severity. Thus, our findings related to alcohol and substance use disorder symptoms should be considered in the light of the majority of our sample endorsing limited use and associated problems. Results should also be considered in the context of the high rate of attrition from Time 1 to Time 2, which – though unrelated to primary study variables or symptom severity – may have impacted the reported associations. Additional limitations were the relatively small sample size used in this study and our reliance on self-report measures, which highlight a need for replication in larger college student samples using multi-modal assessment batteries.

As one of the few longitudinal investigations of young adult mental health during the pandemic, this study highlighted the relative stability of students’ mental health during its early months and pointed to associations of pandemic experiences with mental health symptoms across both the internalizing and externalizing spectra. Pessimism about the future (i.e., how long the pandemic would continue), pandemic-related disruption, worry about COVID-19, and lack of confidence in the government’s response were associated with later internalizing symptoms even after accounting for students’ baseline mental health. Alcohol and substance use were linked with poorer adherence to guidelines, suggesting that alcohol and substance use may represent targets for monitoring and intervention during this time. College students should also be aware of potential associations of media use with mental health: While our findings suggest that using the internet and other sources to check for pandemic-related information may negatively affect mental health, using social media as a way to increase social connection and reduce loneliness may help improve mental well-being (e.g., Marzouki, Aldossari, & Veltri, 2021). Overall, while college student mental health may not have changed significantly during the early months of the pandemic, they endorsed generally elevated levels of anxiety, depression, and stress throughout that time. Furthermore, experiences of the pandemic had a tangible influence on their psychological health, and the primary pandemic experiences impacting mental health, such as predicted duration of COVID-19, have only intensified since the time our data were collected.

Funding

Caitlin Stamatis is supported by a grant from the National Institute of Mental Health (T32MH115882).

Conflicts of interest

All authors declare no conflict of interest.

Author contribution

Caitlin A Stamatis: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Project administration (equal); Writing – original draft (equal); Writing – review & editing (equal). Hannah C. Broos: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Project administration (equal); Writing – original draft (equal); Writing – review & editing (equal). Stephanie E. Hudiburgh: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Project administration (equal); Writing – original draft (equal); Writing – review & editing (equal). Sannisha K. Dale: Supervision (equal); Writing – review & editing (equal). Kiara R. Timpano: Conceptualization (equal); Data curation (equal); Investigation (equal); Methodology (equal); Project administration (equal); Supervision (equal); Writing – review & editing (equal).

Sources

1/ https://Google.com/

2/ https://bpspsychub.onlinelibrary.wiley.com/doi/10.1111/bjc.12351

The mention sources can contact us to remove/changing this article

What Are The Main Benefits Of Comparing Car Insurance Quotes Online

LOS ANGELES, CA / ACCESSWIRE / June 24, 2020, / Compare-autoinsurance.Org has launched a new blog post that presents the main benefits of comparing multiple car insurance quotes. For more info and free online quotes, please visit https://compare-autoinsurance.Org/the-advantages-of-comparing-prices-with-car-insurance-quotes-online/ The modern society has numerous technological advantages. One important advantage is the speed at which information is sent and received. With the help of the internet, the shopping habits of many persons have drastically changed. The car insurance industry hasn't remained untouched by these changes. On the internet, drivers can compare insurance prices and find out which sellers have the best offers. View photos The advantages of comparing online car insurance quotes are the following: Online quotes can be obtained from anywhere and at any time. Unlike physical insurance agencies, websites don't have a specific schedule and they are available at any time. Drivers that have busy working schedules, can compare quotes from anywhere and at any time, even at midnight. Multiple choices. Almost all insurance providers, no matter if they are well-known brands or just local insurers, have an online presence. Online quotes will allow policyholders the chance to discover multiple insurance companies and check their prices. Drivers are no longer required to get quotes from just a few known insurance companies. Also, local and regional insurers can provide lower insurance rates for the same services. Accurate insurance estimates. Online quotes can only be accurate if the customers provide accurate and real info about their car models and driving history. Lying about past driving incidents can make the price estimates to be lower, but when dealing with an insurance company lying to them is useless. Usually, insurance companies will do research about a potential customer before granting him coverage. Online quotes can be sorted easily. Although drivers are recommended to not choose a policy just based on its price, drivers can easily sort quotes by insurance price. Using brokerage websites will allow drivers to get quotes from multiple insurers, thus making the comparison faster and easier. For additional info, money-saving tips, and free car insurance quotes, visit https://compare-autoinsurance.Org/ Compare-autoinsurance.Org is an online provider of life, home, health, and auto insurance quotes. This website is unique because it does not simply stick to one kind of insurance provider, but brings the clients the best deals from many different online insurance carriers. In this way, clients have access to offers from multiple carriers all in one place: this website. On this site, customers have access to quotes for insurance plans from various agencies, such as local or nationwide agencies, brand names insurance companies, etc. "Online quotes can easily help drivers obtain better car insurance deals. All they have to do is to complete an online form with accurate and real info, then compare prices", said Russell Rabichev, Marketing Director of Internet Marketing Company. CONTACT: Company Name: Internet Marketing CompanyPerson for contact Name: Gurgu CPhone Number: (818) 359-3898Email: [email protected]: https://compare-autoinsurance.Org/ SOURCE: Compare-autoinsurance.Org View source version on accesswire.Com:https://www.Accesswire.Com/595055/What-Are-The-Main-Benefits-Of-Comparing-Car-Insurance-Quotes-Online View photos

ExBUlletin

to request, modification Contact us at Here or [email protected]