Contents lists available at ScienceDirect

Psychiatry Research

journal homepage: www.elsevier.com/locate/psychres

Factors associated with depression, anxiety, and PTSD symptomatologyduring the COVID-19 pandemic: Clinical implications for U.S. young adultmental health

Cindy H. Liu (PhD)a,c,d,⁎, Emily Zhang (MA)a,c, Ga Tin Fifi Wong (BA)a,c, Sunah Hyun (PhD)a,c,Hyeouk “Chris” Hahm (PhD)b,c

a Department of Newborn Medicine, Brigham and Women's Hospital, Boston, MA, USAb Department of Psychiatry, Brigham and Women's Hospital, Boston, MA, USAc School of Social Work, Boston University, Boston, MA, USAd Harvard Medical School

A R T I C L E I N F O

Keywords:Psychological stress, LonelinessUniversity health servicesSocial supportEthnicityCOVID-19DepressionAnxietyPTSD

A B S T R A C T

This study sought to identify factors associated with depression, anxiety, and PTSD symptomatology in U.S.young adults (18-30 years) during the COVID-19 pandemic. This cross-sectional online study assessed 898participants from April 13, 2020 to May 19, 2020, approximately one month after the U.S. declared a state ofemergency due to COVID-19 and prior to the initial lifting of restrictions across 50 U.S. states. Respondentsreported high levels of depression (43.3%, PHQ-8 scores ≥ 10), high anxiety scores (45.4%, GAD-7 scores ≥10), and high levels of PTSD symptoms (31.8%, PCL-C scores ≥ 45). High levels of loneliness, high levels ofCOVID-19-specific worry, and low distress tolerance were significantly associated with clinical levels of de-pression, anxiety, and PTSD symptoms. Resilience was associated with low levels of depression and anxietysymptoms but not PTSD. Most respondents had high levels of social support; social support from family, but notfrom partner or peers, was associated with low levels of depression and PTSD. Compared to Whites, AsianAmericans were less likely to report high levels across mental health symptoms, and Hispanic/Latinos were lesslikely to report high levels of anxiety. These factors provide initial guidance regarding the clinical managementfor COVID-19-related mental health problems.

1. Introduction

The COVID-19 pandemic that has upended the lives of individualsworldwide escalated in the U.S. beginning in March of 2020. Althoughresearch on acute and widescale stressors (e.g., natural disasters), de-monstrates severe implications for mental health (Kessler et al., 2008),there is no precedent for understanding the mental health effects due toCOVID-19, as prospective studies investigating the effects of a pan-demic are virtually non-existent. In particular, the identification of riskfactors associated with depression, anxiety, and post-traumatic stressdisorder (PTSD) among U.S. young adults (18-30 years) during thepandemic is urgently needed. Comprising more than one-third of thecurrent U.S. workforce, young adults (often referred to as “Millennials”and “Generation Z”) will be a dominant workforce group for the nextdecade, and our societal functioning depends on how they emerge fromthe pandemic. Understanding their health and well-being now is crucial

as it sets the stage for later outcomes.Certain risk and protective factors are likely to be implicated in

pandemic-related mental health. COVID-19-related worry (e.g., main-taining employment, getting tested for coronavirus) may be linked tomental health symptoms. The early weeks of the pandemic saw rapidchanges in daily routines, with students moving following universityclosures and attending classes remotely, and for other young adults,transitioning to remote work or experiencing loss of work. These dis-ruptions may put an already vulnerable group at greater risk for mentalhealth challenges (Conrad, 2020). Furthermore, loneliness may beparticularly prevalent and devastating during the pandemic given di-rectives for social distancing and isolation. Those under the age of 25already show elevated levels of loneliness (Domagala-Krecioch andMajerek, 2013), and the pandemic may exacerbate these feelings. De-spite the critical role that social support plays in mitigating the risks tomental health problems, directives on social distancing may impede on

https://doi.org/10.1016/j.psychres.2020.113172Received 28 April 2020; Received in revised form 30 May 2020; Accepted 30 May 2020

⁎ Corresponding author.E-mail address: [email protected] (C.H. Liu).

Psychiatry Research 290 (2020) 113172

Available online 01 June 20200165-1781/ © 2020 Elsevier B.V. All rights reserved.

T

one's typical means for obtaining such support.Individual resilience, which refers to one's ability to cope with

stress, and distress tolerance, which describes one's ability to manageand tolerate emotional distress, may be salient characteristics thatprotect against the mental health symptoms that follow major stressors.Individual resilience is a significant protective factor for depression,PTSD, and general health after natural disasters (Kukihara et al., 2014).Findings have generally demonstrated distress tolerance to be asso-ciated with lower symptoms of depression and PTSD following torna-does (Cohen et al., 2016). However, the extent to which these factorsare associated with mental health symptoms during a pandemic is un-known.

This study sought to identify potential factors that contribute tomental health outcomes among young adults during the COVID-19pandemic. The CARES 2020 Project (COVID-19 Adult ResilienceExperiences Study, www.cares2020.com) was launched to track thehealth and well-being of young adults in the U.S. across multiple timepoints in 2020 and 2021. This present analysis assessed depression,anxiety, and PTSD symptomatology, and psychological experiencesincluding distress tolerance, resilience, social support, and loneliness.We included depression and anxiety as these are common mental healthsymptoms among young adults (Blazer et al., 1994; Chen et al., 2019;Eisenberg et al., 2007; Liu et al., 2019; Mojtabai et al., 2016). We as-sessed PTSD symptoms given documented high rates of trauma byyoung adulthood (Costello et al., 2002; Reynolds et al., 2016; Vrana andLauterbach, 1994); a concern was that the pandemic would eithercreate and/or exacerbate symptoms related to prior trauma(Breslau et al., 2008, 1999; Brunet et al., 2001). New items that as-sessed COVID-19-specific concerns were also included. The objective ofthis work is to identify salient psychosocial risks for mental healthsymptoms and to prioritize intervention targets for addressing mentalhealth symptoms among young adults.

2. Methods

2.1. Study population

This present cross-sectional study assessed potential risk and pro-tective factors for mental health outcomes based on preliminary CARES2020 data obtained from Wave 1 data collection (N = 898) conductedfrom April 13, 2020 to May 19, 2020, approximately one month afterthe U.S. declared a state of emergency due to COVID-19 and prior to theinitial lifting of restrictions across 50 U.S. states. Eligible participantswere young adults aged 18 to 30 years currently living in the U.S. orreceiving education from a U.S. institution. Participants were recruitedonline via email list serves, social media, and word of mouth (i.e., listserves and Facebook groups for school organizations or clubs, alumnigroups, classes, churches). This took place initially through organiza-tions from the New England area before additional list serves fromother regions of the U.S. (Midwest, South, and West) were targeted.Respondents were asked to complete a 30-minute online Qualtricssurvey regarding COVID-19-related experiences, risk and resilience,and physical and mental health outcomes. To ensure data quality,human verification and attention checks were implemented throughoutthe survey; the data were further inspected visually for response irre-gularities indicative of bots. Participants were compensated via raffle inwhich one out of every 10 participants received a $25 gift card. Allprocedures were approved by the Institutional Review Board at BostonUniversity.

2.2. Measures

Binary scores were created after calculating the mean or sum ofeach measure. Rather than relying on the sample characteristics tocategorize our data (e.g., mean, median, tertile or quartile split), thedetermination of the cutoff score was based on standard cutoffs from

previous research; when a standard was not available, scale responsedescriptors to determine the cutoffs.

2.2.1. Risk and protective factorsPsychological resilience was measured using the 10-item Connor-

Davidson Resilience Scale (CD-RISC-10, Connor and Davidson, 2003),which assesses one's ability to cope with adverse experiences. Partici-pants indicated how they felt in the past month on a 5-point scale, with0 indicating “not true at all” and 4 indicating “true nearly all the time.”Sum scores were recoded dichotomously into “high resilience” and “lowresilience” with a cutoff score of 30 or greater. This cutoff score char-acterizes responses that tended to be “often true” and “true nearly allthe time,” with those endorsing a score ≥30 considered to be at “veryhigh risk with mental disorders” (Andrews and Slade, 2001; Kessler andMroczek, 1992).

The Distress Tolerance Scale is a 15-item measure that assessesparticipants’ abilities to withstand and cope with emotional distress(Simons and Gaher, 2005). Respondents rated personal attitudes to-wards feelings of emotional distress on a 5-point scale, ranging from 1(“strongly agree”) to 5 (“strongly disagree”), with higher ratings in-dicating greater distress tolerance. A global mean score of distress tol-erance was calculated. We considered the scale descriptors and fol-lowed the cutoffs used for the CD-RISC, which was also a 5-point scale.As such, scores were dichotomously recoded so that global mean scoresless than 4 indicated “low distress tolerance” and scores of 4-to-5 in-dicated “high distress tolerance.”

Perceived social support was measured using the MultidimensionalScale of Perceived Social Support (MSPSS, Zimet et al., 1988), in whichparticipants rated perceived emotional support using a 7-point Likertscale ranging from 1 (“very strongly disagree”) to 7 (“very stronglyagree”). This measure includes three subscales assessing perceivedsupport quality from family, friends, and partners. Because mean scoresgreater than 5 reflected responses indicating “mildly agree,” “stronglyagree,” and “very strongly agree,” each subscale mean scores were re-coded so that scores 5 or greater referred to “high perceived socialsupport,” and scores below 5 were referred to as “low perceived socialsupport.”

Instrumental support was assessed through a 4-item subscale of theTwo-Way Social Support Scale (Shakespeare-Finch and Obst, 2011).Participants indicated the extent of they received instrumental supportbased on a 6-point Likert scale ranging from 0 (“not at all”) to 5 (“al-ways”). Items were summed to create a total score with a possible rangeof 0 to 20. Given scale descriptors, a cutoff score with a sum of 16 orgreater indicated “high instrumental support,” whereas scores lowerthan 16 indicated “low instrumental support.”

Loneliness was measured using an adapted 3-item version of theUCLA Loneliness Scale Short Form (Hughes et al., 2004). Participantsrated lack of companionship, feelings of being left out, and isolationfrom others on a scale of 1-to-3, with 1 as “hardly ever,” 2 as “some ofthe time,” and 3 as “often.” A sum score for loneliness was calculatedwith a total possible range of 3 to 9 and recoded dichotomously; acutoff score of 6 or greater indicated “high loneliness” as used in priorstudies (Lowthian et al., 2016; Tymoszuk et al., 2019).

Severity of COVID-19 pandemic-related worry was assessed using anewly developed measure consisting of 6 items, which included thefollowing concerns: “Having enough groceries during city lockdowns/social distancing protocols”, “obtaining a COVID-19 test if I becomesick”, “getting treated for COVID-19 if I contract it”, “keeping in touchwith loved ones during social distancing protocols”, “maintaining em-ployment during the subsequent economic downturn”, and “havingenough money to pay for rent and buy basic necessities.” Participantswere asked to indicate their level of worry for each item on a scale of 1to 5, with 1 being “not worried at all,” and 5 being “very worried.” Sumscores were calculated with a total possible range of 6 to 30 and re-coded into a dichotomous variable with a cutoff score of 24 or greateras “highly worried.” Cronbach's alpha for measure items was .70,

C.H. Liu, et al. Psychiatry Research 290 (2020) 113172

2

indicating good reliability.

2.2.2. Mental health outcomesDepression was assessed with the 8-item version of the Patient

Health Questionnaire (PHQ-8, Kroenke et al., 2009) which assessedfrequency of depressive symptoms in the past two weeks on a scale of 0(“not at all”) to 3 (“nearly every day”). Sum scores of the PHQ-8 had atotal possible range of 0 to 24 and were recoded dichotomously basedon a cutoff score of 10 or higher (Wu et al., 2019).

Anxiety was assessed with the Generalized Anxiety Disorder Scale(GAD-7, Spitzer et al., 2006) a widely used measure assessing the fre-quency of anxiety symptoms in the past two weeks on a scale of 0 to 3,with 0 being “not at all” and 3 being “nearly every day.” Sum scoresranged from 0 to 21. Following the convention of other studies(Plummer et al., 2016), responses were recoded dichotomously basedon a cutoff score of 10 or higher to determine elevated anxiety.

The PTSD Checklist—Civilian Version (PCL-C), a validated 17-itemmeasure, was administered to assess PTSD symptoms (Weathers et al.,1993). Participants indicated how much they were bothered by pro-blems and experiences in response to stressful life events in the pastmonth, with 1 as “not at all” and 5 as “extremely.” Sum scores of the 17items were calculated and created into a dichotomous variable with acutoff score of 45 or greater, based on the psychometric properties forthe measure and as suggested by the National Center for PTSD(Blanchard et al., 1996).

2.2.3. Statistical analysesThe variables were normally distributed, with predictors indicating

acceptable levels of collinearity (VIF < 5). To identify potential riskand protective factors of mental health symptoms, three logistic re-gression models were performed to examine depression, anxiety, andPTSD symptoms as primary outcomes. Resilience, distress tolerance,perceived social support, instrumental social support, loneliness, andCOVID-19-specific worry were entered as predictors in unadjustedmodels. Age, gender, income, and race were entered in each of the threeadjusted models. All variables were binary with exception to age andincome, which were continuous. Two-tailed p-values were used. Toguard against Type I error, Bonferroni-adjustments were made to con-sider the 8 predictors and 4 covariates used in each model (.05/12=.004). Our results and interpretations are therefore based on asignificance set at p<.004 (note that the significance in the tables re-main unadjusted to provide more rather than less information to thereader). All analyses were performed using SPSS 25.0.

3. Results

Table 1 shows demographic characteristics of our participants anddescriptive data on all predictors and outcomes. The sample was ra-cially and ethnically diverse, with 59.6% White, 21.2% Asian, 5.3%Black, 6.0% Hispanic/Latino, 0.1% AI/NA, 6.2% mixed race, and 1.4%indicating another race. The majority of respondents were women(81.3%), U.S.-born (86.3%), employed (66.7%), students (61.3%), andthose who earned less than $50,000 per year (82.1%). Among thoseidentifying as students, 89.7% were enrolled as full-time and 7.3% wereinternational students. Overall, participants scored as having highloneliness (61.5%), low resilience (72.0%), and low distress tolerance(74.1%). At the same time, the majority of respondents reported havinghigh levels of social support (family, partners, peer, and instrumental).Finally, 43.3% of our sample had high levels of depression (PHQ-8scores ≥ 10), 45.4% had high anxiety scores (GAD-7 scores ≥ 10) and31.8% had high levels of PTSD symptoms (PCL-C scores ≥ 45).

Table 2 displays the associations between predictors and mentalhealth outcomes in each of the three models adjusted for the age,gender, race, and income. The results described here pertain only tosignificance set at p<.004 with Bonferroni corrections. Predictors thatwere significantly associated with depression, anxiety, and PTSD

Table 1Demographic characteristics and variable descriptives from Wave 1 of CARES2020.

Factors Means (range) or %

Age (years) 24.5 (18.0 – 30.9)18-21 28.6 %22-26 34.7 %26-30 36.6 %

GenderMen 14.1 %Women 81.3 %Other gender 4.6 %

RaceWhite 59.6 %Asian 21.2 %Black 5.3 %Hispanic or Latinx 6.0 %American Indian/Native American 0.1 %Mixed 6.2 %Other 1.4 %

U.S.-bornYes 86.3 %No 13.7 %

EmployedYes 66.7 %No 33.3 %

Individual Income (USD/year)No income 11.8 %< $25,000 45.9 %$25,000 – $49,999 24.4 %$50,000 – $74,999 11.6 %$75,000 – $99,999 2.6 %$100,000 – $124,999 2.1 %$125,000 – $149,999 0.3 %$150,000 – $174,999 0.3 %$175,000 – $199,999 0.6 %$200,000 – $249,999 0.2 %≥$250,000 0.2 %

StudentYes 61.3 %No 38.7 %

Student Enrollment Status (students only)Full time 89.7 %Part time 8.7 %Other 1.6 %

International StudentYes 7.3 %No 92.7 %

Loneliness (LS-SF) 6.1 (3.0 – 9.0)<6 38.5 %≥6 61.5 %

COVID-19-specific worry 15.9 (6.0 – 30.0)<24 89.9 %≥24 10.1 %

Resilience (CD-RISC-10) 26.0 (4 – 40)<30 72.0 %≥30 28.0 %

Distress tolerance (DTS) 3.3 (1.0 – 5.0)<4 74.1 %≥4 25.9 %

Family social support (MSPSS) 5.1 (1.0 – 7.0)<5 37.3 %≥5 62.7 %

Partner social support (MSPSS) 5.6 (1.0 – 7.0)<5 26.3 %≥5 73.7 %

Peer social support (MSPSS) 5.7 (1.0 – 7.0)<5 16.9 %≥5 83.1 %

Instrumental social support (2-Way SSS) 16.6 (1.0 – 20.0)<16 30.1 %≥16 69.9 %

Depression (PHQ-8) 9.0 (0 – 24.0)<10 56.7 %≥10 43.3 %

Anxiety (GAD-7) 9.4 (0 – 21.0)<10 54.6 %

(continued on next page)

C.H. Liu, et al. Psychiatry Research 290 (2020) 113172

3

included loneliness (OR range = 1.98 – 2.72), COVID-19-specific worry(OR range = 2.87 – 5.05), and distress tolerance (OR range = 0.22 –0.42). Specifically, those who endorsed high levels of loneliness andworries about COVID-19 and low levels of distress tolerance were morelikely to score above the clinical cutoffs for depression, anxiety, andPTSD. Those with high levels of resilience were less likely to scoreabove the cutoff for depression and anxiety. Those with high levels offamily support were less likely to score above the clinical cutoff fordepression and PTSD (OR = 0.46 and 0.44, respectively). Instrumentalsupport was negatively associated with depression. No associationswere obtained between support from partners and friends.

In analyses of associations between covariates and outcomes, ageand income were not associated with depression, anxiety, or PTSD.With regard to gender, men who identified as transgender were morelikely to report high levels of PTSD (OR = 4.20, CI = 1.62 – 10.89,p=.003); no differences were observed between men and women. AsianAmericans compared to Whites were less likely to report high levels ofdepression (OR = 0.50, CI = 0.33 – 0.76, p=.001) and PTSD(OR = 0.40, CI = 0.25 – 0.64, p<.001). Asians Americans andHispanic/Latinos were less likely to report high levels of anxiety(OR = 0.35, CI = 0.24 – 0.53, p<.001, OR = 0.35, CI = 0.18 – 0.68,p=.00, respectively).

4. Discussion

Our findings highlight major psychological challenges faced by

young adults during the initial weeks of the COVID-19 pandemic. Atleast one-third of young adults reported having clinically elevated le-vels of depression (43.3%), anxiety (45.4%), and PTSD symptoms(31.8%). The rates of depression, anxiety, and PTSD in our study areconsiderably higher compared to prior studies that have used the samecut points (PHQ-8 ≥ 10; GAD-7 ≥ 10; and PCL-C ≥ 45). For instance,PHQ-8 data collected from a study on U.S. adults in 2006 yielded aprevalence of 6.2% among 18-24-year-olds and a prevalence of 13.1%among 25-34-year-olds (Kroenke et al., 2009). Studies using the GAD-7showed the following rates among similar groups: U.S. primary carepatients (23.0%; Spitzer et al., 2006), U.S. college students (21.0%;Martin et al., 2014), and U.S. non-veteran community college students(17.4%; Fortney et al., 2016). Finally, studies using a cutoff of ≥ 45 onthe PCL-C to assess PTSD in trauma survivors showed the followingrates: U.S. patients following hospital discharge from traumatic ortho-pedic injury after one year (22.0%; Archer et al., 2016) and survivorsfrom the Wenchuan, China earthquake also after one year (26.3%;Zhang et al., 2011). The high rates from our sample may reflect ongoingdistress, as we measured the symptoms in the weeks following thegovernment directives for closures. Young adults may have been par-ticularly distressed in managing school or work responsibilities duringthis time while having no sense of certainty regarding the pandemic'send. As well, the high rate of mental health concerns among studyparticipants may be partially attributable to the specific characteristicsof our sample; given that the study was launched on the East Coast, ouryoung adult respondents may have been located at pandemic “hotspots,” with proximity to a greater number of COVID-19 cases poten-tially being an added stressor for our sample.

Strikingly, the majority of respondents reported feeling lonelyduring the first two months of the pandemic, as well as having lowresilience and low ability to tolerate distress. However, the majorityreported having social support from family, partners, and peers, as wellas instrumental support during this time. We note that the absoluterates of low perceived social support seem problematic. For instance,approximately 37% of respondents reported low family support. These

Table 1 (continued)

Factors Means (range) or %

≥10 45.4 %PTSD (PCL-C) 38.3 (17.0 – 85.0)<45 68.2 %≥45 31.8 %

N = 898

Table 2Odds ratios and confidence intervals for mental health outcomes from Wave 1 of CARES 2020.

Factors PHQ-8 – DepressionAdjusted ORa(95% CI) GAD-7 – AnxietyAdjusted ORa(95% CI) PTSD AdjustedAdjusted ORa(95% CI)

Loneliness (LS-SF)<6 1.0 1.0 1.0≥6 2.72 (1.92 – 3.87) ⁎⁎⁎ 1.98 (1.41 – 2.77) ⁎⁎⁎ 2.31 (1.55 – 3.43) ⁎⁎⁎

COVID-19-specific worry<24 1.0 1.0 1.0≥24 2.87 (1.67 – 4.94) ⁎⁎⁎ 4.12 (2.33 – 7.29) ⁎⁎⁎ 5.05 (2.92 – 874) ⁎⁎⁎

Resilience (CD-RISC-10)<30 1.0 1.0 1.0≥30 0.56 (0.38 – 0.83) ⁎⁎ 0.44 (0.30 – 0.64) ⁎⁎⁎ 0.70 (0.46 – 1.07)

Distress tolerance (DTS)<4 1.0 1.0 1.0≥4 0.36 (0.24 – 0.54) ⁎⁎⁎ 0.42 (0.28 – 0.62) ⁎⁎⁎ 0.22 (0.13 – 0.37) ⁎⁎⁎

Family social support (MSPSS)<5 1.0 1.0 1.0≥5 0.46 (0.32 – 0.66) ⁎⁎⁎ 0.64 (0.44 – 0.91)* 0.44 (0.30 – 0.64)⁎⁎⁎

Partner social support (MSPSS)<5 1.0 1.0 1.0≥5 1.26 (0.84 – 1.88) 1.32 (0.89 – 1.96) 1.00 (0.66 – 1.52)

Peer social support (MSPSS)<5 1.0 1.0 1.0≥5 1.05 (0.68 – 1.62) 1.27 (0.83 – 1.96) 0.88 (0.56 – 1.39)

Instrumental social support (2-Way SSS)<16 1.0 1.0 1.0≥16 0.60 (0.41 – 0.86)⁎⁎ 0.67 (0.46 – 0.96)* 0.63 (0.43 – 0.93)*

N = 898⁎ p<.05⁎⁎ p<.01⁎⁎⁎ p<.001 (two-tailed, without Bonferroni adjustment),a Adjusted covariates include age, race, gender, individual income

C.H. Liu, et al. Psychiatry Research 290 (2020) 113172

4

findings highlight major psychological challenges currently faced byyoung adults during the initial weeks of the COVID-19 pandemic.

Our study also identified factors associated with clinical levels ofdepression, anxiety, and PTSD symptoms. High loneliness and lowdistress tolerance levels were consistently associated with high levels ofdepression, anxiety, and PTSD. High levels of resilience were associatedwith low anxiety. Social support from family was associated with lowlevels of depression and PTSD symptoms, whereas support from part-ners or friends was not associated with any mental health outcomes.High levels of instrumental support were associated with low levels ofdepression.

Our data is consistent with findings demonstrating loneliness as arisk factor for mental health (Banerjee et al., 2020; Hawkley andCacioppo, 2010; Okruszek et al., 2020); this is particularly salient withgovernment directives for social distancing and isolation. Feeling cut offfrom social groups may lead one to feel vulnerable and pessimisticabout one's circumstances, altogether producing negative mood statesand anxiety (Muyan et al., 2016) that are further heightened during apandemic. The high levels of reported loneliness in our sample and itsassociation with depression, anxiety, and PTSD symptoms underscorethe severity of experiences of young adults during the pandemic.

Distress tolerance, or one's ability to manage and tolerate emotionaldistress, was strongly associated low levels of depressive and anxiety,and PTSD symptoms; individual resilience was associated with low le-vels of depression and anxiety symptoms, but not PTSD. Individualresilience, which encompasses personal competence and trust in one'sinstincts (Connor and Davidson, 2003), has been associated with lowlevels of depression, anxiety, and PTSD symptomatology after disasters(Blackmon et al., 2017). One's perceived ability to tolerate negative oraversive emotional and/or physical states may be more protective thanthe personal qualities that comprise psychological resilience, especiallyfor those experiencing symptoms of PTSD during a pandemic. Thepandemic is worldwide stressor without a foreseeable endpoint, and theeffects of the pandemic cannot be controlled by a single individual.Furthermore, the pandemic simultaneously impacts various domains(e.g., financial, relational, and health) with this stress potentially ex-acerbating the sensations associated with PTSD symptoms. As such,psychological resilience that is typically associated with overcomingsetbacks may not be sufficient for protecting against PTSD symptomswithin the first several weeks of a widespread pandemic. Interventionsthat target distress tolerance, such as mindfulness-based interventions,may be more effective than cognitive interventions targeting core be-liefs about the self especially for those with PTSD symptoms (Nila et al.,2016). Longitudinal approaches would help to examine this possibilityfurther.

Emotional support from family but not from friends and significantothers was associated with low levels of depression and PTSD. Friendsand significant others may have or are perceived to have less capacityto validate other's emotional experiences during a pandemic, con-sidering that they may be young adults who are experiencing similarstruggles. Emotional support provided by family may be more stableand coupled with the provision of material resources that young adultsmay still receive from parents. Our findings are consistent with priorwork showing that family support but not friend and partner supportmediates the effects of stress on health (Lee et al., 2018). Family sup-port may be more meaningful in providing reassurance to young adults,considering the possible concrete needs during the pandemic.

Instrumental support, or tangible assistance, may be an importantfactor for the mental health of young adults during the immediateweeks of the COVID-19 pandemic onset given that many were facedwith acute disruptions, such as unemployment, financial stress, andrelocation following university campus closures. However, instru-mental support was not significantly associated with any of the out-comes after adjusting the p-value to .004. Additional research is neededto clarify the respective roles on both emotional and instrument supportgiven variations in their potential effects on depression, anxiety, and

PTSD.Our newly developed COVID-19-related worry measure uniquely

predicted mental health symptoms, underscoring how the specific fea-tures of this pandemic give rise to acute stress. The stress resulting fromlifestyle changes due to features of COVID-19 itself may lead to greatermental health concerns distinct from the endorsement of other risks.Our analyses showed that the six items in our measure were reliable,and the total subscale score was significantly associated with thesymptoms assessed in this study; however, additional work is requiredto determine the validity of this measure.

In general, Asian Americans were less likely to report high levels ofmental health symptoms compared to Whites, with Hispanic/Latinxrespondents also being less likely to report high anxiety. Asian andLatinx immigrants compared to those who are born in the U.S. are lesslikely to endorse psychological distress (Dey and Lucas, 2006;Takeuchi et al., 2007). It is possible that other experiences such asethnic identity, social networking, and family cohesion serve as a pro-tective factor for mental health, especially for non-U.S.-born partici-pants (Leong et al., 2013). The under-recognition of distress symptomsmay also be possible among ethnic minorities (Liu et al., 2020). Al-though our sample size of gender minorities was small, men whoidentified as transgender were more likely to report a high level ofPTSD symptoms, consistent with prior research (Reisner et al., 2016;Shipherd et al., 2011). Greater attention to gender differences in mentalhealth symptoms as well as a deeper study regarding the specific ex-periences faced by racial/ethnic and gender minorities during pan-demic is warranted.

The cross-sectional design limits our ability to infer causality in-volved in leading to mental health problems. We used a conveniencesample, and caution must be taken in the generalizability of our find-ings to the broader population of young adults in the U.S. given theuneven sampling of subgroups. The reliance of self-report itself haslimitations, such that it may be prone to misinterpretation. Futureanalyses with the anticipated waves of data collection will enable us toexamine the association of our predictors to outcome measures ofmental health and to adjust for additional confounds. As well, we willhave an opportunity to examine potential moderation effects to un-derstand whether outcomes vary by circumstances or individual char-acteristics, such as socioeconomic capital, social support type, distresstolerance, and resilience.

To our knowledge, our study is the first prospective cohort study toassess mental health outcomes and risk and resilience factors in U.S.young adults during the first several weeks of the COVID-19 pandemic.In our study, one in three U.S. young adults reported clinical cut-offsymptoms of depression, anxiety, and PTSD as well as high levels ofloneliness. We present new evidence that signifies the roles of lone-liness, distress tolerance, family support, and COVID-19-related worryon mental health outcomes during the first month of the COVID-19pandemic. Mental health interventions should incorporate these con-structs to help mediate the impact of COVID-19 on adverse mentalhealth status among U.S. young adults.

CRediT authorship contribution statement

Cindy H. Liu: Conceptualization, Methodology, Formal analysis,Investigation, Writing – original draft, Writing – review & editing,Project administration, Supervision, Funding acquisition. Emily Zhang:Data curation, Writing – original draft, Writing – review & editing,Project administration. Ga Tin Fifi Wong: Data curation, Writing -original draft, Project administration. Sunah Hyun: Writing – review &editing. Hyeouk “Chris” Hahm: Conceptualization, Writing – review &editing, Supervision, Funding acquisition.

Declaration of Competing Interest

There are no conflicts of interest to declare.

C.H. Liu, et al. Psychiatry Research 290 (2020) 113172

5

Acknowledgments

Support for this manuscript was provided through the NationalScience Foundation (2027553) award (to C.H.L. and H.C.H.), a Mary A.Tynan Faculty Fellowship and a NIMH K23 MH 107714-01 A1 award(to C.H.L.), as well as a T32 MH 16259-39 award (to. S.H.).

Supplementary materials

Supplementary material associated with this article can be found, inthe online version, at doi:10.1016/j.psychres.2020.113172.

References

Andrews, G., Slade, T., 2001. Interpreting scores on the Kessler Psychological DistressScale (K10). Aust. New Zealand J. Public Health 25, 494–497. https://doi.org/10.1111/j.1467-842X.2001.tb00310.x.

Archer, K.R., Heins, S.E., Abraham, C.M., Obremskey, W.T., Wegener, S.T., Castillo, R.C.,2016. Clinical significance of pain at hospital discharge following traumatic ortho-paedic injury: general health, depression, and PTSD outcomes at 1 year. Clin. J. Pain32, 196–202. https://doi.org/10.1097/AJP.0000000000000246.

Banerjee, S., Burkholder, G., Sana, B., Szirony, M., 2020. Social Isolation as a predictor formortality: Implications for COVID-19 prognosis. medRxiv 2020.04.15.20066548.https://doi.org/10.1101/2020.04.15.20066548.

Blackmon, B.J., Lee, J., Cochran, D.M., Kar, B., Rehner, T.A., Baker, A.M., 2017. Adaptingto life after hurricane Katrina and the deepwater horizon oil spill: an examination ofpsychological resilience and depression on the Mississippi Gulf Coast. Social WorkPublic Health 32, 65–76. https://doi.org/10.1080/19371918.2016.1188746.

Blanchard, E.B., Jones-Alexander, J., Buckley, T.C., Forneris, C.A., 1996. Psychometricproperties of the PTSD checklist (PCL). Behav. Res. Therapy 34, 669–673. https://doi.org/10.1016/0005-7967(96)00033-2.

Blazer, D.G., Kessler, R.C., McGonagle, K.A., Swartz, M.S., 1994. The prevalence anddistribution of major depression in a national community sample: The NationalComorbidity Survey. Am. J. Psychiatry 151, 979–986. https://doi.org/10.1176/ajp.151.7.979.

Breslau, N., Chilcoat, H.D., Kessler, R.C., Davis, G.C., 1999. Previous exposure to traumaand PTSD effects of subsequent trauma: results from the detroit area survey oftrauma. AJP 156, 902–907. https://doi.org/10.1176/ajp.156.6.902.

Breslau, N., Peterson, E.L., Schultz, L.R., 2008. A second look at prior trauma and theposttraumatic stress disorder effects of subsequent trauma: a prospective epidemio-logical study. Arch. Gen. Psychiatry 65, 431–437. https://doi.org/10.1001/archpsyc.65.4.431.

Brunet, A., Boyer, R., Weiss, D.S., Marmar, C.R., 2001. The effects of initial trauma ex-posure on the symptomatic response to a subsequent trauma. Can. J. Behav. Sci. /Revue canadienne des sciences du comportement 33, 97–102. https://doi.org/10.1037/h0087132.

Chen, J.A., Stevens, C., Wong, S.H.M., Liu, C.H., 2019. Psychiatric symptoms and diag-noses among U.S. college students: a comparison by race and ethnicity. Psychiatr.Serv. 70, 442–449. https://doi.org/10.1176/appi.ps.201800388.

Cohen, J.R., Danielson, C.K., Adams, Z.W., Ruggiero, K.J., 2016. Distress tolerance andsocial support in adolescence: predicting risk for internalizing and externalizingsymptoms following a natural disaster. J. Psychopathol. Behav. Assess. 38, 538–546.https://doi.org/10.1007/s10862-016-9545-y.

Connor, K.M., Davidson, J.R.T., 2003. Development of a new resilience scale: the Connor-Davidson Resilience Scale (CD-RISC). Depression Anxiety 18, 76–82. https://doi.org/10.1002/da.10113.

Conrad, R., 2020. Universities’ response to supporting mental health of college studentsduring the COVID-19 pandemic [WWW Document]. Psychiatric Times URL. https://www.psychiatrictimes.com/article/universities%E2%80%99-response-supporting-mental-health-college-students-during-covid-19-pandemic (accessed 4.26.20).

Costello, E.J., Erkanli, A., Fairbank, J.A., Angold, A., 2002. The prevalence of potentiallytraumatic events in childhood and adolescence. J. Traumatic Stress 15, 99–112.https://doi.org/10.1023/A:1014851823163.

Dey, A.N., Lucas, J.W., 2006. Physical and mental health characteristics of US-and for-eign-born adults: United States, 1998–2003. Adv. Data 369, 1–19.

Domagala-Krecioch, A., Majerek, B., 2013. The issue of loneliness in the period of“emerging adulthood.”. Eur. Scientif. J.

Eisenberg, D., Gollust, S.E., Golberstein, E., Hefner, J.L., 2007. Prevalence and correlatesof depression, anxiety, and suicidality among university students. Am. J.Orthopsychiatry 77, 534–542. https://doi.org/10.1037/0002-9432.77.4.534.

Fortney, J.C., Curran, G.M., Hunt, J.B., Cheney, A.M., Lu, L., Valenstein, M., Eisenberg,D., 2016. Prevalence of probable mental disorders and help-seeking behaviors amongveteran and non-veteran community college students. General Hospital Psychiatry38, 99–104. https://doi.org/10.1016/j.genhosppsych.2015.09.007.

Hawkley, L.C., Cacioppo, J.T., 2010. Loneliness matters: a theoretical and empirical re-view of consequences and mechanisms. Ann. Behav. Med. 40, 218–227. https://doi.org/10.1007/s12160-010-9210-8.

Hughes, M.E., Waite, L.J., Hawkley, L.C., Cacioppo, J.T., 2004. A short scale for mea-suring loneliness in large surveys: results from two population-based studies. Res.Aging 26, 655–672. https://doi.org/10.1177/0164027504268574.

Kessler, R., Mroczek, D., 1992. An update of the development of mental health screening

scales for the US National Health Interview Study. University of Michigan, SurveyResearch Center of the Institute for Social Research, Ann Arbor.

Kessler, R.C., Galea, S., Gruber, M.J., Sampson, N.A., Ursano, R.J., Wessely, S., 2008.Trends in mental illness and suicidality after Hurricane Katrina. Mol. Psychiatry 13,374–384. https://doi.org/10.1038/sj.mp.4002119.

Kroenke, K., Strine, T.W., Spitzer, R.L., Williams, J.B.W., Berry, J.T., Mokdad, A.H., 2009.The PHQ-8 as a measure of current depression in the general population. J. AffectDisord. 114, 163–173. https://doi.org/10.1016/j.jad.2008.06.026.

Kukihara, H., Yamawaki, N., Uchiyama, K., Arai, S., Horikawa, E., 2014. Trauma, de-pression, and resilience of earthquake/tsunami/nuclear disaster survivors of Hirono,Fukushima, Japan. Psychiatry Clin. Neurosci. 68, 524–533. https://doi.org/10.1111/pcn.12159.

Lee, C.-Y.S., Goldstein, S.E., Dik, B.J., 2018. The relational context of social support inyoung adults: links with stress and well-being. J. Adult Dev. 25, 25–36. https://doi.org/10.1007/s10804-017-9271-z.

Leong, F., Park, Y.S., Kalibatseva, Z., 2013. Disentangling immigrant status in mentalhealth: psychological protective and risk factors among Latino and Asian Americanimmigrants. Am. J. Orthopsychiatry 83, 361–371. https://doi.org/10.1111/ajop.12020.

Liu, C.H., Li, H., Wu, E., Tung, E.S., Hahm, H.C., 2020. Parent perceptions of mentalillness in Chinese American youth. Asian J. Psychiatry 47, 101857. https://doi.org/10.1016/j.ajp.2019.101857.

Liu, C.H., Stevens, C., Wong, S.H.M., Yasui, M., Chen, J.A., 2019. The prevalence andpredictors of mental health diagnoses and suicide among U.S. college students:Implications for addressing disparities in service use. Depression Anxiety 36, 8–17.https://doi.org/10.1002/da.22830.

Lowthian, J.A., Lennox, A., Curtis, A., Dale, J., Browning, C., Smit, D.V., Wilson, G.,O'Brien, D., Rosewarne, C., Boyd, L., Garner, C., Cameron, P., 2016. HOspitals andpatients WoRking in Unity (HOW R U?): protocol for a prospective feasibility study oftelephone peer support to improve older patients’ quality of life after emergencydepartment discharge. BMJ Open 6, e013179. https://doi.org/10.1136/bmjopen-2016-013179.

Martin, R.J., Usdan, S., Cremeens, J., Vail-Smith, K., 2014. Disordered gambling and co-morbidity of psychiatric disorders among college students: An examination of pro-blem drinking, anxiety and depression. J. Gambl. Stud. 30, 321–333. https://doi.org/10.1007/s10899-013-9367-8.

Mojtabai, R., Olfson, M., Han, B., 2016. National trends in the prevalence and treatmentof depression in adolescents and young adults. Pediatrics 138, e20161878.

Muyan, M., Chang, E.C., Jilani, Z., Yu, T., Lin, J., Hirsch, J.K., 2016. Loneliness andnegative affective conditions in adults: is there any room for hope in predicting an-xiety and depressive symptoms? J. Psychol. 150, 333–341. https://doi.org/10.1080/00223980.2015.1039474.

Nila, K., Holt, D.V., Ditzen, B., Aguilar-Raab, C., 2016. Mindfulness-based stress reduction(MBSR) enhances distress tolerance and resilience through changes in mindfulness.Mental Health Prevention 4, 36–41. https://doi.org/10.1016/j.mhp.2016.01.001.

Okruszek, L., Aniszewska-Stańczuk, A., Piejka, A., Wiśniewska, M., Żurek, K., 2020. Safebut lonely? Loneliness Mental Health Symptoms COVID-19.

Plummer, F., Manea, L., Trepel, D., McMillan, D., 2016. Screening for anxiety disorderswith the GAD-7 and GAD-2: a systematic review and diagnostic metaanalysis. GeneralHospital Psychiatry 39, 24–31. https://doi.org/10.1016/j.genhosppsych.2015.11.005.

Reisner, S.L., White Hughto, J.M., Gamarel, K.E., Keuroghlian, A.S., Mizock, L.,Pachankis, J.E., 2016. Discriminatory experiences associated with posttraumaticstress disorder symptoms among transgender adults. J. Counsel. Psychol. 63, 509.

Reynolds, K., Pietrzak, R.H., Mackenzie, C.S., Chou, K.L., Sareen, J., 2016. Post-TraumaticStress Disorder Across the Adult Lifespan: Findings From a Nationally RepresentativeSurvey. Am. J. Geriatric Psychiatry 24, 81–93. https://doi.org/10.1016/j.jagp.2015.11.001.

Shakespeare-Finch, J., Obst, P.L., 2011. The development of the 2-way social supportscale: a measure of giving and receiving emotional and instrumental support. J. Pers.Assess. 93, 483–490. https://doi.org/10.1080/00223891.2011.594124.

Shipherd, J.C., Maguen, S., Skidmore, W.C., Abramovitz, S.M., 2011. Potentially trau-matic events in a transgender sample: frequency and associated symptoms.Traumatology 17, 56–67. https://doi.org/10.1177/1534765610395614.

Simons, J.S., Gaher, R.M., 2005. The distress tolerance scale: development and validationof a self-report measure. Motiv. Emot. 29, 83–102. https://doi.org/10.1007/s11031-005-7955-3.

Spitzer, R.L., Kroenke, K., Williams, J.B.W., Löwe, B., 2006. A brief measure for assessinggeneralized anxiety disorder: the GAD-7. Arch. Intern. Med. 166, 1092–1097.https://doi.org/10.1001/archinte.166.10.1092.

Takeuchi, D.T., Zane, N., Hong, S., Chae, D.H., Gong, F., Gee, G.C., Walton, E., Sue, S.,Alegría, M., 2007. Immigration-related factors and mental disorders among AsianAmericans. Am. J. Public Health 97, 84–90. https://doi.org/10.2105/AJPH.2006.088401.

Tymoszuk, U., Perkins, R., Fancourt, D., Williamon, A., 2019. Cross-sectional and long-itudinal associations between receptive arts engagement and loneliness among olderadults. Soc. Psychiatry Psychiatr. Epidemiol. https://doi.org/10.1007/s00127-019-01764-0.

Vrana, S., Lauterbach, D., 1994. Prevalence of traumatic events and post-traumatic psy-chological symptoms in a nonclinical sample of college students. J. Trauma Stress 7,289–302. https://doi.org/10.1007/BF02102949.

Weathers, F.W., Litz, B.T., Herman, D.S., Huska, J.A., Keane, T.M., 1993. The PTSDChecklist (PCL): Reliability, validity, and diagnostic utility, in: Annual Convention ofthe International Society for Traumatic Stress Studies, San Antonio, TX. SanAntonio, TX.

Wu, Y., Levis, B., Riehm, K.E., Saadat, N., Levis, A.W., Azar, M., Rice, D.B., Boruff, J.,

C.H. Liu, et al. Psychiatry Research 290 (2020) 113172

6

Cuijpers, P., Gilbody, S., Ioannidis, J.P.A., Kloda, L.A., McMillan, D., Patten, S.B.,Shrier, I., Ziegelstein, R.C., Akena, D.H., Arroll, B., Ayalon, L., Baradaran, H.R.,Baron, M., Bombardier, C.H., Butterworth, P., Carter, G., Chagas, M.H., Chan, J.C.N.,Cholera, R., Conwell, Y., Ginkel, J.M., de, M., Fann, J.R., Fischer, F.H., Fung, D.,Gelaye, B., Goodyear-Smith, F., Greeno, C.G., Hall, B.J., Harrison, P.A., Härter, M.,Hegerl, U., Hides, L., Hobfoll, S.E., Hudson, M., Hyphantis, T., Inagaki, M., Jetté, N.,Khamseh, M.E., Kiely, K.M., Kwan, Y., Lamers, F., Liu, S.-I., Lotrakul, M., Loureiro,S.R., Löwe, B., McGuire, A., Mohd-Sidik, S., Munhoz, T.N., Muramatsu, K., Osório,F.L., Patel, V., Pence, B.W., Persoons, P., Picardi, A., Reuter, K., Rooney, A.G., Santos,I.S., Shaaban, J., Sidebottom, A., Simning, A., Stafford, L., Sung, S., Tan, P.L.L.,

Turner, A., van Weert, H.C., White, J., Whooley, M.A., Winkley, K., Yamada, M.,Benedetti, A., Thombs, B.D., 2019. Equivalency of the diagnostic accuracy of thePHQ-8 and PHQ-9: a systematic review and individual participant data meta-analysis.Psychol. Med. 1–13. https://doi.org/10.1017/S0033291719001314.

Zhang, Z., Shi, Z., Wang, L., Liu, M., 2011. One year later: Mental health problems amongsurvivors in hard-hit areas of the Wenchuan earthquake. Public Health 125, 293–300.https://doi.org/10.1016/j.puhe.2010.12.008.

Zimet, G.D., Dahlem, N.W., Zimet, S.G., Farley, G.K., 1988. The multidimensional scale ofperceived social support. J. Pers. Assess. 52, 30–41. https://doi.org/10.1207/s15327752jpa5201_2.

C.H. Liu, et al. Psychiatry Research 290 (2020) 113172

7