Abstract
Objectives
An understanding of mental health symptoms during the coronavirus disease 2019 (COVID-19) pandemic is critical to ensure that health policies adequately address the mental health needs of people in the United States. The objective of this study was to examine mental health symptoms among US adults in an early stage of the COVID-19 pandemic.
Methods
We conducted a cross-sectional study in late March 2020 with a national sample of 963 US adults using an online research platform. Participants self-reported state of residence, psychosocial characteristics, and levels of anxiety, depression, anger, cognitive function, and fatigue in the context of COVID-19 using validated patient-reported outcomes scales in the Patient-Reported Outcome Measurement Information System measures. We used analysis of variance and multivariate linear regression to evaluate correlates of mental health symptoms.
Results
Overall, participants reported high levels of anxiety (mean [SD], 57.2 [9.3]) and depression (mean [SD], 54.2 [9.5]). Levels of anger, anxiety, cognitive function, depression, and fatigue were significantly higher among the Millennial Generation and Generation X (vs Baby Boomers), those with not enough or enough (vs more than enough) financial resources, females vs males), those with self-reported disability (vs no self-reported disability), and those with inadequate (vs adequate) health literacy. In adjusted models, being in Generation X and the Millennial Generation (vs Baby Boomer), having not enough or enough vs more than enough) financial resources, and having inadequate (vs adequate) health literacy were most strongly correlated with worse mental health symptoms.
Conclusions
Results suggest that mental health symptoms during the early stages of the COVID-19 pandemic were prevalent nationally, regardless of state of residence and especially among young, psychosocially vulnerable groups.
Keywords: mental health and well-being, psychosocial factors, health disparities, prevention, public health assessment
As of August 2020, coronavirus disease 2019 (COVID-19) had infected more than 5.4 million people in the United States and led to nearly 170 000 deaths,1 more than in any other country worldwide.2 In response, research on the clinical presentation and effect of COVID-19 on physical health has rapidly proliferated. These data have been important to updating clinical guidelines and shaping national, state, and local public health policy.3,4
Conversely, few studies investigating mental health during COVID-19 in the United States have been published.5 Mental health is particularly salient given the ubiquitous social and economic changes caused by the pandemic, in addition to anxiety and stress related to the infection itself. For example, in the United States, stay-at-home orders have disrupted routine life, and more than 40 million people in the United States have filed for unemployment benefits since March 2020.6 These dramatic shifts may lead to a broad range of mental health symptoms including and beyond anxiety and depression, such as anger, fatigue, and cognitive dysfunction.7 Although researchers have described worsening mental health symptoms in many countries with high COVID-19 prevalence globally,8-11 mental health symptoms during the COVID-19 pandemic have yet to be well documented in the United States. Documentation of mental health symptoms in many of these countries has directly informed mental health policies, including the reallocation of resources and improvement of access to diagnosis and treatment.11-13 Thus, evidence about the mental health of US adults is needed to shape health policies that comprehensively and effectively address mental health during the COVID-19 era.
Many global studies have also noted disparities in mental health symptoms during the pandemic.8,9,13 Dramatic social and economic changes may disproportionately affect the mental health of some people more than others in the United States. For example, geographic location may have played a role in mental health outcomes, especially during the early stages of COVID-19 in the United States.7 As of late March 2020, most COVID-19 cases and deaths were concentrated in just a few states. Rural–urban disparities in access to mental health services may also lead to state-level differences in mental health.14 In addition, psychosocial characteristics may influence mental health symptoms. Previous work demonstrated that differences in adherence to public health measures and effective coping in response to public health crises by psychosocial characteristics lead to differences in mental health outcomes.15 Evidence of differences in mental health symptoms, especially among psychosocially vulnerable populations (eg, older adults, racial/ethnic minority groups, people in low socioeconomic groups), will be important to inform mental health policies in the United States that seek to reduce disparities rather than exacerbate them.
The objective of this study was to describe mental health symptoms among US adults in an early stage of the COVID-19 pandemic in the United States and identify correlates of severe symptoms. We hypothesized that psychosocial characteristics and residence in states with a high prevalence of COVID-19 cases, deaths, unemployment, and people living in rural settings would be associated with poor mental health symptoms.
Methods
Study Design and Sample
We conducted a national cross-sectional study from March 24 through April 4, 2020, using Prolific, an online research platform (www.prolific.co). Prolific is a widely used academic research platform that recruits participants using opt-in methods and performs quality checks to ensure participants are human and attentive to questions. Prolific uses a proprietary random-sampling algorithm that uses data from the US Census to ensure representativeness of samples within demographic strata.16 Using Prolific’s algorithm, we recruited a sample of 963 US adults that was representative of the US population on age, sex, race/ethnicity, and geographic location. Participants provided written informed consent and received $15 per hour as compensation. The Weill Cornell Medicine Institutional Review Board approved this study.
Data Collection and Measures
We created a 5-minute online survey that we pretested and revised based on feedback from 5 members of the general public. Using this survey, participants self-reported psychosocial characteristics. They reported data on the following demographic characteristics: age (categorized as Millennial [≤38], Generation X [39-54], and Baby Boomer [≥55]), sex (female, male, nonbinary), and race/ethnicity (non-Hispanic White, racial/ethnic minority [Hispanic and non-Hispanic Black/African American, American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiracial, or other]). They also reported data on socioeconomic characteristics, including highest level of education obtained (≤high school diploma, bachelor’s or associate’s degree, postgraduate degree), financial resources (enough, more than enough, or not enough), and disability status (physical, hearing, vision, other, or none). They also completed the 3-item Brief Health Literacy Screen,17 which asks participants how confident they are filling out medical forms by themselves, how often they have problems learning about a health condition because of difficulty understanding written information, and how often they have someone help them read health materials.
Participants were then asked to report their mental health symptoms in relation to the current COVID-19 pandemic, including anger, anxiety, cognitive function, depression, and fatigue. Participants completed the following short-form Patient-Reported Outcomes Measurement Information System (PROMIS) questionnaires, which ask about perceived severity of symptoms and the impact of symptoms on daily functioning. We used the following surveys: Anxiety 4-item short-form version 1.0, Anger 5-item short-form version 1.1, Depression 4-item short-form version 1.0, Cognitive Function 4-item short-form version 2.0, and Fatigue 4-item short-form version 1.0. The National Institutes of Health developed PROMIS as a gold standard for patient-reported outcomes measurement.18 PROMIS measures have been validated against symptom-specific legacy instruments (eg, Patient Health Questionnaire–9 for depression) and facilitate comparison of the severity of multiple symptoms by using standardized scoring. A mean PROMIS T-score of 50 and a standard deviation (SD) of 10 represents the general US population for each measure.19 The creators of PROMIS scaled the range of possible T-scores for each measure, but scores generally range from 30 to 70. For all measures except cognitive function, a higher T-score indicates worse symptoms. For cognitive function, a lower T-score indicates worse symptoms.
In this analysis, we used state-level data on COVID-19 cases and deaths as of March 30, 2020, from the Centers for Disease Control and Prevention.1 In addition, we accessed 2019 state-level population estimates20 and the population percentage living in rural versus urban communities for 201921 from the US Census Bureau. We calculated case and death rates (per 100 000 population) using state-level population estimates. We accessed the seasonally adjusted percentage of the civilian workforce who filed for unemployment benefits in each state in February and March 2020,22 and we calculated the change in unemployment rates from February to March for each state.
Statistical Analysis
We characterized age as Millennial (aged ≤38), Generation X (aged 39-54), and Baby Boomer (aged ≥55).23 Of 963 participants included in the study, relatively few were in Generation Z (aged ≤23; n = 82) and the Silent Generation (aged ≥75; n = 9). Therefore, these generations were combined into the Millennial and Baby Boomer generations, respectively. We determined participants’ home states using geographic coordinates at the time of survey completion. The geographic representativeness criteria led to a small sample size of participants in certain states; therefore, we categorized the state-level variables by quartile rather than by individual state.
We first conducted basic descriptive statistics and generated US map visualizations using R version 3.6.3 (R Core Team) to describe the geographic distribution of the sample, case and death rates, unemployment rates, rural settings, and average PROMIS scores. We then used analysis of variance (ANOVA) to examine differences in mental health symptoms by each variable, followed by pairwise comparisons using bivariate linear models. We set the significance at P < .05 for ANOVA models. Finally, we used multivariate linear regression models to examine several correlates of worse symptoms and reported β estimates and 95% CIs. We conducted all analyses using R version 3.6.3 (R Core Team).
Results
Participants’ ages ranged from 18 to 83 (mean [SD], 45 [15.8]). Of 963 participants, 368 (38.2%) were in the Millennial generation, 268 (27.8%) were in Generation X, and 327 (34.0%) were in the Baby Boomer generation; 488 (50.7%) were female, 686 (71.2%) were non-Hispanic White, 332 (34.5%) had ≤high school diploma, 287 (29.8%) did not have enough financial resources, 144 (15.0%) reported having a disability, and 271 (28.1%) had inadequate health literacy (Table 1). Of all 5 symptoms measured, the scores were highest for symptoms of anxiety (mean [SD] score, 57.2 [9.3]) and depression (mean [SD] score, 54.2 [9.5]). US map visualizations showed the severity of anxiety and depression nationally (Figure).
Table 1.
Psychosocial characteristics and PROMIS scores of a sample of adults (N = 963) during the coronavirus disease 2019 (COVID-19) pandemic, United States, March–April 2020
| Characteristic | No. (%) |
|---|---|
| Age, y | |
| ≤38 (Millennial) | 368 (38.2) |
| 39-54 (Generation X) | 268 (27.8) |
| ≥55 (Baby Boomer) | 327 (34.0) |
| Gender | |
| Female | 488 (50.7) |
| Male | 465 (48.3) |
| Nonbinary | 10 (1.0) |
| Race/ethnicity | |
| Non-Hispanic White | 686 (71.2) |
| Racial/ethnic minoritya | 277 (28.8) |
| Education | |
| ≤High school diploma | 332 (34.5) |
| Bachelor’s/associate’s degree | 456 (47.4) |
| Postgraduate degree | 175 (18.2) |
| Financial resources | |
| More than enough | 105 (10.9) |
| Enough | 571 (59.3) |
| Not enough | 287 (29.8) |
| Self-reported disabilityb | |
| Yes | 144 (15.0) |
| No | 819 (85.0) |
| Health literacy | |
| Adequate | 692 (71.9) |
| Inadequate | 271 (28.1) |
| PROMISc scores, mean (SD) | |
| Anger | 50.2 (9.9) |
| Anxiety | 57.2 (9.3) |
| Cognitive function | 48.4 (9.6) |
| Depression | 54.2 (9.5) |
| Fatigue | 51.4 (10.3) |
Abbreviations: PROMIS, Patient-Reported Outcomes Measurement Information System; SD, standard deviation.
aIncludes those identifying as Hispanic and non-Hispanic Black/African American, American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiracial, or other.
bDefined as self-reported physical, hearing, vision, or other disability.
cPROMIS measures are validated, standardized, patient-reported outcomes measures assessing the severity of symptoms and their effect on daily functioning. A mean (SD) of 50 (10) represents the general US population; higher scores of anger, anxiety, depression, and fatigue indicate worse symptoms, and lower cognitive function scores indicate worse symptoms.19
Figure.
Sample distribution, Patient-Reported Outcomes Measurement Information System (PROMIS) measures, and state characteristics (change in unemployment and coronavirus disease 2019 [COVID-19] case and death rate per 100 000 population) among a sample of adults (N = 963) during the COVID-19 pandemic, United States, March–April 2020. PROMIS measures are validated, standardized, patient-reported outcomes measures assessing the severity of symptoms and their effect on daily functioning. A mean (standard deviation) of 50 (10) represents the general US population; higher scores of anger, anxiety, depression, and fatigue indicate worse symptoms, and lower cognitive function scores indicate worse symptoms.19
In ANOVA analyses, symptoms differed significantly across nearly all psychosocial characteristics (Table 2) but not by any state-level variables, such as COVID-19 case rates, death rates, and unemployment rates (Table 3; Figure). For example, for participants in the Millennial generation, Generation X, and the Baby Boomer generation, mean (SD) anxiety scores were 58.3 (9.4), 58.1 (9.3), and 55.2 (8.8), respectively, and for depression were 56.2 (9.7), 55.2 (9.4), and 51.1 (8.6), respectively (all P < .001). For participants reporting not enough, enough, and more than enough financial resources, mean (SD) anxiety scores were 60.3 (9.1), 56.1 (9.0), and 54.7 (9.3), respectively, and for depression were 58.1 (9.6), 52.9 (9.0), and 50.4 (8.8), respectively (all P < .001). For participants with inadequate health literacy and adequate health literacy, the mean (SD) anxiety scores were 59.3 (9.0) and 56.4 (9.3), respectively, and depression scores were 56.9 (9.5) and 53.1 (9.3), respectively (all P < .001). We also found significant differences between these groups for anger, cognitive function, and fatigue.
Table 2.
Mean (SD) and analysis of variance of PROMIS measuresa by psychosocial characteristics among a sample of adults (N = 963) during the coronavirus disease 2019 (COVID-19) pandemic, United States, March–April 2020
| Psychosocial characteristic | No. | PROMIS measure | ||||
|---|---|---|---|---|---|---|
| Anger | Anxiety | Cognitive function | Depression | Fatigue | ||
| Generation | ||||||
| Millennial (n = 368) | 368 | 51.7 (10.4) | 58.3 (9.4) | 46.8 (9.4) | 56.2 (9.7) | 53.5 (10.3) |
| Generation X (n = 268) | 268 | 51.1 (9.7) | 58.1 (9.3) | 47.7 (9.9) | 55.2 (9.4) | 51.8 (10.6) |
| Baby Boomer (n = 327) | 327 | 47.8 (9.2) | 55.2 (8.8) | 51.0 (8.9) | 51.1 (8.6) | 48.7 (9.9) |
| P valueb | <.001 | <.001 | <.001 | <.001 | <.001 | |
| Gender | ||||||
| Female (n = 488) | 488 | 51.1 (10.0) | 58.7 (9.1) | 47.6 (9.9) | 55.1 (9.3) | 52.7 (10.4) |
| Male | 465 | 49.1 (9.8) | 55.5 (9.2) | 49.6 (9.0) | 53.1 (9.7) | 49.9 (10.2) |
| Nonbinary | 10 | 55.3 (8.3) | 63.5 (8.8) | 40.9 (10.3) | 60.6 (5.1) | 59.4 (8.1) |
| P valueb | .002 | <.001 | <.001 | <.001 | <.001 | |
| Race/ethnicity | ||||||
| Non-Hispanic White | 686 | 50.2 (9.4) | 57.5 (9.1) | 48.5 (9.4) | 54.3 (9.2) | 51.6 (10.2) |
| Racial/ethnic minorityc | 277 | 50.2 (11.1) | 56.4 (9.8) | 48.3 (10.0) | 53.9 (10.3) | 50.9 (10.9) |
| P valueb | .95 | .08 | .80 | .56 | .34 | |
| Education | ||||||
| ≤High school diploma | 332 | 50.8 (10.1) | 57.2 (9.5) | 48.7 (9.8) | 55.1 (9.8) | 52.3 (10.6) |
| Bachelor’s/associate’s degree | 456 | 49.9 (9.9) | 57.1 (9.0) | 48.6 (9.5) | 53.6 (9.4) | 51.0 (10.5) |
| Postgraduate degree | 175 | 49.7 (9.8) | 57.5 (9.5) | 47.8 (9.2) | 54.0 (9.3) | 50.9 (9.9) |
| P valueb | .17 | .81 | .39 | .11 | .10 | |
| Financial resources | ||||||
| Enough | 571 | 49.2 (9.5) | 56.1 (9.0) | 49.2 (9.4) | 52.9 (9.0) | 50.1 (10.0) |
| More than enough | 105 | 47.5 (10.2) | 54.7 (9.3) | 50.5 (9.7) | 50.4 (8.8) | 48.5 (10.4) |
| Not enough | 287 | 53.1 (10.2) | 60.3 (9.1) | 46.2 (9.5) | 58.1 (9.6) | 55.1 (10.3) |
| P value | <.001 | <.001 | <.001 | <.001 | <.001 | |
| Self-reported disabilityd | ||||||
| No | 819 | 49.8 (9.7) | 56.9 (9.2) | 49.0 (9.4) | 53.7 (9.3) | 50.6 (10.3) |
| Yes | 144 | 52.3 (10.8) | 59.1 (9.7) | 45.6 (10.2) | 56.8 (10.1) | 56.3 (10.0) |
| P valueb | .01 | .01 | <.001 | <.001 | <.001 | |
| Health literacy | ||||||
| Adequate | 692 | 49.2 (9.9) | 56.4 (9.3) | 49.6 (9.5) | 53.1 (9.3) | 50.4 (10.3) |
| Inadequate | 271 | 52.6 (9.6) | 59.3 (9.0) | 45.9 (9.3) | 56.9 (9.5) | 53.9 (10.3) |
| P valueb | <.001 | <.001 | <.001 | <.001 | <.001 | |
Abbreviations: PROMIS, Patient-Reported Outcomes Measurement Information System; SD, standard deviation.
aPROMIS measures are validated, standardized, patient-reported outcomes measures assessing the severity of symptoms and their effect on daily functioning. A mean (SD) of 50 (10) represents the general US population; higher scores of anger, anxiety, depression, and fatigue indicate worse symptoms, and lower cognitive function scores indicate worse symptoms.19 All values are mean (SD) unless otherwise indicated.
bEvaluated using analysis of variance, with P < .05 considered significant.
cIncludes those identifying as Hispanic and non-Hispanic Black/African American, American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiracial, or other.
dDefined as self-reported physical, hearing, vision, or other disability.
Table 3.
Mean (SD) and analysis of variance of PROMIS measuresa by state characteristics among a national sample of adults (N = 963) during the coronavirus disease 2019 (COVID-19) pandemic, United States, March–April 2020
| Quartile | No. | PROMIS measure | ||||
|---|---|---|---|---|---|---|
| Anger | Anxiety | Cognitive function | Depression | Fatigue | ||
| Case rate per 100 000 population in home state | ||||||
| Quartile (no. of cases) | ||||||
| First (22-39) | 241 | 50.3 (9.6) | 57.4 (9.3) | 47.8 (9.6) | 53.9 (9.5) | 52.1 (11.1) |
| Second (40-64) | 242 | 50.5 (10.4) | 57.9 (9.4) | 48.2 (9.9) | 54.9 (9.9) | 52.1 (10.5) |
| Third (65-119) | 247 | 49.4 (10.2) | 56.1 (9.2) | 49.0 (9.7) | 53.5 (9.5) | 50.6 (10.4) |
| Fourth (120-768) | 233 | 50.6 (9.5) | 57.5 (9.1) | 48.8 (9.0) | 54.5 (9.1) | 5.8 (9.6) |
| P valueb | .51 | .18 | .53 | .34 | .23 | |
| Death rate per 100 000 population in home state | ||||||
| Quartile (rate) | 47.8 (9.2) | 55.2 (8.8) | 51.0 (8.9) | 51.1 (8.6) | 48.7 (9.9) | |
| First (0-0.2) | 257 | 49.8 (10.1) | 57.5 (9.1) | 48.7 (9.4) | 53.9 (9.5) | 51.4 (10.8) |
| Second (0.3-0.4) | 262 | 50.9 (10.0) | 57.5 (9.5) | 47.4 (10.0) | 54.5 (9.9) | 52.7 (10.5) |
| Third (0.5-0.6) | 224 | 48.9 (10.4) | 55.9 (9.4) | 49.5 (9.7) | 53.3 (9.5) | 49.6 (10.6) |
| Fourth (0.7-12.0) | 220 | 51.1 (9.1) | 57.9 (9.1) | 484 (8.9) | 55.1 (8.9) | 51.8 (9.4) |
| P valueb | .06 | .13 | .11 | .22 | .01 | |
| Change in unemployment rate in home state, February–March | ||||||
| Quartile (%) | ||||||
| First (−0.4%-0.4%) | 296 | 50.0 (9.3) | 57.2 (9.5) | 48.3 (9.5) | 54.3 (9.3) | 51.9 (9.9) |
| Second (0.5%-1.0%) | 199 | 50.7 (10.3) | 56.7 (9.0) | 49.0 (9.0) | 54.0 (9.4) | 51.0 (10.3) |
| Third (1.1%-1.6%) | 395 | 50.2 (10.3) | 57.4 (9.3) | 48.3 (10.0) | 54.1 (9.8) | 51.4 (10.9) |
| Fourth (1.7%-3.3%) | 73 | 49.7 (9.9) | 57.5 (9.0) | 48.5 (9.2) | 54.8 (9.6) | 50.6 (10.3) |
| P valueb | .97 | .63 | .94 | .97 | .42 | |
| % of population living in rural settings in home state | ||||||
| Quartile (%) | ||||||
| First (0%-10.2%) | 252 | 50.1 (10.4) | 57.4 (9.4) | 48.2 (10.0) | 54.2 (9.9) | 51.3 (10.7) |
| Second (10.3%-15.3%) | 237 | 50.2 (9.8) | 56.9 (9.5) | 48.0 (9.3) | 54.2 (9.6) | 51.9 (10.7) |
| Third (15.4%-25.8%) | 249 | 50.5 (9.9) | 57.5 (9.1) | 48.7 (9.6) | 54.2 (9.4) | 50.9 (10.2) |
| Fourth (25.9%-61.3%) | 225 | 49.9 (9.7) | 57.1 (9.2) | 49.0 (9.3) | 54.1 (9.0) | 51.5 (10.1) |
| P valueb | .91 | .91 | .65 | >.99 | .79 | |
Abbreviations: PROMIS, Patient-Reported Outcomes Measurement Information System; SD, standard deviation.
aPROMIS measures are validated, standardized, patient-reported outcomes measures assessing the severity of symptoms and their effect on daily functioning. A mean (SD) of 50 (10) represents the general US population; higher scores of anger, anxiety, depression, and fatigue indicate worse symptoms, and lower cognitive function scores indicate worse symptoms.19 All values are mean (SD) unless otherwise indicated.
bEvaluated using analysis of variance, with P < .05 considered significant.
In adjusted models (Table 4), younger generations (Millennial and Generation X) versus Baby Boomers, female sex versus male sex, not having enough or having just enough financial resources versus having more than enough financial resources, self-reported disability versus no self-reported disability, and inadequate health literacy versus adequate health literacy were most strongly correlated with worse symptoms of anger, anxiety, cognitive function, depression, and fatigue. In addition, nonbinary sex was strongly correlated with worse anxiety, cognitive function, depression, and fatigue symptoms compared with male sex. Race/ethnicity and education were not strongly correlated with any mental health symptoms.
Table 4.
Associations between psychosocial characteristics and PROMIS measures,a adjusting for state characteristics,b among a sample of adults (N = 963) during the coronavirus disease 2019 (COVID-19) pandemic, United States, March–April 2020
| Characteristic | Change in anger score | Change in anxiety score | Change in cognitive function scorec | Change in depression score | Change in fatigue score |
|---|---|---|---|---|---|
| Generation | |||||
| Baby Boomer (≥55 y) | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Generation X (39-54 y) | 3.39 (1.80 to 4.98) | 2.93 (1.45 to 4.43) | −3.34 (−4.86 to −1.83) | 4.13 (2.63 to 5.63) | 3.19 (1.52 to 4.85) |
| Millennial (≤38 y) | 3.95 (2.47 to 5.42) | 3.06 (1.69 to 4.44) | −4.19 (−5.60 to −2.79) | 5.19 (3.80 to 6.57) | 4.90 (3.36 to 6.43) |
| Sex | |||||
| Male | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Female | 2.05 (0.78 to 3.32) | 3.32 (2.16 to 4.49) | −2.10 (–3.31 to −0.89) | 2.03 (0.82 to 3.24) | 2.88 (1.56 to 4.20) |
| Nonbinary | 6.15 (−0.07 to 12.37) | 8.11 (2.37 to 13.84) | −8.85 (−14.79 to −2.91) | 8.50 (2.56 to 14.44) | 9.62 (3.13 to 16.11) |
| Race/ethnicity | |||||
| Non-Hispanic White | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Racial/ethnic minorityd | 0.06 (−1.35 to 1.47) | −1.18 (−2.49 to 0.13) | −0.07 (−1.41 to 1.28) | −0.43 (−1.78 to 0.91) | −0.71 (−2.18 to 0.77) |
| Financial resources | |||||
| More than enough | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Enough | 1.74 (−0.29 to 3.78) | 1.39 (−0.50 to 3.28) | −1.16 (−3.13 to 0.80) | 2.47 (0.57 to 4.38) | 1.60 (−0.52 to 3.72) |
| Not enough | 5.66 (3.47 to 7.85) | 5.64 (3.61 to 7.67) | −4.24 (−6.35 to −2.13) | 7.74 (5.70 to 9.79) | 6.60 (4.32 to 8.88) |
| Education | |||||
| Postgraduate degree | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Bachelor’s or associate’s degree | 0.27 (−1.47 to 2.01) | −0.44 (−2.06 to 1.18) | 0.81 (−0.85 to 2.48) | −0.38 (−2.05 to 1.28) | 0.02 (−1.81 to 1.84) |
| ≤High school diploma | 1.12 (−0.71 to 2.95) | −0.28 (−1.99 to 1.42) | 0.86 (−0.89 to 2.61) | 1.11 (−0.63 to 2.86) | 1.39 (−0.53 to 3.30) |
| Self-reported disabilitye | |||||
| No | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Yes | 2.56 (0.79 to 4.33) | 2.28 (0.63 to 3.93) | −3.46 (−5.14 to −1.78) | 3.12 (1.43 to 4.80) | 5.76 (3.94 to 7.58) |
| Health literacy | |||||
| Adequate | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Inadequate | 3.49 (2.10 to 4.87) | 2.90 (1.61 to 4.21) | −3.50 (–4.82 to −2.17) | 3.76 (2.44 to 5.09) | 3.57 (2.11 to 5.03) |
Abbreviation: PROMIS, Patient-Reported Outcomes Measurement Information System.
aPROMIS measures are validated, standardized, patient-reported outcomes measures assessing the severity of symptoms and their effect on daily functioning. A mean (standard deviation) of 50 (10) represents the general US population; higher scores of anger, anxiety, depression, and fatigue indicate worse symptoms, and lower cognitive function scores indicate worse symptoms.19 All values are ß estimate (95% CI).
bAll models adjusted for COVID-19 case rate (cases per 100 000 population), death rate (deaths per 100 000 population), unemployment rate (percentage of labor force claiming unemployment benefits), and percentage of population living in rural settings in the participants’ home states as of late March 2020.
cThis PROMIS measure is reverse-coded; lower scores indicate worse outcomes.
dIncludes those identifying as Hispanic and non-Hispanic Black/African American, American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiracial, or other.
eDefined as self-reported physical, hearing, vision, or other disability.
Discussion
In this study, a national sample of 963 US adults representative on age, sex, race/ethnicity, and geographic location reported a high severity of anxiety and depression related to the COVID-19 pandemic. People who were in the Millennial generation or Generation X versus Baby Boomers, female versus male, and facing psychosocial adversity (eg, nonbinary sex or having not enough financial resources, a disability, or inadequate health literacy) compared with no psychosocial adversity reported significantly worse symptoms, whereas we observed no differences by state-level variables.
Several psychosocial characteristics, such as female sex,24 younger generations,25 and financial resources,26 have been associated with worse mental health symptoms. However, the COVID-19 pandemic has created new social, economic, and health care realities that suggest these findings have new implications. For example, the financial toxicity, or the negative effects of high-cost medical care on access to care and overall quality of life, of mental health treatment27 may create further barriers to care for socioeconomically disadvantaged people during the economic downturn. Differences in living arrangements along socioeconomic lines (eg, amount of indoor/outdoor space, number of people in and around one’s home) may also explain the increased mental health symptoms during social-distancing requirements. Younger generations may be more affected by historically high unemployment rates than older adults of retirement age.28 Recent data indicate that women are currently taking on a greater share of caregiving responsibilities than men, often while continuing their professional roles and responsibilities.29 Therefore, providing access to mental health screening and treatment for young adults, women, and those with limited financial resources is critically important during the COVID-19 pandemic.
Researchers have raised concerns that psychosocial and other factors associated with structural racism may cause COVID-19 to disproportionately affect older non-Hispanic Black and Latinx adults as compared with White adults.30 Moreover, national data show that, to date, the number of deaths due to COVID-19 is higher among non-Hispanic Black and Hispanic/Latino adults aged 30-49 than among their non-Hispanic White counterparts.31 However, our study did not find strong associations between race/ethnicity and mental health symptoms. Our findings may reflect differences in reporting of mental health symptoms by race/ethnicity32 or the low proportion of racial/ethnic minority participants in our study. Further research focused on mental health symptoms among racial/ethnic minority groups during the pandemic is needed given the disproportionate prevalence of COVID-19 in this population and the likelihood that structural racism may exacerbate inequalities in access to diagnostic testing and care.33
The lack of differences in mental health symptoms by state-level variables is notable, because the early stages of the COVID-19 pandemic in the United States were largely characterized by state and local government responses.34 Accordingly, a study published in 2020 compared mental health search queries as a proxy for mental health symptoms by state during mid-March 2020, when 11 of 50 states had issued stay-at-home orders, and found that search queries increased before the issuance of orders in each state but dissipated after orders took effect, suggesting that the orders did not have a sustained effect on mental health.7 Conversely, in our study, in which participants directly self-reported symptoms, findings suggest that severe mental health symptoms were prevalent in late March through early April—after most states had issued stay-at-home orders. Moreover, we found these symptoms transcended state lines, with young generations, females, and those reporting inadequate financial resources, inadequate health literacy, and self-reported disabilities reporting the greatest severity of symptoms regardless of their geographic location. However, although unemployment rates were high across the United States when the study was conducted in late March, COVID-19 cases and deaths were concentrated in a few states. Therefore, the lack of differences in mental health symptoms by state could reflect a limited direct effect of COVID-19 on people in most states at the time our study was conducted. Follow-up studies examining how mental health symptoms have changed as COVID-19 cases and deaths have become more prevalent nationally, and whether any differences in symptoms exist by state, are warranted to further ensure mental health resources are appropriately directed.
Although research on mental health symptoms during a sustained period of the pandemic is needed, our findings suggest that unified national mental health policies are needed. From a historical perspective, resources for mental health are not sufficiently available during and after pandemics.15 Policy initiatives may include increasing awareness of maintaining good mental health and access to psychological interventions among people with inadequate financial resources, inadequate health literacy, and self-reported disabilities, who may not have had adequate access even before a pandemic.5 The rapid escalation of telehealth services in response to the pandemic, particularly in mental health services,35 and a temporary waiver of in-state licensure requirements for telehealth providers,36 may have increased access to needed services. However, telehealth usage largely depends on both broadband internet access and health insurance, among other factors.14 Although the recent Coronavirus Aid, Relief, and Economic Security (CARES) Act allocated $100 million to the US Department of Agriculture’s ReConnect Program to bring broadband access to rural areas,37 an estimated $80 billion is needed to bring broadband access to all US residents—including people without broadband access who do not live in rural areas.14 This disparity between the allocated funding and the amount of funding needed to bring adequate broadband access to all US residents suggests that many US residents may be unable to use telehealth services because of a lack of broadband access even after the CARES Act. Moreover, recommendations38 for health insurers to broaden telehealth coverage and reimbursement and limit cost sharing to patients have been variably adopted thus far and do not address people without health insurance—a number that may be as high as 27 million people in the United States,39 because unemployment rates have risen to historical highs. Thus, it is possible that the policy initiatives to improve telehealth access have disproportionately benefited people who remained employed and maintained health insurance during the COVID-19 pandemic, whereas those who lost health insurance as a result of being unemployed were unable to access telehealth services.
Limitations
Our study had several limitations. First, the study may have been subject to biased underreporting related to mental health symptoms among older adults and those in racial/ethnic minority groups, which would explain why they did not report worse symptoms than younger and non-Hispanic White participants, respectively, despite increased susceptibility to COVID-19. Second, to align with Prolific users’ expectations of extremely brief surveys, we did not obtain data on additional variables that may have been relevant to the analysis, such as COVID-19 infection among participants or close family members. COVID-19 testing was unavailable to much of the US population during the study period; therefore, we felt data on self-reported COVID-19 infection would be unreliable. Third, although our sample was representative of the US population on selected variables (age, sex, race/ethnicity, geography), it was not nationally representative because a random sampling strategy was not used. As such, certain groups, such as those with limited internet access, were likely underrepresented in our study. Finally, we excluded participants who did not reside in the continental United States, which limited the generalizability of our findings to people in other US states and territories.
Conclusion
In this study, a national sample of US adults reported severe mental health symptoms, especially anxiety and depression, during the early stages of the COVID-19 pandemic. Participants who were in young generations and female and those who had inadequate financial resources, inadequate health literacy, and self-reported disabilities reported the most severe symptoms. These findings highlight the importance of policies that ensure these groups have adequate access to mental health screening and treatment, including telehealth-based services, during the pandemic.
Footnotes
Authors’ Note: The research presented in this article is that of the authors and does not reflect the official policy of the National Institutes of Health.
Declaration of Conflicting Interests: The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: M.R.T. and J.P. are affiliated and have equity ownership in Iris OB Health Inc.
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute of Nursing Research under grant numbers R00NR016275 (principal investigator R.M.C.) and K99NR019124 (principal investigator M.R.T.).
ORCID iD
Meghan Reading Turchioe, PhD, MPH, RN https://orcid.org/0000-0002-6264-6320
References
- 1.Centers for Disease Control and Prevention Coronavirus disease 2019 (COVID-19): United States COVID-19 cases and deaths by state. Published August 18, 2020. Accessed August 18, 2020 https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/cases-in-us.html
- 2.Johns Hopkins University School of Medicine, Coronavirus Resource Center COVID-19 dashboard. Published 2020. Accessed August 18, 2020 https://coronavirus.jhu.edu/map.html
- 3.Driggin E., Madhavan MV., Bikdeli Bet al. Cardiovascular considerations for patients, health care workers, and health systems during the COVID-19 pandemic. J Am Coll Cardiol. 2020;75(18):2352-2371. 10.1016/j.jacc.2020.03.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ayanian JZ., Buntin MB. Coronavirus and health policy: a JAMA Network interview. JAMA Health Forum. April 29, 2020. Accessed August 18, 2020 https://jamanetwork.com/channels/health-forum/fullarticle/2765522 [DOI] [PubMed]
- 5.Cullen W., Gulati G., Kelly BD. Mental health in the COVID-19 pandemic. QJM. 2020;113(5):311-312. 10.1093/qjmed/hcaa110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.US Department of Labor Unemployment insurance weekly claims. 2020. Accessed August 18, 2020 https://www.dol.gov/ui/data.pdf
- 7.Jacobson NC., Lekkas D., Price Get al. Flattening the mental health curve: COVID-19 stay-at-home orders are associated with alterations in mental health search behavior in the United States. JMIR Ment Health. 2020;7(6):e19347. 10.2196/19347 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Duan L., Zhu G. Psychological interventions for people affected by the COVID-19 epidemic. Lancet Psychiatry. 2020;7(4):300-302. 10.1016/S2215-0366(20)30073-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ni MY., Yang L., Leung CMCet al. Mental health, risk factors, and social media use during the COVID-19 epidemic and cordon sanitaire among the community and health professionals in Wuhan, China: cross-sectional survey. JMIR Ment Health. 2020;7(5):Le19009. 10.2196/19009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sani G., Janiri D., Di Nicola M., Janiri L., Ferretti S., Chieffo D. Mental health during and after the COVID-19 emergency in Italy. Psychiatry Clin Neurosci. 2020;74(6):372. 10.1111/pcn.13004 [DOI] [PubMed] [Google Scholar]
- 11.Wang C., Pan R., Wan Xet al. Immediate psychological responses and associated factors during the initial stage of the 2019 coronavirus disease (COVID-19) epidemic among the general population in China. Int J Environ Res Public Health. 2020;17(5):1729. 10.3390/ijerph17051729 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li W., Yang Y., Liu ZHet al. Progression of mental health services during the COVID-19 outbreak in China. Int J Biol Sci. 2020;16(10):1732-1738. 10.7150/ijbs.45120 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Percudani M., Corradin M., Moreno M., Indelicato A., Vita A. Mental health services in Lombardy during COVID-19 outbreak. Psychiatry Res. 2020;288:112980. 10.1016/j.psychres.2020.112980 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Summers-Gabr NM. Rural–urban mental health disparities in the United States during COVID-19. Psychol Trauma. 2020;12(Suppl 1):S222-S224. 10.1037/tra0000871 [DOI] [PubMed] [Google Scholar]
- 15.Taylor S. The Psychology of Pandemics: Preparing for the Next Global Outbreak of Infectious Disease. Cambridge Scholars Publishing; 2019. [Google Scholar]
- 16.Palan S., Schitter C. Prolific.ac—a subject pool for online experiments. J Behav Exp Finance. 2018;17:22-27. 10.1016/j.jbef.2017.12.004 [DOI] [Google Scholar]
- 17.Chew LD., Bradley KA., Boyko EJ. Brief questions to identify patients with inadequate health literacy. Fam Med. 2004;36(8):588-594. [PubMed] [Google Scholar]
- 18.Cella D., Riley W., Stone Aet al. The Patient-Reported Outcomes Measurement Information System (PROMIS) developed and tested its first wave of adult self-reported health outcome item banks: 2005-2008. J Clin Epidemiol. 2010;63(11):1179-1194. 10.1016/j.jclinepi.2010.04.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.PROMIS adult profile instruments Published 2020. Accessed March 20, 2020 https://www.healthmeasures.net/promis-scoring-manuals
- 20.US Census Bureau State population totals: 2010-2019. Published 2020. Accessed March 31, 2020 https://www.census.gov/data/datasets/time-series/demo/popest/2010s-state-total.html
- 21.US Census Bureau Urban and rural. Published 2020. Accessed March 31, 2020 https://www.census.gov/programs-surveys/geography/guidance/geo-areas/urban-rural.html
- 22.US Bureau of Labor Statistics State employment and unemployment (monthly) news release. May 22, 2020. Accessed September 15, 2020 https://www.bls.gov/news.release/archives/laus_05222020.htm
- 23.Dimock M. Defining generations: where millennials end and generation Z begins. Published January 17, 2019. Accessed August 18, 2020 https://www.pewresearch.org/fact-tank/2019/01/17/where-millennials-end-and-generation-z-begins
- 24.Blanco C., Rubio J., Wall M., Wang S., Jiu CJ., Kendler KS. Risk factors for anxiety disorders: common and specific effects in a national sample. Depress Anxiety. 2014;31(9):756-764. 10.1002/da.22247 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Martin RJ., Usdan S., Cremeens J., Vail-Smith K. Disordered gambling and co-morbidity of psychiatric disorders among college students: an examination of problem drinking, anxiety and depression. J Gambl Stud. 2014;30(2):321-333. 10.1007/s10899-013-9367-8 [DOI] [PubMed] [Google Scholar]
- 26.Prins SJ., Bates LM., Keyes KM., Muntaner C. Anxious? Depressed? You might be suffering from capitalism: contradictory class locations and the prevalence of depression and anxiety in the USA. Sociol Health Illn. 2015;37(8):1352-1372. 10.1111/1467-9566.12315 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Greenberg PE., Fournier AA., Sisitsky T., Pike CT., Kessler RC. The economic burden of adults with major depressive disorder in the United States (2005 and 2010). J Clin Psychiatry. 2015;76(2):155-162. 10.4088/JCP.14m09298 [DOI] [PubMed] [Google Scholar]
- 28.Hoffman J. Young adults report rising levels of anxiety and depression in pandemic. The New York Times. August 13, 2020. Accessed August 18, 2020 https://www.nytimes.com/2020/08/13/health/Covid-mental-health-anxiety.html
- 29.World Economic Forum COVID-19 highlights how caregiving fuels gender inequality. Published April 24, 2020. Accessed May 31, 2020 https://www.weforum.org/agenda/2020/04/covid-19-highlights-how-caregiving-fuels-gender-inequality
- 30.Garcia MA., Homan PA., García C., Brown TH. The color of COVID-19: structural racism and the pandemic’s disproportionate impact on older racial and ethnic minorities [published online August 5, 2020]. J Gerontol B Psychol Sci Soc Sci. 10.1093/geronb/gbaa114 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Centers for Disease Control and Prevention Demographic trends of COVID-19 cases and deaths in the US reported to CDC. Published 2020. Accessed August 10, 2020 https://www.cdc.gov/covid-data-tracker/index.html#demographics
- 32.Blumberg SJ., Clarke TC., Blackwell DL. Racial and ethnic disparities in men’s use of mental health treatments. NCHS Data Brief. 2015;206:1-8. [PubMed] [Google Scholar]
- 33.Fortuna LR., Tolou-Shams M., Robles-Ramamurthy B., Porche MV. Inequity and the disproportionate impact of COVID-19 on communities of color in the United States: the need for a trauma-informed social justice response. Psychol Trauma. 2020;12(5):443-445. 10.1037/tra0000889 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Home Committee on the Budget The response to COVID-19 requires a stronger federal, state, and local partnership. Published April 27, 2020. Accessed May 31, 2020 https://budget.house.gov/publications/report/response-covid-19-requires-stronger-federal-state-and-local-partnership
- 35.Torous J., Jän Myrick K., Rauseo-Ricupero N., Firth J. Digital mental health and COVID-19: using technology today to accelerate the curve on access and quality tomorrow. JMIR Ment Health. 2020;7(3):e18848. 10.2196/18848 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Federation of State Medical Boards U.S. states and territories modifying requirements for telehealth in response to COVID-19. Published 2020. Accessed May 31, 2020 https://www.fsmb.org/siteassets/advocacy/pdf/states-waiving-licensure-requirements-for-telehealth-in-response-to-covid-19.pdf
- 37. Rural Utilities Service Broadband Pilot (ReConnect) Program. Fed Regist. 2020;85:20240. [Google Scholar]
- 38.Weigel G., Ramaswamy A., Sobel L., Salganicoff A., Cubanski J., Freed M. Opportunities and barriers for telemedicine in the U.S. during the COVID-19 emergency and beyond. Published May 11, 2020. Accessed May 31, 2020 https://www.kff.org/womens-health-policy/issue-brief/opportunities-and-barriers-for-telemedicine-in-the-u-s-during-the-covid-19-emergency-and-beyond
- 39.Garfield R., Claxton G., Levitt L. Eligibility for ACA health coverage following job loss. Published May 13, 2020. Accessed May 31, 2020 https://www.kff.org/coronavirus-covid-19/issue-brief/eligibility-for-aca-health-coverage-following-job-loss