A National Study of Zoom Fatigue and Mental Health During the COVID-19 Pandemic: Implications for Future Remote Work

Introduction

The COVID-19 pandemic has affected social interactions with family and friends, and workers in certain types of employment have been required to reconfigure responsibilities previously held in the office or worksite to virtual environments in their own residences.^1,2 Videoconferencing platforms such as Microsoft Teams, Skype, Zoom, and many more have become pivotal to maintaining employment and sustaining business operations. As one commentator notes, these innovations make it “possible to continue some semblance of business as usual during quarantine, allowing people to move their lives online while maintaining physical distance to stop the spread of the virus.”³

However, these technological adaptations needed to be made rapidly. In addition to mechanical aspects employees were expected to learn, videoconferencing for work was occurring in the context of an employee's household. Thus, employees have needed to coordinate with other household members who themselves were videoconferencing for work or school as well as navigate through household chores and childcare, all the while adjusting to new technical aspects of virtual work life.^4–6 The phrase “Zoom fatigue” was coined to describe the mental and physical effects of staring face-to-face with peers over a computer monitor for long periods of time, which has been postulated to lead to exhaustion, tiredness, anxiety, and fatigue.^2,3,6–9 Many employers plan to offer remote work options beyond the pandemic. Videoconferencing may be an essential part of work for many employees, even after the conclusion of the COVID-19 pandemic.^10,11

To our knowledge, there have been only a few empirical studies of “Zoom fatigue.” One study found significant associations between Zoom fatigue and videoconference meeting frequency, duration, and “burstiness” (i.e., shorter time between meetings) and a significant link between nonverbal mechanisms (e.g., mirror anxiety, feeling physically trapped) and Zoom fatigue.⁸ Surveys have identified five dimensions of fatigue, including general, social, emotional, visual, and motivational fatigue.¹² A cross-sectional study of medical school students showed lower Zoom fatigue when classes used a problem-based learning approach in which students learn about a subject by working in groups to solve an open-ended problem and facilitated more active participation among students.¹³

In this regard, it has been hypothesized that one potential factor related to Zoom fatigue is loneliness, distress due to feeling alone, and social isolation, the state of being alone.¹⁴ Loneliness and social isolation are related yet distinct: studies consistently report a significant but modest correlation between loneliness and isolation.^15–18 Several theoretical approaches have been applied to loneliness, one of the most studied is the cognitive approach to loneliness.^14,19

This cognitive approach to loneliness, which Perlman and Peplau pioneered, posits that cognitive processes mediate the association between one's social environment and loneliness.^14,19 Models adhering to the cognitive approach typically take into account (1) characteristics of the social network, (2) relationship standards, (3) personality, and (4) background characteristics.¹⁴ Applied to COVID-19 and increased use of videoconferencing technology, these theoretical models of loneliness would postulate that differences in how people perceive videoconferencing would shape their level of loneliness in the context of social distancing.

Because of the recency of the COVID-19 pandemic, empirical investigation of Zoom fatigue is scant. Although coined during the COVID-19 pandemic, Zoom fatigue will likely be relevant in post-pandemic “hybrid” remote/on-site working arrangements and thus deserves research attention.^5,6,20 To address this gap in the literature, we conducted a survey of psychosocial well-being among a large national sample of middle- and low-income U.S. adults in August 2020 amid the nationally implemented social distancing efforts intended to mitigate the spread of COVID-19.

Based on theoretical suppositions, we hypothesized that Zoom fatigue would be correlated with greater loneliness and number of weekly videoconferencing calls. In addition, in terms of potential impact of videoconferencing on reported mental health problems, we hypothesized that Zoom fatigue would be associated with depressive symptoms, even when controlling for covariates.

Materials and Methods

Participants

A total of 902 middle- and low-income adult residents in the United States were surveyed using Amazon's Mechanical Turk (MTurk) platform.^21,22 Selection criteria included a minimum age of 22 years, a reported annual gross income of $75,000 or below, and a reported workplace shift to Zoom or other videoconferencing platforms.

Measures

Participants were asked whether they experienced a shift in work to videoconferencing, and if they answered affirmatively, they were included in the current analysis. Quantity of videoconferencing was measured: “In the past 3 months, how many videoconference calls (e.g., Zoom) did you have on average each week? (enter numerical number only).”

Then, participants were presented the Zoom Fatigue Scale, developed for this study, consisting of three questions measured on a 5-point Likert scale ranging from 1 (not at all) to 5 (very much), summed to yield a total score (higher scores indicate greater videoconference-related fatigue): “Do you feel sad or depressed that remote videoconference calls have replaced in-person meetings,” “do you feel isolated and disconnected from others because remote videoconference calls have replaced in-person meetings,” and “do you feel that remote videoconference calls increase your level of stress and anxiety?” These three items showed good internal consistency for the whole sample (Cronbach's α = 0.85).

To examine how the scale performed across different subgroups, we conducted multiple analyses by stratifying the sample into demographic groupings, including gender (female and male), age (>median and ≤median), ethnicity (Hispanic and non-Hispanic), and race (white and nonwhite). For example, we ran analyses for Cronbach's alpha for female participants and separately for male participants, for younger participants and separately for older participants, etc. Across these multiple analyses, Cronbach's alpha scores ranged from 0.83 to 0.87, indicating that the scale demonstrated good internal consistency across gender, age, race, and ethnicity.

Demographic variables measured included age, race, ethnicity, gender, education level, marital status, employment, and history of military service. Psychosocial variables measured included annual income, whether they owned their own home, number of children in home, and having enough money to cover food expenses. Social support variables included loneliness, measured with the University of California, Los Angeles (UCLA) Loneliness Scale—Short Form²³ assessing feeling left out, isolated, and lacking companionship. Each item was scored on a scale of 1 (hardly ever) to 3 (often), summed to yield a total score (higher scores indicate greater perceived loneliness). Internal consistency was good in the sample (Cronbach's α = 0.81). Also, perceived social support was measured with the Medical Outcomes Study (MOS) Social Support Survey (Short Form)²⁴ that consists of six items measured on a 5-point Likert scale ranging from 1 (none of the time) to 5 (all of the time), summed to yield a total score (higher scores indicate greater social support). Internal consistency was excellent in the current sample (Cronbach's α = 0.90).

Mental health history was assessed with the question, “Has a doctor or nurse ever told you that you have any of the following conditions (check all that apply)?” and lifetime severe mental illness was operationalized as endorsing a lifetime history of psychotic disorder, bipolar disorder, or major depressive disorder. Participants also indicated whether they had over their lifetime been diagnosed with a pain condition or an anxiety disorder. The three-item version of the Alcohol Use Disorders Identification Test (AUDIT-C)²⁵ was used to assess alcohol use quantity/frequency. Participants were also asked about current drug use. The Patient Health Questionnaire-2 (PHQ-2) was used to assess depressive symptoms, specifically frequency of depressed mood and anhedonia over the past 2 weeks.²⁶

Procedure

Participants were recruited online through Amazon MTurk for a study designed to “understand COVID-19's impact on people's health.” After providing informed consent, participants completed a brief screener about age and income before being directed to a third-party website (Qualtrics) to complete a 15- to 20-minute online survey. All tasks were completed on the same day, and participants were compensated $0.75 for completing the study. Data were collected in August 2020—8 months after the first Centers for Disease Control and Prevention (CDC)-confirmed U.S. COVID-19 case and 2 months after the World Health Organization declared COVID-19 a global pandemic. This study was approved by the institutional review boards at Duke University and the University of Texas, San Antonio.

Analyses were conducted using SAS 9.4. To maximize the generalizability of our findings, we used an SAS Macro for raking procedures²⁷ to create sample weights representative of the U.S. population using data from the U.S. Census Bureau consistent with our inclusion criteria for age (>22 years) and income (≤$75,000) to more precisely weight the sample using U.S. census data on age, geographic region, gender, race, and ethnicity.

Results

Sample characteristics are depicted in Table 1. Our sample was 43.4% male and 24.3% nonwhite, with 11.2% identifying as Hispanic. Approximately 57.7% were married, 6.9% were military veterans, and 96.5% had at least a high school education. Mean age was 37.7 years old and mean annual gross income was $45,908. In total 16.4% reported a lifetime history of severe mental illness, 12.2% reported current drug use, and 22.7% and 31.6% reported having a lifetime pain or anxiety condition, respectively.

Weighted bivariate Spearman correlations were estimated to determine bivariate associations between Zoom fatigue and study variables and between depressive symptoms and study variables. Table 2 shows that Zoom fatigue had small-to-moderate positive correlations with the following study variables: nonwhite race, lifetime severe mental illness, loneliness, food insecurity, and number of videoconference calls per week. Small-to-moderate negative correlations with depressive symptoms were found for two demographic variables: age and ethnicity, two psychosocial variables: social support and annual income.

Small to strong positive correlations with depressive symptoms were found for all clinical variables as well as the psychosocial variables of loneliness, number of children in home, and food insecurity. Relevant to the COVID-19 pandemic, a moderate correlation was found between depressive symptoms and Zoom fatigue.

Multivariable linear regression and weighted were conducted in which Zoom fatigue was regressed onto demographic, clinical, and social variables. Table 3 depicts multivariable regressions. The following variables were statistically significant predictors of total scores on the Zoom Fatigue Scale: marital status, lifetime severe mental illness, education level, loneliness, food insecurity, and number of weekly videoconference calls. Finally, analyses showed that the following variables were statistically significant predictors of depressive symptoms: nonwhite race, military status, loneliness, social support, food insecurity, all clinical variables, and the total score on the Zoom Fatigue Scale.

Figure 1 depicts scores on the Zoom Fatigue Scale as a function of high versus low scores on loneliness and number of weekly videoconferencing calls divided by quartile, demonstrating additive effects of these two variables.

Discussion

This national survey of U.S. adults found that higher scores on a three-item scale measuring Zoom fatigue were significantly related to several factors including being married, nonwhite race, post-high school education, and severe mental illness; concurrent psychosocial stressors including greater loneliness, lower social support, and lacking money for food; and greater number of videoconference calls per week. Zoom fatigue was also significantly associated with greater depressive symptoms, even when adjusting for demographic and clinical covariates.

Although we cannot determine the directionality of this association, depression and Zoom fatigue may share some common symptomatology. Post hoc analyses examining predictors of Zoom fatigue separately among male and female participants yielded slightly different findings, with loneliness emerging as the only statistically significant correlate in both groups (number of weekly conference calls was significantly associated among male participants but only marginally significant among female participants).

The study yielded new findings with important implications for future remote work. As other commentators have discussed,^3,5,6,10 we found multiple psychosocial stressors associated with Zoom fatigue: depression, loneliness, and food insecurity. First, Zoom fatigue was associated with depressive symptoms even after adjusting for demographic, clinical, and psychosocial covariates, suggesting Zoom fatigue was uniquely related to depressive symptoms in this national sample above and beyond social support and loneliness. This finding has important clinical implications, indicating it may be helpful for health professionals to inquire whether patients are experiencing stress, social isolation, or depression related to their videoconferencing technology use.^15,20

Although some of these issues may naturally resolve upon future lifting of social distancing measures end at the conclusion of the COVID-19 pandemic, this issue will remain important in the future as it is likely that many workplaces will adopt a hybrid model of in-person and videoconferencing meetings.^1,27 The current findings, including those illustrated in Figure 1, do suggest that fewer videoconference calls per week may help reduce Zoom fatigue and, by proxy, depression symptoms. Therefore, future research should explore the optimal balance of in-person and virtual workplace communication to minimize the potential impact on mental health.

Second, higher levels of Zoom fatigue correlated with higher levels of loneliness. This finding is consistent with the cognitive model of loneliness, which posits that cognitive processes mediate the relationship between the social environment and loneliness.^14,19 However, the mechanisms underlying the correlation between Zoom fatigue and loneliness remain unclear. A third factor may be driving this relationship; for example, there may be individual differences in the ability to find a sense of connection and belonging in virtual interactions, which may, in turn, increase both Zoom fatigue and loneliness.

Alternatively, increased Zoom fatigue could impinge on the ability to engage meaningfully in virtual interactions^1,4,7,8 thereby increasing loneliness. In contrast, increased loneliness may lead to depressive symptoms, shaping a person's appraisal of videoconferencing. Additional longitudinal and experimental research is necessary to understand what is driving the relationship between Zoom fatigue and loneliness.

Third, food insecurity was an unexpected variable associated with Zoom fatigue that should be closely considered. During the COVID-19 pandemic, there were numerous reports of food insecurity on the rise. The U.S. Census Bureau, which tracks the economic impact of the pandemic, estimated that over 1 in 10 Americans were experiencing food insecurity, along with one-half experiencing loss in employment income and over a quarter of Americans experiencing housing insecurity.^28,29

If a worker is expected to change communication by using videoconferencing but cannot make ends meet financially from that work, then it would be logical for that worker to feel greater stress associated with these calls, as suggested by the current findings. Furthermore, considering that that videoconferencing can be cognitively taxing, employees without proper nutrition may have additional difficulty focusing and concentrating during videoconferencing calls. We also found food insecurity to predict depressive symptoms significantly, and depression is associated with lower work productivity and higher absenteeism from work.³⁰

Accordingly, ensuring employees' basic needs are met financially during health crises may be a critical way to protect an organization's financial health. Although these hypotheses should be tested in future research, the current findings underscore the need for employers concerned for their workers' welfare to keep in mind that food insecurity could impact an employee's mental health and work engagement.

In addition, we found a potentially modifiable factor with Zoom Fatigue: frequency of videoconference calls. A greater number of weekly videoconference calls was associated with higher scores on the Zoom Fatigue Scale (in both bivariate and multivariable analyses). As depicted in Figure 1, this correlation should be considered in the context of whether respondents also indicated higher levels of loneliness. Nevertheless, to our knowledge, this is the first study to empirically demonstrate a direct link between the number of weekly videoconference calls and perceived depression, anxiety, and social isolation associated with videoconference calls.

Although we did not examine whether this related to productivity at work,³⁰ analyses presenting videoconference call frequency quartiles showed that Zoom fatigue was highest for participants reporting seven or more videoconference calls per week. This is consistent with many commentators^6–9,31,32 who suggest that cognitive processes (e.g., attention/concentration, nonverbal overload) mediate the relationship between social distancing and the increased use of videoconferencing technology during the COVID-19 pandemic.

Still, it is essential to note that number of weekly videoconference calls was not, in multivariable analysis, associated with depressive symptoms in our study. Future research may examine whether Zoom fatigue mediates a link between the number of videoconference calls and poor health outcomes. It is also worth exploring whether the number of videoconference calls leads to subsequent increases in loneliness, which in turn increases depressive symptoms.

Although our study had many strengths and yielded new findings, several limitations warrant mention. As we did not assess participants at multiple time points, our cross-sectional survey design precludes drawing inferences about the directionality or causality of detected associations. Nor do our data speak of Zoom fatigue and associated mental health impacts further into the pandemic (e.g., after the first anniversary of the first domestic case in the United States). Our survey also relied on self-report, such that participants' responses may have been subject to interpretation or recall biases.

Although we recruited a heterogeneous sample of adults residing in all major geographic regions of the United States, our analyses did not use stratified sampling. Accordingly, although statistical raking procedures²⁷ improved the generalizability of study findings, our sample may not represent all U.S. adults, so prevalence and epidemiological values derived from this study should therefore be interpreted with caution.

We also developed the Zoom Fatigue Scale for this study, as there was no validated measure available at the time of data collection. As a result, we cannot speak in-depth about its psychometric properties. We also assessed respondents' experiences with videoconferencing and virtual work implemented due to the COVID-19 pandemic. We did not assess job type (e.g., tradesperson, food service worker, educator, health care worker). Future studies should examine whether Zoom fatigue and associated impacts on mental health and vocational functioning differ based on experience with virtual work, the employment sector, and other job-related variables.

We asked participants for their gender and provided male, female, and other as responses; future research should differentiate between gender at birth and self-identified gender. We did not have pre-pandemic information on study variables for survey participants. Future research with pre-pandemic vocational, financial, and social functioning data would help examine factors associated with exacerbation, initial onset, and perhaps alleviation of social isolation resulting from pandemic-related social distancing measures.

Conclusions

The goal of this study was to take a preliminary step toward examining the intricate links between videoconferencing for work and mental health problems in the context of the COVID-19 pandemic. Overall, the national survey data supported hypotheses that Zoom fatigue would emerge as an independent construct related to social isolation, depressive symptoms, and number of weekly work videoconference calls. Zoom fatigue was also associated with sociodemographic variables, including marital status and food insecurity.

The results indicate that employers and employees should consider a complex array of individual-level and environment-level factors when assessing the degree to which videoconferencing at work may engender stress, social isolation, and emotional exhaustion, which could adversely impact mental health, work productivity, and quality of life.^5,30

Disclaimer

The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the United States Government or Department of Veterans Affairs (VA).

Author Disclosure Statement

No competing financial interests exist.

Funding Information

S.C.G. and S.M.B. were each supported by a VA Office of Academic Affiliations Advanced Fellowship in Mental Illness Research and Treatment. S.M.B. is now at Research Triangle Institute International.