Secondary Logo

Journal Logo


Effects of Sleep, Exercise, and Leadership Support on Resilience in Frontline Healthcare Workers During the COVID-19 Pandemic

Kaye-Kauderer, Halley MD; Loo, George DrPH; Murrough, James W. MD, PhD; Feingold, Jordyn H. MD; Feder, Adriana MD; Peccoralo, Lauren MD; Ripp, Jonathan MD; Pietrzak, Robert H. PhD

Author Information
Journal of Occupational and Environmental Medicine: May 2022 - Volume 64 - Issue 5 - p 416-420
doi: 10.1097/JOM.0000000000002460

In the Spring of 2020, the COVID-19 pandemic struck the United States, leading to unprecedented and widespread death, overrun hospitals, and lasting physical and psychological consequences for the entire population. With more than 23,000 confirmed deaths1 during just the first wave, New York City was hit particularly hard and frontline health care workers (FHCWs) experienced a significant psychological burden.2

While numerous studies have demonstrated the negative psychological impact of COVID-19 on FHCWs,2–5 none have examined the potential protective effects of specific personal behaviors and systemic factors on their mental health and wellbeing. A number of restorative personal behaviors, including adequate sleep and physical activity, have been associated with increased professional fulfillment and decreased burnout, depression, and anxiety among health care workers.6,7 Importantly, through the practice of behaviors that may protect against negative psychological consequences, individuals may feel more agency in achieving small, well-being-oriented goals, even during a crisis. Examining such behaviors during the pandemic may help inform interventions designed to enhance well-being and mitigate risk for adverse psychological sequelae.

While individual behaviors are important, systemic factors may also affect psychological outcomes in a crisis. Prior to the pandemic, the impact of systemic factors on the well-being of health care workers was well established. For example, hospital leaders had recognized the importance of creating cultures in which workers were empowered to pursue both personal and collective goals.8,9 In a recent prospective survey-based study of over 3000 FHCWs in New York City during the pandemic surge from April to May 2020, higher levels of perceived leadership support and feeling valued by hospital leadership were associated with lower likelihood of screening positive for major depressive disorder (MDD), generalized anxiety disorder GAD), and/or COVID-19-related posttraumatic stress disorder (PTSD) symptoms.2 While other work has similarly emphasized the importance of hospital leadership in decreasing psychological distress, empirical data examining how systemic factors may interact with individual behaviors are scarce.10–12

To address these gaps, we examined the direct and interacting effects of modifiable personal behaviors (eg, sleep, physical activity, engagement in hobbies, and various coping strategies) and systemic factors, such as leadership support, on pandemic-related psychological distress in a large cohort of FHCWs during the acute pandemic surge in New York City.



An electronically administered anonymous survey was delivered to a purposefully selected group of FHCWs at The Mount Sinai Hospital, an urban tertiary care hospital in NYC. Responses were collected between April 14 and May 11, 2020. This timeframe included the peak and downward slope of the hospitals epidemic curve, based on COVID-19 inpatient census data. Participants were offered a $25 gift card after finishing the survey. The study was approved by the Institutional Review Board at the Icahn School of Medicine at Mount Sinai.

The eligible study population included health care workers directly involved in the care of patients with COVID-19, either as a result of their standard practice or redeployment within the study period. The sample included attending's and house staff from various departments (eg, Internal Medicine, Anesthesiology, Pediatrics, Emergency Medicine, Surgery, and Psychiatry) as well as nurses, physician assistants, hospital chaplains, social workers, and dietitians. The total eligible population included 6026 presumed FHCWs.


COVID-19-Related Psychologic Distress (C19-Distress)

The presence of C19-distress was defined as a positive screen for symptoms of COVID-19-related post-traumatic stress disorder (PTSD), major depressive disorder (MDD), and/or generalized anxiety disorder (GAD). Symptoms of COVID-19-related-PTSD were assessed using a 4-item PTSD-Checklist for DSM-5 (4-item PCL-5),13 with questions altered to inquire about symptoms related to the COVID-19 pandemic (eg, “Over the past two weeks, how often were you bothered by repeated, disturbing, and unwanted memories of your experiences related to the COVID-19 pandemic?“) A positive screen was defined by a score of 8 or higher, which has the highest efficiency (90.4%; sensitivity = 0.81, specificity = 0.94) in diagnosing PTSD (14) [Cronbach's α = 0.85]. Symptoms of MDD were assessed using the Patient Health Questionnaire-8 (PHQ-8),14 an eight-item measure that assesses symptoms of MDD experienced over the previous 2 weeks. A positive screen was defined by a score of 10 or higher, which yields comparable estimates of the prevalence of current major depressive or other depressive disorders relative to diagnostic interviews [Cronbach's α = 0.89]. Symptoms of GAD were assessed using the Generalized Anxiety Disorder-7 (GAD-7),15 a seven-item measure that assesses symptoms of GAD experienced over the past 2 weeks. A positive screen was defined by a score of 10 or higher, which has a sensitivity of 0.89 and specificity of 0.82 in diagnosing GAD [Cronbach's α = 0.91]. Variables examined in relation to positive screens for C19-distress are elaborated further in Supplemental Table I,

COVID-19 Stressors

An exploratory factor analysis of COVID-19-related stressors was conducted to generate a composite measure of COVID-19 stressors. These stressors included number of hours worked onsite, redeployment to a different unit, availability of personal protective equipment (PPE), availability of testing for workers and patients, number of coworkers infected with COVID-19, coworker hospitalization or death, sum of patient and personal COVID-19 exposures, perceived personal medical risk, having to make difficult decision(s) prioritizing some patients over others, feeling torn between patients and loved ones, and fear about infecting loved ones.

Personal Behaviors

Sleep hours were assessed with the question: “Atpresent, on average how many hours per day do you sleep (out of 24 hours) on a typical workday?” Physical activity (exercise, sports, etc.), mindfulness activities (mediation, deliberate breathing, etc.), artistic or creative activities (art, music, journaling, etc.), consume media content (listen to podcasts, reading, watch TV or movies, etc.), and other hobbies and games (cooking, board games, video games, etc.) were assessed with the question: “Atthis time, during the pandemic, how many days per week do you do the following?” (response options: 0–7 days with 0–7 point-scale).

Team and Hospital Support

Clinical team camaraderie and hospital leadership support were assessed with the question: “In your opinion, what is the current level of: (1) camaraderie/team spirit among your group of co-workers, in your own clinical practice team or setting, and (2) support from your hospital leadership.” Response options were Low, Medium, and High for all questions.

Perceived Value at Work

Perceived value at work was assessed with the question: “In your opinion, to what extent do you feel valued by: (1) your immediate supervisors (team leader, service chief, etc.), and (2) hospital leadership.” Response options were Not at all valued, Slightly valued, Moderately valued, and Very much valued.

Data Analysis

A multivariable binary logistic regression analysis was conducted to identify personal behaviors and systemic factors associated with screening positive for C19-distress. Demographic characteristics and professional role, number of COVID-19-related stressor exposures, personal behaviors and systemic factors, and interaction terms between personal behaviors and systemic factors were included in this model. The interaction terms evaluated whether associations between personal behaviors (eg, sleep hours) and C19-distress differed in magnitude as a function of systemic factors (eg, perceived leadership support). To illustrate significant interactions, probabilities of screening positive for C19-distress for different levels of personal behaviors were plotted as a function of different levels of significant moderating systemic variables. A relative importance analysis was conducted to determine the relative contribution of each significant variable and interaction term(s) in predicting screening positive for C19-distress. This analysis partitions the explained variance in a dependent variable that is explained by each independent variable, while simultaneously accounting for inter-correlations among these independent variables. We also identified thresholds of personal behaviors and interactions between personal behaviors and systemic factors associated with a probability of screening positive for C19-distress that was significantly lower than that observed in the full sample (ie, below the lower limit 95% confidence interval [CI]).


Sample Characteristics

Six thousand twenty-six health care workers were eligible to complete the survey and 3360 (55.8%) responded. Of those, 497 (14.8%) were excluded because they denied being directly engaged in clinical activities. An additional 284 (10.0%) were excluded because their surveys lacked adequate information about the outcome variables. The final sample included 2579 FHCWs. One thousand four hundred seven (54.6%) were under the age of 35, 1897 (73.6%) were female, and 1821 were married/partnered (70.6%). In terms of professional breakdown, 1082 (42.0%) were registered nurses, 541 (21.0%) medical residents or fellows, 398 (15.4%) attending physicians, 394 (15.3%) physician assistants or advanced practice registered nurses, and 164 (6.4%) other (eg, social workers, psychologists, and chaplains). The median number of years in practice was 6.0 (interquartile range [IQR] = 8.0); median number of hours working onsite was 37.5 (IQR = 10.3); and median number of COVID-19 patients treated was 30.0 (IQR = 48.0).


In total, 1005 healthcare workers (39.0%, 95%CI = 37.140.9%) met the prespecified cutoff values for significant symptoms of C19-distress with descriptive data shown in Table 1. As shown in Table 2 and detailed further in Feingold et al,2 results of a multivariable regression model predicting C19-distress revealed that lower exposure to COVID-19-related stressors, professional role other than registered nurse (RN), male gender, being married or partnered, higher number of years in practice, engagement in more days of physical exercise and hobbies/games, and greater team camaraderie were associated with decreased likelihood of screening positive for C19-distress. Sleep-by-leadership support was the only significant interaction associated with this outcome.

TABLE 1 - Descriptive Characteristics of the Sample
Personal Factors Total (Percentage)
 18–34 1,408 (54.6%)
 35–44 581 (22.5%)
 45–54 330 (12.8%)
 >55 260 (10.1%)
 Female 1,897 (73.6%)
 Male 682 (26.4%)
Relationship status
 Single/divorced/widowed 741 (29.4%)
 Married/partnered 1776 (70.6%)
 RN 1082 (42.0%)
 Residents/fellows 541 (21.0%)
 Attending MD/DO 398 (15.4%)
 PA/NP 394 (15.3%)
 Other 164 (6.4%)
Median years in practice 6.0 (interquartile range [IQR] = 8.0]
Median hours working onsite 37.5 (IQR = 10.3)
Median number of COVID-19 Patients treated 30.0 (IQR = 48.0)
DO, doctor of osteopathic medicine; MD, medical doctor; NP, nurse practitioner; PA, physician assistant; RN, registered nurse.

TABLE 2 - Results of Multivariable Logistic Regression Model Predicting Positive Screen for C19-Distress in Frontline Healthcare Workers
Wald X 2 P OR (95%CI)
Age 2.68 0.10 0.82 (0.65–1.04)
Male gender 16.09 <0.001 0.63 (0.50–0.79)
Married/partnered 16.95 <0.001 0.66 (0.54–0.80)
RN vs other profession 8.07 0.005 1.34 (1.10–1.64)
Years in practice 10.35 0.001 0.98 (0.97–0.99)
COVID-19 stressor exposure index 92.32 <0.001 1.63 (1.48–1.80)
Sleep hours 0.003 0.96 1.01 (0.80–1.27)
Physical exercise 7.48 0.006 0.93 (0.88–0.98)
Mindfulness activities 2.02 0.16 1.04 (0.99–1.10)
Artistic/creative activities 0.527 0.47 0.98 (0.92–1.03)
Hobbies/game activities 5.27 0.02 0.95 (0.90–0.99)
Consuming content 0.012 0.91 1.00 (0.95–1.05)
Team camaraderie 10.21 0.001 0.78 (0.66–0.91)
Hospital leadership support 1.973 0.16 1.68 (0.82–3.45)
Valued by leadership 1.487 0.22 0.93 (0.83–1.05)
Valued by close supervisor(s) 2.48 0.12 0.90 (0.79–1.03)
Sleep hours × hospital leadership support 4.19 0.04 0.89 (0.80–0.99)
C19-distress, COVID-19-related psychological distress; OR, odds ratio; RN, registered nurse; COVID-19 stressor exposure index: factor scores reflecting COVID-19-related stressor exposures.2

Relative importance analysis revealed that of the significant personal behaviors and systemic factors shown in Table 2, the sleep- by-leadership support interaction explained the majority of variance in positive screen for C19-distress (19.4% relative variance explained [RVE]), followed by team camaraderie (6.4% RVE), physical exercise (4.9% RVE), and hobbies (3.2% RVE). As shown in Figure 1, for the interaction between sleep and leadership support, among workers who slept a median of 6 hours/night, high leadership support was associated with lower probability of screening positive for C19-distress; among workers who slept a median of 7 or 8 hours/ night, medium-to-high levels of leadership support were associated with greater resilience to C19-distress in a dose response” manner (Fig. 1). As shown in Figure 2, the threshold for number of exercise days per week associated with a probability of positive screen for C19-distress below the 95%CI observed in the full sample was 4 days (N= 522, 20.2%) with even further reductions in probability of this outcome for FHCWs who reported exercised 5 and 6 to 7 days per week.

Interaction between sleep hours and leadership support during pandemic surge in predicting probability of screening positive for C19-distress. Note. Error bars represent 95% confidence intervals (95%CI). Black line represents mean probability and 95%CI of screening positive for C19-Distress in the full sample.
Frequency of physical exercise during pandemic surge and probability of screening positive for C19-Distress. Note. Error bars represent 95% confidence intervals (95%CI). Black line represents mean probability and 95%CI of screening positive for C19-Distress in the full sample.


To the best of our knowledge, this is the first study to examine the main and interactive effects of personal restorative behaviors and systemic factors in relation to screening positive for C19- distress in FHCWs during the initial peak of the COVID-19 pandemic. In our analysis of 2579 FHCWs, a greater number of sleep hours and high leadership support, as well as higher levels of team camaraderie and greater engagement in physical exercise and hobbies during the pandemic surge, were associated with a significantly lower likelihood of screening positive for C19-distress. These results underscore the importance of encouraging and enabling the practice of healthy personal behaviors (eg, sleep, physical exercise, and hobbies), paired with system-level supports (eg, leadership support and team connectivity) as potential prevention and early intervention targets to help promote psychological resilience in FHCWs in the midst of a crisis.

Results of this study are consistent with prior studies, which found that sleep and physical exercise impact overall well-being and distress, both in the general population, as well as among healthcare workers.16–19 Notably, while physical exercise and engagement in hobbies were independently and significantly associated with decreased probability of screening positive for C19-distress, sleep hours were only significantly associated with C19-distress in the context of moderate to high leadership support. Specifically, among FHCWs who reported getting what would be considered an adequate amount of sleep, those who reported low leadership support were significantly more likely to screen positive for C19-distress, even higher than the average prevalence for the full sample, relative to those with medium and high levels of support. Of note, our findings complement previous work suggesting specific thresholds for sleep (6–6.5 hours per night in the context of high leadership support; 7–12 hours per night in the context of moderate-to-high leadership support) and exercise (4 or more days per week) that are associated with greater resilience to the negative psychological effects of the pandemic in FHCWs.20,21 This discrepancy underscores the importance of promoting not just individual restorative behaviors such as sleep, but also a highly supportive work environment through thoughtful and engaged leadership.

While sleep, physical exercise, and engagement in hobbies are all theoretically modifiable by an individual health care worker, one's ability to engage in behavior change may be augmented by workplace interventions through system-wide and direct leadership. When leaders serve as role models and consciously create cultures that support behaviors such as exercise, attention to nutrition, health, and restoration, their constituents may be more likely to engage in such behaviors.22,23 Conversely, devaluing these behaviors has been shown to be detrimental to team members.22 For example, lack of sleep among leaders has been shown to have downstream impacts on employees, leading to sleep-related impairment for those they supervise.24,25 Importantly, prior work has shown that physicians experience their organizations through the prism of their direct unit leaders,26 suggesting that immediate supervisors might be the ideal group in which to promote the alignment of personal-organizational values. Thus, leaders closest to staff can aim to cultivate an environment that encourages self-care through role modeling and specific initiatives, such as limiting work hours, minimizing work done at home (ie, no emailing between certain hours) hiring additional healthcare workers to meet the demand, and creating safe spaces for staff to discuss work-life balance challenges.

There are a number of limitations in this study. First, the study design was cross-sectional. While our analysis revealed several robust associations between modifiable personal restorative behaviors and systemic factors, and C19-distress, we cannot establish causal relationships between these behaviors and GAD, MDD, and PTSD symptoms. Additionally, our sample population was limited to a single hospital in NYC, limiting the ability to generalize our findings across diverse populations in NYC and beyond. While the survey remained open for several weeks, our findings represent only a limited time period during a highly dynamic pandemic. Finally, this research did not capture a broad range of organizational leadership styles and cultures.

Despite these limitations, results of this study provide new insights into the role of personal restorative behaviors and systemic factors in contributing to psychological distress during the acute phase of the COVID-19 pandemic. These findings stress the importance of sleep, physical exercise, and leadership support in fostering resilience during times of stress. Further research is needed to generalize these findings to other FHCW cohorts, examine the role of personal and systemic factors in predicting trajectories of pandemic-related distress over time, and evaluate the efficacy of prevention and both individual- and system-directed interventions to promote resilience to pandemic-related stress in this population.


The authors wish to thank all of the participants at the Mount Sinai Hospital who contributed to this work. They wish to also thank Chi Chan, Carly Kaplan, Dennis Charney, Jaclyn Verity, Alicia Hurtado, Larissa Burka, Shumayl Syed, and Steven Southwick for their guidance and help with this project.


1. Main data page: casese, hospitalizations, and deaths: City of New York, Department of Health.
2. Feingold JH, Peccoralo L, Chan CC, et al. Psychological impact of the COVID-19 pandemic on frontline health care workers during the pandemic surge in New York City. Chronic Stress 2021; 5:2470547020977891.
3. Que J, Le Shi JD, Liu J, et al. Psychological impact of the COVID-19 pandemic on healthcare workers: a cross-sectional study in China. Gen Psychiatry 2020; 33:e100259.
4. Giusti EM, Pedroli E, D’Aniello GE, et al. The psychological impact of the COVID-19 outbreak on health professionals: a cross-sectional study. Front Psychol 2020; 11:1684.
5. Hennein R, Mew EJ, Lowe SR. Socio-ecological predictors of mental health outcomes among healthcare workers during the COVID-19 pandemic in the United States. PLoS One 2021; 16:e0246602.
6. Lindwall M, Gerber M, Jonsdottir IH, Börjesson M, Ahlborg G Jr. The relationships of change in physical activity with change in depression, anxiety, and burnout: a longitudinal study of Swedish healthcare workers. Health Psychol 2014; 33:1309–1318.
7. Trockel MT, Menon NK, Rowe SG, et al. Assessment of physician sleep and wellness, burnout, and clinically significant medical errors. JAMA Netw Open 2020; 3:e2028111–e12028111.
8. Shapiro DE, Duquette C, Abbott LM, Babineau T, Pearl A, Haidet P. Beyond burnout: a physician wellness hierarchy designed to prioritize interventions at the systems level. Am J Med 2019; 132:556–563.
9. Shanafelt TD, Noseworthy JH. Executive leadership and physician wellbeing: nine organizational strategies to promote engagement and reduce burnout. Mayo Clin Proc 2017; 92:129–146.
10. Shanafelt TD, Ripp J, Brown M, Sinsky CA. Creating a Resilient Organization. 2020.
11. Kishore S, Ripp J, Shanafelt T, et al. Making the case for the chief wellness officer in America's health systems: a call to action. Health Affairs Blog 2018; 10:1377.
12. Ripp J, Peccoralo L, Charney D. Attending to the emotional well-being of the health care workforce in a New York City Health System during the COVID- 19 pandemic. Acad Med 2020; 95:1136–1139.
13. Geier TJ, Hunt JC, Hanson JL, et al. Validation of abbreviated four- and eight-item versions of the PTSD checklist for DSM-5 in a traumatically injured sample. J Trauma Stress 2020; 33:218–226.
14. Shin C, Lee SH, Han KM, Yoon HK, Han C. Comparison of the usefulness of the PHQ-8 and PHQ-9 for screening for major depressive disorder: analysis of psychiatric outpatient data. Psychiatry Investig 2019; 16:300–305.
15. Spitzer RL, Kroenke K, Williams JB, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006; 166:1092–1097.
16. Kancherla BS, Upender R, Collen JF, et al. Sleep, fatigue and burnout among physicians: an American Academy of Sleep Medicine position statement. J Clin Sleep Med 2020; 16:803–805.
17. Johnson DA, Billings ME, Hale L. Environmental determinants of insufficient sleep and sleep disorders: implications for population health. Curr Epidemiol Rep 2018; 5:61–69.
18. Mead GE, Morley W, Campbell P, Greig CA, McMurdo M, Lawlor DA. Exercise for depression. Cochrane Database Syst Rev 2008; CD004366.
19. Olson SM, Odo NU, Duran AM, Pereira AG, Mandel JH. Burnout and physical activity in Minnesota internal medicine resident physicians. J Grad Med Educ 2014; 6:669.
20. van Dam NH, van der Helm E. There's a proven link between effective leadership and getting enough sleep. Harvard Bus Rev Dig Articles; 2016; 2–5.
21. Dyrbye LN, Satele D, Shanafelt TD. Healthy exercise habits are associated with lower risk of burnout and higher quality of life among U.S. Medical Students. Acad Med 2017; 92:1006–1011.
22. Shanafelt T, Trockel M, Rodriguez A, Logan D. Wellness-centered leadership: equipping health care leaders to cultivate physician well-being and professional fulfillment. Acad Med 2021; 96:641.
23. Trockel MT, Hamidi MS, Menon NK, et al. Self-valuation: attending to the most important instrument in the practice of medicine. Mayo Clin Proc 2019; 94:2022–2031.
24. Shanafelt TD, Makowski MS, Wang H, et al. Association of burnout, professional fulfillment, and self-care practices of physician leaders with their independently Rated leadership effectiveness. JAMA Netw Open 2020; 3:e207961–e1207961.
25. Guarana CL, Barnes CM. Lack of sleep and the development of leaderfollower relationships over time. Organ Behav Hum Decis Process 2017; 141:57–73.
26. Shanafelt TD, Wang H, Leonard M, et al. Assessment of the association of leadership behaviors of supervising physicians with personal-organizational values alignment among staff physicians. JAMA Netw Open 2021; 4:e2035622–e12035622.

COVID-19; health care workers; modifiable behaviors; psychological distress; well-being

Supplemental Digital Content

Copyright © 2022 American College of Occupational and Environmental Medicine