The compassion and dedication of health care workers are vital assets that have been deemed elemental to "effective and empathetic" patient care.1 But our complex health care environment makes it increasingly difficult for clinicians to hold onto those assets. Clinicians face growing pressure to improve care quality while containing costs, even as they must care for larger numbers of sicker patients. The demands of their jobs often exceed the available resources.2 As a result, clinicians are often left overwhelmed and stressed by health care structures and environments, wherein many patients are critically ill, clinicians' time is constrained, and there is constant pressure to do more with less. These factors have contributed to an epidemic of burnout among health care workers.3-5 When the coronavirus disease 2019 (COVID-19) pandemic emerged, it taxed clinicians in new and unexpected ways. It's imperative that we understand how the pandemic affects health care workers, so that leaders can develop effective strategies to support and promote the overall well-being of this workforce.
Study purpose. The aim of this study was to identify modifiable environmental factors in the workplace that affect the well-being and resilience of health care workers during the COVID-19 pandemic.
BACKGROUND
Well-being is perhaps best understood as a complex interplay among various domains of individual health as well as environmental, organizational, and psychosocial factors. A high level of well-being is characterized by positive psychological functioning, a sense of personal fulfillment, and engagement with work that leads to joy in practice.3, 6, 7 In contrast, poor well-being can adversely affect mental, physical, emotional, and psychosocial health; if not addressed, this can lead to burnout, mental health conditions, addiction, and even suicide.8-10 Research indicates that health care workers experience higher degrees of burnout than the general population.2, 3, 9 Up to half of U.S. nurses and physicians exhibit substantial symptoms of burnout, including exhaustion, cynical or callous responses to others, and feelings of incompetence.2 Although most of this research has focused on nurses and physicians, there is evidence that burnout also poses a significant problem for other health care workers.2 Furthermore, clinician burnout presents a significant threat to patient safety and health care outcomes.2, 8, 9, 11
In a recent report, the National Academies of Sciences, Engineering, and Medicine recognized that clinician well-being is influenced not only by individual considerations, but also by health care systems.2 Thus a systems perspective is optimal for understanding what shapes health care workers' well-being and targeting areas for improvements within systems. From this perspective, well-being is affected by the relationships between job demands and job resources.2, 3, 12
Research suggests that resilience, which can be defined as the ability to cope with and adapt positively to adversity, is an important contributing factor to well-being.2, 13-16 Resilience appears to help health care workers handle the complexities of their day-to-day jobs, thereby mitigating negative effects of health care systems on well-being.
The COVID-19 pandemic has further stressed an already vulnerable health care workforce. In the context of many unknowns, health care systems have been forced to rapidly create or modify policies, procedures, and practices in order to provide care for patients. Besides disrupting the usual workflow, the pandemic has also created great uncertainty in the work environment. Such stressors threaten the well-being of health care workers.
Conceptual model. In 2017, a National Academy of Medicine (NAM) action collaborative created a working group of representatives from across health care disciplines. The group was tasked with creating a comprehensive, multidisciplinary conceptual model of factors that affect clinician well-being and resilience.3, 17 The model includes five broad categories of external factors (society and culture, rules and regulations, organizational factors, learning and practice environment, health care responsibilities) and two broad categories of individual factors (personal factors, skills and abilities). This model served as the theoretical basis for the conceptual model in our study (see Figure 1). The NAM model's five categories of external factors were used to select and operationalize study variables and to address this research question: During the COVID-19 pandemic, how do environmental factors in the workplace affect the well-being and resilience of health care workers?
METHODS
Study design, sample, and setting. The study used a cross-sectional survey design. The convenience sample consisted of nurses (LPNs and RNs), advanced practice providers (NPs, physician assistants, and certified nurse midwives), certified registered nurse anesthetists, respiratory therapists, health care technicians, and therapy service professionals (physical, occupational, and speech therapists). An initial 6,120 participants were recruited from nine hospitals, behavioral health facilities, physical rehabilitation facilities, and numerous outpatient settings within a large health care system in the southeastern United States.
Data collection was conducted using a self-administered online survey. Potential participants were sent an invitation e-mail that notified them of the study and requested their participation; this was followed by three reminder e-mails. Each e-mail included the survey link. Recruitment flyers were also posted in clinical areas, and clinical leaders shared information about the study with their staff. Participation was voluntary and all responses were anonymous. Completion of the survey constituted implicit consent. Data were collected and managed electronically through Research Electronic Data Capture (REDCap), a secure, web-based application with survey and data management capability. Data collection took place from June 1 through July 17, 2020. The study was reviewed and approved by the health care system's institutional review board.
Study measures. The survey questionnaire included three instruments and 14 single-item questions used to measure the study variables; in total, there were 44 items. Well-being was measured with the Well-Being Index (WBI), a nine-item tool designed to assess the domains of fatigue, stress, anxiety, depression, burnout, quality of life, and work-life integration.18 The WBI has been used in large studies conducted among various health care professionals, and has demonstrated good construct validity and utility.18-20 It consists of seven yes-no questions and two items with Likert-scale response options. Possible final scores range from -2 to 9, with higher scores indicative of greater distress or poorer overall well-being.21 Based on earlier research among nurses, for our study, we set the threshold score at 2.18, 20 Participants with WBI scores of 2 or higher were considered to have "at-risk" well-being; those with scores lower than 2 were not.
Resilience was measured with the 10-item Connor-Davidson Resilience Scale (CD-RISC 10).22 It consists of 10 statements with Likert-scale response options ranging from 0 (not true at all) to 4 (true nearly all the time). Possible total scores range from 0 to 40, with higher scores indicative of greater resilience. Surveys of the general U.S. population suggest that mean scores range from 31 to 32.23 The CD-RISC 10 has been used in studies with various health care professionals,24, 25 and has demonstrated good construct validity and reliability.22
Psychological safety, which may be defined as the degree to which an individual feels the environment is conducive to vulnerability and interpersonal risk-taking,26, 27 was measured with the Psychological Safety Scale (PSS).26 The original seven-item PSS was based on the theoretical constructs of the model of work-team learning.26 We used a four-item version similar to one used by earlier researchers28; it consisted of four statements about psychological safety with Likert-scale response options ranging from 1 (strongly disagree) to 5 (strongly agree). We also adapted item wording to more broadly reflect the team participants worked with or the unit where participants were employed (for example, "If I make a mistake within my team or unit, it tends to be held against me"). The mean score for each participant, which was obtained by averaging responses to the four statements, ranged from 1 to 5.
Work environment factors experienced during the COVID-19 pandemic were evaluated using 14 items. These focused on the following topics: organizational emotional support of health care workers; use of well-being and resilience resources; workload; COVID-19-related communications; perceived prioritization of health care workers' safety; availability of personal protective equipment (PPE); perceptions of COVID-19 policies; COVID-19 educational support; leadership support from direct managers; staffing; redeployment of frontline staff to areas with greater needs; and increased use of virtual care or telehealth services (see Table 1). One item, assessing organizational emotional support ("My organization understands that health care workers may need help to process and resolve the effects of caring for patients during COVID-19") was modified from the Second Victim Experience and Support Tool, which has demonstrated validity and reliability.29 We sought to strengthen all items' content validity by using the NAM model as the theoretical foundation and having external experts assess for face validity. Time constraints prevented us from testing these questions prior to survey administration. Most items offered response options on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Two additional questions asked participants what percentage of time they spent providing care through telehealth or virtual care services before and since the start of the COVID-19 pandemic. Answers to these two questions were used to create an ordinal variable ranging from 1 (no increase in these services) to 5 (76% to 100% increase in these services).
Further questions concerned sample characteristics, including health care worker role, employment setting (inpatient versus outpatient), employment area (department or service), and employment location (hospitals, outpatient facilities, or other [such as behavioral health or rehabilitation services]). We also asked, "In your role, have you provided care to suspected or confirmed COVID-19-positive patients?" Each of these characteristics were categorical variables.
Data analysis. Descriptive statistics were computed for sample characteristics and included differences in well-being and resilience by characteristic.
Univariate logistic regression was used to evaluate the significance of within-group differences for WBI scores of 2 or higher and each of the sample characteristics. For variables with sparse data, Firth's penalized likelihood was used to adjust for that lack. Analysis of variance (ANOVA) was used to evaluate within-group differences for resilience. Logistic regression models were used to evaluate the significance of univariate relationships between work environment factors, sample characteristics, and at-risk well-being; for variables with sparse data, Firth's penalized likelihood was used. Linear regression models were used to evaluate the significance of univariate relationships between work environment factors, sample characteristics, and resilience.
Multivariable models for well-being and resilience were then constructed via backward elimination of significant variables from the univariate analyses. We chose backward elimination because it's been shown to be an acceptable, efficient approach to model construction.30 Variables with P values greater than 0.05 were eliminated; thus the final combined models included only variables that significantly contributed to well-being and resilience. Multiple logistic regression was used to identify factors that significantly affected at-risk well-being. Multiple linear regression was used to identify factors that significantly affected resilience.
Analyses were conducted with SAS software version 9.4.
RESULTS
Sample. Of the 6,120 health care workers who were recruited, we received responses from 2,848 participants (response rate, 46.5%). We excluded 389 because they did not provide direct patient care; therefore, the final analyses involved data from 2,459 participants. Because some participants did not answer all items or complete all scales, sample sizes for specific analyses are varied; analyses were based on the actual number of respondents to a given item.
Sample characteristics and differences in well-being and resilience by characteristic are shown in Table 2. Most of the health care workers who participated in the study were nurses (58.3%) or advanced practice providers (23.9%). More than two-thirds worked in inpatient settings (67.2%) and had provided care to patients who tested positive for COVID-19 (71.6%). Only 21.5% of participants indicated that they'd used well-being or resilience resources since the start of the pandemic. WBI scores ranged from -2 to 9, and nearly half the participants (43.9%) had at-risk scores. There were significant within-group differences for at-risk well-being based on employment setting, employment location, and caring for COVID-19-positive patients. Resilience scores ranged from 8 to 40, and the mean resilience score was 30.6. There were significant within-group differences for resilience based on role, employment setting, and employment location.
At-risk well-being. Univariate logistic regression models identified numerous factors as having a statistically significant relationship with well-being (see Table 3). In these analyses, only frequency of support resources use and virtual care or telehealth use did not significantly increase the odds of at-risk well-being.
Multiple logistic regression analysis identified several factors that significantly increased the odds of at-risk well-being. These included lower levels of resilience, use of support resources, feeling that the organization doesn't understand the emotional support needs of health care workers during the pandemic, believing the workload was greater, believing there is insufficient PPE, believing that staffing is inadequate, and poorer psychological safety (see Table 4).
Resilience. Univariate regression models identified several work environment factors that were significantly associated with resilience. In these analyses, the only factors that lacked significance were use of support resources, frequency of support resources use, and virtual care or telehealth use. Results of these analyses are reported in Table 5.
Multiple linear regression analysis identified several factors that, after controlling for role and employment location, were significantly associated with higher resilience. These factors included feeling that the organization understands the emotional support needs of health care workers during the pandemic, believing that sufficient educational resources were available to help workers safely care for COVID-19-positive patients, having positive perceptions of leadership support from direct managers, believing that staff redeployment to critical areas was necessary to meet patient and team needs, and greater psychological safety (see Table 6).
DISCUSSION
This study identified several work environment factors that have affected the well-being and resilience of health care workers during the COVID-19 pandemic. These findings have practical implications for health care leaders, policy makers, and researchers.
The study was conducted during June and July 2020, roughly three months into the pandemic. While some of the initial shock and panic had lessened by then, we found that 43.9% of health care workers-more than one in three-had at-risk well-being. As such, the research indicates these workers are more likely to experience burnout, severe fatigue, poor quality of life, below-average job performance, greater intent to leave, and higher risk of patient care errors.20
In this study, well-being was significantly and adversely affected by lower levels of resilience, use of support resources (such as meditation apps, employee assistance programs, and counseling), feeling the organization didn't understand the emotional support needs of staff during the pandemic, perceiving workload had increased, believing PPE was insufficient, believing staffing was insufficient, and poor psychological safety. Study findings also indicated that, after controlling for role and employment location, resilience was significantly affected by workers' perceptions of the organization's understanding of the emotional support needs of staff during COVID-19, the availability of COVID-19 educational resources, leadership support from direct managers, organizational staff redeployment practices, and psychological safety. Specific implications of these findings warrant a closer look, as follows.
PPE availability. Shortages of PPE place frontline health care workers at greater physical risk for contracting COVID-19, and the perception that PPE is insufficient also has an adverse effect on well-being. At the practice level, our findings underscore the importance of effective crisis communication. Clear and coordinated messages regarding PPE availability and any supply chain challenges must be shared with frontline workers if they are to believe the organization is committed to their protection.31 Leaders should communicate about PPE not only via e-mail but also through other means (such as team huddles, one-on-one meetings, leader rounds, and "town hall" teleconferences). Using multiple means of communication allows leaders to better assess workers' perceptions of, and share current information about, PPE availability and effectiveness.32
The Centers for Disease Control and Prevention has developed a framework that health care facilities can use to optimize their supply of PPE during shortages (available at http://www.cdc.gov/coronavirus/2019-ncov/hcp/ppe-strategy/index.html). This framework may also help health care leaders to develop plans that take PPE availability and actual use into account. The PPE shortages seen during the COVID-19 pandemic and their effect on health care workers also have lasting implications with regard to local, state, and national policies. Future pandemics and other disasters are certain, and will require government investment in a strong public health system that can better address the needs of workers as well as patients.31 Supply chain deficiencies must be corrected, and innovations pursued, so that we can adequately meet future PPE demands.
Work environment. Our findings suggest that leaders can take crucial steps toward optimizing workers' well-being by paying careful attention to workload and staffing, creating a culture of psychological safety within teams and units, and recognizing and actively addressing the unique challenges posed by the pandemic.
Workload and staffing concerns have long been challenging for our health care systems, and the COVID-19 pandemic has only amplified the challenges. When infection rates increase and hospitalizations rise in a given region, nursing leaders may have to adjust skill mixes or nurse-to-patient ratios. Hospitals near or at surge capacity might have to hire contract employees (such as travel nurses) to maintain adequate staffing, which could require an increased staffing budget. More creative strategies may also be necessary. For example, some hospitals have used redeployment, moving health care workers from low census or low patient acuity areas to areas of greatest need. While this takes careful coordination and planning, it can be an effective strategy that both ensures optimal patient care and supports staff strained by heavy workloads. Another strategy, when feasible, might be to give workers more flexible staffing options, such as working four 10-hour shifts versus five eight-hour shifts, with the former allowing more time away from the work setting.
Psychological safety. Another factor this study assessed was psychological safety within one's team or unit, as this can affect both well-being and resilience. Much of past research has studied psychological safety at the organizational or team level in corporate settings; recently there has been growing interest in psychological safety specifically in health care settings. At the team level, a culture of psychological safety is created through the efforts of both the team leader and individual team members.33, 34 To this end, team leaders can invite and encourage input from all team members, listen actively to members' concerns, model vulnerability by acknowledging the limits of their own knowledge, and provide positive reinforcement for innovative efforts, even when these fail.33 Indeed, a "fail fast" approach allows teams to rapidly test new ideas and to quickly implement those that are successful and drop those that are not. Similarly, individual team members can foster psychological safety by demonstrating a willingness to ask for help, speaking up when problems are observed and offering solutions, and listening actively to and supporting their fellow teammates.33
Resilience. Prior research indicates that resilience is important to well-being because it lessens the effects of workplace stressors.2 This study considered resilience among health care workers in the context of the COVID-19 pandemic. While we didn't specifically examine how resilience affects well-being (such as potential mediating or moderating effects), our findings have implications for individual clinicians and leaders seeking to improve the well-being of health care workers, especially during such times of collectively elevated stress.
One concerning study finding was that only 21.6% of participants indicated that they'd used resources to support their well-being and resilience since the pandemic began. We also found that those who did so were actually at greater risk for poor well-being. This suggests that workers who seek out such resources might already be feeling highly stressed, less resilient, and in need of help rather than simply seeking to protect their well-being. It's important for leaders to make certain that health care workers at all organizational levels are aware of the available resources and to ensure that these can be easily accessed, especially by frontline staff. It's also important to emphasize that support resources can be used not only to address eroded well-being but also to help prevent it.
Findings from this study further suggest that resilience is influenced at multiple levels within health care systems. At the organizational level, perceptions about the organization's support of its employees, the availability of educational resources to help workers safely care for COVID-19-positive patients, and the organization's redeployment policy were all factors that significantly affected resilience. At the team or unit level, direct manager support and psychological safety within the team significantly affected resilience. These findings suggest that any efforts to improve worker resilience will require a concerted effort at multiple organizational levels. Because of the scale and complexity of organization-wide initiatives, such efforts will take ongoing planning and actions that extend beyond the COVID-19 pandemic.
Taken together, the study findings indicate that the work environment during this pandemic is contributing to poor well-being among health care workers. Long-lasting solutions will require more than supportive measures. Careful evaluation of and improvements to organizational structures and processes are vital in order to ease clinicians' burdens and protect their well-being.
Limitations and research recommendations. This study was conducted in a single health care system. Some of the work environment factors investigated could be unique to this system. Although the survey response rate was robust, over 50% of recruited workers chose not to participate. The generalizability of the findings may be limited, as there might be distinct differences between those who responded and those who did not. Another possible limitation is the study's cross-sectional design. Longitudinal studies will allow us to better understand how well-being and resilience are affected by various circumstances, over time, and for the duration of the pandemic. There is a profound need for research that explores ways to foster and support well-being and resilience and yields evidence-based strategies that address at-risk well-being in the workplace. Lastly, the study questionnaire assessed well-being and resilience among health care workers during the COVID-19 pandemic. Although many of the factors we examined are likely associated with health care workers' well-being and resilience in general, caution is warranted in applying the study findings outside the context of the pandemic. Further research that more broadly explores work environment factors and their effects on health care workers' well-being and resilience is essential.
CONCLUSIONS
Well-being among health care workers is significantly affected by the structures and environments in which care is delivered. The COVID-19 pandemic has greatly tested the health care system, increasing the demands placed on workers and straining resources. These added stressors have further threatened health care workers' well-being. This study identified several modifiable work environment factors that negatively affect well-being, as well as several factors that positively influence resilience. The insights gained from this study can help health care leaders to target these factors and develop strategies that allow organizations to better support well-being and resilience among clinicians.
REFERENCES