*Article* **Challenges in Working Conditions and Well-Being of Early Childhood Teachers by Teaching Modality during the COVID-19 Pandemic**

**Kyong-Ah Kwon 1,\*, Timothy G. Ford 1, Jessica Tsotsoros 2, Ken Randall 2, Adrien Malek-Lasater <sup>3</sup> and Sun Geun Kim <sup>1</sup>**


**Abstract:** While a global understanding of teacher well-being during the COVID-19 pandemic is beginning to emerge, much remains to be understood about what early childhood teachers have felt and experienced with respect to their work and well-being. The present mixed-method study examined early care and education (ECE) teachers' working conditions and physical, psychological, and professional well-being during the COVID-19 pandemic using a national sample of 1434 ECE teachers in the U.S. We also explored differences in working conditions and well-being among inperson, online, and closed schools, given the unique challenges and risks that ECE teachers may have faced by teaching in these different modalities. From the results of an online survey, we found that in the early months of the pandemic, many ECE teachers faced stressful, challenging work environments. Some were teaching in new, foreign modes and formats, and those still teaching in person faced new challenges. We found many common issues and challenges related to psychological and physical well-being across the three teaching groups from the qualitative analysis, but a more complicated picture emerged from the quantitative analysis. After controlling for education and center type, we found that aspects of professional commitment were lower among those teachers teaching in person. Additionally, there were racial differences across several of our measures of well-being for teachers whose centers were closed. Upon closer examination of these findings via a moderation analysis with teacher modality, we found that Black and Hispanic teachers had higher levels of psychological well-being for some of our indicators when their centers were closed, yet these benefits were not present for Black and Hispanic teachers teaching in person.

**Keywords:** COVID-19 impact; early care and education; early childhood teachers; well-being; job demands; teaching modality; racial disparity

#### **1. Introduction**

It has been well documented that the majority of early care and education (ECE) teachers reported high levels of satisfaction with and commitment to their work before COVID-19 [1–3]. However, other barriers and challenges—disparities in wages, benefits, resources, and challenging working conditions—outweigh satisfaction and commitment and serve as job stressors [2,4–6]. Unfortunately, the COVID-19 pandemic has likely exacerbated challenges to teachers' work and well-being [7,8]. The pandemic necessitated rapid changes in teaching and student support—the demands of which fell squarely on the shoulders of teachers and leaders—and were intensified by a shifting landscape as schools and communities were pressed to re-open [9–12]. Such demands have resulted in

137

**Citation:** Kwon, K.-A.; Ford, T.G.; Tsotsoros, J.; Randall, K.; Malek-Lasater, A.; Kim, S.G. Challenges in Working Conditions and Well-Being of Early Childhood Teachers by Teaching Modality during the COVID-19 Pandemic. *Int. J. Environ. Res. Public Health* **2022**, *19*, 4919. https://doi.org/10.3390/ ijerph19084919

Academic Editor: Paul B. Tchounwou

Received: 10 March 2022 Accepted: 15 April 2022 Published: 18 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

unprecedented stress, threatening the short- and long-term well-being of teachers, many of whom are coping with similar stress and demands in their personal lives [9].

Papaioannou et al. [13] have referred to the COVID-19 period as a triple pandemic, calling attention to not only the disease itself, but to the physical inactivity and mental illness that have followed. Numerous studies conducted during the pandemic have slowly revealed the deleterious effects of social restrictions, "shelter-at-home", and online learning on adults [14]. The general consensus of the empirical research on P-12 teachers conducted in various countries around the world, such as the U.K., the U.S., Brazil, Mexico, Australia, Spain, and Portugal, is that teachers have suffered a significant impact to their psychological, physical, and professional well-being [8,15–18]. For example, Swigonski et al. [8] found that physical and behavioral symptoms of stress among early childhood teachers in the U.S. were 2–3 times higher during the COVID-19 pandemic, and this is significantly higher than in the community at large. Similarly, Alves et al. [15] found that the pandemic has reduced teachers' perceptions of professional well-being, leading them to feel more uncertain about the future of their career.

The role of teachers and ECE centers/schools becomes even more significant in crisis situations such as the COVID-19 pandemic as students and families look to teachers for more psychosocial support [19]. For example, Ozmiz-Etxebarria et al. [19] found that preschool and primary grade teachers working in a university nursery school showed the highest ratings of psychological symptoms such as anxiety. Teachers and ECE centers/schools are expected to serve as the "great equalizer", providing additional socialemotional learning and educational opportunities for vulnerable students who are more likely to be harmed in such crises [20,21]. Thus, it is critical to broadly examine ECE teachers' working conditions and well-being during the pandemic, but holistic studies of working conditions and well-being remain sparse [15], particularly for U.S. teachers and U.S. ECE teachers more specifically [11,22]. However, few studies with national samples are available.

The onset of the COVID-19 pandemic meant that many ECE centers and schools transitioned to online learning, some closed altogether, and others remained open, precipitating substantial upheaval in the nature and scope of teachers' professional lives and work. While evidence shows that teachers' well-being has suffered due to the pandemic (e.g., declining well-being and satisfaction) [15], some studies are have revealed that these impacts may vary by different contextual and individual factors, such as teaching modality and the demographic characteristics of teachers. During the early months of the pandemic, some teachers suddenly found themselves employed without work, while others experienced increased workloads [23,24]. Teachers who remained teaching in person during the pandemic were at higher risk for contracting COVID-19 and at higher risk of severe illness if they did because of the chronic health conditions that accompany this vulnerable workforce [24,25]. The unique challenges and circumstances that surrounded the in-person work of ECE teachers likely influenced these risks as well. For instance, having a group of children in a small space, or having children in close contact during routine activities (e.g., diapering, feeding), likely made social distancing a challenge [23–25]. For ECE teachers, the additional tasks of taking precautions and constantly reminding young children to keep separated, wear masks, and wash hands was undoubtedly exhausting and stressful. For those one-in-four teachers at risk for severe illness, not remaining protected could have had lasting consequences, yet many ECE teachers may have felt financially compelled to continue to teach in person, given the low wages that many teachers receive, despite the risks [25].

Outside of worrying about their own health risks, some evidence suggests that teachers teaching online may have faced a unique set of stressors. For example, a study by Besser and his colleagues [26] demonstrated that a sudden transition to online teaching was related to higher levels of psychological stress among teachers. An international sample of 600 language K-12 teachers involved in online teaching also revealed high levels of stress related to additional workload, family health, loss of control over work decisions, blurred professional and personal lines, concerns about their colleagues and most vulnerable students, social isolation, and the stress of online teaching itself [27]. Similarly, Allen et al. [28] found high levels of stress for K-12 teachers in the U.K. teaching online during the beginning of the pandemic; however, those teaching in person were shown to have higher levels of stress for longer periods of time. Although ECE teachers teaching online are likely to face similar stressors, there is limited evidence in how these conditions affected them.

Some teachers worked in schools that were completely closed for a time during the pandemic, although this group of teachers rarely received attention. Teachers reported worries related to financial and job security as well as the sustainability of their center during closure [8,17]. Other teachers were worried about their students: how they were managing the pandemic, how they were being taken care of, and how their social development was being impacted. These types of worries could be associated with higher levels of poor well-being such as stress and vicarious/secondary trauma [17].

The picture of teachers' working conditions and well-being differing by modality may be further complicated by teacher demographic characteristics and other contextual factors. Based on the well-documented fact that COVID-19 has disproportionately impacted people who are vulnerable, including children and adults from minoritized and under-resourced groups [29–31], we expected that teachers from underrepresented (e.g., racial/ethnic minority groups) and under-funded groups (e.g., private childcare centers, family childcare centers) were more likely to suffer from poorer working conditions and well-being during the pandemic. However, most extant studies during this period address racial/ethnic health disparities or educational inequality among underrepresented groups of children or the public—not teachers (see Souto-Manning and Melvin [32] for an exception). Limited empirical evidence has been collected across states documenting the shared and divergent impacts of the COVID-19 pandemic on the ECE workforce by teaching modality and the complex interplay among individual and contextual factors on teachers' work and various aspects of well-being. In particular, research on ECE teachers' physical well-being during the pandemic is scarce.

As one of the few exceptions, Collie's study examined these issues, but in primary and secondary education settings [9]. In this study, Australian teachers' gender, work experiences, and teaching modality were associated with differences in teacher working conditions and well-being. For example, teaching half remotely and half in person was related to greater stress. While teachers who have reduced work hours may have more time to address challenges in their teaching, it may also precipitate other concerns (e.g., concern about financial issues and job security due to reduced hours [9]). Despite its contribution, this study was conducted in primary and secondary school settings in Australia, which may not capture the unique demographic characteristics of the early childhood workforce (e.g., no variation in gender; more racial/ethnic diversity; varied educational level; more limited resources) in various settings (e.g., family childcare homes, private childcare, Head Start, public school, private school). In the U.S., Souto–Manning and Melvin [32] conducted an in-depth multi-method study to address these gaps by examining racial, occupational, and environmental factors on physical and psychological well-being among early childhood teachers of color in New York. However, the scope was limited to ECE teachers of color in one city.

#### *The Present Study*

While a global understanding of teacher well-being during the COVID-19 pandemic is beginning to emerge, much remains to be understood about what ECE teachers have experienced and felt with respect to their working conditions and well-being—particularly considering the wide range in responses of ECE centers/schools to the pandemic in the U.S. Most of the existing evidence is limited in scope (e.g., confined to one region or one particular aspect of well-being), rarely addressing issues specific to ECE teachers who were already vulnerable prior to the pandemic. Furthermore, given that there was substantial variation in how teachers were teaching during this time, it is highly probable that teachers teaching in different modalities experienced different challenges in their working conditions and well-being and that these experiences might have differed across teachers' race, educational level, and program type.

Thus, in this study, we report on the challenges the COVID-19 pandemic via quantitative and qualitative analysis of national online survey data from 1434 ECE teachers in the U.S. We asked three research questions: (a) How did ECE teachers' working conditions and well-being differ by teaching modality (i.e., in-person, online, closed school)? (b) Were there teacher demographic and center-based differences in teacher well-being? and (c) Did teaching modality moderate the relationship between demographic and centerbased differences and teacher well-being? The findings of this study could offer important implications for the field given the ongoing and unpredictable nature of the pandemic and its lingering effects.

### **2. Method**

#### *2.1. Participants and Settings*

A total of 1434 ECE teachers serving children ages 0 to 5 (including Kindergarten) in 46 states in the United States completed an online survey in late spring to mid-summer of 2020. The overall racial/ethnic composition of the sample is similar to the population of early childhood teachers nationally [33], with a somewhat higher percentage of Hispanic teachers. The sample includes 58% White, 21% Hispanic, 14% Black, 3% American Indian or Alaska Native, 2% biracial, and 1% Asian. The vast majority of teachers in the sample were women (98.3%). The average age of the participants was 42 (*Range =* 17 to 80). The majority of teachers in the sample were fully paid (83%), but some were only partially paid (12%) or not paid at all (5%). Among those who were paid, the annual salary for fifty-one percent of the teachers was below USD 30,000. Fifteen percent of teachers received some form of public support, such as Medicaid, food stamps, or childcare subsidies. Seventy-three percent of participating teachers held an associate degree or higher, followed by some college but no degree (20%), and high school diploma or general education diploma (GED) (5%).

Participating teachers worked in Head Start programs (43%), childcare centers (34%), public schools (14%), family childcare homes (6%), and private schools (3%). They served infants and toddlers (24%), preschool or pre-K (38%), Kindergarten (6%), and children in multiage groups (31%). Teachers in the sample served children and families of diverse socio-economic status (SES): predominately low SES (54%), middle SES (15%), upper SES (7%), and mixed SES classrooms (24%). Of the 1434 early childhood teachers in the sample, approximately 29% reported that they were teaching in person, 28% were teaching online, and the remaining 43% were not teaching due to their centers/schools being closed.

#### *2.2. Research Procedure and Analysis*

After receiving Institutional Review Board (IRB) approval, we recruited early childhood teacher participants via various social media platforms (e.g., Facebook, Twitter). To ensure responses from various states and types of settings, such as private childcare centers, public schools, Head Start programs, and family childcare homes, stratified sampling (by state and setting type) was also integrated into the recruitment approach. This procedure involved producing a sample frame of ECE settings in each U.S. state and, from this, developing a contact list, which was first proportional to state population and then sought to preserve U.S. representativeness by setting type. These centers/schools were then first contacted to participate in the study. Once this procedure was complete, we also emailed state and national ECE organizations and agencies to distribute our survey more broadly.

Our interdisciplinary research team developed questions for an online survey that asked about their personal and professional background and the teachers' experiences and well-being at work during the COVID-19 pandemic. The questions included: (a) demographic and background information, including teacher race, education, income; (b) teaching modality (i.e., teaching online, teaching in person, school closed); (c) whether they experienced

changes in their work and well-being (i.e., negative change, positive change, no change); (d) if they experienced any change, what change they experienced (open-ended response); and (e) what they needed for improvement in their work (open-ended responses) and well-being (multiple choice). The online survey also consisted of previously validated scales to assess teachers' well-being. On average, it took 25–30 min to complete the online survey. Among teachers who completed the survey and who requested to participate, sixty teachers were randomly selected to receive a USD 50 electronic gift card.

#### *2.3. Measures*

Below, we briefly describe the key measures used for the quantitative analysis of the study. More detailed descriptions and the psychometric properties of each measure are organized in Table 1.

#### 2.3.1. Working Conditions

This was assessed using three subscales from the Job Content Questionnaire (JCQ) [34]. The three subscales used in this study were physical job demands (related to the physical demands of the job), skill discretion (related to the variety of skills used for the job), and decision authority (related to the amount of job control). We also included questions asking teachers to report how they were paid (i.e., fully paid, partially paid, not paid) and if they had health insurance provided by their employer during the pandemic.

#### 2.3.2. Teacher Well-Being

**Psychological Well-Being**. Teachers' psychological well-being was operationalized via ECE teachers' perceptions of: (a) depressive symptoms, (b) stress, (c) resiliency, (d) life satisfaction, and (e) secondary trauma. Teacher depressive symptoms were assessed with the 10-item Center for Epidemiologic Studies of Depression Short Form (CES-D-10) [35]. Perceived stress was measured using the Perceived Stress Scale (PSS) [36]. Teachers' resiliency was measured using the Brief Resilience Scale [37], and life satisfaction was assessed on the Satisfaction with Life Scale [38]. Lastly, teachers' secondary trauma was assessed using one of the subscales of the Professional Quality of Life Scale [39] We used a total score of each measure for data analysis.

**Physical Well-Being**. Teachers' physical well-being constituted a measure of (a) general health condition, (b) ergonomic pain, (c) food security, (d) Body Mass Index (BMI) [40] and (e) physical activity. To examine their general health status, we used a composite score of various dichotomous doctor-diagnosed symptoms that teachers reported. Ergonomic pain was assessed using the modified version of the Work-Related Musculoskeletal Disorders Scale (WMDS) [41]. The Short Form of the Food Security Survey Module was modified to assess food insecurity. Body Mass Index (BMI) [40] was calculated by dividing weight in kilograms by the square of height in meters. To measure physical activity, direct questions about how many days and hours were spent on vigorous physical activities and how much time was spent sitting on a weekday were asked.

**Professional Well-Being**. Professional well-being was assessed via two constructs: work commitment and intent to leave. Work commitment was measured using the Early Childhood Job Satisfaction Survey (ECJSS) [42]. Intent to leave the field/profession was measured via three items.


**Table 1.**

Description

 of measures used in the study.

#### *2.4. Data Analysis*

We conducted an analysis of both qualitative and quantitative responses. For the analysis of qualitative data (i.e., open-ended responses in the survey), five members of the research team conducted a content analysis of the qualitative data (i.e., teachers' openended responses in the online survey). The first author served as master coder and assigned questions to four other researchers for analysis. She met with individual coders, conducted open coding, and developed the initial codes. We (the first author and four other coders) met multiple times to refine codes and categories and discuss any discrepancy until reaching consensus. We conducted reliability checks with 10–15% of cases per question and established an inter-coder reliability ranging from 90 to 100 percent agreement before independent coding. Cohen's Kappa ranged from 0.65 to 0.85 across the categories and coders. We compared categories of challenges in and needs for job demands and well-being during the pandemic across the three teaching modalities. While the response consolidation process resulted in more than ten categories, in this paper, we present the five most frequently reported responses from the content analysis. Similarly, the analysis of the quantitative data began with a descriptive analysis, and sample descriptive statistics are displayed in Table 2. Finally, an Ordinary Least Squares (OLS) regression analysis of the main effects of teachers' education level, race, and type of setting on the various studied aspects of physical, psychological, and professional well-being was conducted. This was followed up by a test of the moderation of teaching modality upon the effects of teacher race and well-being.


**Table 2.** Psychological and physical well-being of 1434 ECE teachers during pandemic.


Note. ª is an indicator of the product of both number of affected areas (max of 5 areas) and severity of pain (from 0 = no pain to 4 = unbearable pain), for a max range of 20.

#### **3. Results**

Of the teachers surveyed, approximately 29% reported teaching in person, 28% teaching online, while the remaining 43% were not working due to their centers being closed. As data were collected in the early phase of the COVID-19 pandemic, most of the teachers fit in one of these categories. Teaching modality significantly differed by type of setting: the majority of teachers in family childcare homes (82%) and in childcare centers/preschools (55%) taught in person, while 81% of public school teachers taught online. Fifty-four percent of Head Start teachers reported that their centers were closed during the pandemic and 39% of Head Start teachers taught online.

We found that in the early months of the pandemic, while many ECE teachers were committed to their work, moderately resilient, and tried to have a positive outlook, they still faced stressful, challenging work environments. Some were teaching in new, foreign modes and formats, and those still teaching in person faced new challenges. An understanding of these unique challenges begs an understanding of how they might have affected their overall well-being. Below, we examine this question with a specific focus on the status of ECE teachers' job demands, as well as their physical, psychological, and professional well-being and how it differed by teaching modality.

#### *3.1. Working Conditions and Needs by Teaching Modality during the Pandemic*

From the content analysis (see Table 3), we found that, regardless of teaching modality, early childhood teachers experienced significant challenges working with young children in the early days of the pandemic, characterized by limited resources and a lack of clear guidelines and regulations. Although some challenges were common, we found distinctive challenges faced between the in-person teaching group and the online teaching group. As the school closed group was not working during the pandemic, our analysis focused only on the in-person and online teaching groups.


**Table 3.** Content analysis summary: challenges in work and well-being.

Respondents teaching in person reported many challenges at work: (a) challenges teaching and interacting with children and families in person during the pandemic; (b) financial hardships; (c) fears and uncertainty of becoming infected and passing it to their family; (d) additional job demands, and (e) too frequent changes in regulations and circumstances.

First, the major issue that many teachers encountered was related to the various challenges that arose in supporting and properly working in person with children and families during the pandemic while having to wear masks, implement social distancing, and frequently wash their hands. They also reported concerns about the current method of teaching not being optimal and developmentally appropriate for young children. It was difficult for teachers to help young children understand the situation and the importance of keeping masks on. One teacher stated:

*I work with infants. Wearing a mask makes it difficult to show them facial expressions. Infants love to see people smile at them* ... *The infants don't know the new teachers and don't feel as safe as they would with their original teacher*.

In addition, teachers were concerned that their interactions and communications with families were limited or restricted because of health concerns during the pandemic. At the time, many centers did not allow families to enter the building so teachers were not able to have any face-to-face communication with families.

The second most frequently mentioned challenge was financial in nature. Low enrollment seemed to be a major cause of the financial issues of the center/school. While lower enrollment led to some positive changes for teachers (such as reduced workload and more time and attention given to each child), it also led to reduced working hours and compensation, which raised concerns for teachers. Furthermore, centers/schools were required to purchase additional resources such as cleaning supplies and personal protective equipment, which were limited in availability. This resulted in additional financial burdens and stress on the center as well as teachers and was particularly the case in family childcare homes and private childcare centers that relied heavily on tuition as a main source of income. One teacher commented, "*The biggest challenges of teaching during the COVID-19 is not having enough kids to stay open, but being shut down for a while, because we had no kids*".

Third, although they tried to take all possible precautions, teachers who worked in person faced the great risk of becoming infected and passing it to their family. They reported that they were fearful and concerned about the possibility of contracting COVID-19 at work. Teachers mentioned that although they were aware of the lower risk of children being infected with COVID-19, they were afraid that a child could be infected without symptoms and inadvertently infect them. In addition, they were concerned that the supplies and resources provided were not often appropriate or sufficient to protect them from infection. One teacher said:

*I have to go through and to breathe and communicate clearly with the face mask on. The gloves I feel are acceptable but the mask that the company provided is not appropriate. It is a t-shirt that someone made., It is super thin almost see-through if you put up to the light. When I have it on and breathe the clothing goes into my mouth. Not sure if it is protective enough*.

Fourth, additional job demands were identified as a major challenge. These mostly consisted of additional tasks related to the new health and safety measures for in-person teaching, such as difficulty finding appropriate cleaning, personal protective equipment, and materials (as they were often out of stock or in low supply); having to follow many new regulations regarding cleaning/safety, and having to do constant cleaning. Pick-ups and drop-offs to ensure safety became challenging and stressful for children, families, and teachers, which is evidenced in one teacher's response: "*It is hard because we have one or two teachers constantly running out to receive children from the car or take them to the car*".

Fifth, teachers reported that they experienced too many changes to their routines, staffing, and grouping and were never certain of what to expect the next day. Due to

low enrollment and staff shortages, centers/schools had to lower teacher–child ratios and merge different groups (resulting in challenges teaching multiple age groups). Feelings of being overwhelmed and uncertain sometimes stemmed from a lack of clear communication and guidance at the program, district, state, and national levels.

In response to the challenges they experienced, the in-person teaching group reported that they needed: (a) more resources, supplies, staff, and testing to cope with COVID-19 related challenges; (b) more financial support for the center/school, including better or additional (hazard) pay and benefits for teachers; (c) clearer, more equitable, and more consistent regulations and communication; (d) more emotional support, such as appreciation, respect, and acknowledgement; and (e) more positive attitudes and hope for their situation.

The online teachers also experienced many challenges, but these were substantially different from the challenges reported from the in-person teachers. They identified the following as their major challenges: (a) difficulty supporting children's learning via online teaching; (b) difficulty with parent involvement; (c) technology issues; (d) social isolation/feeling of disconnection; and (e) barriers to resources and preparation for online teaching.

Similar to the in-person teaching group, this group of teachers experienced challenges supporting children through online teaching. Many teachers in this group mentioned that it was difficult to get children to participate and engage in the lessons, and the overall rate of attendance and participation was low. Teachers also found it difficult to make activities engaging and developmentally appropriate through the online platform. These challenges often led to concerns about whether and how much children were learning through this format and how children were doing at home (e.g., they might miss signs of neglect or abuse). Some teachers commented that online teaching made it even more challenging to engage and support dual-language learners and children with special needs and expressed concern that online teaching would undermine equity and exacerbate learning gaps for children from marginalized groups. This was clearly evidenced in this teacher's response:

*Connecting with young children over the screen is HARD. We are not able to address their needs/goals regarding behavior or social emotional play skills. It's all artificial. We are missing out on teaching them and addressing their needs during critical windows of development. Many of these kids just started getting services (special needs) and are going to kindergarten in the fall. It's just yucky all around*.

Second, difficulty with parent involvement was a common issue. Teachers acknowledged that the online teaching format relied heavily on parent involvement and empathized with parents that this added a significant burden for them. They often found it difficult to engage families already distressed from various hardships and additional work demands in their children's learning at home. One teacher expressed this concern:

*Parents were totally overwhelmed by more demands and school expectations and they weren't able to do all the Zoom meetings and lesson activities. I felt like I tried to focus most of my support on the parents and let them know that I believe they are doing their best and that is OK*.

Third, unlike in-person teaching, the availability of and access to technology was an inevitable concern for online teaching. The technology-related issues that teachers experienced included limited access to the internet, unstable internet connectivity—especially for children living in rural or impoverished areas—and access to an appropriate computer and electronic devices necessary for online learning. As one teacher put it, "*The biggest challenge for me is attendance because they do not have access to the internet. Because I teach in a low socio-economic area, many of my students do not have access to technology and/or the internet*".

Fourth, many teachers found it difficult to engage children in social interactions and felt disconnected from their children in the online teaching format. They mentioned that it was difficult not to be physically present for children, show affection (e.g., they cannot hug, make direct eye contact, and play with children), and build and foster relationships. Lastly, teachers noted limited resources and a general lack of preparation necessary for quality online teaching. Although there were some online trainings offered to them, teachers still

felt unprepared and ill-equipped to teach online. They felt that it was difficult to find the right resources, feel effective, and maintain accountability.

To address these challenges, online teachers requested the following for improvement: (a) better access and recourse for online teaching, including technology and internet accessibility; (b) more parent involvement and better ways of engaging them; (c) improved curriculum and format optimized for online learning; (d) clearer and more consistent guidelines and communication, and (e) more training on online teaching. Analysis of the quantitative data (see Table 4 below) revealed no differences among our measures of working conditions by teaching modality.

**Table 4.** Main effects regression of psychological, physical, and professional well-being by modality and teacher race.


Note. All outcome variables standardized. Standard errors in parentheses. \*\* *p* < 0.01, \* *p* < 0.05.

#### *3.2. Early Childhood Teachers' Well-Being by Teaching Modality during the Pandemic*

Overall, a substantial number of teachers reported poor psychological and physical well-being. This is evidenced in both quantitative and qualitative responses. The content analysis identified the most frequently reported responses about perceived changes in well-being (see Table 3). Overall, early childhood teachers experienced remarkably similar psychological and physical well-being-related issues across the three teaching modalities. Across questions on psychological and physical well-being, five themes emerged: (a) more stress, anxiety, and fear of becoming ill with COVID-19; (b) weight gain and lack of physical activity; (c) increased feeling of social disconnection, depression, and sadness; (d) increased concerns about other illnesses as an existing or new health condition; (e) financial concerns.

First, stress, anxiety, and fear of becoming ill with COVID-19 were the most common across the three groups. Those who experienced negative changes in well-being reported high levels of anxiety and stress, often related to fear and uncertainty due to COVID-19. Anxiety and stress were replete and by far the most common challenge to well-being among our sample of teachers. Regardless of teaching modality, it was common that teachers were anxious and fearful of the possibility of contracting COVID-19. Stress was also frequently reported from teachers who were now juggling online teaching while also helping their own children with online learning. A teacher in the online group reported:

*There is constant anxiety. Not knowing if I will lose more families; if I will be able to replace the ones I did lose any time soon; if I or someone in my family gets the virus and unemployment errors cause delays, will we survive financially*?

Second, weight gain and lack of physical activity were also prevalent concerns among early childhood teachers during the pandemic. This was common for all three groups, but it was especially prevalent for online teachers. They reported weight gain resulting from stress eating, unhealthy food choices, being in close proximity to food all the time, and increases in sedentary behavior. As one teacher noted:

*I have gained about 20 pounds (during the pandemic). The added weight has affected my mobility. I have an increased level of discomfort with what formally was very mild aches and pains. I haven't been to the doctor, but I'm sure my blood pressure is out of whack. These may be contributing to what I believe is mild depression*.

Teacher weight gain was also partially due to increased sedentary behavior and a lack of physical activity, and this was a concern across modalities. As one teacher in this online group remarked, "*During the school year, I feel as if I prioritized my diet and exercise more. By spending 7.5 h a day (or more online), my body is very sore after sitting all day*".

Third, feelings of social disconnection, depression, and sadness were the next most prevalent theme across the groups. Teachers felt socially disconnected from their coworkers, students, friends, and family. Social disconnectedness was described as being lonely and feeling distant from loved ones (e.g., not being able to meet their family and friends). Missing their students was common for online teachers and teachers whose sites were closed. One teacher stated, "*I miss the children, I miss the job that I love, I miss my routine, and I miss interacting with my coworkers*".

Depression and hopelessness were commonly reported well-being-related challenges, especially for the online and school closed groups. Teachers connected these feelings of sadness and depression to being out of their routine, missing their kids at school, and feeling trapped—a finding that was not shared by the in-person teachers. A teacher teaching online wrote:

*I'm a very productive person typically. Being cooped up at home has been very challenging to my happiness. Some days, I don't want to get off the couch. Some days, I am very productive and get everything done! I try to spend time in my yard and growing things in my garden. But when I run out of things to do, I get sad* ... *It is difficult for me to put my foot down and refuse to go out when pressured by others as well*.

Fourth, increased concerns about other illnesses as an existing or new health condition were often noted as a challenge in physical well-being during the pandemic. Illnesses incurred pre-pandemic were exacerbated, and new ones arose. One teacher reported, "*I have Type 1 diabetes, and the stress causes my blood sugars to run high*". Another teacher says, "*stress has brought on migraines, which I had not previously experienced, and I've been having frequent chest pains*". Other illnesses included insomnia, reduced energy, restless sleep, and poor sleep patterns.

Fifth, financial concerns were also identified and there were slight differences among the three teaching modalities. Financial concerns seemed to be a more common issue for the in-person group and the school closed group. One teacher puts it, "*It's very frustrating financially. I know many people who are making about double in unemployment benefits than what we are making having to go to work every day and risk our health. My friend is getting over \$1000 a week from unemployment, and we won't be getting any raises this year due to COVID. Teachers deserve competitive pay!*". However, financial issues were not as prevalent for the online teaching group.

In general, the quantitative analysis corroborated the findings of the content analysis on the range of challenges that teachers experienced related to psychological and physical well-being described above. With respect to psychological well-being, our quantitative data revealed that 31% of teachers in the sample reported doctor-diagnosed anxiety and 23% reported doctor-diagnosed depressive symptoms (see Table 2), with 35% having depressive symptom scores reaching clinical levels (based upon a recommended cut-off score of 11 for the shorter CES-D instrument). Forty-eight percent of teachers experienced somewhat or mostly negative changes in psychological well-being during the early months of the pandemic.

Regarding physical well-being, 20% of teachers reported experiencing somewhat or mostly negative changes in their physical health, with substantial numbers of teachers reporting chronic conditions such as being overweight or obese (72%), having high blood pressure (28%), or asthma (22%; see Table 2). With regard to physical activity, 63.56% indicated that they performed an average of 1.37 h of moderate to vigorous physical activities every week. From an ergonomic perspective, 78% of the study participants reported having at least one area of work-related pain, and half (50.29%) indicated that pain interfered to some degree with their work.

Our regression analysis of the main effects of teacher race, center type, and modality, which is displayed in Table 4, reveal that ergonomic pain was highest for in-person teachers, as compared to teachers whose sites were closed, *β* = 0.155, *SD* = 0.076, *p* < 0.05. Regarding professional well-being, more than fifteen percent of teachers (15.40%) in the sample reported that they wanted to leave as a result of the current situation. Furthermore, work commitment was lower for in-person teachers, *β* = −0.298, *SD* = 0.075, *p* < 0.01, than teachers whose centers were closed, *β* = 0.203, *SD* = 0.081, *p* < 0.05, and there were similar differences among these two groups for intent to leave, with in-person teachers reporting higher intent to leave, *β* = 0.219, *SD* = 0.075, *p* < 0.01, than their closed counterparts, *β* = −0.128, *SD* = 0.081, *p* = n.s. Teachers teaching virtually showed no distinct differences with respect to professional well-being as compared to teachers whose schools were closed.

Among those who wanted to leave, the major reasons for thinking of leaving were health concerns and a fear of contracting the COVID-19 or passing it to others (41.50%). The second most frequent reason was dissatisfaction with their current teaching assignment (e.g., they did not want to teach online or in person as it is difficult to work with young children in a developmentally appropriate way under these circumstances, etc.) and/or additional job demands (16.98%). Further, teachers wanted to leave because they did not feel that their field had provided enough job security amidst the crisis and were interested in looking for other career opportunities (15.85%).

With respect to teacher psychological well-being, while there were no differences between teachers by modality, we did find racial differences among teachers and these effects held in our moderation analysis (see Table 5), which we discuss below. First, the pattern of difference in psychological well-being was pronounced for Black and Hispanic teachers whose centers were closed. On average, Black teachers had lower personal stress, *β* = −0.545, *p* < 0.01, depressive symptoms, *β* = −0.458, *p* < 0.01, and secondary trauma, *β* = −0.299, *p* < 0.05, as compared to White teachers whose centers were closed, *β* = 0.170, *p* < 0.05, *β* = 0.284, *p* < 0.01, and *β* = −0.085, n.s., respectively. The same was true for Hispanic teachers in comparison to White teachers: Hispanic teachers reported lower personal stress, *β* = −0.306, *p* < 0.01, depressive symptoms, *β* = −0.355, *p* < 0.01, and secondary trauma, *β* = −0.395, *p* < 0.01, and reported higher life satisfaction than White teachers whose centers were closed, *β* = 0.404, *p* < 0.01. White teachers whose centers were closed had the poorest psychological well-being of the different racial groups—in particular, they had the highest levels of personal stress, *β* = 0.170, *p* < 0.05, and depressive symptoms, *β* = 0.284, *p* < 0.01, and Black and White teachers shared similarly low levels of life satisfaction, *β* = −0.35, *p* < 0.05, for White, *β* = 0.003, n.s., for Black, respectively.

However, these patterns shift when we look at those teachers teaching in person and virtually. Those Black teachers teaching in person exhibited large differences in personal stress, *β* = 0.425, *p* < 0.05, and depressive symptoms, *β* = 0.362, *p* < 0.10, from Black teachers whose schools were closed while, for White teachers, being in person or closed did not make a difference: personal stress, *β* = 0.001, n.s., depressive symptoms, *β* = −0.087, n.s., secondary trauma, *β* = 0.104, n.s., and life satisfaction, *β* = 0.059, n.s. Hispanic teachers teaching in person were similar to Black teachers, showing significantly lower life satisfaction than those whose centers were closed, *β* = −0.381, *p* < 0.05, and brief resiliency for those teaching virtually (*β* = −0.405, *p* < 0.05). Other notable findings for psychological well-being were the fact that teachers with an associate's degree or higher reported marginally higher life satisfaction, *β* = 0.239, *p* < 0.01, and lower depressive symptoms, *β* = −0.111, *p* < 0.10. Public school teachers, as compared to childcare center/pre-K teachers, had higher life satisfaction, *β* = 0.279, *p* < 0.01, but also marginally higher secondary trauma, *β* = 0.366, *p* < 0.01.


**Table 5.** Moderation of teacher race and psychological, physical, and professional well-being outcomes by modality.

Note. All outcome variables standardized. Standard errors in parentheses. \*\* *p* < 0.01, \* *p* < 0.05, † *p* < 0.10.

With respect to professional well-being, there were also some notable findings. First, White teachers teaching in person had significantly lower work commitment, *β =* −0.197, *p* < 0.05, and marginally higher intent to leave, *β* = 0.153, *p* < 0.10. There were also a few differences with respect to physical well-being. As mentioned above, Black and Hispanic teachers at closed centers had marginally lower ergonomic pain than White teachers, *β* = −0.238, *p* < 0.10, *β* = −0.229, *p* < 0.05, *β* = 0.059, n.s., respectively, but Hispanic teachers teaching in person experienced sharply higher ergonomic pain as compared to White teachers, *β* = 0.443, *p* < 0.05, *β* = 0.059, n.s., respectively.

Finally, with respect to job demands by modality or race (or the interaction between the two), there was only one difference: teachers teaching in person reported higher physical job demands across the board, *β* = 0.163, *p* < 0.10, with no differences among racial groups. Other differences, not surprisingly, broke down by education level and center type; teachers with higher education levels reported marginally lower physical job demands but higher skill discretion and decision authority, *β* = −0.116, *p* < 0.10, *β* = 0.249, *p* < 0.01, *β* = 0.201, *p* < 0.01, respectively. Family childcare home teachers had high decision authority and Head Start teachers lower as compared to childcare centers/Pre-K, *β* = 0.826, *p* < 0.01, *β* = −0.339, *p* < 0.01, respectively. Public school teachers had higher skill discretion than childcare center/Pre-K teachers, *β* = 0.390, *p* < 0.01.

#### *3.3. Addressing Challenges: Teachers' Reported Needs for Support*

In concluding our analysis, teachers were also asked to rank the top three items needed to support their well-being out of a list of 22 possible choices. We combined some categories in arriving at 12 overall themes/responses. Among these, the five most frequently listed responses included: (a) higher wages (16.81%); (b) more resources for health and wellbeing (15.84%); (c) more coaching, mentoring, and professional development (including comprehensive safety training, 12.58%); (d) more daily breaks and paid leave (11.17%), and (e) more support for children with behavioral challenges and special needs (10.54%).

#### **4. Discussion**

Dramatic shifts in working conditions occurred for ECE teachers during the pandemic, as some remained teaching in person, while others taught online, and still others whose sites were closed were not teaching at all. Even before the pandemic, the ECE workforce was often characterized as a marginalized group because of their exposure to poor working conditions and their heightened risk of diminished well-being [1,4,8]. We sought to understand if and how these conditions changed during the early days of the pandemic and the role of teaching modality as well as teacher and center characteristics in any differences we found. Thus, this study examined the challenges, risks, and needs for early childhood teachers' work and well-being primarily as a result of these shifts in their work during the COVID-19 pandemic in the U.S. We sought answers to the following three research questions: (a) How did ECE teachers' working conditions and well-being differ by teaching modality (i.e., in-person, online, closed school)? (b) Were there teacher demographic and center-based differences in teacher well-being? and (c) Did teaching modality moderate the relationship between demographic and center-based differences and teacher well-being? The following discussion is organized along the lines of these three questions.

#### *4.1. Differences in Working Conditions and Well-Being by Modality*

Overall, while our findings are in line with those in recent COVID-19 studies mostly focused on K-12 teachers [8,9,18,19,27,28], they extend our understanding of the pandemic's effects on work and well-being by highlighting the unique effects that it has had on the early childhood teacher workforce. This was one of the first studies, to the authors' knowledge, to take a holistic view of ECE teacher work and well-being (i.e., psychological, physical, and professional well-being) using a large sample of teachers serving birth through Kindergarten from nearly all states in the U.S. Our mixed-method approach provided both the depth and breadth needed to fully explore this complex issue and the story we captured is a complicated one—both in terms of the nature of the challenges faced across the three teaching groups as well as the differences between racial groups we found.

Generally speaking, our sample of ECE teachers experienced challenges working with children both in person and online, but most of the challenges they experienced were different in type and intensity due to the unique circumstances of these different teaching formats. One clear distinction between these two modalities was the continued physical demands for in-person teachers and increased demands for new skills to teach virtually for online teachers. In addition to the increased physical job demands and constantly changing guidelines and regulations, in-person teachers had more concerns about financial restrictions due to low enrollment, the concomitant possibility of school closure, and fear of contracting COVID-19 at work. In-person teachers' financial concerns may be related to the fact that they are more likely to work in settings where income sources are not stable or are highly tuition-dependent (e.g., family childcare homes, private childcare centers). This echoes previous findings that, while the pandemic compromised ECE teachers' financial stability across all settings during the pandemic, it is possible that it disproportionately affected childcare teachers [8,51].

Conversely, as expected, online teachers had different concerns about technological issues, social isolation, as well as a lack of resources and support needed for online teaching. Early childhood teachers often use technology as a teaching tool to some extent (e.g., showing a video) [52], but the challenges in learning this new mode of teaching brought increased new skill demands as they faced challenges engaging children and families in an online format. The quantitative data also corroborate these findings in that in-person teachers perceived more physical demands while online teachers reported that their jobs required new skills and training (e.g., technology integration for online teaching). Due to these unique challenges and demands, the perceived needs of in-person versus online teachers were necessarily different.

These findings are consistent with previous studies on the challenges that teachers have experienced during the pandemic [11,17,28,53,54]. However, the current study adds to the extant literature by demonstrating that these challenges can greatly differ by teaching modality. In addition, we extended the findings of previous studies by investigating teachers' needs directly. For example, the current study found that in-person teachers and online teachers needed vastly different types of resources in order to perform their teaching tasks (e.g., more cleaning supplies and financial support for in-person teaching vs. better technology, resources, curriculum, and training for online teaching). To provide appropriate resources and support, it is important to first assess what the challenges are for teachers, centers, and schools who are having potentially very different pandemic working experiences. Resources and support will need to be tailored to better meet the needs of teachers working under these very different circumstances.

While approximately half of the participating teachers perceived negative changes in psychological well-being, only approximately 20 percent of teachers reported negative changes in their physical health. This indicates that the COVID-19 pandemic appeared to more acutely impact psychological rather than physical aspects of well-being for most teachers. This was noteworthy given that COVID-19 is a major health-related issue. What is clear from our data, however, is that serious concerns remain about the overall health and physical well-being of ECE teachers. Among those teachers who experienced negative changes during the pandemic, teachers across all three teaching modalities shared anxiety and fear of COVID-19 and social isolation/disconnection as the greatest challenges related to their well-being. Concerns about weight gain and lack of physical activity as well as chronic health conditions were prevalent and common among those who reported a change during the pandemic among the three groups—chronic health conditions that, in some cases, were shared by the majority of our sample, including obesity, high blood pressure, and asthma, and that place them at higher risk for contracting and becoming very ill from COVID-19 [25,55]. This is remarkable, particularly for those teachers who continue to teach in person. These findings support and add to the limited extant evidence on teachers' concerns about diminished physical well-being during the pandemic [22,25,56].

Teachers in all three groups also had other issues in common, such as feelings of social disconnection, depression, and sadness, financial concerns, and additional job demands, although there were slight differences across the three groups. For example, financial concerns, additional job demands, and lack of resources and support were more prevalent challenges for the in-person teaching group than the others. Loss of purpose was more unique to online teachers, and concern for students was more common for school closed teachers, which may be related to fact that they were not able to directly interact with the children—an aspect of the job that has traditionally been a strong attractor to the profession.

### *4.2. Differences in Well-Being by Teacher and Center Characteristics, Modality*

We found many common issues and challenges related to psychological and physical well-being across the three teaching groups from the qualitative analysis, but a more nuanced picture emerged from the quantitative analysis. Upon further examination of the effects of well-being by modality, we found them, to a degree, to be moderated by teacher race—in particular, for psychological and professional well-being. Our findings show that White teachers' psychological well-being was poorer on average and, across the measures of well-being, tended differ less by teaching modality. Black and Hispanic teachers, on the other hand, experienced large differences in psychological well-being indicators such as life satisfaction, stress, depressive symptoms, and secondary trauma depending on whether their centers were closed or they were teaching in person. When their centers were closed, Black and Hispanic teachers had better well-being, yet looked similar to White teachers when they were teaching in person. In contrast, White teachers experienced lower professional well-being outcomes (lower work commitment and higher intent to leave) when teaching in person versus closed, but Black in-person and virtual teachers experienced even larger negative effects on work commitment than White teachers. These findings corroborate and extend prior work [29,31,32] in that the COVID-19 crisis has had a disproportionate impact on vulnerable people, yet our findings reveal that this

impact was unique and complex in its effects on minoritized groups. Because few studies have examined race differences in well-being during the pandemic, there is little precedent for understanding these findings. What seems clear is that more concerted effort should be made to examine the differential effects of race on the pandemic ECE teaching experience.

### *4.3. Limitations*

This study has some limitations. First, although the data were collected from a large group of teachers from 46 states in the U.S., they may not be representative of the national sample as the sample was not randomly drawn. It was also a one-time, cross-sectional study. Teachers, early childhood programs, and schools experienced and continue to experience rapid changes in COVID-19 prevalence, regulations, and guidelines as the pandemic continues, and this means that teachers' experiences may have changed drastically since our study was conducted. The present study only captured a snapshot of the early phase of the pandemic. The field would benefit from rigorous, longitudinal studies of the impacts of the pandemic on early childhood well-being to examine the "real impact of the COVID-19 pandemic and real changes".

Second, we did not examine hybrid teaching as a separate modality; teachers in our study were asked to select one of the three categories. Because our data were collected in the early phase of the pandemic, the hybrid option was not as prevalent, but became more so in the later phases of the pandemic. Including this group would likely reveal yet another set of unique challenges, which would warrant further investigation. Third, we relied solely on teacher reports on their perceptions of work and well-being during the pandemic. Although it is valid to use self-reports in this case, adding an objective measure such as a direct assessment or a doctor's report on health conditions would provide a more accurate picture of teachers' experiences and well-being. Fourth, our intent in this study was to examine challenges in working conditions and well-being; however, in order to offer clearer implications for improving teacher well-being, future studies should include a larger set of center-level context and climate variables such as effective communication, leadership, and professional development support.

#### *4.4. Implications for Practice and Policy*

The COVID-19 pandemic has disrupted all sectors of the workforce, particularly those considered as the frontline workers. Our study provides ample evidence of the various challenges and high demands that early childhood teachers face that would have a negative impact on their well-being and work during the pandemic. This calls for additional resources and support to address the urgent needs. Stressed, overworked, and depressed teachers have little hope of meeting the needs of similarly stressed, traumatized, or otherwise needy young children. Support and resources to improve teachers' psychological and physical well-being could include a physical wellness program, mental health services, and increased time for breaks and physical activity.

Regarding concerns about a high rate of obesity, weight gain, and lack of physical activity and energy, the American College of Sports Medicine [57] recommends that healthy adults aged 18–65 years should engage in moderate-intensity physical activity that increases the heart and respiratory rates for a minimum of 30 min five days per week. Given the evidence from recent studies, it is imperative for early childhood teachers to not only encourage physical activity in children but engage in it themselves throughout the day. This will be extraordinarily difficult for ECE teachers, who are typically not afforded time in the day to even take a 15 min restroom break. While additional funding will not solve these challenges alone, it could allow for the hiring of additional staff to help cover classrooms so that teachers may have time for breaks, exercise, or other leisure activities.

What is clear is that these supports cannot be monolithic; they need to be tailored to the unique challenges, demands, and needs at multiple levels and across various settings (e.g., centers, family childcare homes, public schools, online settings) and sectors of the early childhood workforce. This also includes increased training and support to prepare teachers for different teaching modalities. In particular, support needs to prioritize teachers teaching in person, who face more job demands and health risks and likely come from family childcare homes and private childcare centers that do not have stable funding sources. Finally, we need to more closely investigate the differential impacts of pandemic working conditions on the well-being of different racial groups, as this study revealed some evidence that these differences were consequential.

More important than the realization of the incredible stress and strain we have placed upon the teachers of our young children is the need to acknowledge the sacrifices for the greater good that these teachers have given and provide these much-needed supports as a gesture of appreciation for these incredible sacrifices. In doing this, we also need to better recognize these frontline workers as essential—those workers who in many cases have risked, and continue to risk, their well-being and health to support children and families during these difficult times. This shift in perception and recognition, coupled with program, policy, and funding changes, can help to prioritize the needs of schools and ECE teachers, which will support their work and well-being.

#### **5. Conclusions**

While we see signs that the pandemic is waning—vaccinations are increasing and COVID-19 variants have receded for the moment—inviting a return to "normal", this does not absolve the field from ensuring that teachers are better prepared in the future to meet unique needs, whether pandemic-related or otherwise. For example, there has been an influx of online teaching in ECE due to the pandemic and these learning formats will likely persist long after the pandemic has subsided [7]. In our present circumstances, educators have learned how to adjust to weather this tumultuous time. It may not be a question as to if but when conditions worsen that they need to once again put these skills to action. For teachers working in person or online moving forward, we must continue to help them to address the physical and emotional demands of the profession. Even prior to the pandemic, we knew that teachers were suffering from a number of physical and health-related ailments due to stressful and strenuous working environments, poor wages and health benefits, and a lack of breaks [2,7]. Supporting teachers with plentiful and appropriate resources is an important pre-emptive step to ensuring that acute demands do not push our teachers and the profession over an edge from which there is little hope of recovery. The time is ripe to begin to seriously address the needs of this workforce to ensure that our educational systems are prepared for other challenges.

**Author Contributions:** Conceptualization, K.-A.K., T.G.F., J.T., K.R. and A.M.-L.; Data curation, K.-A.K. and T.G.F.; Formal analysis, K.-A.K., T.G.F., J.T., K.R., A.M.-L. and S.G.K.; Funding acquisition, K.-A.K.; Investigation, K.-A.K.; Methodology, K.-A.K., T.G.F., J.T. and K.R.; Project administration, K.-A.K.; Validation, T.G.F.; Writing—original draft, K.-A.K., T.G.F., J.T., K.R. and A.M.-L.; Writing—review and editing, K.-A.K., T.G.F., J.T., K.R. and A.M.-L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by Susan Sisson's operational funds at the University of Oklahoma Health Science Center.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of University of Oklahoma (IRB#:12045, date of approval, 5 November 2020).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy/confidentiality issues.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Chang-Ho Jihn 1,†, Bokyoung Kim 2,† and Kue Sook Kim 3,\***


**Abstract:** This study aimed to identify the factors that influence the components of burnout emotional exhaustion (EE), depersonalization (DP), and personal accomplishment (PA)—among hospital health workers, including doctors and nurses, during the COVID-19 pandemic. We analyzed 200 healthcare workers' responses to the Employee Health Promotion Survey conducted at a general hospital in Seoul with over 200 hospital beds. The questionnaire included items about COVID-19-related burnout and its influencing factors. We performed three different multiple regression analyses using EE, DP, and PA as the dependent variables. The results show that sex, marital status, workload of treating suspected COVID-19 patients, fear of COVID-19 infection, anxiety, and depression predicted EE. The predictors of DP were job category, consecutive months of work in the current department, satisfaction with work environment, anxiety, and depression. The predictors of PA were the workload of directly interacting with patients, socioeconomic status, and job stress. For EE and DP, burnout was found to be worse in doctors and nurses than in other health workers; moreover, burnout was worse among nurses than among doctors across all three aspects of burnout. The findings can be used to establish tailored policies to address each burnout component.

**Keywords:** burnout; COVID-19; hospital health worker; doctor; nurse; emotional exhaustion; depersonalization; personal accomplishment; Maslach burnout inventory

### **1. Introduction**

According to a report by the World Health Organization (WHO), as of 15 September 2021, there were approximately 230 million confirmed cases of COVID-19 worldwide. In South Korea, 277,989 confirmed cases and 2380 deaths have been reported [1]. The WHO defines health workers (HWs) as "all people engaged in actions whose primary intent is to enhance health" [2]. These HWs constitute the core workforce when encountering infectious diseases such as COVID-19. As a result of their role in managing and maintaining medical services at the front line during the spread of infectious diseases, HWs—such as doctors, nurses, midwives, paramedical staff, hospital administrators, support staff, and community workers—face a higher risk of infection than the general public. Furthermore, these workers are exposed to risks such as psychological distress, fatigue, and stigma [3]. According to a report from the International Council of Nurses (ICN), as of February 2021 [4], the average infection rate across the ICN dataset ranges between 6% and 10% at different points in time and HW infection rates of up to 30% have been reported. In South Korea, 565 health practitioners tested positive for COVID-19 while treating patients between February 2020 and June 2021. Of these, 20.0% were doctors and 73.5% were nurses, with the higher number of the latter likely due to the distinctive nature of nursing tasks in the field of disease prevention and patient care [5]. High infection and death rates among HWs can affect the maintenance of the healthcare system.

**Citation:** Jihn, C.-H.; Kim, B.; Kim, K.S. Predictors of Burnout in Hospital Health Workers during the COVID-19 Outbreak in South Korea. *Int. J. Environ. Res. Public Health* **2021**, *18*, 11720. https://doi.org/10.3390/ ijerph182111720

Academic Editors: Paul B. Tchounwou and Nygård Clas-Håkan

Received: 29 September 2021 Accepted: 3 November 2021 Published: 8 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Amid the prolonged COVID-19 pandemic, HWs are complaining about accumulated fatigue and mental stress, and are suffering from burnout due to constant labor shortages and insufficient benefits [6]. Burnout refers to "a psychological syndrome of emotional exhaustion (EE), depersonalization (DP), and reduced personal accomplishment (PA) that can occur among individuals who work with other people in some capacity" [7]. EE is the depletion of emotional resources, DP is developing a cynical attitude toward patients, and reduced PA is a negative evaluation of oneself [7]. The head of the Korean Health and Medical Workers' Union reported that HWs frequently quit their jobs because of extreme fatigue and exhaustion, and that many workers suffer from extreme physical and mental stress, which leads to depression and trauma. Moreover, HWs experiencing burnout can impact patients and colleagues as they have an increased risk of making wrong decisions [8]; burnout is therefore not an issue that only pertains to individual HWs. However, sufficient measures have not been taken to resolve the poor working conditions and heavy workloads for HWs [9].

Previous studies that analyzed burnout after the outbreak of COVID-19 limited their research subjects to single job categories, such as doctors [10–12] and nurses [13–15]. Such fragmentary analysis of burnout, with a focus on a particular job category, may be helpful for making policy decisions pertaining to a particular job category. However, these studies overlooked the actual clinical setting where multiple job categories are organically interconnected within one hospital and that burnout during a pandemic, such as COVID-19, affects the entire hospital. Although one study investigated the burnout of HWs from multiple hospitals [16], it was difficult for the researchers to build three individual models for the levels of EE, DP, and PA under the same conditions in one hospital and to identify relevant influencing factors.

For this reason, we aimed to differentiate our study from previous studies by including all HWs working at the same hospital. Departing from the previous pattern of analyzing job-oriented burnout, we employed an exhaustive analysis method focusing on the organization. This analytical approach can assist hospital personnel in charge of healthcare policy to gain a comprehensive understanding of the challenges confronting hospital HWs and to make efficient decisions. In addition, by implementing an employee burnout prevention policy that is applicable to all job categories, hospital HWs can perform their jobs more cost effectively. Thus, we aimed to provide a foundation and baseline data for establishing hospital-level policies on burnout prevention right at the early stages when encountering a pandemic such as COVID-19.

The overarching research question was: "What factors affect hospital HWs' burnout during the COVID-19 outbreak in terms of the individual burnout components?" The hypotheses were:

**Hypothesis 1 (H1):** *Sociodemographic characteristics and variables related to COVID-19, work overload, psychological conditions, and hospital resources will affect the EE of hospital HWs during the COVID-19 outbreak.*

**Hypothesis 2 (H2):** *Sociodemographic characteristics and variables related to COVID-19, work overload, psychological conditions, and hospital resources will affect the DP of hospital HWs during the COVID-19 outbreak.*

**Hypothesis 3 (H3):** *Sociodemographic characteristics and variables related to COVID-19, work overload, psychological conditions, and hospital resources will affect the PA of hospital HWs during the COVID-19 outbreak.*

#### **2. Materials and Methods**

#### *2.1. Study Design*

This retrospective descriptive study explored the factors affecting burnout among HWs at a Korean hospital during the COVID-19 pandemic.

#### *2.2. Participants and Sample Size Calculation*

This study's subjects included the participants of the 2020 Employee Health Promotion Survey for HWs working at a general hospital in Seoul equipped with over 200 hospital beds during the COVID-19 pandemic. Data were collected during the comprehensive wellness check from 6 January 2020, to 28 February 2020. The total number of employees in the hospital is 311, of which 210 responded to the questionnaire. The overall response rate was 64%. After excluding 10 people who did not submit their responses within the collection period, only 200 questionnaires were used as valid data in this study. The 200 responses that had already been collected were used for the final analysis. Using the G\*Power 3.1.9.7 program [17], the minimum sample size for multiple regression analysis was calculated to be 183 based on a previous study [18] with a significance level of 0.05, a median effect size of 0.15, a power of 0.90, and a number of predictors at 18. Therefore, a sample size of 200 for this study was appropriate.

The composition of the 200 participants is as follows: 48 doctors, 83 nurses, 6 pharmacists, 28 health workers, 7 managers, 3 technical workers, and 25 service workers. The response rates by occupational group were: 100% for doctors, 52% for nursing, 100% for pharmaceutical workers, 100% for health workers, 26% for managers, 38% for technical workers, and 72% for service workers.

#### *2.3. Instruments*

#### 2.3.1. General Characteristics

The questionnaire on general characteristics consisted of 13 items as follows: sex, age, education level, marital status, job category, working duration at the current job, working duration in the current department, number of rotating shifts within a month, workload of directly interacting with patients, workload of treating suspected COVID-19 patients, socioeconomic status, satisfaction with work environment, and current health condition. Age details were collected as an ordinal variable. Using Likert scales, we categorized and measured the following variables: education level, workload of directly interacting with patients, workload of treating suspected COVID-19 patients, socioeconomic status, satisfaction with work environment, and current health condition. Nominal variables with *c* classes were represented by *c* − 1 dummy variables, each taking on the values 0 and 1.

#### 2.3.2. Fear of COVID-19 Infection

Fear of COVID-19 infection was measured using a revised scale based on the fear of MERS-CoV infection [18] scale. This scale consisted of one item: "I am afraid of being infected with COVID-19", which was measured using a 10-point Likert scale, whose values ranged from 1 (not at all afraid) to 10 (unbearably afraid). A higher score indicated a stronger fear of COVID-19 infection.

#### 2.3.3. Job Satisfaction

Job satisfaction was measured using the Minnesota Satisfaction Questionnaire, which was developed by the Minnesota Industrial Relation Center [19] and translated into Korean by Lee and Park [20]. This questionnaire consisted of 20 items: 10 items about intrinsic factors and 10 about extrinsic factors. Using a 5-point Likert scale, responses were measured from 1 (very dissatisfied) to 5 (very satisfied). A higher score indicated a higher level of job satisfaction. Cronbach's α, the reliability indicator of the scale, was 0.88 in Lee and Park's [20] study, and 0.97 in this study.

#### 2.3.4. Hospital Anxiety and Depression

Hospital anxiety and depression levels were measured using the Hospital Anxiety and Depression Scale (HADS) developed by Zigmond and Snaith [21], and standardized by Oh et al. [22]. The HADS consists of 14 items: 7 items about anxiety and 7 items about depression. Responses were measured on a 4-point Likert scale ranging from 0 (never) to 3 (frequently). A higher score indicated a higher level of anxiety or depression. As 8 was suggested as the cut-off score in a previous study [22], we regarded scores above 8 as a manifestation of anxiety or depression. The Cronbach's α of the scale was 0.89 for anxiety and 0.86 for depression in Oh et al.'s study [22]. In this study, it was 0.687 for anxiety and 0.76 for depression.

#### 2.3.5. Job Stress

Job stress levels were measured using the Korean Occupational Stress Scale Short Form (KOSS-SF) [23]. The KOSS-SF consists of seven subscales and 24 items as follows: job demand (4 items), insufficient job control (4 items), interpersonal conflict (3 items), job insecurity (2 items), occupational system (4 items), lack of reward (3 items), and organizational climate (4 items). Using a 4-point Likert scale, the scores were reversecoded from 1 (strongly disagree) to 4 (strongly agree). The stress score of the KOSS-SF was calculated by converting the scores of the seven subscales into a 100-point scale and averaging them. The Cronbach's α for each subscale at the time of KOSS's development was as follows: 0.71 for job demand, 0.66 for insufficient job control, 0.67 for interpersonal conflict, 0.61 for job insecurity, 0.82 for occupational system, 0.76 for lack of reward, and 0.51 for organizational climate [23]. In this study, the overall reliability of the scale was 0.89, and that of each subscale was: 0.56 for job demand, 0.77 for insufficient job control, 0.77 for interpersonal conflict, 0.74 for job insecurity, 0.89 for occupational system, 0.81 for lack of reward, and 0.82 for organizational climate.

#### 2.3.6. Hospital Resources for the Treatment of COVID-19

Hospital resources for the treatment of COVID-19 were measured using a revised scale based on Kim and Choi's scale of "Hospital Resources for the Treatment of MERS-CoV" [18]. That is, we replaced the term MERS-CoV used in their scale with COVID-19 as follows: "My hospital is equipped with facilities sufficient for preventing the spread of COVID-19", "My hospital applies the best infection control guidelines for preventing the spread of COVID-19", and "My hospital discusses how to prevent COVID-19 regularly". Based on a briefing on the supply management plan for COVID-19 by the Central Disaster and Safety Countermeasures Headquarters [24], the following two items were added: "My hospital supplies facemasks steadily" and "My hospital steadily supplies personal protective equipment (gloves, bodysuit, goggles, hood, etc.)". The scale consisted of five items, which were scored on a 4-point Likert scale ranging from 1 (strongly disagree) to 4 (strongly agree). A higher number indicated a greater availability of hospital resources for encountering COVID-19. The Cronbach's α was 0.81 in a previous study [18], and 0.84 in this study.

### 2.3.7. Support from Family and Friends

Support from family and friends was measured using a revised scale based on Kim and Choi's scale of "Support from Family and Friends", which focused on the MERS-CoV epidemic [18]. Based on their scale, our scale was revised by replacing the term MERS-CoV with COVID-19 as follows: "My friends will avoid me if they find that I have cared for COVID-19 patients", "My friends will support me caring for COVID-19 patients", "My family will avoid me if they find that I have cared for COVID-19 patients", and "My family will support me caring for COVID-19 patients". The scale consisted of four items and was measured using a 4-point Likert scale ranging from 1 (strongly disagree) to 4 (strongly agree). A higher score indicated stronger support from family and friends. The Cronbach's α was 0.80 in a previous study [18], and 0.57 in this study.

#### 2.3.8. Working Overtime and Compensation Related to COVID-19

The scale measuring working overtime and compensation related to COVID-19 consisted of 2 items as follows: "My hospital requires me to work overtime because of COVID-19" and "My hospital pays extra for working overtime because of COVID-19". Using a 4-point Likert scale, the responses were measured from 1 point (strongly disagree) to

4 points (strongly agree), and additional pay was reverse-coded. A higher score indicated a higher frequency of working overtime without proper compensation.

#### 2.3.9. Burnout

Burnout was measured using the Korean version [25] of the Maslach Burnout Inventory Scale Human Services Survey (MBI-HSS) developed by Maslach and Jackson [26]. The MBI-HSS differs from MBI for medical personnel (MBI-HSS (MP)) [27] in terms of the choice of words in the items: "recipients" vs. "patients". Accordingly, we replaced the corresponding words with the Korean translations. Similarly, Jung's [25] Korean MBI-HSS version was used in a study of HWs at community health centers [28].

Both MBI-HSS and MBI-HSS (MP) consist of three independent subscales: EE, DP, and PA. Each questionnaire comprises 22 items: 9 items about EE, 5 items about DP, and 8 items about PA. The responses were measured using a 7-point Likert scale as follows: 0 points for "Never", 1 point for "Less than once a year", 2 points for "Less than once a month", 3 points for "2–3 times a month", 4 points for "Once a week", 5 points for "2–3 times a week", and 6 points for "Every day". High EE and DP with low PA scores indicated a higher level of burnout. The commonly used cut-off points of EE, DP, and PA are 27, 10, and 33, respectively [29].

Following the MBI advice that the sum of all subscale scores is not an ideal indicator of burnout [26,27], we calculated the scores per subscale and interpreted them separately. Cronbach's α, the test score reliability indicator for each subscale at the time of scale development by Maslach and Jackson [30], was 0.90 for EE, 0.79 for DP, and 0.71 for PA. The reliability of the subscales in this study was 0.92 for EE, 0.84 for DP, and 0.90 for PA.

#### *2.4. Ethical Considerations and Data Collection*

Before data collection, we obtained approval from the institutional review board (IRB) of the Seoul Medical Center regarding adherence to ethical guidelines and permission to view the responses (no. SEOUL 2021-01-002-003). The collected responses did not contain any information that could identify the participants. To avoid data leakage, we viewed the data only in the office of the hospital of the employee health promotion team. Moreover, the document file was encrypted with a password and saved on a computer that could only be accessed by the researcher. A total of 200 responses were retrieved and used in the analysis. The research data file will be disposed of three years after the research is complete.

#### *2.5. Data Analyses*

We used R and SPSS for Windows (version 26.0; IBM Corp., Armonk, NY, USA) to analyze the data. The data set included the following variables: general characteristics of the participants, burnout related to COVID-19, fear of COVID-19 infection, job satisfaction, anxiety, depression, job stress, available hospital resources, additional compensation for overtime work, and support from family and friends. The reliability of the scales was measured using Cronbach's α. Regarding the general characteristics of the participants, we calculated the frequency, percentage, mean, and standard deviation. To analyze the differences in burnout components (EE, DP, and PA) according to the general characteristics, we used the following tests: Student's *t*-test, Welch's *t*-test, analysis of variance, Kruskal– Wallis test, and Scheffe's post hoc test. Pearson's correlation test was used to analyze correlations. These test results were used to ensure that only informative variables were selected to avoid the curse of dimensionality. To explore the factors that affected burnout related to COVID-19, we performed three different multiple regression analyses, where the explanatory variables were the statistically significant variables from the difference tests and correlation analysis, and the dependent variables of burnout were the EE, DP, and PA scores.

### **3. Results**

#### *3.1. General Characteristics*

Most of the participants were women, college graduates, and health practitioners. In terms of satisfaction with the work environment and health condition, they perceived it to be above average. The number of female participants was approximately three times higher than that of male participants. Regarding age, the ratio of participants under 40 and above 40 years was almost the same. Moreover, the number of married people was 1.6 times higher than that of single people. As for job type, nurses accounted for 41.5% and doctors accounted for 24.0%. The other health workers (OHWs) accounted for 34.5% of the participants. Of all participants, 73.5% belonged to a job category that required interaction with patients. Participants who were likely to interact with suspected COVID-19 patients accounted for 31.0%. Of all participants, 72.0% perceived themselves as having a middle or high socioeconomic status. Regarding satisfaction with the work environment, the ratio of satisfied and dissatisfied was approximately 6:4 (Table 1).




**Table 1.** *Cont.*

#### *3.2. Burnout and Other Variables*

The mean EE, DP, and PA scores were 26.48, 11.03, and 28.47, respectively. These were worse than the means of a medicine group (*n* = 1104) provided for comparison in the fourth edition of the MBI manual [27], where the mean EE, DP, PA scores were 22.19, 7.12, and 36.53, respectively. The median fear of COVID-19 infection was 6 and its interquartile range (IQR) was 4. Regarding job satisfaction, the mean was 50.92, which was between "satisfied" and "slightly satisfied". Regarding anxiety and depression, the means were 7.49 and 8.87, respectively, almost or above the cut-off point (8 points). Regarding job stress, the mean was 49.09, which was close to the median. As for hospital resources for COVID-19 and support from others, the mean was close to the median and between "slightly disagree" and "slightly agree" (Table 1).

#### *3.3. Difference Testing, Correlation Analysis, and Multiple Regression Analysis*

Based on the mean difference test and correlation analysis of EE scores, the variables that were found to be statistically significant were as follows: sex, age, education level, marital status, job category, number of rotating shifts within a month, workload of directly interacting with patients, workload of treating suspected COVID-19 patients, satisfaction with work environment, current health condition, fear of COVID-19 infection, job satisfaction, anxiety, depression, job stress, and hospital resources for COVID-19. Based on the mean difference test and the correlation analysis of DP scores, the variables that were found to be statistically significant were as follows: sex, job category, working duration in the current department, workload of treating suspected COVID-19 patients, satisfaction with work environment, current health condition, fear of COVID-19 infection, job satisfaction, anxiety, depression, job stress, and hospital resources for COVID-19. Based on the mean difference test and the correlation analysis of PA scores, the variables that were found to be statistically significant were as follows: age, education level, doctor, workload of directly interacting with patients, socioeconomic status, job stress, and working overtime and

compensation. For all the mean difference tests and correlation analyses, the significance level of 0.05 was used.

We used the Durbin–Watson test to detect the presence of autocorrelations. The heteroscedasticity and autocorrelation consistent (HAC) covariance matrix estimation were used to overcome autocorrelations [31]. The variance inflation factor was used to test for multicollinearity. All variables had variance inflation factor (VIF) values between 1.118 and 2.465, indicating no multicollinearity.

Based on the multiple regression analysis of EE, the variables that were found to be statistically significant included sex, marital status, workload of treating suspected COVID-19 patients, fear of COVID-19 infection, anxiety, and depression. Significant variables for DP included nursing, working duration in the current department, and anxiety. When adjusting the significance level up to 0.10, satisfaction with the work environment and depression were included. Significant variables for PA included workload of directly interacting with patients, socioeconomic status, and job stress. The regression models of EE, DP, and PA explained 52.08%, 34.98%, and 14.94%, respectively, of the variance in their scores for COVID-19-related burnout (Table 2).


**Table 2.** Multiple linear regression analysis for burnout (*N* = 200).


**Table 2.** *Cont.*

Note: Ordinal variables: education level, workload of directly interacting with patients, workload of treating suspected COVID-19 patients, satisfaction with work environment, current health condition, fear of COVID-19 infection, hospital resources for COVID-19. Integers 0 to *n* − 1 are assigned to the values of ordinal variables, where *n* is the number of values. Reference groups for dummy variables were sex (male), nursing (no), and doctor (no).

#### *3.4. Hierarchical Regression Analysis*

We performed three sets of hierarchical regression analyses (HRA) to see whether adding four different groups of variables significantly improved the model's ability to predict the burnout. For each analysis, the predictor variables were entered within four successive steps. The order in which groups of variables were added is as follows: sociodemographic group, COVID-19 group, work overload group, and psychological group. The groups for EE, DP, and PA were constituted by statistically significant variables in Table 2 for each component type. Thus, the composition of each group varies depending on the component type. A sociodemographic group can be constituted of sex, age, education level, marital status, socioeconomic status, current health condition, nursing, or doctor. A COVID-19 group can be constituted of fear of COVID-19 infection, workload of treating suspected COVID-19 patients, or hospital resources for COVID-19. A work overload group

can be constituted of working duration in the current department, number of rotating shifts within a month, workload of directly interacting with patients, satisfaction with work environment, job satisfaction, or working overtime and compensation. A psychological group can be constituted by anxiety, depression, or job stress.

Regarding EE, in the first step, the sociodemographic group accounted for 28.0% of variance in burnout (*p* < 0.001). In the second step, the COVID-19 group explained an additional 9.9% of the variance (*p* < 0.001). In the third step, the work overload group explained an additional 4.8% of the variance (*p* < 0.05). In the fourth step, the psychological group explained an additional 13.4% of the variance (*p* < 0.001).

Regarding DP, the group of the first step accounted for 14.8% of variance in burnout (*p* < 0.001). The groups of the second, third, and fourth step, respectively, explained an additional 4.7% (*p* < 0.001), 6.4% (*p* < 0.05), and 12.9% (*p* < 0.001) of variance in burnout.

Regarding PA, the COVID-19 group was not considered for the HRA since there is no statistically significant variable of the COVID-19 group in Table 2. In the first step, the sociodemographic group accounted for 10.6% of variance in burnout (*p* < 0.001). The work overload group in the second step and psychological group in the third step, respectively, explained an additional 3.5% (*p* < 0.05) and 3.8% (*p* < 0.05) of the variance in burnout.

#### **4. Discussion**

This study is significant in that it investigated the burnout of all HWs working at the same hospital during the early stages of treating the COVID-19 pandemic, and analyzed the factors of burnout from multiple perspectives in terms of its three components (EE, DP, and PA). This section explains burnout based on the significant variables of each component and proposes concrete strategies to mitigate burnout.

#### *4.1. Situation of Hospital Health Workers in South Korea Compared to Other Countries*

According to a survey of possible burnout for 2707 healthcare professionals in 60 countries in April 2020, 51% of respondents reported burnout [32]. In addition, in a systematic review analyzing 11 studies of healthcare professionals' burnout conducted mainly in April–May 2020, the prevalence of overall burnout was 49.3% to 58% [33]. In Korea, as a result of a regular survey of about 67,000 health and medical workers conducted by the Korean Health and Medical Workers' Union (KHMWU) in February 2019, before the outbreak of COVID-19, 70.6% of respondents complained of physical and mental burnout. This is even higher compared to the reported burnout rate during COVID-19 in other countries. As a result of a regular survey conducted in March 2021 by KHMWU in about 43,000 health and medical workers, 69.6% of the respondents complained of physical exhaustion, and 65.8% of them were mentally exhausted. Of the total respondents, 78.7% answered that their daily life had deteriorated, and 70.6% of the respondents answered that their psychological state also deteriorated [34,35]. As such, Korean HWs report chronic burnout every year, and they endure daily physical exhaustion and emotional labor.

#### *4.2. Burnout Level of Korean Hospital Health Workers during the COVID-19 Outbreak*

The results show that the Korean participants in this study had much greater risks for burnout in all three aspects (EE, DP, and PA) than other countries. Regarding the means of each component during the COVID-19 outbreak, the mean of the participants' EE was 26.48 points out of 54 points, DP was 11.03 points out of 30 points, and PA was 28.47 points out of 48 points. By comparison, a study conducted among frontline healthcare professionals who worked during the peak of the COVID-19 pandemic in Italy showed the following: 22.7 points in EE, 6.1 points in DP, and 37.5 points in PA [36]. In another study carried out among health professionals in Italy [37], the means were 22.3 points in EE, 4.7 points in DP, and 33.7 points in PA. Compared with these findings, the participants in this study scored higher in EE and DP and lower in PA, which may have resulted from the exceptional surge in COVID-19 infections in Korea. To illustrate, after the first confirmed patient was found in Korea in January 2020, the number of patients grew rapidly across the

local community due to a large-scale group infection centering on a religious organization in late February [38,39]. This led to an increase in the number of screening sites and the establishment of strong measures against the spread of COVID-19. Therefore, the surge in patients may have influenced the burnout levels of HWs in this study.

In EE and DP, burnout was found to be worse for doctors and nurses than OHWs; it was worse for nurses than doctors in all three aspects of burnout. The burnout levels of nurses are worse than those from large-scale studies in other countries. To illustrate, a study conducted among 2014 Chinese nurses in February 2020 showed 23.44 in the mean score of EE, 6.77 in the mean score of DP, and 34.83 in the mean score of PA [13]. Likewise, another large-scale study carried out among 12,596 Chinese HWs in April 2020 showed 19.1 in the mean score of EE, 5.5 in the mean score of DP, and 29.0 in the mean score of PA [15]. In addition, a study carried out among HWs in Saudi Arabia [16] showed that nurses scored high in EE (24.70 points) and DP (8.37 points) compared to other job categories including doctors, who scored 22.85 points in EE and 7.36 points in DP. In PA, nurses scored low (34.22 points) compared to doctors (35.70 points).

#### *4.3. Factors Influencing Emotional Exhaustion (EE)*

The results of this study show that the factors affecting EE include sex, marital status, workload of treating suspected COVID-19 patients, fear of COVID-19 infection, anxiety, and depression. The level of EE is likely to increase under the following conditions: being female, single, having frequent contact with suspected COVID-19 patients, a strong fear of COVID-19 infection, and serious levels of anxiety and depression. In the hierarchical regression analysis, the psychological group demonstrated the greatest predictive power of EE. Similarly, the results of a study conducted among medical and administrative staff at a tertiary hospital in Italy during the COVID-19 pandemic [40] showed that the factors affecting the levels of EE, which was measured using the MBI-GS (General Survey), included sex, living condition, workplace, length of working experience, occupation, preexisting psychological problems, COVID-19-related traumatic events, increased conflict among colleagues, additional task assignment, increased workload, and interpersonal avoidance. According to the study, females and people living alone had a higher level of EE compared to males and people living with family/other relatives. Comparatively, people working in administration, non-COVID wards, and frontline services caring for patients with COVID-19 had lower levels of EE than those working in an intensive care unit (ICU), frequently interacting with COVID-19 patients, and having a higher probability of infection. The results of the present study are consistent with another study conducted in Italy [37], which showed that being female, being in contact with COVID-19 patients, and fear of COVID-19 infection predicted increases in EE. Furthermore, the results of the present study are consistent with the findings of family division and trait anxiety in a study conducted among physicians and nurses in northern Italy [40]. A study conducted in Saudi Arabia [16] also showed that direct involvement in the management of COVID-19 patents increased the levels of EE.

Among the three job categories, there was a statistically significant difference in the EE scores. The results of the post hoc test show that the EE score for nurses was the highest. It is necessary to identify factors that increase nurses' EE, such as anxiety, depression, fear of infection and death [13], and come up with measures to strengthen factors that decrease nurses' EE, such as self-efficacy and resilience [13].

#### *4.4. Factors Influencing Depersonalization (DP)*

The factors that affected DP as a sign of burnout included nursing, working duration in the current department, satisfaction with work environment, anxiety, and depression. DP was worse for nurses than for doctors and OHWs. DP increased when the period of continuous service in the department was longer and when the satisfaction with the work environment was higher. Furthermore, DP levels increased as the levels of anxiety and depression increased. In the hierarchical regression analysis, the psychological group

demonstrated the greatest predictive power of DP. A study conducted in Saudi Arabia [16] also found an association between DP and satisfaction with work environment and showed that DP levels are higher when working more than 8 h during the COVID-19 pandemic, when performing on-call duties, and when job duties are changed. In a study among health professionals in Italy [37], work hours were found to be one of the predictors of DP during the COVID-19 pandemic. In a study conducted in northwest Italy [41], DP during COVID-19 increased as the level of anxiety increased. The findings of this study are consistent with meta-analysis research [42], which reported that "greater experience through years worked" is one of the factors that decreases HWs' risk of adverse psychological outcomes during virus outbreaks.

#### *4.5. Factors Influencing Personal Accomplishment (PA)*

The factors that affected PA included the workload of directly interacting with patients, socioeconomic status, and job stress. The levels of PA increased when interacting more frequently with patients and when socioeconomic status was higher. Meanwhile, PA decreased when job stress was high. In the hierarchical regression analysis, the psychological group demonstrated the greatest predictive power of PA. The results of this study are similar to the findings of a study conducted in Saudi Arabia [16], which suggested that PA levels during the COVID-19 pandemic are positively associated with direct involvement with the care of COVID-19 patients. The findings of this study are also supported by research conducted among emergency workers in Italy [43], which reported that emotional stress and cognitive stress decreased PA. According to a meta-analysis study [42], lower household income was included as one of the risk factors that increased HWs' adverse psychological outcomes during virus outbreaks.

As it is ambiguous whether reduced PA is a result of or a manifestation of burnout [44], it is necessary to first consider the strategies that can be applied to both EE and DP. In this study, anxiety and depression were found to be predictors of both EE and DP. Therefore, first, it is necessary to discover the factors with the greatest impact on both anxiety and depression and, second, to establish and implement practical measures to relieve them. From an administrative standpoint, it is important to change the work environment so that hospital staff can feel that they are managing well autonomously without being overwhelmed by their jobs. This also helps them improve their work competency, job satisfaction, and mental health [45,46].

#### *4.6. Comprehensive Strategies Considering the Three Components of Burnout*

In addition to mitigating anxiety and depression, which are required for improving the levels of EE and DP, job stress is one of the factors that affects PA. Therefore, measures to lower psychological stress, simultaneously with anxiety and depression, should be considered. As preventive interventions, resilience, and social support were reported as mediator variables for psychological problems among HWs during the COVID-19 pandemic [47], it is essential to establish a policy that reflects these variables. To this end, it is imperative for individual employees and hospitals to collaborate and establish measures that can analyze and regulate the factors that cause anxiety, depression, and job stress during a pandemic by accurately diagnosing mental health conditions through counseling with professionals and operating a tailored counseling program.

It is also imperative to explore strategies to strengthen the resilience of hospital HWs to mitigate EE and DP, as well as to improve PA. Resilience can be defined as maintaining or recovering mental health during significant adversity, such as a potentially traumatic event, challenging life situation, major life change, or physical illness [48]. Furthermore, hardiness is a trait of resilience [49]. During the COVID-19 pandemic, resilience works as a protective factor and elicits a positive outcome by reducing burnout and stress [43] while improving performance, productivity, and satisfaction [50]. Therefore, by proactively introducing educational programs to improve and strengthen HWs' hardiness and by offering these programs regularly, it will be possible for hospital staff to proactively counter and regulate

their stress, stemming from work-related burden and interpersonal relationships, during a traumatic event such as COVID-19.

#### *4.7. Response Policies at the National and Hospital Level*

Currently, Korea is monitoring the mental health of frontline pandemic responders at the national level. The South Korean government launched the National Center for Disaster Trauma (NCDT) and began to operate the "COVID-19 Integrated Psychological Support Team" within a month after the COVID-19 pandemic began in January 2020. As the number of complaints about burnout related to COVID-19 increased among HWs and government officials, the NCDT and Korean Society of Traumatic Stress Studies published guidelines on psychosocial care for infectious disease management [51] and distributed them in March 2020. In addition to the guidelines, the Ministry of Health and Welfare and NCDT announced plans in May 2020 to provide burnout prevention education, counseling, and burnout management programs based on tailored consultation with individuals and organizations to prevent and monitor the job stress and burnout of COVID-19 responders [52].

In this study, satisfaction with the work environment was found to be a predictor of the levels of DP. Similarly, a safe and healthy working condition was found to mitigate DP in a study conducted among nurses in Malaysia [53], and a large-scale study conducted among nurses in China during COVID-19 also showed a significant statistical difference in DP depending on the level of work safety [13]. Moreover, a systematic review listed many attributes of the work environment that can mitigate DP as follows: pleasantness of tasks, value/meaning of work, emotional reward, and making a difference [54]. Thus, organizational management can mitigate the levels of DP by focusing on the following aspects: work hours, work environment, work patterns, work speed, work autonomy, education and training, communication within the organization, violence and discrimination, accidents and diseases, and work incentives [55].

According to this study's findings, the workload of direct patient care was found to be a conflicting factor in improving EE and increasing PA. A high frequency of direct encounters with suspected COVID-19 patients increased EE, while a heavier workload of direct patient service improved PA. Thus, by expanding minimum contact systems and equipment, the burden of direct contact with suspected COVID-19 patients can be reduced. However, PA decreased because of a decrease in the workload of the direct patient service. Therefore, to maintain proper levels of PA, it is important to imitate and take advantage of direct patient services by providing contactless channels for HWs to sufficiently communicate with patients. For example, South Korea creatively developed and operated drive-through screening centers and walk-through screening centers [56,57]. The walk-through system is an efficient system that uses a booth to minimize the physical contact between the HWs and patients, protecting the medical staff with minimal protective equipment and shortening the testing time. Such a minimum contact system can help prevent passive treatment behavior due to a decrease in contact because of fear of infection. Nevertheless, it cannot be denied that reduced contact between HWs and patients can pose a danger to patients. Therefore, the degree of non-contact should be determined in line with the patient's safety.

Burnout is a critical problem that adversely effects not only HWs, but also patients and the overall healthcare environment. When employee burnout is not properly monitored, it leads to intention to leave, reduced job performance, missed care, general health problems, mental health problems, and reduced job satisfaction; from the patients' perspective, it can lead to poor quality of care, poor patient safety, adverse events, negative patient experience, medication errors, infections, and patient falls [58]. Therefore, preemptive burnout management is essential for hospital employees as well as for the effective treatment, care, and safety of patients. It is necessary to provide the opportunity for COVID-19 responders to debrief on their psychological experience and manage their mental health from the early

stage of an outbreak of an infectious disease, since previous studies have shown that HWs display psychological trauma caused by epidemic outbreaks [59–61].

As such, the central and local governments' countermeasures against the COVID-19 pandemic, which have continued from its early stages, are timely policies for reducing the burnout of COVID-19 responders and further preventing post-traumatic stress disorder (PTSD). These findings are in line with the WHO guidelines that highlight mental health and psychosocial support for HWs during the COVID-19 pandemic [62], as well as with the U.S. CDC guidelines [63] on how to cope with stress and build resilience for healthcare personnel and first responders.

#### *4.8. Limitations and Suggestions for Follow-Up Studies*

This study had the following limitations. First, this study limited its scope to the situation at one medical institution in one country during the COVID-19 outbreak, so caution should be exercised in generalizing the findings. Although this study's findings cannot be applied to hospitals with different sizes and conditions, the research method used in this study can help identify the influencing factors of burnout components at different hospitals. Second, this study excluded some factors that can influence the burnout of hospital employees during a pandemic such as COVID-19. For example, other variables such as job demand, job control, value congruence, role conflict, decision latitude [58], or the presence of psychological comorbidities [33] should be considered in a follow-up study.

Based on the study findings, we suggest a stepwise approach that identifies the predictors of burnout components (EE, DP, and PA), selects the most vulnerable job category based on the identified predictors, and then manages the target job category, preferentially improving predictors that can exert a favorable influence on other job categories. We also suggest that follow-up studies should identify biomarkers and somatization in the workforce responding to infectious diseases by referring to the psychosomatic symptoms of burnout [64], biomarkers such as salivary cortisol, or biochemical parameters such as HbA1C [65].

#### **5. Conclusions**

During a pandemic of a novel infectious disease such as COVID-19, the government and hospital healthcare policy managers should consider the potential for burnout in HWs who first encounter patients and provide treatment. The results of this study show that, in the early stages of the response to COVID-19, the burnout (EE, DP, PA) levels of doctors and nurses at a general hospital were worse than that of other hospital HWs in EE and DP; the burnout levels of nurses were worse than those of doctors in all three aspects. The factors that affected EE related to COVID-19 were sex, marital status, fear of COVID-19 infection, anxiety, and depression; DP was affected by nursing, working duration in the current department, anxiety, and depression; and PA was affected by workload of directly interacting with patients, socioeconomic status, and job stress.

This study is significant for several reasons. First, this study referred to results from an exhaustive survey conducted among employees at one medical institution, who experienced the early stage of COVID-19; therefore, it was possible to identify and explain burnout mechanisms based on the characteristics of diverse job categories and work environments. Second, this study performed three different multiple regression analyses using EE, DP, and PA as the dependent variables, and identified significant factors for each component to enable the establishment of tailored policies according to the burnout component. The multi-perspective approach of this study can help establish macroscopic and comprehensive countermeasures at the institution level.

**Author Contributions:** Conceptualization, C.-H.J., B.K. and K.S.K.; methodology, C.-H.J., B.K. and K.S.K.; software, C.-H.J.; validation, C.-H.J. and B.K.; formal analysis, C.-H.J.; investigation, K.S.K.; resources, K.S.K. and C.-H.J.; data curation, K.S.K.; writing—original draft preparation, B.K., C.-H.J. and K.S.K.; writing—review and editing, C.-H.J. and B.K.; supervision, K.S.K.; project administration, C.-H.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was approved by the institutional review board of Seoul Medical Center (protocol code: no. SEOUL 2021-01-002-003, date of approval: 16 February 2021).

**Informed Consent Statement:** The institutional review board waived the requirement for written informed consent from subjects since we used the collected responses from the "2020 Employee Health Promotion Survey". The institution also excluded information that could identify survey respondents. To avoid data leakage, all responses were viewed only at the office of the employee health promotion team. Therefore, written informed consent was waived.

**Data Availability Statement:** The datasets generated and analyzed during the current study are not publicly available due to personally sensitive records but are available from the corresponding author upon reasonable request.

**Acknowledgments:** We would like to thank Jae-Hong Lee for R programming. We are also grateful to the reviewers for offering valuable advice that substantially improved the quality and readability of the publication.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

