Setting

This study is part of a larger cross-sectional study of HWs’ psychosocial functioning across the US. During May 2020, we distributed an online, cross-sectional survey to HWs at 25 medical centers across the US. During this period, the total confirmed cases of COVID-19 peaked in the US, exceeding 1.74 million cases and 100,000 deaths [10]. Hospitals were chosen based on their geographic location to include HWs in the Northeast, Midwest, South, and West regions of the US. The Yale Institutional Review Board approved the study, and participants provided informed consent. The 32-item checklist from the Consolidated Criteria for Reporting Qualitative Research was employed to direct study reporting [11].

Eligibility and recruitment

We employed convenience sampling whereby hospital department chairs and division chiefs were contacted at each hospital to forward our recruitment email and online survey to their staff. The recruitment email stated that our purpose was to understand HWs’ experiences within and outside the hospital during the COVID-19 pandemic. Individuals were eligible if they were at least 18 years of age and were HWs. HWs were defined as those who work in a hospital or clinic, including but not limited to attending physicians, medical trainees (i.e. residents and fellows), physician assistants, nurses, medical assistants, nurse assistants, research coordinators, social workers, administrators, lab technicians, health technicians, psychologists, and respiratory, occupational, and physical therapists.

Data collection

Demographic data collected included: age, gender, race/ethnicity, marital status, occupation, and geographic location. One question was used to ascertain HWs’ front line status, i.e. if they worked directly in COVID-19 rooms.

The survey included previously validated scales for mental health outcomes. These patient-reported outcome measures were included to give context of the psychological burden of the pandemic in the sample. The Patient Health Questionnaire-9 (PHQ-9) [12, 13], Generalized Anxiety Disorder-7 (GAD-7) [14], Primary Care-PTSD (PC-PTSD) [15, 16], and Alcohol Use Disorders Identification Test-Concise (AUDIT-C) [17] were used to assess MD, GAD, PTSD, and AUD, respectively. The cutoff values for probable clinical depression, GAD, and PTSD are PHQ-9> = 10, GAD-7> = 10, and PC-PTSD> = 3, respectively. For AUDIT-C, a score of 4 or more is considered probable AUD for men and a score of 3 or more is considered probable AUD for women. Patient-reported outcome measures and risk factors will be further explored in a separate study.

To assess HWs’ unique experiences during the COVID-19 pandemic, we included two open-ended questions asking respondents to recount their most upsetting and most hopeful experiences during the COVID-19 pandemic and how it made them feel. We chose these two questions based on the literature that has demonstrated substantial variability in psychological responses to trauma, including severe and persistent trauma-related symptoms and experiences of psychological growth [18, 19]. A meta-analysis of 63 studies including 26,951 patients found that trauma survivors sometimes experience both post-trauma symptoms and psychological growth simultaneously [20]. We therefore asked two-open ended questions–one focusing on negative experiences and another focusing on positive experiences–to understand the range of factors that account for such variability in psychological outcomes within and between participants. Mixed methods studies have used similar approaches, including assessing positive and negative experiences among survivors of Hurricane Katrina [21, 22].

Our survey was piloted in two iterations with one attending physician and one resident physician to receive their feedback regarding appropriateness of questions, interpretation of the open-ended questions, and survey length. Their responses were excluded from our analysis.

We defined data saturation as the point at which new themes were no longer emerging in subsequent responses [23]. Data saturation was reached after 50 responses. However, to better understand the frequencies of the positive and negative experiences among the sample, we continued to code all available responses.

Analysis

Responses to both open-ended questions were coded in two rounds by one researcher (RH, MD-PhD candidate, six years of qualitative analysis experience, female), and the code tree was reviewed and validated by another researcher (SL, MA, PhD, Assistant Professor and Clinical Psychologist, Trauma and Mental Health content expert, over decade of qualitative analysis experience, female). Both researchers reviewed and discussed sample quotes during the coding process to ensure that code definitions were consistent and appropriately applied to participant responses. Using an inductive approach, codes were first developed that represented themes that were present across responses [24].

Once the code tree was finalized after two rounds of inductive analysis, codes were grouped using an abductive approach [25] based on the Social Ecological Model and analyzed using thematic content analysis [26]. The Social Ecological Model holds that an individual’s mental wellbeing and behavior is influenced by interactions of individual, interpersonal, organization, community, and public policy levels [27–29]. The individual level includes factors related to their attitudes, beliefs, knowledge, and behavior. Factors within the interpersonal level include families, friends, and social networks. The organization level includes factors related to their workplace and businesses, which largely focused on the hospital or health system in which the respondents work for our study. Community level factors relate to the culture and communications within the population. Lastly, public policy level factors include interactions with the national, state, or local government and policies.

To give an example of our analysis process, inductive analysis was employed that identified various codes, including ‘childcare changes,’ ‘homeschooling,’ ‘household finances and job security,’ and ‘familial bonds.’ Using thematic content analysis, we grouped these codes within a single theme called ‘Family Functioning.’ Using an abductive approach and the Social Ecological Model, we mapped the ‘Family Functioning’ theme to the interpersonal ecological level.

Qualitative responses were coded using Microsoft Excel to facilitate integration of responses and codes with the quantitative dataset. Quantitative analysis was performed using SPSS statistical software (IBM Corp). We computed mean scores for each mental health indicator in the overall sample. We also determined the prevalence of probable MD, GAD, PTSD, and AUD in the sample using the pre-defined cutoff values. Assessing risk factors for mental health conditions using multivariable logistic regression was beyond the scope of the current study; however, we aim to conduct those analyses in a separate paper.

Sample

A total of 1,132 participants completed the survey. Descriptive information on participants is provided in Table 1. Most respondents were female (71.4%), were married (64.4%), and identified as White (74.6%). Of the 1,132 respondents, 31.3% were physicians, 17.6% were medical trainees, 18.9% were nurses, 5.9% were physician, nursing, or medical assistants, 8.1% were health technologists or technicians, 8.2% had another clinical job, and 9.7% had a non-clinical job. About half of the respondents participated in direct care with patients with COVID-19 (51.9%). Most respondents were from the Northeast region of the US (53.7%), followed by Midwest (23.7%), West (12.7%), and South (9.5%). The average scores for depression, GAD, PTSD, and alcohol use disorder were 4.8 (95% CI: 4.5–5.1), 4.9 (95% CI: 4.6–5.2), 1.4 (95% CI: 1.4–1.5), and 2.5 (95% CI: 2.4–2.6), respectively. Based on the cutoff values for the mental health scales, 14.0% of the sample had scores indicative of probable MD, 15.8% probable GAD, 23.1% probable PTSD, and 42.6% probable AUD.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Participant characteristics.

https://doi.org/10.1371/journal.pone.0240646.t001

Social ecological model

Thematic content analysis revealed individual-, interpersonal-, organization-, community, and public policy-level themes that impact HWs’ negative and positive experiences in and out of the hospital. Fig 1 organizes these themes within a Social Ecological Model. Table 2 depicts these themes, subtheme factors, and representative quotes organized by social ecological level.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Social ecological model for health workers’ wellbeing during the COVID-19 pandemic.

https://doi.org/10.1371/journal.pone.0240646.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Social ecological model themes, factors, and associated quotes for the most upsetting and hopeful experiences reported by health workers during the COVID-19 pandemic.

https://doi.org/10.1371/journal.pone.0240646.t002

Within the individual level, the respondent’s health status and ability to practice self-care behaviors were themes present in both the most upsetting and most hopeful answers. In the interpersonal level, respondents reported that social support, the health of their social circle, and family functioning were themes that related to both their positive and negative experiences during the pandemic. In the organization level, themes identified that relate to the negative and positive experiences of the HWs include hospital resources, routine, management, team, and patient care experiences. In the community level, availability of scientific knowledge about COVID-19, compliance with social distancing, community morale, support of HWs, and the media were themes related to the positive and negative experiences of the respondents. Lastly, in the public policy level, the US government and health system leadership and shelter-in-place policy were themes that contributed to the positive and negative experiences of the HWs.

Individual level themes

Among the sample, 40 (3.5%) and 124 (11.0%) respondents reported that their most upsetting and hopeful experiences, respectively, occurred in the individual level. Individual level themes pertained to the HWs’ health status and ability to practice self-care behaviors. Within the personal health theme, factors included COVID-19 infection and accessing non-COVID-19 medical care. Regarding COVID-19 infection, many respondents expressed that getting infected with COVID-19 was their worst experience during the pandemic. Other HWs reported that when they tested negative for COVID-19, they were hopeful and were put “at ease.” The respondent’s health status and ability to pursue medical care for medical conditions not related to COVID-19 also contributed to their most upsetting and hopeful experiences. For example, difficulties accessing obstetric, surgical, and physical rehabilitation care were among the most upsetting experiences for some HWs. On the other hand, many respondents suggested that beginning psychotherapy was their most hopeful experience during the pandemic, as it improved their mental health.

Another theme in the individual level was HWs’ ability to pursue self-care behaviors, which included two factors: time to process stress and self-care activities. Some respondents suggested that their work hours had increased without additional days off, which prevented them from engaging in self-care behaviors and process the added stress of the pandemic. For example, one physician recounted when she “supervised residents who had to pronounce death five times in 3 hours,” without the ability to fully cope with the experience “because no one will help us find time in our lives to process it.” Regarding hopeful experiences, others recounted that having more time off from clinical duties enabled them to have more time for spiritual practices, hobbies, exercise, healthy cooking, and self-reflection.

Interpersonal level themes

Next, 437 (38.6%) and 352 (31.1%) respondents suggested that their most upsetting and hopeful experiences, respectively, occurred within the interpersonal level. For example, respondents expressed that the health of their immediate social circle, including family members and close friends, contributed to their most upsetting and hopeful experiences. Within the health of social circle theme, factors included: fear of infecting family with COVID-19, COVID-19 infection or death, compliance with COVID-19 safety precautions, and medical care unrelated to COVID-19. Many respondents suggested that they feared that their household members would get COVID-19 from them given their exposure at the hospital. Respondents also wrote that their most upsetting experiences were related to their family members thinking that they had COVID-19, testing positive for COVID-19, or dying from COVID-19. Many respondents were upset that their immediate social circle was not taking COVID-19 safety precautions seriously, i.e. shelter-in-place and wearing masks, or believed that COVID-19 was a “hoax.” One medical trainee expressed that it is “infuriating” to her when her family “goes about their daily lives without masks or consideration for those who they might kill, or that they might get sick themselves.” However, some respondents said that they felt hopeful when their family members were compliant with COVID-19 safety precautions. Other HWs suggested that the health of their social group unrelated to COVID-19 caused them significant distress. Many respondents were upset that they could not visit their family members who had a serious medical condition out of fear of infecting them. Others recounted hope for their social circle’s health, such as a healthy pregnancy.

The second theme within the interpersonal level related to social support, which included the factors of physical isolation, receiving social support during isolation, and thankfulness. Many HWs reported feeling upset about physical isolation during the pandemic and made them feel “useless” when wanting to help their family. Furthermore, many recounted that their ability to receive social support was altered during the pandemic. For example, many expressed that decreased interactions with family and friends due to social distance restrictions made it difficult for them to receive support during this high stress time. However, others reported that they received more virtual and social-distanced support, which took the form of letters, care packages, and video calls from their friends, family, and neighbors. Respondents also felt hope when their social circle thanked them for their work as a HW, which made one health technician feel “immense love.”

Third, family functioning was a theme within the interpersonal level and included the factors: childcare, family stress, bonding, and finances. Many HWs recounted hardships caused by the closure of childcare and schools during shelter-in-place. One nurse reported that she felt “overworked between career job and motherhood.” HWs who are parents also suggested that there was added stress within their families as their children were processing the pandemic. Respondents who had to isolate themselves from family members expressed that this negatively impacted their family bonding. One medical trainee recounted that having to isolate from her 2-year-old son made her fear that “bonds we’ve worked so hard to foster are being broken down by this time apart.” However, many respondents who were able to shelter-in-place with their family suggested that increased time together strengthened their bonds. Many suggested that they were grateful that their children’s activities were canceled so that their family could slow down and spend more quality time together. Some HWs suggested that their families were facing financial hardships due to job loss; however, others suggested that their families were able to save money during this time and focus on non-materialistic values.

Organization level themes

Within the organization level, 457 (40.4%) and 445 (39.3%) participants reported upsetting and hopeful experiences, respectively. Five organization level themes were identified in the data, including hospital resources, routine, patient care experiences, team, and management. First, many respondents reported that the most upsetting aspect of the pandemic has been the lack of hospital resources, which includes the subtheme factors of personal protective equipment (PPE), testing, and human resources. Respondents suggested that the lack of essential resources made them feel unsafe, dehumanized, and “as if they didn’t care if I contracted COVID.” On the other hand, some respondents suggested that increased testing contributed to their hope.

Many respondents reported that there was a change in their hospital routine due to the pandemic, including changes in work hours, schedule, duties, wearing PPE, and work settings. For example, increased workload and hours were reported as being the most upsetting aspect of the pandemic, while decreased workload and hours were reported as being the most hopeful. Many respondents also reported that they felt distressed when their duties shifted to provide care to units outside of their specialty with minimal training and resources, which made one physician assistant feel “overwhelmed and unsure of myself.” Notably, some respondents suggested that they enjoyed their new duties at work. Some HWs reported challenges with wearing masks as part of their new routine, including difficulties breathing and providing patient care. Lastly, many HWs reported that they were grateful for being able to conduct telehealth in their home as part of their new routine. Some HWs, however, reported that there were difficulties in having meaningful patient interactions over the phone. Still, the majority suggested that telehealth was the most hopeful aspect of the pandemic. One social worker reported that being able to work via telehealth made her “have no fear of becoming infected and/or getting sick.”

The third theme within the organization level related to patient care and included the following factors: COVID-19 patient outcomes, isolated patients, non-COVID-19 medical care, and patient and family appreciation. Many HWs recounted “heartbreaking” and unprecedented experiences witnessing COVID-19 patients declining and dying. One emergency medical technician recounted a story of picking up a COVID-19 patient that was unlike any case she had seen in 30 years. “As we pulled out of the patient’s driveway, we left his entire family sitting in front of their house grieving, scared, wanting to believe there was a possibility we could save their father, knowing there was a good chance that that was the last time they would see him alive.”

On the other hand, HWs recounted specific stories of when their patients with COVID-19 improved, which gave them hope that their efforts were working. Due to hospital policies prohibiting visitors, many HWs reported that witnessing “patients die alone is traumatic to watch.” However, some HWs felt hopeful when they were able to connect with patients and their families virtually. Other HWs also reported ways in which patients without COVID-19 were negatively impacted in the hospital by the pandemic. For example, HWs reported that patients who had other diseases delayed coming to the hospital due to fear of COVID-19, which resulted in poor health outcomes. Lastly, HWs were hopeful when patients and their families expressed their appreciation for their work. One medical trainee wrote that family members’ appreciation “make[s] it worth it, and bring[s] me some hope and motivation to keep going.”

The fourth theme within the organization level related to hospital teams, including the subtheme factors of colleague diagnosed with COVID-19, teamwork, emotional support, inspiration, and supervisor relationships. First, HWs reported distress when they saw their colleagues getting treated for COVID-19 in the hospital and reported feeling “sad, angry, and scared.” Many HWs also reported upsetting experiences that related to teamwork. Most commonly, this conflict was concerning the HWs’ perception that their colleague did not want to work with COVID-19 patients out of fear of getting exposed, which made them feel “angry.” Despite these issues, many HWs reported that their most hopeful experience pertained to teamwork and having “shared experiences, a common enemy, and coming together to overcome adversity.” Many HWs also reported that they received more emotional support from their colleagues, which made them feel “closer.” HWs also shared that the inspirational, selfless work of their colleagues made them feel hopeful and proud. Lastly, some HWs suggested that getting positive feedback and appreciation from supervisors made them feel hopeful, while other HWs suggested that feeling like supervisors are not “open or sensitive to feedback” was upsetting.

Lastly, hospital management was a theme within the organization level, which included the factors of PPE policies, leadership, job security, modified schedules, and hazard pay. For example, many HWs reported that unclear, non-transparent, and transient hospital policies regarding PPE made them feel unsafe at work. One nurse even reported that his hospital’s insufficient PPE policy caused him to “never feel safe at [his] own job” and ultimately caused him to contract COVID-19. As a result, he “felt taken advantage of, disposable, and like [he] didn’t matter.” However, some respondents felt hopeful when hospital policies were changed to protect HWs. Many respondents also suggested that their hospital leadership upset them, as they did not provide open and clear communication. Other HWs that stated that their hospital did have effective leadership highlighted the importance of disseminating updates and important information to stay safe while at work. Many respondents also commented on the fear of job insecurity, which made them feel “disposable.” On the other hand, receiving hazard pay and modified schedules to decrease work hours made them feel “acknowledged.”

Community level themes

Next, 94 (8.3%) and 175 (15.5%) HWs reported community level factors that influenced their negative and positive experiences, respectively. Five themes were present within the community level: community behavior, media, community morale, community support of HWs, and scientific knowledge of treating and preventing COVID-19.

First, many HWs reported that the community behavior during the shelter-in-place mandate contributed to their positive and negative experiences. HWs reported two factors within the community behavior theme: protesting the shelter-in-place order and compliance with COVID-19 safety precautions. For example, many HWs reported that the protests against the shelter-in-place order made them upset. Several HWs stated that the lack of social distancing and mask-wearing among community members made them “feel like our sacrifices in the frontlines (personally and to my family) were unappreciated and dismissive of patients suffering/ losses.” On the other hand, some HWs suggested that they felt “proud” that the majority of community members were compliant with the shelter-in-place order.

The second theme within the community level related to the media, including subtheme factors of misinformation and positive news. Many HWs expressed their frustration with the misinformation perpetuated by the media. HWs suggested that dealing with this “infodemic” amidst a pandemic had been “exhausting.” Some HWs reported that they felt hopeful when the media highlighted positive news stories, however.

Next, community morale was a theme at the community level and included the subtheme factors of discrimination against people of Asian descent, fear, and community altruism. For example, some HWs suggested that the “increased/blatant racism towards healthcare workers of Asian descent” within and outside of the hospital made them feel upset. Others described that the general fear in the community contributed significantly to their negative experiences. On the other hand, many HWs reported that the altruism within the community gave them hope and was “really touching.” For example, they felt hope when they heard stories of community members volunteering to deliver groceries to those in need and supporting local businesses.

Community support of HWs was another theme at the community level and included the factors of conspiracy theories, stigmatization of HWs, recognition of HWs, and donations to HWs. First, the conspiracy theories that HWs and hospitals were profiting off of the pandemic were the most upsetting aspects of the pandemic for several respondents and made them feel “so unappreciated” for their work. Others reported feeling stigmatized while in the community for being a HW, making them feel like the community is “afraid” and “disgusted” with them. However, many HWs reported that they felt support from the community through increased recognition, such as community members thanking them or doing nightly “healthcare hero” celebrations. HWs reported that donations of food and PPE to the hospitals made them “feel like we are all a team,” “like we are not alone and being seen,” and “appreciated.”

The last theme within the community level related to the scientific knowledge of treating and preventing COVID-19, and included factors of uncertainty, hope for a cure, and hope for a vaccine. HWs reported that the uncertainty of COVID-19 prevention, treatment, and how long the pandemic would last was upsetting. However, participants reported that the prospect of increased scientific knowledge to find a cure and vaccine was providing them hope.

Public policy level themes

Overall, 56 HWs (4.9%) reported upsetting experiences and 24 (2.1%) reported hopeful experiences in the public policy level. Two public policy-level themes were present in the data: the US government and health system leadership and the shelter-in-place national and local policy. First, several respondents suggested that the US national government preparedness for the pandemic and lack of a unified response caused them significant distress. Many respondents were frustrated by the lack of scientifically driven policy to end the pandemic and thought that it was “a national disgrace.” Others reported that the misinformation being perpetuated by the national government, such as “suggesting people should inject bleach or Lysol,” made them “fearful.” One physician reported that “the lies that come out of the White House make my job more difficult and I am concerned about a resurgence when we already have had to treat 2 ICU units housing 30 COVID-19 patients.” However, a couple respondents felt hope that the pandemic would shine light on the issues with healthcare delivery in the US to spur reform.

The second theme within the public policy level related to the shelter-in-place policy and included factors of shutting down services, economic recession, flattening the curve, advocacy, and environmental benefits. First, some respondents expressed being upset that services were shut down during the shelter-in-place order, including gyms and restaurants. A few HWs reported being upset of the economic impact of shutting down services. However, many HWs reported that the shelter-in-place order was hopeful to them. Many suggested that the policy was helping to flatten the curve. Many even suggested unanticipated benefits of this policy that gave them hope, including improved environmental impact.

Strengths and limitations

Our study has some notable strengths and limitations. First, our sample included 1,132 HWs across all US regions who were recruited during the peak of COVID-19 transmission. The open-ended questions enabled us to learn about novel factors related to HW wellbeing that have not been studied explicitly previously. These insights can be used to further explore risk and protective factors of HW mental health during the COVID-19 pandemic and future outbreaks. Due to the anonymous nature of the survey, HWs were able to provide open and honest answers to minimize response bias. However, we were unable to analyze differences between HWs who responded and who did not respond, and our study is at risk of nonresponse bias. Our study also included significantly more females than males, and the majority of participants identified as White; future research should strive for more diverse samples, for example by oversampling racial and ethnic minority HWs. The majority of participants were also located in the Northeast, and future work should strive for better representation of HWs from other regions within the US. Yet, our study aimed to capture a range of HW positions, including physicians, nurses, health technicians/ technologists, medical assistants, medical trainees, social workers, and non-clinical workers.

In sum, this is the first study to our knowledge that has used a hybrid inductive-abductive analysis of the experiences and wellbeing of HWs during the COVID-19 pandemic in the US. Notably, we found novel factors that negatively impact HWs’ wellbeing during the pandemic that should be further explored, including national government policies, misinformation disseminated by political leaders, discrimination against people of Asian descent, and community non-compliance with social distancing. These findings situated within a Social Ecological Model can inform interventions and policies that aim to support HWs and foster their resiliency during the COVID-19 pandemic and future outbreaks.