Introduction

The public health workforce started the response to the COVID-19 pandemic with a critical deficit, having lost 20% of its workforce between 2018–2019 [1]. To address the needs created by the pandemic, public health practitioners shouldered expanded responsibilities, took on more job roles, and worked more hours per day and days per week [2–5]. Despite reports from the de Beaumont Foundation that at least 80,000 additional public health workers were needed to ensure that state and local health departments could provide minimally acceptable levels of public health services to their respective communities, there are still critical staffing and funding shortages [6].

Burnout is defined by Maslach and Leiter (2016) as “a psychological syndrome emerging as a prolonged response to chronic interpersonal stressors on the job,” with three areas of effect, including 1) exhaustion, 2) cynicism and detachment, and 3) feelings of being ineffective or unaccomplished [7]. Lack of control over job circumstances, unclear expectations, challenging workplace dynamics, chaotic or extreme levels of work-related activity, lack of social support, and work-life imbalance also contribute to occupational burnout [8]. The COVID-19 pandemic response–in addition to the existing workforce and funding deficits–generated circumstances that exacerbated many of these factors, leading to high rates of burnout across many job sectors. While significant research has examined burnout among patient-facing healthcare providers such as nurses and physicians, the burnout experiences of the public health workforce have been underexplored [9,10].

Burnout and associated sequelae are a hindrance to the effectiveness of public health preparedness and response capacity. For example, in a U.S. public health workforce study conducted in September 2020, 66.2% of respondents reported burnout [5], and burnout was one of the most frequently cited reasons for considering leaving the public health workforce during the COVID-19 response [11]. Successful response to future public health emergencies is contingent upon a well-equipped and adequately supported public health workforce. Accordingly, we conducted a follow-up survey of public health workers in summer 2021 to assess changes in the prevalence of, and risk factors for, burnout among the U.S. public health workforce.

Material and methods

Results

Demographic and work-related characteristics comparing baseline and follow-up cohorts

For quality control, demographics were checked for consistency between baseline and follow-up surveys. Basic demographics, well-being, and work-related factors were comparable between the baseline and follow-up surveys. Descriptive statistics and comparisons for baseline and follow-up cohorts are summarized in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Demographic characteristics of respondents in the initial survey and follow-up survey, including χ2 p-values comparing proportions of given demographic characteristics for respective survey waves.

https://doi.org/10.1371/journal.pmen.0000100.t001

Experience of burnout among follow-up cohort

The proportion of respondents with burnout in the baseline survey (60.4%) was significantly lower than the proportion reporting burnout at both the baseline and follow-up survey (χ2 = 4.37; p-value = 0.0367). Individuals reporting burnout in the baseline survey were more than 7 times as likely to report burnout at follow-up when compared to those reporting no baseline burnout (OR = 7.26; 95% CI: 2.43–21.71; McNemar’s χ2 = 4.54; p-value = 0.03). Over half of respondents to both surveys reported burnout in both; 20% reported burnout at only baseline survey; 7.5% reported burnout at follow-up but not baseline survey; and 21.25% reported did not report burnout in either survey.

Follow-up survey respondents reported feeling emotionally exhausted from work, frustrated with work, and burnout. Nearly all (97.87%) of individuals reporting burnout reported being exhausted at the end of the day; they also noted feeling exhausted every hour of the day (72.34%) and in the morning (87.23%). Those with and without burnout significantly differed in the experience of tiredness every hour of the day (OR: 6.02; 95% CI: 2.26–16.02) and exhaustion in the morning (OR: 10.51; 95% CI: 3.48–31.75). Nearly 90% of individuals without burnout reported having enough energy for family and friends during leisure time compared to only 70% of those reporting burnout (OR: 0.33; 95% CI: 0.10–1.10). Follow-up survey respondents were 76% more likely to report burnout at the baseline survey than those only participating in the baseline survey (OR = 1.76; 95% CI: 1.03–3.01). Findings are summarized in Table 2.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Burnout-related variables of interest for the follow-up cohort of public health workers.

https://doi.org/10.1371/journal.pmen.0000100.t002

A summary of the multivariable logistic regression model is found in Table 3. Participants with five or more years of experience in the public health workforce were less likely to report burnout at follow-up (AOR = 0.45; 95% CI: 0.13–1.53) when controlling for baseline burnout, hours worked per week, and sleep quality. Those reporting frequently or occasionally getting enough sleep vs. rarely or never getting enough sleep had lower adjusted odds of burnout (AOR: 0.12, 95% CI: 0.03–0.52) when adjusting for baseline burnout, experience, and hours worked per week. When adjusting for baseline burnout, experience, and sleep quality, there was no difference in burnout at follow-up for respondents working less than 40 hours and 41–49 hours per week (AOR = 0.96; 95% CI: 0.16–5.82); those working 50+ hours per week were three times as likely to report burnout than colleagues working 40 or less hours per week (AOR = 3.02; 95% CI: 0.80–11.48). Baseline burnout was predictive of burnout at follow-up, where individuals without baseline burnout were more likely to continue to report no burnout at follow-up compared to those with baseline burnout when adjusting for experience, sleep quality, and hours worked per week (AOR: 7.76; 95% CI: 2.03–29.65).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Summary of point estimates burnout at the follow-up survey from the multivariable generalized linear regression assessing the impact of wellbeing, workplace, and demographic variables on burnout at follow-up.

As noted above, odds ratios and 95% confidence intervals represent the odds of burnout described when controlling for other variables included in the model.

https://doi.org/10.1371/journal.pmen.0000100.t003

Discussion

There is limited literature on burnout among the public health workforce–predominantly composed of non-patient-facing practitioners–specifically. A 2024 meta-analysis of studies from the U.S., South Korea, Japan, Malaysia, China, and India, found that nearly 40% of public health practitioners were experiencing burnout [15]. In our study, the prevalence of burnout was high in both baseline (165 of 273) and follow-up surveys (47 of 80), with roughly 60% of respondents reporting burnout. Individuals with burnout at follow-up reported experiencing symptoms in-line with the Maslach and Leiter’s “three key dimensions” of burnout–exhaustion, detachment, and ineffectiveness [7]. For example, a higher proportion of individuals who reported burnout at follow-up indicated that work was emotionally exhausting, causes burnout, and causes frustration, compared to those not reporting burnout. Work was more tiring, and workdays were more exhausting at both the beginning and end of the day for those reporting burnout. Further, individuals experiencing burnout had less energy for their family, friends, and hobbies. These findings are consistent with other cross-sectional studies that reported high prevalences of burnout among healthcare workers. For example, 67% of healthcare workers in a multinational study conducted between March and June of 2020 reported burnout [16]. Other longitudinal studies examining the trajectory of burnout later in the course of the COVID-19 pandemic also noted that healthcare workers continued to report high levels of psychological distress and mental health impacts as the pandemic progressed [17,18]. The persistence of burnout was seen among respondents in our study.

Experience, job roles, and hours worked can be important risk factors for burnout. Hours worked could be an important indicator of ability to maintain work-life balance, which has been identified as a key risk factor for burnout during the COVID-19 response [2,19]. In a study of Intensive Care Unit staff, burnout increased nearly 80% for physicians and nurses between late 2019 and June 2020, with the highest increase among those working the most overtime and those caring for patients with COVID [20]. Working more than 60 hours per week has consistently been shown to lead to workplace burnout [21]. Our study showed that odds of burnout were lower for respondents working less than 50 hours a week, although there was not an appreciable difference in burnout experience for those working less than 40 and those working 40–49 hours per week. Emerging literature highlights the potential value of flexible scheduling for mitigating burnout [22]. Other studies, conducted predominantly in European populations, have shown both positive and negative aspects of remote and hybrid work environment for addressing burnout [23,24]. Acceptable overtime and capped work hours can help health departments to maintain productivity while also protecting their workforce against burnout and additional stressors. Both flexible scheduling and hybrid/remote work options should be further explored as approaches for controlling work hours, giving employees occupational autonomy, and effectively reducing burnout.

Studies have found burnout to be highest in trainees and residents and to decrease with age and experience [25,26], with some bimodal findings for age and burnout [27]. In our study, respondents with five or more years of public health work, while not statistically significant, were less likely to report burnout. This could be attributable to a range of factors. For example, public health practitioners took on a range of additional job roles throughout the pandemic [5,28]. Individuals with more years of experience could have skills developed through prior experiences and utilize more adept professional coping strategies compared to those with less time and experience in the workforce. Mentoring, which is often a function of longevity in the workforce, has been shown to improve experience of burnout for both mentors and mentees [29]. For mentors in-specific, this relationship could improve feelings of burnout such as limited effectiveness and detachment. Tailored supportive measures across career stages should be explored as preemptive steps for reducing risk of developing and severity of burnout.

Prior to the COVID-19 pandemic, a small-scale survey found that insufficient sleep was a predictor of burnout [30]. Sleep can also mitigate the negative mental health impacts of stressful situations and work environments, while a lack of sleep can augment feelings of burnout [31]. This was reflected among our study respondents. However, without additional information about individual sleep habits and changes in sleep habits due to the pandemic, it is difficult to discern the role sleep played in feelings of burnout among our cohort. Future research should more closely address the relationship between subjectively or objectively measured sleep quality and burnout.

Our choice of burnout measure could have implications for study findings. The non-proprietary burnout measure we chose for our study has been validated across a number of occupations within the health sector and compared against the MBI:EE [12]. The MBI:EE is a subscale within the 22-item Maslach Burnout Inventory (MBI)–the gold standard of burnout measures [32]. There are several considerations for why we chose to use the non-proprietary measure rather the MBI or MBI:EE. First, the measure used in our study is non-proprietary while the MBI is proprietary. Second, and most importantly, a 22-item scale was less feasible for self-reported surveys. Given that the non-proprietary single question, 5-item measure has been validated for use in health-related professions and has a sensitivity of 83.2% and specificity of 87.4%, we felt that this measure’s performance and ease-of-use made it the best fit for our methods, study population, and research questions. There could be measurement errors and resulting biases stemming from our choice of measure.

There are several important limitations to this study. Our study has a relatively small sample size, in part due to needs and challenges specific to the different phases of pandemic response at the time of both baseline and follow-up survey deployment. Small cell counts made more granular analysis challenging. The follow-up cohort was appreciably smaller than the baseline cohort, with a response rate of only 33%. This concern was mitigated by no significant differences in the distribution of key demographic variables. However, selection bias could have influenced our findings if those most affected by COVID-19-associated burnout were more likely to respond to the follow-up survey (see Experience of burnout among follow-up cohort). Conversely, response rates could vary by work-related stressors, with those who were less affected by these stressors having more time to complete the survey than those experiencing higher levels of burnout. Our survey respondents overrepresented female, White, and younger demographics compared to the overall public health workforce. While the public health workforce is predominantly White and female [13], the overrepresentation of these demographics in our data render our findings not generalizable to the workforce at large. Additionally, the mechanisms that underpin burnout and individual experiences of burnout are complicated and multimodal. Our study accounts for some, but not all, of the potential confounders that could be influence the development of burnout among respondents.

Future studies should better represent the workforce at large. This will require a better census of the public health workforce, including contact information being made accessible through professional bodies such the National Association of County and City Health Officials (NACCHO). Qualitative research, building on our team’s prior work using key informant interviews, could add context to findings in this study [3,4]. Despite these limitations, this study adds value to the existing literature and gives context to the discussion of burnout experiences of the public health workforce throughout the COVID-19 pandemic.

Conclusions

Broadly, the burnout experiences of the public health workforce have been underexplored. Our study reported high levels of burnout among public health practitioners at multiple time points throughout the COVID-19 pandemic. Individuals experiencing burnout were, as expected, more likely to report experiencing work-related exhaustion. Participants with five or more years of experience and working less than 50 hours per week were less likely to report burnout. Findings regarding sleep quality and burnout were inverse from the expected relationship and garner further investigation. Workforce mental health and wellbeing must be considered to address workforce deficits and bolster public health preparedness and response capacity for future emergency response. Further investment in occupational research among public health practitioners is needed to identify and implement evidence-informed policies to address burnout within the workforce.

Supporting information