Background
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections were first reported in China in 2019. 1,Reference Chan, Yuan and Kok2 The virus and its associated illness, COVID-19, rapidly became a global pandemic by early 2020. Reference Cucinotta and Vanelli3 In subsequent years, several variants of the virus have emerged, including Delta and Omicron lineages in 2021, the latter persisting today as its subvariants continue to evolve. Reference Mlcochova, Kemp and Dhar4,5 COVID-19 has caused substantial morbidity and mortality in North Carolina, with over 175,000 hospitalizations and 29,000 deaths through May 2023. 6,7 This heavy disease burden has caused extreme pressure on healthcare facilities in both caring for patients and protecting healthcare personnel. Reference Emanuel, Persad and Upshur8
Controlling the spread of infectious illnesses like COVID-19 in the healthcare setting is important not only for the wellbeing of HCP and patients but also for maintaining quality healthcare services for communities. Reference Danzmann, Gastmeier, Schwab and Vonberg9,Reference Widera, Chang and Chen10 A critical strategy for mitigating COVID-19 transmission among HCP is discouraging working while ill, a practice called “presenteeism”. Reference Aronsson11 Presenteeism can lead to occupational transmission to other HCP and also risks exposing patients, many of whom may be at high risk for severe COVID-19. Reference Stern, Charness and Gupta12,Reference Tori, Dinh and Mather13
Research on presenteeism with acute respiratory illness (ARI) symptoms has largely focused on influenza, with limited data from the COVID-19 era. Reference Daniels, Wei and Han14,Reference Babcock, Beekmann and Pillai15 This gap in research is important, given the unique challenges posed by the pandemic including high transmission rates, resource strain, HCP burnout, and rapidly changing quarantine and isolation guidelines. Previous studies have identified various motivators for presenteeism among HCP, including heavy workloads, staffing challenges, employment consequences, mild illness, and a sense of professional duty. Reference Daniels, Wei and Han14–Reference Webster, Liu, Karimullina, Hall, Amlôt and Rubin16 However, it remains unclear which factors are the most important drivers of presenteeism that should be prioritized for interventions. Furthermore, it is unknown whether motivations, symptoms, or frequency of presenteeism changed during the pandemic. Understanding these shifts is crucial for addressing presenteeism in the context of ongoing and future public health crises.
We conducted a study on ARI presenteeism in HCP across eras of the COVID-19 pandemic. Our aims were to determine the prevalence of ARI presenteeism, and to examine motivations and workplace-related attitudes that influence the behavior.
Methods
We conducted two voluntary, anonymous surveys of HCP at the University of North Carolina Medical Center (UNC-MC) to assess motivations for, and demographic factors associated with, self-reported acute respiratory illness (ARI) presenteeism (Appendix 1). All HCPs at UNC-MC (n = 11,429) were eligible. Participants were recruited through digital employee newsletters. The first survey was conducted from November 1 through December 27, 2021 (the “pre-Omicron” survey), and the second survey was conducted from April 1, 2022 through May 31, 2022 (the “Omicron” survey). Although minor modifications were made to the Omicron survey, the question structure and topics were consistent.
Survey overview
We posed 24 total questions on a range of presenteeism behaviors, knowledge, and attitudes toward COVID-19, workplace infection risk and transmission, as well as job-related and personal demographics. No identifying information was collected. The study was approved by the UNC IRB (study #20-2533).
Participants reported ARI presenteeism, defined as “working despite feeling feverish plus a cough and/or sore throat,” at frequencies of never, once, twice, or three or more times during the designated recall period for each survey. For the pre-Omicron survey, the recall period was March 2020 through the time of response in November/December 2021 (depending on when the survey was taken). For the Omicron survey, the recall period was from January 1 through April/May of 2022 (depending on when the survey was taken). Participants were also asked to report on any symptoms they may have had at work during the recall period, choosing from a provided list of common acute illness symptoms. We refer to this as “any-symptom” presenteeism.
Motivations for any-symptom and ARI presenteeism were presented as a list of sixteen choices created by a team of infection preventionists, epidemiologists, and clinicians with feedback from nursing staff. Participants identified all motivators applicable to them, as well as the most important of those (“primary” motivator”). We also measured how often respondents called out sick or were released from work due to ARI symptoms.
We used a Likert scale to measure knowledge of workplace absence policies, attitudes toward work demands, colleague behaviors, and COVID-19 risk perception. During the Omicron survey, we measured pandemic fatigue and burnout using two items from a validated survey on COVID-19 fatigue. Reference Cuadrado, Maldonado, Tabernero, Arenas, Castillo-Mayén and Luque17
Survey demographics included age, gender (female, male, non-binary, self-described), child caregiver status, education level, race, and ethnicity. We assessed occupation using questions about the participant’s professional role, interaction with patients, and work location (COVID-19 unit, non-COVID unit, non-patient care areas). We measured COVID-19 vaccination status by number of doses. Vaccines were required for all HCP at UNC-MC as of September 2021. We obtained internal institutional data to evaluate the representativeness of our sample relative to the total employee population. 18
Analytical approach
We compared reported levels of ARI presenteeism (never vs at least once) and the distribution of motivator categories between surveys and across demographic variables using two-tailed Fisher’s exact tests employing Monte Carlo simulated p-values, with the R standard 2000 replicates. 19 We used log-binomial regression to model relationships between ARI presenteeism and attitudes related to workplace policies and culture, COVID-19 risk perception, COVID-19 prevention burnout, and pandemic fatigue. Assuming that these attitudes and perceptions predated ARI presenteeism behavior, we estimated prevalence ratios adjusted for age, gender, work role, and patient interaction. Surveys were conducted using Qualtrics, and statistical analysis was performed in R 4.3.1 (R Core Team 2023). Figure 3 was created using the likert R package (v2.0.0, Bryer, 2022).
Presenteeism symptoms reported by pre-Omicron (2021) and Omicron BA.1 (2022) survey respondents.
Results
The pre-Omicron survey received 582 responses and the Omicron survey received 341 responses. In both surveys, a large majority of respondents were female, identified as white, were full-time employees, had paid leave available, and were fully vaccinated per the current recommendations at the time they took the survey (Table 1).
Demographic characteristics of University of North Carolina Medical Center survey respondents, Pre-Omicron (2021) and Omicron BA.1 (2022) era surveys
*Clinical frontline healthcare personnel (HCP) includes nurses and advanced practice providers (physicians, physician assistants and nurse practitioners); **Other HCP includes paraprofessionals such as occupational, physical or respiratory therapists, and nursing and medical assistants.
Representativeness of survey respondents
Relative to the total employee population, non-White participants were underrepresented in both surveys, with Black and Hispanic respondents representing approximately half or less of their share of the total employee population. However, presenteeism did not vary considerably by race (Table S1).
The pre-Omicron survey had a higher proportion of administrative HCP relative to all employees, and fewer “other” HCP in non-frontline roles. In contrast, Omicron had a much larger share of clinical frontline HCP than was seen in the total employee population. The Omicron participant age distribution was well-matched to the total employee population, while pre-Omicron had more HCP over 40. The total employee population had fewer females than each survey, at slightly less than 80%.
Symptoms reported and ARI presenteeism frequency
When presented with a list of 11 acute illness symptoms (Figure 1), 60.7% of pre-Omicron participants and 70.6% of Omicron participants reported attending work with at least one symptom. Headache, sinus congestion, cough, and sore throat were the most common presenteeism symptoms reported in both surveys. In both surveys, less than 10% of respondents reported measured fever, chills, or loss of taste/smell while working. Notably, compared to pre-Omicron, there was a higher prevalence of presenteeism with every symptom in the Omicron era, except for chills. Excluding runny nose from the Omicron set for comparability, the mean number of symptoms ever reported at work over the recall period was significantly higher during the Omicron era (2.18) compared with the longer pre-Omicron period (1.80) (Student’s t-test, p = 0.002) (Table 2).
Distribution of primary motivators for acute respiratory illness presenteeism during pre-Omicron (2021) and Omicron BA.1 (2022) COVID-19 waves among healthcare personnel at University of North Carolina Medical Center.
Reported acute respiratory illness presenteeism, any-symptom presenteeism, and staying home with acute respiratory illness (ARI) symptoms among healthcare personnel at University of North Carolina Medical Center, pre-Omicron (2021) and Omicron BA.1 (2022)
* Symptoms surveyed included chills, cough, diarrhea, feeling feverish, fever >100°F, headache, loss of appetite, loss of taste and smell, muscle aches, sinus congestion, sore throat and vomiting. Runny nose was measured in the Omicron survey only and was excluded from these summaries for comparability between eras.
† Frequency of ARI presenteeism and staying home with ARI were surveyed as separate questions to all participants. Those with no ARI during the recall period responded “never” to both questions.
In the pre-Omicron survey, 24.4% of participants reported at least one instance of ARI presenteeism during the recall period (Table 2). In the Omicron survey, 33.7% of respondents reported at least one instance of ARI presenteeism in the recall period. More respondents stayed away from work with ARI symptoms pre-Omicron (45.1%) than during Omicron (34.0%).
Demographics and presenteeism
Measuring ARI presenteeism across demographic categories (Table S1) revealed some differences in age, gender, role, and patient interaction. Respondents under 40 were more likely to report ARI presenteeism compared to those over 40 (pre-Omicron PR 2.18 [95% CI, 1.63–2.94], Omicron 2.07 [95% CI, 1.43–3.13]). Pre-Omicron, ARI presenteeism was more common in females (PR 3.82 [95% CI, 1.82–10.5]) and other gender identities (PR 7.89 [95% CI, 2.49-24.3]) than males. No gender associations were found during Omicron. HCP with direct patient interaction were more likely to report ARI presenteeism than those not primarily providing direct care (pre-Omicron PR 1.74 [95% CI, 1.26–2.47], Omicron 2.00 [95%CI, 1.31–3.29]). Clinical frontline HCP were more likely to report ARI presenteeism than non-frontline roles (pre-Omicron PR 1.45 [95% CI, 1.08–1.93], Omicron 1.59 [95% CI, 1.13–2.32]). We found no significant differences in ARI presenteeism in either survey by education, race/ethnicity, caregiver status, full vs part-time status or paid leave availability.
Motivators between survey eras
The distribution of primary motivators among those reporting ARI presenteeism differed between surveys. Pre-Omicron, the leading motivator categories were employment impact (27%) and sick leave (22%), while workplace culture predominated during Omicron (32%) (Figure 2). Low COVID-19 risk perception was an important influence during both eras. In particular, having mild symptoms was one reported motivator among 51% of pre-Omicron and 68% of Omicron participants. However, low risk was not the primary motivator for most HCP; only ∼15% of participants cited this as their primary motivation (Figure 2).
Knowledge and attitude Likert-scale items from pre-Omicron (2021) and Omicron BA.1-era (2022) surveys.
The three leading primary motivators pre-Omicron were employment record concerns (a disciplinary “occurrence”) at 17%, needing sick leave for other purposes (16%), and mild or allergy-like symptoms (16%), while Omicron participants were most concerned with increasing colleague workload (18%) (Figure 2, Table S2). Among the motivator options added to the Omicron survey, 11.8% cited a recent negative COVID-19 test, while very few individuals cited being cleared to return by Occupational Health Services, or “financial reasons”, intended to reflect monetary bonuses provided to some HCP completing a series of shifts.
As a sensitivity analysis utilizing a broader definition of presenteeism, we also assessed motivator categories among individuals who reported any-symptom presenteeism (Figure S1). With this approach, a higher proportion of HCP were primarily motivated by perceived low COVID-19 risk (40% pre-Omicron, 37% Omicron).
Attitudes and perceptions
We surveyed all participants on workplace perceptions regardless of reported presenteeism. Most HCP in both surveys reported feeling a professional obligation to come to work despite feeling ill, and during Omicron, reported feeling a sense of duty to return even with mild symptoms (Figure 3). Despite consistent procedures for calling in sick to work across eras, fewer Omicron participants reported that they clearly understood those processes. Most pre-Omicron participants (53%) agreed that given workplace precautions, the risk of spreading respiratory illness was low, but only 37% agreed with this sentiment during Omicron. Of note, institution-wide precautions at both times included universal masking, a vaccine mandate and environmental controls. Half of Omicron participants endorsed COVID-19 fatigue, but only 36% reported that they were less careful than the beginning of the pandemic.
We found several associations between workplace perceptions and ARI presenteeism. Feeling a professional obligation to come to work increased the likelihood of ARI presenteeism in both surveys (Table 3). Not understanding call-out procedures predicted ARI presenteeism during Omicron but not pre-Omicron. In both surveys, hearing colleagues speak negatively of others who called out sick increased the likelihood of ARI presenteeism. Presenteeism as a social norm was further evident in that HCP who frequently observed colleagues engaging in presenteeism were more likely to report ARI presenteeism themselves. During Omicron, HCP who reported experiencing either COVID-19 fatigue or burnout, or both, had a higher ARI presenteeism prevalence compared to those reporting neither.
Adjusted prevalence ratios 1 (PR) of acute respiratory illness (ARI) presenteeism across Likert-scale knowledge and attitudes items, pre-Omicron (2020-2021) and Omicron BA.1 (2022)
*Model covariates include age, role, patient interaction; all other models adjusted for age, gender, role, and patient interaction.
1 In developing these models, we assumed that perceptions of workplace culture preceded an individual’s decision to come to work sick. Thus, these models include covariate adjustment for important demographic factors (age, gender, role, patient interaction) associated with presenteeism which were considered for their relevance to each predictor. Because this data is cross-sectional with long recall periods, this assumption may not hold for every individual. As a sensitivity analysis, we relaxed this assumption and removed covariates from the models. We found prevalence ratios similar in both direction and magnitude. Specifically, 12 of the 14 statistically significant PRs here moved slightly away from the null, one became non-significant (Omicron sick policy) and one moved toward the null (Delta procedures).
Missing data
Although some participants skipped all end-of-survey demographic questions, missingness on these items was low overall at 10%–12%. A sensitivity analysis revealed no substantial change in model results when excluding individuals with missing demographics.
Discussion
A majority of respondents (pre-Omicron 59.7%, Omicron 70.8%) engaged in broadly defined any-symptom presenteeism including at least one symptom listed in Figure 1. A substantial proportion of surveyed HCP reported at least one instance of febrile acute respiratory illness (ARI) presenteeism; ∼25% during the first 21 months of the pandemic. This figure rose to >30% during the first four months of 2022, a period containing the highest to-date hospitalization and community case rates. 6,7,20 With fewer non-frontline personnel and an older age distribution than the total employee population, our pre-Omicron survey data may underestimate the true employee population prevalence of ARI presenteeism during that period.
Because of the range of ways presenteeism prevalence is measured in the literature, direct comparison of our results to other studies is challenging. Nevertheless, one other study during the COVID-19 pandemic reported 49.8% of COVID-19-positive HCP came to work with symptoms over a two-week period. Reference Linsenmeyer, Mohr, Gupta, Doshi, Gifford and Charness21 One systematic review found prevalence of presenteeism among HCP with infectious illness ranged from 37%–97%. Reference Webster, Liu, Karimullina, Hall, Amlôt and Rubin16 Another review reported respiratory illness presenteeism across industries and time frames to range between 7% and 100% among symptomatic and 14% to 55% among clinically confirmed influenza and influenza-like illness cases. Reference Daniels, Wei and Han14
Our results are particularly concerning as they revealed high levels of presenteeism despite many workplace supports at UNC-MC including access to testing and health insurance coverage for all staff, repeated communication discouraging presenteeism, and institutional responsibility for staffing coverage. In more resource strained or organizationally challenged settings, rates of presenteeism might be higher than reported here.
Our data showed younger HCP and female HCP were more likely to report ARI presenteeism, in agreement with two previous studies that examined these relationships. Reference Mossad, Deshpande, Schramm, Liu and Rothberg22,Reference Ablah, Konda, Tinius, Long, Vermie and Burbach23 Previous data is inconclusive on presenteeism across HCP roles. Reference Gudgeon, Wells, Baerlocher and Detsky24–Reference Imai, Hall, Lambert and Merollini26 Our assessment demonstrated that HCP working in clinical frontline roles and those providing direct patient care are more likely to report ARI presenteeism than other HCP types. This is likely due to the necessity of performing duties in-person whereas administrative professionals may be able to work remotely and have more flexibility to remain at home when experiencing symptoms. Also more likely to report ARI presenteeism, younger individuals may be more likely to be employed in these in-person, direct care roles at earlier stages of a career or work experience, and therefore have less flexibility to remain outside the workplace while symptomatic. This connection between direct care roles and presenteeism highlights the serious risks of presenteeism, in that the very workers who are most likely to be present are those most likely to be in contact with vulnerable patient populations.
Even so, since only 15% of participants were primarily motivated by low Covid-19 risk perception, effective strategies to combat ARI presenteeism must move beyond communication that focuses only symptoms and potential risk to others. The pre-Omicron data suggests that employers may have direct leverage over ARI presenteeism in their disciplinary action and benefits policies. Half of respondents were primarily motivated by concerns about disciplinary consequences or sick leave constraints, both institutionally regulated. In contrast, the Omicron data suggests that reducing presenteeism may require strategies addressing more complex issues in workplace culture and staffing, which remains challenging nationally. The timing of each survey may underlie these differences, with the pre-Omicron survey conducted at the very end of the calendar year, when more sick leave is likely to have been used, while the Omicron survey recall period covered the beginning of the year. Further, the CDC implemented COVID-19 return-to-work guidance changes at the end of December 2021 in anticipation of crisis-level staffing during the impending Omicron wave. 27,28 While UNC-MC policy did not change during this period, these highly publicized CDC guidance changes may have caused confusion on calling out sick and increased operational pressures. Maintaining and communicating clear procedures for calling out sick are important, particularly during periods with high community case rates, changes in institutional policy, or changes in public health guidance.
The workplace attitudes findings emphasize important influences on presenteeism that should be considered by employers, including culture, operations issues (adequate staffing), and COVID-19 fatigue and burnout. Presenteeism can become a social norm, as evident here with HCP who observed presenteeism engaging in the behavior more often themselves. Creating a culture in which presenteeism is discouraged, and promoting a social norm of staying home when sick may be helpful. Importantly, these cultural norms may be most effective in tandem with strong institutional support for CDC-recommended HCP furloughs, including appropriate staffing redundancy and eliminating penalties for required absences, particularly for specialized and critical units such as those providing extracorporeal membrane oxygenation.
While some desensitization to a pandemic threat over time may be inevitable, addressing burnout and fatigue should not be dismissed. It is particularly concerning that those less motivated to engage in COVID-19 precautions due to fatigue and burnout are more likely to report presenteeism, increasing the risk of occupational outbreaks or transmission to patients.
Strengths
This study is one of few assessments of presenteeism during the COVID-19 pandemic era. To our knowledge, this is the largest and most comprehensive and thorough study of HCP presenteeism motivators in the COVID-19 era. Additionally, ours is one of two presenteeism studies in the COVID-era where eligibility was not restricted to confirmed COVID-positive HCP.
Another advantage of this study is measurement at multiple time points, allowing for comparison of presenteeism across two waves of the pandemic. Different phases of a pandemic (early days vs higher infection rates) may require different approaches to reduce presenteeism. In addition, having obtained demographic data on the total employee population, we were able to contextualize our results despite low response rates.
Limitations
This study has several limitations. First, accurate measurement of presenteeism is challenging in any survey-based research due to potential recall and self-report bias. Here, the recall period for the pre-Omicron survey spanned 20–21 months and the Omicron 4–5 months. The latter, more limited time range may lead to more accurate self-estimates and therefore differential recall bias between surveys could exist.
Secondly, social desirability bias may have limited participants’ likelihood to report ARI presenteeism. This underreporting, combined with long recall periods, suggests that our results likely underestimate the true prevalence of ARI presenteeism. Measures of workplace perceptions may have also been subject to social desirability bias, with participants inferring a “preferred” response to provided statements.
Third, accurate measurement of motivations is challenging. While we delineated a wide range of motivations for presenteeism, additional drivers of this behavior may be unmeasured. We devised categories of motivators to represent important systems and incentive structures that may influence presenteeism, but these may not precisely align with participant interpretations.
Fourth, while this survey was distributed on the largest e-newsletter platforms available, responses captured a small fraction of all employees, and selection bias may be present. Engagement with a COVID-19 survey may indicate more COVID-19 concern than the general HCP population, and consequently those surveyed may be more risk-averse, and therefore report less presenteeism than might be observed in the broader HCP population.
Finally, generalizability of these results may be limited by the study setting. The most common motivations for presenteeism for HCP in a pandemic may differ from experiences in non-crisis times and in other work environments. Still, it is particularly important to study presenteeism among HCPs because of their role in caring for vulnerable populations. In addition, the broad themes and even specific motivations explored here may inform approaches to presenteeism reduction in other occupational settings, specifically other high-pressure, understaffed, and/or caregiving-oriented work environments.
In summary, our findings demonstrate that presenteeism among healthcare personnel is a persistent issue through recurring waves of a respiratory illness pandemic. While reasons for this behavior are manifold, employers may have leverage to reduce presenteeism through supportive institutional policies and culture. As the COVID-19 landscape evolves and the risk of future pandemics remains, future work should continue to focus on motivations and solutions to accommodate HCP needs.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ice.2025.69
Acknowledgments
The authors would like to acknowledge the Infection Prevention Department and Occupational Health Services at UNC-MC for their assistance in creating and modifying the questionnaires used for this study.
Financial support
None reported.
Competing interests
All authors report no conflicts of interest relevant to this article.
