Health Care Personnel Workdays Lost and Direct Health Care Salary Costs Incurred due to COVID-19 Infection in the Age of Widespread Vaccine Availability

Abstract

The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, has resulted in significant mortality and morbidity since it was first identified in December 2019 [1]. This has placed an unprecedented burden on societal systems, including the health care setting, with economic consequences [2]. The economic impact of the COVID-19 pandemic was magnified by the lack of global pandemic preparedness, resulting in reactive governmental and institutional responses [3]. The impact of the pandemic on the health care system is multifaceted. While the most obvious impact is due to the large numbers of patients admitted with symptomatic COVID-19, this is exacerbated by health care worker (HCW) illness and workdays lost due to SARS-CoV-2 infection. HCWs are at increased risk of SARS-CoV-2 infection compared to the general population [4, 5]. HCW illness and workdays lost combined with increased patient attendances has previously been associated with adverse patient outcomes [6].

HCW impacts due to COVID-19 have been mitigated by the introduction of personal protective equipment and limitations on time spent in contact with COVID-19 patients [7, 8]. The introduction of vaccination was a significant advance in preventing symptomatic infection and reducing the risk of onward transmission of SARS-CoV-2 in HCWs [9, 10]. Prior to the introduction of vaccination, there were significant health care system burdens associated with COVID-19–related HCW illness and workdays lost [11], with 7% of exposed HCWs having COVID-19–associated absences [12]. This is not unique to COVID-19; previous studies have examined the effects of influenza infection on workdays lost. These demonstrated that nurses and allied health professionals accounted for the majority of related costs [13]. Younger age is also associated with higher workdays lost [14]. High uptake of influenza vaccination has been shown to mitigate HCW sick leave and workdays missed [15].

The economic impact of infection due to vaccine-preventable respiratory infections such as influenza and the mitigating effects of vaccine program implementation have previously been described [16, 17]. The effect of vaccination programs on HCW absences and the subsequent impact on the health care personnel costs related to COVID-19 is understudied. The Prevalence of Antibodies to SARS-CoV-2 in Irish HCWs (PRECISE) study began as a multicenter cross-sectional study of viral respiratory infections in HCWs, and has now evolved into a prospective HCW study [18, 19]. We aimed to evaluate the incidence of staff illness and workdays lost as a result of SARS-CoV-2 infection more than 2 years after widespread COVID-19 vaccination availability.

METHODS

Study Setting, Participants, and Vaccine Access

This was a multisite prospective study in 2 hospital sites in Ireland, namely St James's Hospital (SJH) in Dublin and University Hospital Galway (UHG). These sites were chosen as they are in geographical areas with contrasting COVID-19 seroprevalence [20]. Data collection took place over an 8-month period from January to August 2023. Both sites are tertiary referral centers, with approximately 4700 and 4400 staff, respectively. HCWs were invited to participate via email, text message, and internal hospital communications. These communications were issued by both the study team and the hospital management groups. The recruitment procedure was similar at both sites. HCWs were defined as any person employed to work on site in the hospital, and recruitment was not limited to those with direct patient contact. Mandatory absence in the event of COVID-19 infection at both sites for the duration of the study was 5 days from day of symptom onset or positive test, whichever was shorter.

There was no vaccine mandate for HCWs in Ireland at any point during the pandemic, nor did HCWs have to disclose their vaccine status to their employer. However, vaccination was made freely available throughout all hospital sites, with walk-in clinics across the hospitals. Vaccination was promoted throughout the hospital setting, with emails, webpages, and posters conveying information on how to access vaccines, as well as information stands staffed by peer leaders at various points on the hospital campus. While mandates for HCWs were not introduced, there were societal restrictions in place for any citizen who did not receive vaccination. Only those who possessed a European Union Digital COVID Certificate or Health Service Executive Vaccination Card were permitted to enter pubs, restaurants, or cafes in the Republic of Ireland for a 6-month period from August 2021.

Ethical approval for the study was obtained from the local ethics committees at SJH and UHG (application No. TUH/SJH REC 2022-Nov-23002300 and GCREC 15/09/2022 C.A. 2860, respectively). Electronic informed consent was obtained from all participants.

Data Collection, and Dependent and Independent Variables

Data collection was paperless, with participants completing online questionnaires. Two dependent variables were identified for the statistical analysis: (1) HCW workdays lost; and (2) direct health care personnel costs. The independent variables included age, sex, ethnicity, occupational status, site, vaccination status, and month. An initial enrolment questionnaire collected demographic data, HCW occupation role, medical history, and SARS-CoV-2 vaccine history. COVID-19 vaccination history was confirmed by cross-checking with COVAX, the Irish national COVID-19 immunization system. Vaccination status was characterized as 4 groups: unvaccinated/incomplete primary course, primary course only, primary course and 1 booster, and primary course and 2 boosters (Supplementary Table 1). Monthly questionnaires over the 8-month study period recorded intercurrent COVID-19 infections as well as days with symptomatic COVID-19 and HCW workdays lost as a result of acute infection. Participants who failed to complete their monthly questionnaire were contacted by a member of the study team and reminded to do so. COVID-19 diagnosis was self-reported, with dates of positive self-taken antigen test or laboratory-performed polymerase chain reaction (PCR) test for SARS-CoV-2. All participants had access to rapid antigen testing via the workplace, while PCR testing was accessed on a case-by-case basis for HCWs at both sites, in line with current occupational health guidance. Participants were asked if they had the presence of any symptoms characteristic of COVID-19, but a symptom severity scale was not used. Characteristic COVID-19 symptoms in the absence of a reported positive diagnostic test were not considered to represent COVID-19 infection. These participants did not undergo further testing for alternative causes as part of this study. There was no mandated periodic testing of HCWs, apart from those working directly in high-risk areas, such as bone marrow transplant cohorts. Testing was generally self-directed, either as a result of symptoms or notification regarding close contact of a known case. Some participants may have also been involved in smaller active surveillance studies, during which periodic PCR testing was performed. Direct health care personnel costs were calculated using published HCW salary scales from the Health Service Executive, Ireland (Supplementary Table 2) [21]. The salary band assigned to each HCW was in line with the guidelines for the economic evaluation of health technologies in Ireland [22].

Statistical Analysis

Descriptive statistics are reported as means with standard deviations (SD) and medians with interquartile ranges (IQR), as appropriate. Univariate analysis and multivariable linear regression were used to analyze factors associated with HCW workdays lost and direct health care personnel costs. The independent variables in the analyses included age, sex, ethnicity, occupational role, site, vaccination status, and month of the year. For the multivariable analysis of HCW workdays lost, a negative binomial regression analysis was employed. For the health care personnel cost analysis, a generalized linear model adopting a gamma variance function and log-link function was employed [23]. Models were examined for multicollinearity by computing variance inflation factors and visually examining residual-versus-fit models. For the purposes of the analysis, statistical significance was considered at the .05 level. All analysis was performed using Stata version 18.0 (Stata Statistical Software).

RESULTS

Participant and Salary Characteristics

This study enrolled n = 1218 participants; n = 637 (52%) from SJH and n = 581 (48%) from UHG. This represented 14% and 13%, respectively, of the eligible HCWs at each site. A breakdown of the proportion of staff grades in PRECISE relative to their total staff number across both sites is shown in Supplementary Table 3. Data were collected over an 8-month period, or 242 consecutive days. The characteristics of these participants are shown in Table 1. Participants at SJH were older and had higher median annual salary, while there were also differences in the ethnicity of participants across both sites. The cohort had a high level of vaccination, with 98% receiving at least a primary vaccine course. Of the HCW participants, 208 (17%) missed at least 1 day of work due to COVID-19 illness. HCW personnel salary costs in the cohort as a whole were higher for men (median, €72 560; IQR €61 337–€96 152) than women (median, €70 976; IQR, €61 337–€83 218; z = 3.35; P = .001) and with increasing age (r² = 0.09, P = .002), while being in a clinical role was also associated with a higher salary (z = 6.05, P < .001). Salary differences were seen between participants based on vaccination status, with higher salary associated with increased number of vaccines received (r² = 0.03, P = .04).

Associations With Workplace Absences

There were 266 COVID-19 infections reported in n = 258 individuals during the study period, with participants reporting COVID-19 symptoms for 712 days. No participant required hospitalization. Over the course of the study, 1191 workdays were lost, equivalent to 4.5 days lost per infection. The burden of illness and absenteeism varied from month to month. The monthly changes in days spent off work and days symptomatic with COVID-19 are shown in Figure 1, along with the direct health care personnel cost per month. The number of workdays missed per participant each month did not differ significantly (r² = 0.04, P = .39; Table 2). However, being unvaccinated was associated with higher number of days missed, while male sex was also associated with increased days missed (Table 2).

Figure 1.

Trends of cumulative monthly totals of days spent symptomatic (blue/bottom line) and days spent absent from work (grey/middle line) on the left y-axis, as well as trends of cumulative monthly costs for health care worker absenteeism (yellow/top line) on the right y-axis.

Open in new tabDownload slide

A multivariate linear regression negative binomial model was used to further interrogate the factors associated with lost workdays due to COVID-19 (Table 3). This demonstrated seasonal variation, with June having significantly fewer days of absence compared to January. Receipt of primary COVID-19 vaccination course was associated with fewer days missed, while no significant difference was seen with subsequent booster doses.

Direct Health Care Personnel Costs

The total direct health care personnel cost of the 1191 workdays lost during the study period was €397 974, equating to €1645 per day. Univariate analysis of factors associated with COVID-19–related absenteeism demonstrated that increased costs were associated with male sex and vaccination status (Table 4). These associations were further investigated using a generalized linear regression model. Direct health care personnel costs were associated with vaccination status. Receiving a primary vaccination course as well as any boosters had lower associated costs than unvaccinated HCWs. This was independent of the other variables included in the model (Table 5). Male sex and older age were also associated with increased health care personnel costs. Seasonal variation was seen, with higher costs in March and lower costs in June.

DISCUSSION

This study demonstrates the impact and burden of HCW workplace absences due to COVID-19 infection in the setting of widespread rollout of vaccination and also evaluates the associated direct health care personnel cost. The factors associated with HCW workdays lost and its related costs identify potential areas that can be targeted to mitigate COVID-19–related HCW absences as well as other vaccine-preventable seasonal illnesses, to allow for appropriate workforce planning and reduce the economic burden of COVID-19 infection. These interventions are of increasing importance, given the waning uptake of SARS-CoV-2 booster vaccines and ongoing high levels of health care utilization and costs, posing continued workplace challenges [24–26].

Our cohort of HCWs demonstrated good engagement with initial SARS-CoV-2 vaccination programs, with 98% completing the primary vaccination course and 88% receiving at least 1 booster dose. This is reflective of the health care service in 2023, with high vaccine uptake reported across multiple settings [27, 28]. Despite this, there was still a significant level of COVID-19 absence, with more than 1000 workdays missed over the 8-month study period. Unvaccinated HCWs were significantly more likely to miss days due to COVID-19 than HCWs who received vaccination. The number of workdays lost due to COVID-19 is in excess of those reported due to similar respiratory viral illnesses such as influenza [29]. This may in part be due to mandatory minimum periods of absence due to COVID-19, which in Ireland is currently 5 days from symptom onset. These mandated periods of absence are likely to persist for the foreseeable future.

Increased direct health care personnel costs were associated with male sex and increasing age. These differences were seen in both univariate analysis of salaries and multivariate analysis. This may reflect increased salary costs with increasing age as well as sex salary disparities, which are well-recognized issues within health care salary structures [30–32]. The direct personnel costs associated with COVID-19–related HCW absences were also higher in unvaccinated participants. This is particularly striking, as salary costs in the study population were lowest in those who were unvaccinated. This demonstrates the significant impact of vaccination on mitigating absences and associated costs. Receipt of a primary SARS-CoV-2 vaccine course was associated with both significantly fewer days missed and with lower costs, irrespective of receiving subsequent booster doses. The cost per day was higher than those costs reported previously for other vaccine-preventable infections [13].

Another important finding of this study is the changing pattern of HCW workdays lost and associated cost with time of the year. As expected, workdays lost peaked in the winter months and decreased in summer, with associated significant changes in cost burden. This provides important insights for workforce planning. Staff absences coincide with the busiest periods for health care systems, with the surge of respiratory viral infections such as COVID-19, influenza and respiratory syncytial virus placing increased stress on the health care service [33–35]. Seasonal impact of these viruses has previously been associated with increased health care costs [36, 37].

We identify a potentially modifiable risk factor for increased HCW workdays lost and increased direct health care personnel costs, namely SARS-CoV-2 vaccination. The associated significant reductions in both absences and costs with receipt of primary vaccination course highlights the importance of targeted vaccination campaigns, with focus needed on the small cohorts yet to receive any vaccine. The need to increase base levels of HCW vaccination to mitigate costs has previously been highlighted in the context of influenza [38, 39]. Vaccine hesitancy amongst HCWs has previously been associated with age, female sex, and role, both within this study cohort and others [24, 40]. The reasons underlying hesitancy within these groups require investigation to inform targeted vaccination campaigns, aiming to vaccinate all eligible HCWs with at least a primary vaccine course. The absence of additional booster vaccine doses influencing workdays lost may indicate that a more targeted approach to delivering booster doses may be needed. Identification of these cohorts in both HCWs and the general population is an important next step. These findings can also be adapted to other vaccine programs.

There are several limitations of this study worth noting. The dependence on self-reported data gives rise to the potential for recall bias. Given the study design, the representativeness of the study sample to the HCW population may be questioned. This is an increasingly important issue for HCW epidemiology studies at this stage of the pandemic, with significant drop off in participation seen in many COVID-19 studies. However, there was representation across all HCW groups within this study. We did not quantify all direct health care costs or any indirect costs associated with unexpected short-notice HCW absences, including overtime payments and short-term locum cover. It is therefore likely that the true cost of HCW workdays lost is higher than we report. Similarly, we did not evaluate the impact of postinfection conditions such as long COVID, which cause significant morbidity and direct and indirect costs [41, 42]. Another variable not quantified is the indirect cost to patient care and outcome, including but not limited to cancellation of clinics and theatre lists, delays in patient movement from point of hospital access to a ward bed, and the consequences of staff cross-cover outside areas of their usual scope of practice. These variables are difficult to quantify, but should be considered in any prospective model being implemented to track workdays missed.

Overall, we demonstrate ongoing significant health care system burden relating to COVID-19, while identifying possible mechanisms to mitigate these effects. These findings should help inform annual vaccination programs and workforce planning, and highlight the need for new approached to tackle vaccine hesitancy amongst HCWs.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Acknowledgments. The authors thank all participants in this study. We also thank the contributions of the Occupational Health Department at University Hospital Galway and the support received from the Health Protection Surveillance Centre in Ireland.

PRECISE Study (Prevalence of COVID-19 in Irish Healthcare Workers) Steering Group Members and Affiliations: Lorraine Doherty, National Clinical Director for Health Protection, Health Service Executive (HSE)-Health Protection Surveillance Centre (HPSC), Dublin, Ireland, and Chair of Steering Group; Lisa Domegan, HPSC, Dublin, Ireland; Niall Conlon, Consultant Immunologist, St James's Hospital, Dublin, Ireland; Greg Martin, HPSC, Dublin, Ireland; Cillian de Gascun, Director, UCD National Virus Reference Laboratory, University College Dublin, Dublin; Joan Gallagher, Programme Manager, Office of the National Clinical Director for Health Protection, HSE-HPSC, Dublin; Mary Keogan, Consultant Immunologist, Beaumont Hospital and Clinical Lead, National Clinical Programme for Pathology, HSE; Noirin Noonan, Consultant in Occupational Medicine, St James's Hospital, Dublin; Cliona O’Farrelly, Department of Immunology, Trinity College Dublin; David Byrne, St James's Hospital, Dublin.

Author contributions. L. T., J. McG., C. K., C. B., and C. F. contributed to conceptualization and study design. L. T., P. G., C. B., and C. F. contributed to the analysis plan. L. T., P. G., J. McG., C. B., and C. F. contributed to the preparation of study data for analysis. L. T. and P. G. performed the data analysis and visualization, as well as manuscript preparation. All authors reviewed and approved the final manuscript prior to submission.

Data availability. Data will be made available from the corresponding author upon reasonable request.

Disclaimer. The funder had no role in the design of the study, the data analysis, or the decision to publish.

Financial support. This work was supported by the Irish Health Service Executive COVID-19 budget.

References

Author notes