Modelling the economic burden of SARS-CoV-2 infection in health care workers in four countries

The COVID-19 pandemic presented both a health and economic crisis. While the overall economic impact is widely acknowledged, the specific economic burden of SARS-CoV-2 infections among healthcare workers (HCWs) remains less understood. The individual burden on HCWs includes medical costs, lost earnings, and reduced productivity. During a pandemic, these costs are amplified due to increased viral transmission within healthcare settings, households, and the wider community. High infection rates among HCWs also disrupt essential health services, impacting areas such as cancer care, dialysis, and critically, maternal and child health. This disruption results from HCW illness, isolation, and death, but also from the diversion of resources and personnel to manage the surge of COVID-19 patients. These disruptions have severe consequences, particularly in low- and middle-income countries (LMICs) with already limited healthcare resources. Existing economic analyses often focus on the overall pandemic cost or HCW absenteeism, neglecting the broader societal impact of HCW infections, secondary transmissions, and service disruptions. This study aims to comprehensively estimate the societal-wide economic burden of SARS-CoV-2 infections in HCWs across four LMICs during the first year of the pandemic, highlighting the need for prioritizing HCW protection and pandemic preparedness.

Previous research has highlighted the higher risk of SARS-CoV-2 infection among HCWs compared to the general population across various settings. Studies have attempted to quantify the economic costs associated with HCW infections, primarily focusing on absenteeism and direct medical expenses. One study in Iran estimated the cost of absenteeism among HCWs to be \$1.3 million. Another study in Greece included absenteeism, presentism, and direct medical care costs. However, a comprehensive estimation of the societal-wide economic burden in LMICs, encompassing direct and indirect costs, secondary transmission, and service disruptions, has been lacking. This gap in understanding the economic costs hinders effective pandemic response financing and preparedness strategies.

This study employed a cost-of-illness (COI) approach to model the economic burden of SARS-CoV-2 infections among HCWs across five sites: Kenya, Eswatini, Colombia, and two South African provinces (Western Cape and KwaZulu-Natal). The study analyzed three pathways: 1. **Pathway 1:** Direct costs associated with primary SARS-CoV-2 infections and deaths among HCWs, including medical, non-medical (e.g., travel, meals), and indirect costs (lost productivity). 2. **Pathway 2:** Costs of secondary SARS-CoV-2 infections and deaths resulting from transmission from infected HCWs to close contacts (household members and inpatients). Population attributable risk (PAR) was calculated to estimate secondary infections. 3. **Pathway 3:** Economic costs of excess maternal and child mortality due to health service disruptions caused by HCW illness, absence, or death. The impact on maternal mortality rate (MMR) and under-five mortality rate (U5MR) was assessed using elasticities relative to HCW density. Data were collected through primary data collection from national/provincial health authorities, supplemented by data from World Development Indicators, the Johns Hopkins University COVID-19 database, peer-reviewed publications, and the research team's assumptions. Cost estimations were performed in 2020 US dollars, incorporating direct and indirect costs, and accounting for disease severity. Three scenarios (low, moderate, high impact) were created by varying key parameters, including the share of inpatients as close contacts, the reduction in HCW productivity, and the elasticities of MMR and U5MR. Stochastic sensitivity analysis with 10,000 iterations was used to obtain 95% confidence intervals for the moderate impact scenario. The analysis was conducted using R (version 4.2.2).

The study found that HCWs experienced significantly higher COVID-19 incidence rates than the general population in all sites except Colombia. In all sites except Colombia, transmission from infected HCWs resulted in substantial secondary infections and deaths. The number of deaths due to secondary infections significantly exceeded HCW deaths from primary infection. Kenya experienced the highest estimated excess maternal and child mortality due to health service disruptions. The total economic burden of SARS-CoV-2 infections in HCWs varied significantly across the sites, ranging from \$16.19 million in Eswatini to \$544.64 million in KwaZulu-Natal, South Africa. These losses represented a substantial share of total health expenditure, ranging from 1.51% in Colombia to over 8% in both South African provinces. Indirect costs were the most substantial component of the total economic burden in many sites. Analysis of three pathways revealed that in Kenya, excess maternal and child mortality (Pathway 3) constituted the largest share of the economic burden, while in other sites (Eswatini, Colombia, and both South African provinces), secondary infections (Pathway 2) contributed most significantly. Scenario and sensitivity analyses showed that the economic burden was sensitive to various parameters, particularly the proportion of inpatients considered close contacts and the impact of infection on HCW productivity.

This study's findings demonstrate the substantial societal-wide economic costs associated with SARS-CoV-2 infections among HCWs, particularly in settings where HCW infection rates significantly exceed those of the general population. The economic losses emphasize the preventable nature of these costs through effective implementation of infection prevention and control (IPC) measures, including adequate personal protective equipment (PPE), improved sanitation and hygiene, and comprehensive IPC training. The significant contribution of secondary infections and health service disruptions, especially to maternal and child health, underscores the interconnectedness of HCW protection and broader public health outcomes. The high costs incurred through excess maternal and child mortality highlight the need for strategies to protect HCWs and safeguard essential services, particularly in settings with high baseline maternal and child mortality rates. The study’s results are consistent with other research showing significant economic losses due to HCW shortages in LMICs. The study's findings should inform policy decisions to prioritize investments in robust IPC measures and building resilient health systems to better protect HCWs and mitigate the economic consequences of infectious outbreaks.

This study demonstrates the significant and largely preventable economic burden of SARS-CoV-2 infections among HCWs. The findings underscore the importance of robust infection prevention and control measures to minimize risks to HCWs and the consequent economic impacts on societies, particularly in LMICs. Future research should explore the long-term consequences of HCW infections on health systems, including workforce pipeline impacts and the mental health effects on HCWs. Further investigation into specific interventions to mitigate economic losses associated with health service disruptions is also crucial.

This study has several limitations. First, the estimation of secondary infections relied on a PAR approach based on data from a high-income country, which might not perfectly reflect the transmission dynamics in LMICs. Second, the data on HCW infections and deaths from South Africa did not fully include private sector data, potentially underestimating the overall economic burden. Third, some data were missing, requiring assumptions and extrapolations from available information. Fourth, the model may have underestimated the economic burden by not fully capturing costs such as long-term absenteeism, presenteeism, and mental health impacts. Fifth, the study did not fully account for mitigating actions implemented by countries to reduce service disruptions. Finally, the estimation of changes in HCW productivity relies on assumptions due to data limitations. These limitations suggest that the economic burden may be even larger than estimated in this study.