The economic impact of the COVID-19 pandemic on health systems worldwide is a major concern, and there is an emergent need for additional resources and financial investments. The available hospital capacity, including hospital facilities, equipment, supplies and health professionals, have been significantly increased. Economic evaluations are thus essential to ascertain the health care-related resources and costs required for treating patients with this new disease.1

Available data suggest that 5–20% of patients with COVID-19 require hospitalization and between 14 and 20% of them require intensive care unit (ICU) admission.2-4 Underlying medical conditions and patient characteristics have already been shown to be associated with disease severity and outcomes. Comorbidities also affect resource and treatment requirements as they can result in longer length of hospital stay, mechanical ventilation or dialysis, among others.2,5,6

The setting of this study was the Hospital das Clínicas of the University of São Paulo Medical School (HCFMUSP), the largest public hospital complex in Latin America, with 2400 beds, located in the state of Sao Paulo, Brazil. The healthcare system in Brazil is composed of both private (paid by health insurance or out-of-pocket) and public (paid directly by the government) providers. The Unified Health System (UHS) represents the public provider, with universal access, and it is the single healthcare provider of more than 70% of the population.7 HCFMUSP is part of the UHS and plays an important role in quaternary care and high complexity emergencies in the state of São Paulo´s healthcare network, which was already operating at full capacity before 2020.8

HCFMUSP was located in the national epicenter of the pandemic in the first months of 2020,9 and the 900 beds of its Central Institute, including 300 ICU beds, 206 of which were newly installed in response to the pandemic, were entirely dedicated to COVID-19 patients referred for admission from the state's public healthcare network; HCFMUSP has delivered tertiary hospital care to more than 4000 severe COVID-19 patients.10-12 Due to the large number of hospitalized patients, the fast response to the disaster, the academic characteristics of the institution and its importance to the healthcare system, the results of the studies conducted at HCFMUSP are of great interest to national health policy makers. The centralization of care in a single institute provided a unique opportunity for a precise and homogeneous economic evaluation of these admissions. The aim of this study was to describe the direct and fixed hospital costs related to the treatment of hospitalized patients with COVID-19 as well as the main cost components and related clinical factors to support health policy makers in optimizing the use of financial resources.

Understanding the hospital costs of COVID-19 patients is essential to evaluating the economic impact of the pandemic on healthcare, providing important information for preparedness against and response planning for future risks, and improving knowledge regarding economic evaluation of global health emergencies.1 Since a significant number of patients with COVID-19 require hospitalization,2-4 an economic evaluation from the hospital's perspective is warranted and assesses an important part of the economic impact of the COVID-19 outbreak on the healthcare system.

Herein we report an average hospital admission cost of US$12,637.42, which is almost double the China's published cost of US$6827 from 70 empirically observed cases19 but is similar to the US$12,547 reported in a study from Saudi Arabia.20 However, when interpreting average admission costs, the disease severity profile of the population must be considered. In our study, the average cost of admission was US$20,002.80 for the hospitalized patients with severe disease who were admitted to the ICU17 (who accounted for 51.7%) and US$4839.57 for the other patients. In a study of COVID-19 patients admitted to two hospitals in New York City, 22% of the patients were considered critically ill.2 The population in our study resembles this group since 41% required invasive mechanical ventilation, 19% required hemodialysis, and almost 29% resulted in death.2,5 This increased severity is due to the existing risk-stratified healthcare network that was reinforced as part of the state´s COVID-19 pandemic response, with shelter hospitals for hospitalized patients with mild disease and referral units, such as HCFMUSP, designated for patients with severe conditions. Therefore, regulators transferred critically ill COVID-19 patients and those with complex underlying conditions requiring specialized care to HCFMUSP, where the ICU capacity was increased.10,11 As expected, in our study, the highest cost/day was incurred by ICU stay.

The correlation between pre-existing health conditions, age and disease severity was established previously.21,22 Here, we describe how these variables are also related to higher costs. An increasing trend was observed with the number of comorbidities, where compared to no comorbidities (US$10,522.72), 2 or 3 comorbidities (n = 1301) elevated the average admission cost by 16%, and >3 comorbidities (n = 452) increased it by 19%. These findings are also similar to Shandong´s report, with a 40% increase in hospital admission cost for patients with any pre-existing disease.19 Laboratory confirmation of SARS-CoV-2 infection was also independently associated with higher admission costs and may be related to disease severity, with a higher viral load being present in more severe cases.23,24 Death was the admission outcome related to lower costs, probably due to limiting the LOS by occurring early in the disease course (Supplementary Figure S2).

Human resources accounted for the greatest proportion of costs in all hospital sectors. In the ED, this component substantially elevated the cost/day, which can be partly explained by the ED needing to have medical and surgical specialties available 24/7 to assist with the specific urgent needs of inpatients from all sectors. The COVID-19 pandemic response increased the need for healthcare professionals since they were essential to capacity expansion. Competitive salaries had to be paid given the limited number of professionals, and this required a 46% increase in shift wages, which also increased overhead costs.25

Analyzing the costs of COVID-19 patient admissions and comparing them to costs incurred during other epidemics puts the costs of COVID-19 hospital admissions into a historical context. The average hospital admission cost of US$12,637.42 is similar to the model means of US$12,264, which was predicted using data for severe pneumonia hospitalizations due to H1N1 in the USA,26 and US$12,947, which was based on MERS-CoV data from Saudi Arabia.20 The reported hospital costs of seasonal influenza infections are higher; a study described them to be US$34,743 for patients aged between 20 and 64 years.27 Data about other epidemics showed that the cost of treatment of patients with pertussis ranged from $412–$55528, and the estimated cost for treating 11 patients with Ebola virus disease during the 2014 outbreak in West Africa was US$1.2 million in a treatment center in the USA.29,30

Economic evaluations are usually impacted by local conditions related to resource availability and the local market. However, in the COVID-19 pandemic, some aspects were due to global concerns, such as the simultaneous worldwide need for PPE and mechanical ventilators.31 Our study reinforces the impact of PPE on costs, as they totaled US$951,526.50.

Despite being a population with a high proportion of critically ill patients and some of them with complex underlying conditions, analyzing the costs of a large hospital sample according to detailed patient clinical and demographic characteristics makes it possible to apply this analysis to other populations and to other hospital settings since critically ill patients requiring mechanical ventilation, for example, were extensively present worldwide.2-4 In addition, the centralization of care in a single institute where no other disease was being treated during the study period ensured that the cost analysis of these admissions was precise, homogeneous and less subject to errors.

The UHS is known to be underfinanced, with gaps between the amount paid to providers and hospital costs.32 During the pandemic, hospital capacity expansion and treatment of patients with COVID-19 required additional financing,25 which could be described and analyzed in future studies since it was not the focus of our study.

Our study has some limitations since it took place in a single quaternary reference center, which could limit its generalizability, and was performed from the hospital perspective, excluding indirect costs and additional investment in fixed capacity expansion and equipment acquisition. No sensitivity analysis was performed, which limits the comparison of cost components to the few studies available to date about COVID-19 hospital costs. However, the study findings were analyzed with established, standardized, reproducible methods with the aim of supporting referral hospitals’ emergency preparedness. Therefore, the prospective collection of detailed clinical data, sector itineraries, outcomes, resources used and costs for a large sample of patient hospitalizations are strengths of this study and make the results applicable to other populations and settings.

In this single quaternary reference center, human resources accounted for the largest cost component, and ICUs accounted for the costliest sector. Total costs and average and daily estimated costs increased by 50% for older age strata, by 10-24% according to the number of comorbidities and by 24%-200% when additional therapeutic procedures were required; these costs decreased by 24% when the outcome was death.

Understanding the hospital costs related to COVID-19 admission and the economic impact of the disease across different population subgroups can support health policy makers in developing a comprehensive approach to hospital preparedness, decision-making and planning for future risk management.

This study was conducted without funding.

Deidentified participant data and a data dictionary defining each field in the set will be made available on request. They will be sent by the corresponding author after approval of a proposal with a signed data access agreement. Data will be available on publication.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.