22

Abstract
The Coronavirus disease 2019 (COVID-19) affect-

ed almost all activities worldwide. The medical sec-
tor was one of those which were most significantly 
impacted because the medical infrastructure was 
not sized for such a high scale shock, specialized 
human resources and medical infrastructure prov-
ing to be much undersized and with slow growth 
potential. Many changes were required, important 
financial resources being mobilized in order to mo-
tivate medical staff, offer treatments for the most 
severely affected patients, but also to create new fa-
cilities where the increasing number of sick persons 
could be cured. 

In our research we want to offer a hospital cost 
perspective based on empirical analysis of the 
COVID-19 impact on different categories of expens-
es made by Romanian hospitals that treated patients 
with COVID-19 in different stages of their disease. 
The period analyzed was January 2019 to December 
2020 on a monthly basis. Our results showed that 
expenses with goods and services, drugs, reagents 
and human resources are influenced by COVID-19 in 
a significant manner.

Keywords: hospital costs, COVID-19 impact, resil-
ience in hospitals, expenses.

THE IMPACT OF COVID-19 ON 
THE ROMANIAN HOSPITALS’ EXPENSES. 
A CASE STUDY TOWARD THE FINANCIAL 
RESILIENCE AFTER THE PANDEMIC 

Attila GYÖRGY 
Iliana SIMIONESCU

Attila GYÖRGY (corresponding author)
Associate professor, Finance Department, Bucharest 
University of Economic Studies and State Secretary, 
Ministry of Finance, Bucharest, Romania
Tel.: 0040-21-319.1900
E-mail: attila.gyorgy@fin.ase.ro

Liliana SIMIONESCU
Associate professor, Finance Department, Bucharest 
University of Economic Studies, Bucharest, Romania 
Tel.: 0040-21-319.1900
E-mail: liliana.simionescu@fin.ase.ro

DOI: 10.24193/tras.SI2021.2
Published First Online: 12/15/2021

Transylvanian Review 
of Administrative Sciences, 

Special Issue 2021, pp. 22–36



23

1. Introduction

The Coronavirus disease 2019 (COVID-19) affected unexpectedly the current ac-
tivities worldwide. In a short time, governments had to react in order to offer treat-
ment to those affected by this virus and mitigate its effects in almost all fields of ac-
tivity. Thus, governments introduced gradual measures as they managed to identify 
levers presumed to be effective; these measures ranged from the simplest to the most 
drastic lockdowns for quite significant periods of time.

Under the aforementioned unpredictable and uncertain environment, different ac-
tivities from various domains had to change and adapt in order to survive, to further 
offer the goods and services for which they were established, and to be in line with 
the needs of clients and beneficiaries.

The medical sector was one of those which were most significantly impacted be-
cause the source of the problem was the pandemic of coronavirus disease. The med-
ical infrastructure was not sized for a high scale shock, specialized human resources 
and medical infrastructure proving to be much undersized and with slow growth 
potential in comparison with the evolution of specialized medical needs. Hospitals 
were the frontrunners in this field, taking into account the needs for long-lasting 
treatment and recovery, in many cases with connectivity to oxygen facilities for in-
vasive or non-invasive ventilation. Thus, activities specific to infectious diseases have 
exploded, being offered in spaces that were not long ago used to treat other types of 
diseases. In order to realize such shifts, important financial resources were reoriented 
in this regard.

In our research we want to offer a slightly different perspective and fill a gap in 
the literature because ‘a detailed analysis of hospital cost structure remains an un-
explored area in the literature’ (Bai and Zare, 2020, p. 2807). Thus, we analyze the 
impact of COVID-19 over the main categories of expenses of all Romanian hospitals 
which were designated to officially hospitalize patients with symptomatic COVID-19 
and provide treatment for moderate, severe and critical forms of the disease in or-
der to explain the variance of hospital expenses, helping to understand how medical 
sector resilience can be faster and better fulfilled and offer a tool for predictability 
of expenses. In this regard, in the second section we presented the relevant liter-
ature which focuses on hospital costing, especially during pandemics. In the third 
and fourth sections we presented two case studies which highlight the impact of 
COVID-19 on different variables that explain the variance of hospital expenses using 
the OLS method with fix effects. In the fourth section we selected for each hospi-
tal, on a monthly basis, the expenses related to compensation of employees, use of 
goods and services, depreciation of fixed capital, drugs, sanitary materials, reagents, 
disinfectants, and laboratory materials, in order to compare with the monthly num-
ber of sick cases due to COVID-19, respectively with the number of deceased due to 
COVID-19.



24

2. Literature review

International Health Regulations adopted by the World Health Organization (WHO) 
define officially what public health emergencies of international concerns are (WHO, 
2016, p. 9). Since its revision in 2007, nine health events were counted: the influenza 
A (H1N1) pandemic, the Middle East respiratory syndrome coronavirus (MERS-CoV) 
outbreak, the international spread of poliovirus, the West Africa Ebola virus disease 
outbreak, the Zika virus outbreak, yellow fever, the 9th and 10th Ebola virus disease 
outbreak, and the on-going epidemic of COVID-19 (Mullen et al., 2020, p. 2).

COVID-19 generated a change in the activity of hospitals, the number of patients 
they treated, the ailments which needed medical support. All these caused changes in 
the level and structure of the expenses; all the main categories (staff remuneration, 
consumables, and fixes assets) were affected, but each category reacted differently.

In periods of severe pandemics, authorities can decide that selected hospitals will 
cease all day-to-day activities and shift completely to fight against the challenge rep-
resented by the new virus. As in the case of the influenza pandemic, this is happening 
because ‘hospitals will need to allocate limited healthcare resources in a rational, 
ethical, and organized way so as to do the greatest good for the greatest number of 
people. This can be done by deferring nonemergency care and, if necessary, institut-
ing alternative patient care routines’ (Toner and Waldhorn, 2006, p. 400).

Staff remuneration expenses increased in some cases because in many countries 
new health workers were hired (Hernández-Quevedo et al., 2020, p. 42), and in other 
cases bonuses were approved for existing workers. 

From the perspective of medical supplies, COVID-19 required larger quantities of 
oxygen for ventilation, specific drugs recommended for treatments, but also differ-
ent sanitary materials, reagents, disinfectants, and laboratory materials. For example, 
Al-Gheethi et al. (2020, p. 10) pointed out that ‘the survival of the virus on surfaces 
requires effective disinfection to ensure that the virus has become inactive’, a process 
which generates costs.

During COVID-19, respiratory therapy was an important component of healing 
many hospitalized patients. A similar situation was also due to 2009 H1N1 influenza, 
when Wiesen et al. (2012, p. 7) compared the consumption of resources, as measured 
by hospital charges, in the case of patients with acute lung injury (ALI) or acute re-
spiratory distress syndrome (ARDS) confirmed or suspected H1N1 infection vs. ALI/
ARDS arising from other etiologies (non-influenza group); the authors concluded 
that ‘respiratory charges are more likely a reflection of duration of mechanical venti-
lation rather than the degree of ventilator support necessary’. But, absolute intensive 
care unit (ICU) ‘charges for room and board, blood products, pharmacy, radiology, 
average daily charge, and overall charge per patient were larger in the noninfluenza 
group. ICU charges for blood products in the noninfluenza group were greater by a 
factor of four, and pharmacy charges double that of the H1N1 group. This finding 
is likely a reflection of the higher prevalence of underlying comorbid medical con-



25

ditions in the noninfluenza group, such as malignancy and cirrhosis, which require 
expensive medications and predispose to anemia. Moreover, the high mortality in 
this cohort likely precluded even higher hospital charges. Nevertheless, the H1N1 
cohort amassed charges of similar magnitude to the most ill and expensive patients in 
the ICU, indicating the abundant health care resources consumed by severe pandemic 
influenza infection’ (Wiesen et al., 2012, p. 7).

Fixed assets are the main component of the infrastructure used by medical units 
to offer services. Catană (2020, p. 172) highlighted the fact that the historical invest-
ments in the healthcare system in EU countries did not lead to a limitation of the 
number of deaths. 

In regard to the efficiency of measurement of medical outputs in a pandemic, 
it could be based on the number of sick cases and the number of deceased due to 
COVID-19. This approach was also used by Dan et al. (2009, p. 1911) when they quan-
tified how the virulence or case-fatality rate of a respiratory viral infection had a 
serious impact on the hospital infection control response using the actual number of 
deaths and ill persons.

3. COVID-19 impact on expense aggregates

In this section, our goal is to reveal the impact of the recent sanitary crisis on hos-
pitals, respectively on the expenses of Romanian hospitals most involved in treating 
COVID-19 patients. In this sense, our sample consists of all 10 healthcare units which 
hospitalized patients with symptomatic COVID-19 and provided treatment for moder-
ate, severe and critical forms of the disease, in accordance with the designation made 
by the Ministry of Health through Order no. 555 from April 3, 2020. The 10 hospitals 
included in our sample are representative for Romania, being the only hospitals con-
sidered by the Ministry of Health prepared from all perspectives to deal with moder-
ate, severe and critical forms of the disease. These hospitals are located in Baia Mare, 
Brașov, Bucharest (two hospitals), Cluj-Napoca, Constanța, Craiova, Iași, Suceava, and 
Timișoara. Nine of them are infectious disease hospitals, while the tenth was serv-
ing the area of Suceava declared in lock-down for a long period during 2020. From 
a geographical point of view, all regions were covered in a balanced way, although 
each hospital had to take cases from any region if some of these hospitals reached 
their maximum capacity. These hospitals were best prepared to fight COVID-19, being 
mainly situated in important medical university centers, having strong teams special-
ized in infectious diseases, and more than 4,000 beds in these hospitals (in some peri-
ods of time dedicated mainly, or even exclusively, to treat COVID-19 cases).

A limitation of our study regards the number of ICU beds due to the lack of data 
for each hospital. Although, the Ministry of Health named the aforementioned hospi-
tals to treat most affected patients with COVID-19, the number of ICU beds available 
for COVID-19 patients is absent. Thus, it was impossible to construct a correlation 
between the registered cases of COVID-19 and the number of ICU beds during the 
sanitary crisis.



26

3.1 Research methodology

Our case study starts by analyzing the impact of the COVID-19 evolution on dif-
ferent expense categories in the previously mentioned COVID-19 hospitals. The pe-
riod analyzed is on monthly basis starting from January 2019 until December 2020. 

The dependent variables of this research are different expense items classified 
taking into account the methodological guidance offered by the latest edition of IMF’s 
Government finance statistics manual (IMF, 2014). We offer a separate analysis on 
large aggregates composing total expenses (adjusted total expenses, constant human 
resource expenses, use of goods and services, and depreciation of fixed capital), but 
also distinct analyses for components of human resource expenses and different med-
ical supplies. 

The adjusted total expenses include all the expenses made by the hospital in the 
reference period, but exclude non-depreciable assets (non-depreciable assets in hos-
pitals are represented mainly by goods that belong to the public domain and lands). 
The exclusion is justified because these items are recorded with the whole value as 
expenses in a single month, although they are used on a long term basis. Due to this 
accounting treatment, the values from certain months may present significantly de-
viated values compared to the previous and following months.

Constant human resource expenses, use of goods and services, and depreciation of 
fixed capital were proposed as the main components of the adjusted total expenses. 
We opted for those aggregates because they are reflecting costs with very different 
types of resources used in the activity of hospitals. Besides, these elements are con-
secrated in financial research, since in the medical sector it is well known that ‘key 
hospital inputs are salaries, supplies and funding for capital investment’ (Hassan, 
2005, p. 131).

Medical human resources are very important in fighting COVID-19. For this rea-
son we analyzed the correlation of the main categories of expenses with the compen-
sations of hospital’s staff. Under ‘Constant human resource expenses’ we included all 
payable amounts in cash or in kind offered in return for work performed (salaries or 
resident doctor scholarships, including social security schemes and vouchers), while 
only the special allowance for treating COVID-19 were considered separately. This 
allowance was introduced in mid-2020 and it was paid only in certain months de-
pending on the available budgetary resources.

The quality of healthcare services is closely linked to the usage of medical supplies. 
In this regard we looked for the correlation of expenses with the consumptions of 
drugs, sanitary materials, reagents, disinfectants, and laboratory materials. We opted 
to emphasize and detail this structure having in regard the specific needs generated 
by COVID-19 from the perspective of personal protective equipment, respectively 
the perspective of the patients. In the category of procurements of personal protec-
tive equipment WHO issued very detailed technical specifications regarding gloves, 
goggles, face shields, masks, scrubs, aprons, gowns, hand rubs and so on (WHO, 2020, 
pp. 4–6). In the case of treatments, numerous drugs were recommended during the 



27

pandemic, a full treatment might have varied between less than 1 USD and several 
thousands of USD (Hill et al., 2020, p. 63).

The econometric method employed is OLS on panel data. The general equation of 
the model is as given bellow in equation 1:

Y
it
=α+βX1+δ

it
         (1),

where:
 – i = number of hospitals, thus i=1...10;
 – t = period analyzed, respectively monthly data, thus t=1…24;
 – Y – is the dependent variable consisting of different expense items; 
 – α – intercept term; 
 – β – coefficient of Xs;
 – X – independent variable, respectively proxy for COVID-19;
 – δ – error term.

All the data has been collected from the hospitals’ balance sheets prepared on a 
monthly basis according to the official accounting standards using accrual account-
ing principles. The variables used in the regression models are listed in Table 1.

Table 1: Variables used in the regression models and their acronyms

Variables Acronym Type of variable
Dependent variables

Adjusted total expenses LogATE Dependent
Compensation of employees LogCE Dependent
Use of goods and services LogUGS Dependent
Depreciation of fixed capital LogDFC Dependent
Constant human resource expenses LogCHRE Dependent
Special allowance for treating COVID-19 LogSA Dependent
Expenses with drugs LogED Dependent
Expenses with sanitary materials LogESM Dependent
Expenses with reagents LogER Dependent
Expenses with disinfectants LogEDI Dependent
Expenses with laboratory materials LogELM Dependent

COVID-19 variables

Type of hospital TH

Dummy (if in the analyzed period there was a 
hospital providing healthcare to patients tested 
positive for COVID-19 it had a value of 1 or zero 
otherwise) and independent

Number of monthly sick cases due to COVID-19 LogNB Independent
Number of monthly deceased due to COVID-19 LogND Independent

Source: Made by authors



28

Table 2 reveals the descriptive statistics for the variables considered in the regres-
sion models. Thus, the number of observations is between 86 and 90, while the mini-
mum value for COVID-19 variables (LogNB, LogND) are between 3.8 and 2.8 whereas 
their maximum values are 5.3 and 4.1 with a standard deviation of 0.5 and 0.4.

Table 2: Descriptive statistics

Variable Obs. Mean Std. Dev. Min. Max.
LogNB 90 4.544 0.541 3.846 5.369
LogND 90 3.540 0.426 2.841 4.197
LogATE 90 0.483 0.342 -0.132 1.764
LogCE 90 0.326 0.227 -0.172 1.219
LogUGS 90 0.915 0.843 -0.410 3.936
LogDFC 90 0.415 1.096 -8.771 2.326
LogCHRE 90 0.272 0.213 -0.151 1.196
LogED 90 0.251 0.734 -0.824 2.687
LogESM 90 11.612 10.967 -19.554 54.544
LogER 88 4.509 9.723 -1.111 82.640
LogEDI 89 5.341 14.541 -0.163 108.896
LogELM 86 1740.763 16065.76 -4.134 148995.6

Source: Made by authors

Table 3 shows the correlation matrices for variables considered in the regression 
models. In this paper we considered that coefficients with a value higher than 0.7 are 
highly correlated and therefore will not be used in the same regression model. 

Table 3: Pearson correlation matrices

LogNB LogND LogATE LogCE LogUGS LogDFC LogCHR LogED LogESM LogER LogEDI LogELM

LogNB 1
LogND 0.948 1
LogATE 0.45 0.44 1
LogCE -0.04 0.01 0.53 1
LogUGS 0.52 0.50 0.88 0.19 1
LogDFC -0.10 -0.13 -0.34 -0.12 -0.33 1
LogCHRE 0.08 0.12 0.52 0.89 0.26 -0.16 1
LogED 0.63 0.60 0.63 0.06 0.73 -0.22 0.12 1
LogESM 0.04 0.01 0.38 0.29 0.30 -0.10 0.32 0.01 1
LogER 0.19 0.19 0.42 0.10 0.55 -0.25 0.15 0.30 0.15 1
LogEDI -0.04 -0.05 -0.05 0.13 -0.07 0.06 0.13 -0.08 0.09 -0.05 1
LogELM -0.14 -0.21 -0.003 -0.02 -0.04 0.09 -0.04 -0.04 0.26 -0.007 0.01 1

Source: Made by authors



29

3.2 Regressions results and discussion

The aim of this research is to examine from an empirical point of view the impact 
of COVID-19 on different variables that explain the variance of hospital expenses us-
ing OLS method with fix effects. In this sense we considered a sample of 10 hospitals 
where patients with COVID-19 symptoms received treatment for different stages of 
the disease. The analyzed period was January 2019 to December 2020. 

When we consider the impact of COVID-19 on different categories of expenses 
in Table 4, the type of hospital has a negative impact on the adjusted total expenses. 
The number of deceased and sick cases due to COVID-19 virus has a positive impact 
on total expenses (model 1 and 2). The first 2 models from table 3 have an R-squared 
of 0.54, respectively 0.44. This means that the variation of type of hospital — TH and 
number of monthly deceased due to COVID-19 — LogND, explains the variation of 
adjusted total expenses in a proportion of 54%, whereas the variation of TH and the 
number of monthly sick cases due to COVID-19 — LogNB explains the variation of 
adjusted total expense in a proportion of 44%. In the models with Depreciation of 
fixed capital as dependent variable, TH did not reveal any statistical significance to 
explain the impact of COVID-19 on adjusted total expenses (models 7 and 8). More-
over, LogNB also showed no statistical significance whereas logND is statistically sig-
nificant at 10%. Therefore, the number of monthly deceased due to COVID-19 influ-
ences positively the depreciation of fixed capital. The model 6 explains the variation 
of dependent variables of 61% by the variation of independent variables. Moreover, 
the number of monthly deceased and the type of hospital are strongly significant, re-
spectively at 1%. Furthermore, TH has a negative impact on the use of goods and ser-
vices while LogND have a positive impact. The Governments around the world took 
measures in order to prevent the spread of COVID-19. The fact that our results reveal 
a positive impact between COVID-19 patients and hospital expenses, is underlined 
by the financial effort of the Romanian government. As our sample comprises all 
hospitals approved by the Romanian Government to treat patients affected by mod-
erate, severe and critical forms of COVID-19, our outcomes show that, in line with 
each country around the word, Romania also took measures to counteract the spread 
of this new pandemic virus, among which the increase of all the expenses within a 
hospital are imperative during sanitary crisis.

In table 5 we analyzed the impact on human resources expenses. We found a posi-
tive and statistically significant relation of TH on Constant human resource expenses 
and on Special allowance for treating COVID-19 as well. Models having as dependent 
variable Special allowance for treating COVID-19 have the greatest impact when TH 
and number of monthly deceased due to COVID-19 virus was considered (model 11). 
Furthermore, model 11 reveals that the LogND has a negative impact on LogSA. The 
number of sick cases due to COVID-19 has a negative and statistical significance on 
Special allowance for treating COVID-19, but the impact is less than LogND which 
showed a greater impact on dependent variables. The expense with human resources 
is statistically significant as in 2020 the Ministry of Health required Romanian gov-



30

ernment to allocate special allowance for medical personnel that work in the hospital, 
respectively departments that treat patients affected by COVID-19 due to risk expo-
sure. Subsequently, these results reinforce the allocation of this special allowance as 
well as law enforcement.

Table 5: Fixed-effects regression outcomes regarding the impact of COVID-19 on human resources expenses

(9) (10) (11) (12)
Log CHRE Log CHRE Log SA Log SA

TH
0.065* 0.073** 0.079* 0.085**

(2.341) (3.291) (1.240) (2.803)

LogND
0.012 -1.143**

(0.760) (-3.804)

LogNB
0.003 -0.704**

(0.521) (-3.197)

_cons
6.619*** 6.635*** 9.093*** 8.715***

(186.460) (372.172) (10.040) (9.088)
F stat 9.490*** 18.52*** 14.48*** 10.22***

R-squared 0.177 0.274 0.475 0.390
Obs. 100 110 27 27
N 10 10 10 10

Note: t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01

Source: Made by authors

Table 4: Fixed-effects regression outcomes regarding the impact of COVID-19 on category of expenses

(1) (2) (3) (4) (5) (6) (7) (8)
Log ATE Log ATE Log CE Log CE Log UGS Log UGS Log DFC Log DFC

TH
-0.059* 0.032 0.104*** 0.107*** -0.252*** -0.051 -0.053 0.006
(-2.098) (1.148) (3.871) (5.025) (-5.165) (-0.967) (-0.777) (0.120)

LogND
0.139*** -0.003 0.317*** 0.093*

(8.318) (-0.217) (11.013) (2.321)

LogNB
0.036*** -0.002 0.088*** 0.023
(4.368) (-0.374) (5.560) (1.419)

_cons
6.650*** 6.822*** 6.651*** 6.648*** 5.902*** 6.279*** 4.921*** 5.044***

(186.087) (306.858) (197.203) (391.888) (95.845) (148.928) (56.984) (114.933)
F stat 52.34*** 37.99*** 15.08*** 30.03*** 70.14*** 31.27*** 3.645*** 3.081**

R-squared 0.543 0.437 0.255 0.380 0.614 0.390 0.0765 0.0591
Obs. 100 110 100 110 100 110 100 110
N 10 10 10 10 10 10 10 10

Note: t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01

Source: Made by authors



31

From the point of view of the expenses regarding the use of medical supplies, the 
regression results having as dependent variable consumption of drugs, sanitary ma-
terials and reagent showed an R-squared of 55% (model 13), 51% (model 16) and 24% 
(model 18). These results are listed in table 6. In the last two models mentioned ear-
lier, correspondingly 16 and 18, LogNB is statistically significant at 1%, respectively 
10%. Thus, an increase in the number of sick cases due to COVID-19 would increase 
the consumption of sanitary materials and reagent. On the other hand, the expenses 
with sanitary materials are proven to have a higher impact than reagent as the co-
efficient of LogNB is greater in model 16. The model with the highest R2 from Table 
6, namely model 13, showed that LogND has a positive statistically significance on 
drugs consumption. Therefore, an increase of logND would enlarge logED with ap-
proximately 49%. Likewise, if the number of COVID-19 cases increases the expenses 
with sanitary materials, drugs and reagent enhances as more of this supply is used in 
order to treat and to prevent the virus.

Table 6: Fixed-effects regression outcomes regarding the impact of COVID-19 
on expenses concerning the use of medical supplies

(13) (14) (15) (16) (17) (18) (19) (20) (21) (22)
LogED LogED LogESM LogESM LogER LogER LogEDI LogEDI LogELM LogELM

TH
-0.539*** -0.223* 0.0401 0.102 -0.0007 0.168 -0.041 -0.020 0.343 0.343
(-6.590) (-2.542) (0.3572) (1.0772) (-0.005) (1.362) (-0.477) (-0.263) (1.410) (1.410)

LogND
0.496*** 0.141* 0.266** 0.0428 -0.146 -0.146

(10.278) (2.123) (3.075) (0.829) (-1.002) (-1.002)

LogNB
0.106*** 0.152*** 0.0814* 0.0821***

(3.988) (5.329) (2.194) (3.569)

_cons
5.198*** 5.934*** 5.332*** 5.015*** 5.009*** 5.293*** 4.640*** 4.377*** 4.317*** 4.317***

(50.418) (84.580) (37.674) (66.430) (27.121) (53.877) (42.035) (71.746) (13.803) (13.803)
F stat 53.87*** 8.462*** 6.288*** 51.78*** 10.38*** 15.36*** 0.363 15.17*** 0.996 0.996
R-sq 0.550 0.147 0.125 0.514 0.193 0.241 0.008 0.236 0.024 0.024
Obs. 100 110 100 110 99 109 100 110 92 92
N 10 10 10 10 10 10 10 10 10 10

Note: t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01

Source: Made by authors

4. COVID-19 severity impact on expense aggregates

This section proposes to emphasize the changes of hospitals’ expense items con-
sidering the severity of COVID-19. In order to take into account seasonality, the as-
sessment was made by comparing a certain month with the similar month of the pre-
vious year. The analyzed hospitals are the same as the ones described in the previous 
section. 



32

4.1 Research methodology

Data is presented on a monthly basis. The period analyzed covers April to Decem-
ber 2020 (months in which COVID-19 patients were treated in hospitals), percentage 
changes being computed having as reference April to December 2019. Dependent 
variables are the percentage changes of expense items described in the previous case 
study.

Table 7: Variables used in the regression models and their acronyms

Variables Acronym Type of variable
Dependent variables

Change in adjusted total expenses %ATE Dependent
Change in compensation of employees %CE Dependent
Change in use of goods and services %UGS Dependent
Change in depreciation of fixed capital %DFC Dependent
Change in constant human resource expenses %CHRE Dependent
Change in expenses with drugs %ED Dependent
Change in expenses with sanitary materials %ESM Dependent
Change in expenses with reagents %ER Dependent
Change in expenses with disinfectants %EDI Dependent
Change in expenses with laboratory materials %ELM Dependent

COVID-19 variables
Number of monthly sick cases due to COVID-19 LogNB independent
Number of monthly deceased due to COVID-19 LogND independent

Source: Made by authors

4.2 Regressions results and discussion

In order to examine if there were changes in hospitals expense categories, we ran 
regressions by comparing with 2019. Table 8 shows that LogNB influence positively 
%ATE and %UGS (model 1 and 5). The number of monthly deceased due to COVID-19 
variable has a positive impact on %ATE and %UGS in models 2 and 6, while the num-
ber of monthly sick cases due to COVID-19 has a negative impact on %CE and %DFC 
is not statistically significant (model 3 and 8). The increased expenses are associated 
with the number of sick and deceased persons due to COVID-19 as the Romanian 
Government made a great financial effort to stop (as much as possible) the spread of 
the virus. Therefore, the results confirm the Government financial effort in order to 
protect the population from getting infected with COVID-19. 



33

Table 8: Fixed-effects regression outcomes regarding the impact of COVID-19 
on category of expenses for changes in 2020 vs. 2019

(1) (2) (3) (4) (5) (6) (7) (8)
%ATE %ATE %CE %CE %UGS %UGS %DFC %DFC

LogNB
0.286*** -0.0144 0.806*** -0.195

(5.9400) (-0.3553) (7.5828) (-1.0003)

LogND
0.360*** 0.0102 0.984*** -0.318

(5.8652) (0.1969) (7.0954) (-1.2937)

_cons
-0.817*** -0.791*** 0.392* 0.290 -2.749*** -2.567*** 1.299 1.541†

(-3.7062) (-3.6153) (2.1085) (1.5787) (-5.6486) (-5.1930) (1.4598) (1.7582)
F stat 35.28*** 34.40*** 0.126 0.0388 57.50*** 50.34*** 1.001 1.674
R-sq 0.309 0.303 0.00160 0.000490 0.421 0.389 0.0125 0.0207
Obs 90 90 90 90 90 90 90 90
N Hospitals 10 10 10 10 10 10 10 10

Note: t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01

Source: Made by authors

In Table 9, the result regarding the impact of COVID-19 on human resource ex-
penses and use of medical supplies for 2020 vs 2019 are shown. The models with the 
highest R-squared are 11 and 12, respectively between 0.44–0.39. Variable LogNB and 
LogND are statistically significant at 1% level of significance and positively influence 
‘Change in medical expenses with drugs’ — %ED. Our results are supported by the 
literature in the field as well. Dalu et al. (2021) analyzed the Italian region Lombardy, 

Table 9: Fixed-effects regression outcomes regarding the impact of COVID-19 
on human resource expenses and use of medical supplies for 2020 vs 2019

(11) (12) (13) (14) (15) (16) (17) (18) (19) (20)
%ED %ED %ESM %ESM %ER %ER %EDI %EDI %ELM %ELM

LogNB
0.857*** -0.193 3.515* 0.204 -1012.1
(7.891) (-0.097) (2.081) (0.077) (-0.309)

LogND
1.029*** -1.590 4.449* -0.363 -1711.3
(7.173) (-0.632) (2.087) (-0.109) (-0.410)

_cons
-3.644*** -3.392*** 12.49 17.24† -11.44 -11.23 4.414 6.627 6340.6 7801.6
(-7.329) (-6.629) (1.3754) (1.9241) (-1.482) (-1.478) (0.367) (0.558) (0.424) (0.524)

F stat 62.27*** 51.45*** 0.009 0.400 4.332*** 4.357*** 0.006 0.011 0.096 0.168
R-squared 0.441 0.394 0.0001 0.005 0.053 0.053 0.00007 0.0001 0.001 0.002
Obs. 90 90 90 90 88 88 89 89 86 86
N Hospitals 10 10 10 10 10 10 10 10 10 10

Note: t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01

Source: Made by authors



34

respectively the Luigi Sacco Hospital which had, in the highest influx of COVID-19 
patients, a number of 280 beds out of which 30 in the Intensive Care Unit (ICU). The 
authors’ research underlines the importance of medical attention and drugs during 
the pandemic. Therefore, the increased expenses with drugs in the face of such dis-
ease which spreads exponentially is crucial in order to decrease the number of sick or 
deceased persons due to COVID-19.

5. Conclusions

Resilience after a huge shock such as COVID-19 and the predictability of expenses 
in hospitals is vital for a better financial management. If recent balance sheet data 
can be used in order to estimate future changes that will help hospitals prepare the 
organization for the future challenges that are expected to happen. 

Our study presented, for the first time as far as we are aware, the structural modi-
fication of different expense categories in public hospitals as a result of shifting their 
activity to fight exclusively against a pandemic. Having in regard that any expense 
category reacts differently and presents different correlation strength, we captured 
the connection of the most important expense categories with indicators reflecting 
the magnitude of the pandemic.

We found that the Romanian hospitals used in our sample, dedicated to treating 
the COVID-19 cases showed, from a statistical point of view, that adjusted total ex-
penses is positive and stronger related with the number of monthly deceased persons 
than with the number of monthly sick cases due to COVID-19. Similarly, a positive 
and statistical significant relation has been identified between the use of goods and 
services with the number of monthly deceased due to COVID-19. From the human re-
sources expense point of view, the independent variable that underlines the number 
of monthly deceased due to COVID-19 confirms the previous results, respectively it 
has been shown to be statistically significant at a 5% level but it has a negative impact 
on the expenses with special allowance. In these models both variables that are used 
for COVID-19, namely the number of sick persons and deceased due to the pandemic 
are statistically significant and have a negative impact on the special allowance re-
ceived by the hospital personnel. 

However, the model where the number of deceased persons due to COVID-19 had 
been considered better explains the variance of special allowance expenses, respec-
tively with approximately 48%, representing the highest value.

Regarding the medical supplies, only expenses with drugs and sanitary materi-
als seem to be strongly significant with COVID-19 outputs, while expenses with re-
agents, disinfectants, and laboratory materials were not directly influenced by the 
magnitude of the pandemic.

Furthermore, when we compared the changes in expenses, the impact of Covid-19 
over the period of 2020 vs. 2019, our results were reinforced by the previous out-
comes, respectively the expenses by category where the number of monthly sick 



35

cases and deceased due to COVID-19 has a positive impact on %ATE and %UGS at 
1% level of statistically significance. Also, the outcomes of the regressions revealed 
that the models with the monthly sick cases variable better explain the variance of 
expenses (adjusted total expenses and the use of goods and services). On the other 
hand, the results concerning the medical supply expenses hold only for drugs where-
as the models with sanitary materials as dependent variable barely showed values for 
R-squared (0.01% and 0.5%). Instead, models having as dependent variable expenses 
with reagents are influenced by the number of monthly sick cases and deceased at the 
same level of statistically significance and R-squared.

Lastly, the present paper reveals the measures and financial effort made by the Ro-
manian Government in order to prevent the sanitary crisis. The empirical outcomes 
may be biased since the selected balance sheets refer to entire hospitals and not only 
the departments and/or sections concerning COVID-19. However, the aggressive 
transmission of the virus by air led to the decision that, in the vast majority of time, 
the sampled infectious disease hospitals were dedicated exclusively to the treatment 
of COVID-19.

References:

1. Al-Gheethi, A., Al-Sahari, M., Malek, M.A., et al., ‘Disinfection Methods and Survival of 
SARS-CoV-2 in the Environment and Contaminated Materials: A Bibliometric Analysis’, 
2020, Sustainability, vol. 12, no. 18, art. no. 7378. 

2. Bai, G. and Zare, H., ‘Hospital Cost Structure and the Implications on Cost Management 
during COVID-19’, 2020, Journal of General Internal Medicine, vol. 35, no. 9, pp. 2807–2809.

3. Catană, Ş.A., ‘Coronavirus (Covid-19) Pandemic – Effects and Correlations in the Europe-
an Union Countries’, 2020, SEA - Practical Application of Science, vol. 8, no. 23, pp. 169–174. 

4. Dalu, D., Rota, S., Cona, M.S. et al., ‘A Proposal of a “Ready to Use” COVID-19 Control 
Strategy in an Oncology Ward: Utopia or Reality?’, 2021, Critical Reviews in Oncology/
Hematology, vol. 157, pp. 1–6.

5. Dan, Y.Y., Tambyah, P.A., Sim, J., et al., ‘Cost-effectiveness Analysis of Hospital Infection 
Control Response to an Epidemic Respiratory Virus Threat’, 2009, Emerging Infectious 
Diseases, vol. 15, no. 12, pp. 1909–1916.

6. Hassan, M.K., ‘Management Accounting and Organizational Change: An Institutional Per-
spective’, 2005, Journal of Accounting & Organizational Change, vol. 1, no. 2, pp. 125–140.

7. Hernández-Quevedo, C., Scarpetti, G., Webb, E. et al., ‘Effective Contract Tracing and the 
Role of Apps: Lessons from Europe’, 2020, Eurohealth, vol. 26, no. 2, pp. 40–44.

8. Hill, A., Wang, J., Levi, J., Heath, K. and Fortunak, J., ‘Minimum Costs to Manufacture New 
Treatments for COVID-19’, 2020, Journal of Virus Eradication, vol. 6, no. 2, pp. 61–69.

9. International Monetary Fund, Government Finance Statistics Manual, Washington, D.C.: 
International Monetary Fund, 2014. 

10. Mullen, L., Potter, C., Gostin, L.O., Cicero, A. and Nuzzo, J.B., ‘An Analysis of International 
Health Regulations Emergency Committees and Public Health Emergency of International 
Concern Designations’, 2020, BMJ Global Health, vol. 5, art. no. e002502. 



36

11. Toner, E. and Waldhorn, R., ‘What Hospitals Should Do to Prepare for an Influenza Pan-
demic’, 2006, Biosecurity and Bioterrorism: Biodefense Strategy, Practice, and Science, vol. 4, 
no. 4, pp. 397–402.

12. Wiesen, J., Komara, J.J., Walker, E., Wiedemann, H.P. and Guzman, J.A., ‘Relative Cost and 
Outcomes in the Intensive Care Unit of Acute Lung Injury (ALI) Due to Pandemic Influen-
za Compared with other Etiologies: A Single-center Study’, 2021, Annals of Intensive Care, 
vol. 2, no. 41, pp. 1–8.

13. World Health Organization, International Health Regulations (IHR), Geneva: WHO, 2016.
14. World Health Organization, Technical Specifications of Personal Protective Equipment for 

COVID-19, Geneva: WHO, 2020.