The Economic Impact of Carbapenem Resistant-Non Lactose Fermenter and Enterobacteriaceae Infections on Hospital Costs in Dr. Soetomo General Academic Hospital Surabaya, Indonesia

Background: Carbapenem resistant-non lactose fermenter (CR-NLF) and Carbapenem resistant-Enterobacteriaceae (CR-E) bacterial infections are likely to be a global threat to people’s health. However, studies on the economic impacts according to the hospital setting are very scarce. The study aimed to explore the impact of CR-NLF (Acinetobacter baumannii = CRAB) & Pseudomonas aeruginosa = CRPA) and CR-E (Escherichia coli = CREC) & Klebsiella pneumoniae = CRKP) infections on hospital costs from a payer perspective among patients admitted to Dr.Soetomo Hospital, Surabaya, Indonesia. Methods: In the retrospective case-control study, medical records of all included patients hospitalized during 2018–2021 were reviewed for CRAB, CRPA, CREC, CRKP, and carbapenem sensitive (CSAB, CSPA, CSEC, CSKP) were collected. We retrieved the data of age, gender, clinical specimen, dates of admission, and discharge status. The outcomes of interest were hospital length of stay and hospitalization cost. Results: The cost for CR-NLFs infections was higher than carbapenem sensitive, $3026.24 versus $1299.28 (p < 0.05). There was no significant difference between CR-E against carbapenem sensitive. It showed that the highest impact of the cost was CRAB, followed by CRPA, CRKP, and CREC. The bed, antibiotics, pharmacy, and diagnostic costs of CR-NLFIs were significantly higher than CR-E. Conclusion: This study showed that the hospital cost and expenditure of CR-NLFs per patient were higher than CS. The hospital cost per patient for CR-NLF was higher than CR-E.


Introduction
Carbapenem resistant-Enterobacteriaceae (CR-E) and carbapenem resistant Acinetobacter baumannii (CRAB) are urgent threats [1] that cause severe healthcare-associated infections (HAIs) especially in critically ill and immunocompromised subjects [2], associated with significant morbidity, mortality [3], and increased healthcare costs [4]. Acinetobacter baumannii with minor frequency put a noteworthy burden on hospitals. The annual cost of CR-E is little was known about hospital outcomes and costs associated with HAIs, particularly CR-NLFs, Carbapenem sensitive-non lactose fermenters (CS-NLFs), and CR-E, Carbapenem sensitive-Enterobacteriaceae (CS-E). Therefore, this study was aimed to estimate and explore the impact of the carbapenem-resistant and sensitive non-lactose fermenter and Enterobacteriaceae infections on LoS and costs; and to estimate the impact of the COVID-19 pandemic on LoS and hospitalization costs related to carbapenem-resistant and sensitive non-lactose fermenter & Enterobacteriaceae infections among patients admitted in the Dr. Soetomo Hospital, Surabaya, Indonesia.

Results
The study started from January 2021 until August 2021. The data were extracted from the electronic medical record from March 2018 until February 2021. The study included 810 cases, and 270 (33.3%) were found carbapenem-resistant. There were 431 females, and 31.1% were resistant, while among males, 35.9% were resistant. This difference between two genders was insignificant (35.9% vs. 31.1%; p = 0.149). The median age was 55.0 (44.0-64.0) for sensitive cases and 53.0 (42.0-61.0) years for resistant cases, and the difference was insignificant, with a p-value of 0.070. Most common specimen were urine n = 567, 70% followed by blood n = 188, 23%, pleural fluid n = 22, 2.7% and CSF (cerebrospinal fluid) n = 15, 1.9%, and peritoneal n = 10, 1.2% and then few cases of others. Among clinical specimens, carbapenem-resistant cases were more frequent in CSF (46.7%), followed by blood (37.8%), urine (32.8%), peritoneal fluid (30.0%), and pleural fluid (13.6%). However, the frequency of carbapenem-resistant cases was not significantly different across different specimens, with a p-value of 0.054. The distribution of carbapenem-resistant cases was also not significantly different between organisms and their subtypes (Table 1). Both carbapenem-resistant and sensitive cases were categorized into four groups each, i.e., carbapenem-resistant EC + KP, carbapenem-resistant PA + AB, carbapenem-sensitive EC + KP, and carbapenem-sensitive PA + AB. All were found significantly different among four groups with p < 0.05. The carbapenem-resistant group with PA + AB had a significantly higher median cost of $3026.2 as compared to the carbapenem-sensitive group and had a significantly higher median cost as compared to the carbapenem-resistant group EC + KP with p-values < 0.001. When hospital cost was compared between two groups, the prepandemic (March 2018-February 2020) (n = 700) and pandemic (March 2020-February) (n = 110), irrespective of carbapenem status, the pandemic median cost of PA + AB and EC + KP was higher than the pre-pandemic but we did not find any significant difference in the cost ( Table 2). The median LoS was 15.0 (IQR, 9.0-26.0) days for carbapenem-resistant cases, which was significantly more prolonged than 11.0 (IQR, 7.0-20.0) days for carbapenem-sensitive cases (p < 0.001). The carbapenem-resistant cases' median bed cost was $249.1, for antibiotics $45.4, for pharmacy $601.5, for diagnostic $597.4, and $1536.4 was the total hospital cost and all were significantly higher as compared to sensitive cases with p-values < 0.05. The other costs were insignificantly different between the two groups, with a p-value of 0.933. (Table 3).
Both carbapenem-resistant and sensitive cases were categorized into two groups each, i.e., carbapenem-resistant EC + KP, carbapenem-resistant PA + AB, carbapenemsensitive EC + KP, and carbapenem-sensitive PA + AB. Median LoS and various costs were compared, and all of them were found to be significantly different between the four groups with p-values < 0.05. The carbapenem-resistant PA + AB had significantly higher median costs of bed $827.1, antibiotics $77.6, pharmacy $1147.3, diagnostics $783.8 and total treatment cost $3026.2 as compared to carbapenem-resistant EC + KP, carbapenem-sensitive PA + AB and carbapenem-sensitive EC + KP with p-values < 0.05. The cost of antibiotics for carbapenem-resistant EC + KP was also significantly higher than for carbapenem-sensitive EC + KP, with a median cost of $32.3 vs. $16.7 (Table 4).   The median costs of beds, antibiotics, pharmacy, diagnostics, and total hospital costs increased with the increase in LoS. The median total cost for a stay ≤5 days was $580.4, which reached $799.8 for a stay of 6-10 days, $1284.7 for 11-15 days, and 2802.7 for a stay of more than 15 days (Table 5) When the hospital cost was compared between two groups, i.e., COVID-19 prepandemic cases and pandemic cases, irrespective of carbapenem status, the median cost of antibiotics was $13.3 in pandemic cases, which was significantly lower compared to $26.8 in pre-pandemic cases (p = < 0.001). However, the median cost of diagnostics of $599.4 in pandemic cases was significantly higher than $479.6 in pre-pandemic cases (p = 0.034). Similarly, the median cost of other costs was $2.7 in pandemic cases, which was significantly higher than $2.4 in pre-pandemic cases (p = 0.004). Among carbapenem-sensitive, the median cost of antibiotics was significantly lower in pandemic cases than in pre-pandemic cases (p = < 0.001). Among carbapenem-resistant, the median costs of diagnostics (p = 0.019) and others (p = 0.048) were significantly higher in pandemic cases than the pre-pandemic cases. (Table 6).

Discussion
Carbapenem was recognized two decades ago as a very effective broad-spectrum antibiotic with fewer side effects. It is commonly used for multidrug-resistant strains treatment as the last choice [35,36]. In the β-lactam family, the carbapenems are important in anti-bacterial and distinctively steady in most β-lactamases [37]. The impact of carbapenemresistance (CR) has been examined earlier in GNI patients. More studies are needed to independently determine the economic cost of NLF (Acinetobacter bamuanai and Pseudomonas aeruginosa) and CR-E (Escherichia coli and Klebsiella pneumoniae) to show a more inclusive picture of infection cost. As compared to the west, Asia experienced more burden of Gram negative-resistant [38]. At the beginning of the 21st century, Enterobacteriaceae was reportedly identified in the USA and later on globally [39]. According to US Premier Healthcare Database (2014-2019), an analysis of CR demonstrated an increased cost of hospitalization than CS [40] and is associated with increased mortality, costs, and prolonged stay [41]. According to our knowledge, this is the first study conducted to evaluate outcome and hospital cost statistically to compare the financial burden of carbapenem-resistant versus carbapenem-sensitive GNIs in hospitalized patients in Indonesia. Therefore, this study aimed to estimate and explore the impact of carbapenem-resistant and sensitive Enterobacteriaceae and non-lactose fermenter infections on hospital outcomes and costs among patients admitted to the Dr. Soetomo Hospital Surabaya Indonesia. The present study showed that LoS and hospitalization costs were high in carbapenem-resistant versus sensitive. And CR-NLF infections LoS and hospitalization costs were higher than the CR-E. These results are in agreement with the findings reported in other studies. In a recently published systematic review and meta-analysis, Avendano et al. concluded that carbapenem-resistant GNIs in high-risk patients were associated with increased mortality, emphasizing antimicrobial stewardship and infection control in hospitals [42]. Likewise, Zhen et al. reported that CRKP, CRPA, and CRAB were associated with significantly higher hospital costs and prolonged hospital stay [10]. We found that some patients infected with Pseudomonas aeruginosa or Acinetobacter baumannii as a group of non-lactose fermenters were primarily a late infection or patients with chronic diseases and more therapeutic medicine. This also showed in this study that LoS patients infected with CRAB/CRPA are longer than the sensitive group. In the same way, Priyendu et al. found that the costs of overall treatment and hospital LoS were higher in the CRKP group than in the CSKP group. The authors concluded that carbapenem resistance could lead to a greater burden of treatment costs for patients [43]. Lemos et al. also calculated higher treatment costs in the CRAB group as compared to the CSAB group. However, LoS in the ICU and hospital mortality were not significantly different between CR and CS groups [44]. In the study, carbapenem-sensitive cases were more frequent in urine, followed by blood, pleural fluid, CSF, and peritoneal fluid. However, the frequency of carbapenem-resistant cases was not significantly different across different specimens. Luke et al. reported that the most frequent infections were respiratory infections, followed by urinary tract, systemic, and skin infections in overall CRKP and CSKP infections [45]. Among demographic characteristics evaluated in the study, age was not significantly different between the two groups by carbapenem status. However, gender male was significantly associated with carbapenem resistance. Whereas Zhen et al. found significant age and gender differences between CRAB versus CSAB and CRPA versus CSPA; these differences were not significant between CRKP versus CSKP [17]. When the costs were compared between cases before the COVID-19 pandemic and during the pandemic, it was observed that the antibiotic cost was lower during the pandemic than before the pandemic. Likewise, in Scotland, a study showed the same trends of decreased antibiotic prescribing during 2020 compared to pre-pandemic [46]. Comparing the bed, pharmacy, and total costs was lower, similarly Best et al. found that the revenue of American hospitals drops $16.3 to $17.7 billion due to canceled hospital services because of the COVID-19 pandemic [47]. Our study has several notable strengths. The present study is the first in Indonesia to examine the economic impacts of carbapenem resistance among patients with NLFs and Enterobacteriaceae. Comparatively, a study done in 2020 was a retrospective study, inherently resulting in bias and confounding. PSM (propensity score matching) was conducted to balance potential confounding factors to minimize the risk of bias, but the PSM ignores unmeasured confounders that could potentially impact outcomes [17]. But our study was case-control. We used case-control in this study design. A case-control study is always retrospective because it starts with an outcome and then traces back to investigate exposure allowing for meaningful comparison of CREC with CSEC, CRKP with CSKP, CRAB with CSAB, and CRPA with CSPA. Our study strengthens the case-control match study by using this method to eliminate the effects of confounding factors. However, the main potential benefit of matching in case-control studies is a gain in efficiency [48]. Some limitations are worth noting. Our study was conducted in a tertiary referral hospital and A-type hospital. The results may not be generalizable to other medical institutions and hospitals like B, C, and D types. It is necessary to expand this study to different types of hospitals in various areas in the future and also include clinical outcomes to provide inclusive sight related to infections cost and the impact of carbapenem resistance who were infected with NLF (Acinetobacter baumannii and Pseudomonas aeruginosa) and CR-E (Escherichia coli and Klebsiella pneumoniae).

Design and Population Study
We performed a retrospective case-control study with cost analysis using a payer perspective in a tertiary referral hospital Dr. Soetomo Hospital, Surabaya, Indonesia, with 1514 beds and 26 departments serving eastern Indonesia. The ethical committee of Dr. Soetomo Hospital approved this study (No. 0188/KEPK/IV/2021) dated 29th April 2021. Patients were identified from the records of the clinical microbiology laboratory patients. Data on medical/pharmaceutical/operating costs were obtained from electronic medical records, accounting department, and finance department records from March 2018 to February 2021.
We expected that NLF and Enterobacteriaceae infections due to resistance compared with sensitive bacteria were the independent exposure of interest. We estimated their impact on excess LOS and excess hospital charges.

Criteria for Selecting Patients
The criteria included those aged 18 years and above, including males and females, admitted to surgical, medical, and ICU wards. Patients with clinical samples positive for CRAB, CRPA, carbapenem-resistant Escherichia coli (CREC), and CRKP infection were included in the case group. We collected the control group from all carbapene-sensitive strains of patients, namely CSAB, CSPA, carbapenem sensitive Escherichia coli (CSEC), and CSKP. Patients with COVID-19 were excluded.

Data Collection
The data source was the patient's medical records. We included patient demographics such as age, sex and clinical specimen, hospital events admitting service, surgical and medical services, and hospital and ICU admission/discharge dates.

Cost Calculation
The retrospective data included age, sex, and hospital events (admitting service, surgical & medical services, dates of admission, discharge from hospital, and ICU). The study outcomes included hospital LoS and hospitalization costs such as bed charges, antibiotics, pharmacy, laboratory, radiology procedures, other medical facilities costs, and total costs covered by health insurance. Bed costs included hospital administration fees, daily room service, nursing and medical staff care, and technicians' services. We referred to the Pharmacy department to extract the antibiotics costs. Pharmacy costs covered many expenditures, including drugs, fluids, blood products for transfusion, disposable devices, mechanical ventilators, oxygen therapy, and pharmacy services. The service charge for Physiotherapists and consultancy fees were under bed service costs. The costs of other medical facilities were including of administration, ticket costs, ambulance, patient transfer, and others. The currency exchange rate 2021 (US $1 = 13,308.15 IDR) was used, values of the economic cost convert Indonesian Rupiahs (IDR) into US Dollars (US $) [49].

Outcomes
The primary outcomes were to estimate extended LoS due to CRAB, CRPA, CREC, and CRKP compared with sensitivity CSAB, CSPA, CSEC, and CSKP bacteria from the payer perspective and compare the differences in the results between cases and control.

Statistical Analysis
The data analysis was conducted using the IBM Statistics SPSS version 20.0. Gender, organism, specimen, and other drug sensitivity were shown with frequency and percentages. A comparison between two groups by carbapenem resistance was performed using Chi-square and likelihood ratio test. We used the median (IQR) for groups by carbapenem resistance to describe the status data of age, length of stay, and costs for different factors. The analysis between groups used the Mann-Whitney U test. The comparison among four groups by carbapenem status, organism type, and LoS using Kruskal-Wallis followed by Mann-Whitney for pairwise comparison. Comparison of the cost between pre-pandemic and pandemic period for all cases, between carbapenem-resistant and sensitive cases, used the Mann-Whitney U test. p-value < 0.05 was considered significant.

Conclusions
The present study concluded that length of stay and hospitalization cost was high in carbapenem-resistant versus sensitive and higher in CR-NLF than in CR-E infections. Furthermore, the costs of diagnostics and others were higher in pandemic cases than in pre-pandemic cases, irrespective of their carbapenem status among patients admitted to the Dr. Soetomo Hospital Surabaya Indonesia.