The daily direct costs of isolating patients identiﬁed with highly resistant micro-organisms in a non-outbreak setting

Background: Isolation precautions are recommended when caring for patients identiﬁed with highly resistant micro-organisms (HRMOs). However, the direct costs of patients in isolation are largely unknown. Aim: To obtain detailed information on the daily direct costs associated with isolating patients identiﬁed with HRMOs. Methods: This study was performed from November until December 2017 on a 12-bed surgical ward. This ward contained solely isolation rooms with anterooms. The daily direct costs of isolation were based on three cost items: (1) additional personal protective equipment (PPE), measured by counting the consumption of empty packaging materials; (2) cleaning and disinfection of the isolation room, based on the costs of an outsourced cleaning company; and (3) additional workload for healthcare workers, based on literature and multiplied by the average gross hourly salary of nurses. A distinction was made between the costs for strict isolation, contact-plus isolation, and contact isolation. Findings: During the study period, 26 patients were nursed in isolation because of HRMO carriage. Time for donning and dofﬁng of PPE was 31 min per day. The average daily direct costs of isolation were the least expensive for contact isolation (gown, gloves), V 28/$31, and the most expensive for strict isolation (surgical mask, gloves, gown, cap), V 41/$47. Conclusion: Using a novel, easy method to estimate consumption of PPE, we conclude that the daily direct costs of isolating a patient


Introduction
Patients admitted to a hospital can be placed in protective or source isolation during their admission. Protective isolation is used to protect immunocompromised (e.g., prolonged neutropenia) patients from getting a fungal or viral infection which is transmitted by air [1]. In this study, we focus on source isolation, which is used to prevent transmission of highly resistant micro-organisms (HRMOs) from carriers to other patients [1,2]. The HRMO involved determines the type of source isolation, including the recommended additional infection prevention and control (IPC) precautions (e.g., personal protective equipment (PPE)) [3e5]. In general, a distinction is made between universal IPC precautions that need to be applied to all patients, regardless of the presence or absence of HRMOs, and transmission-based precautions (IPC precautions on top of universal precautions), which differ per HRMO and thus per type of source isolation.
The World Health Organization (WHO) and Centres for Disease Control and prevention (CDC) formulated guidelines with recommendations on transmission-based precautions, including the selection and use of PPE in healthcare settings [1,6]. Local guidelines are usually derived from these international guidelines, but are often adapted to the local situations (e.g., design of the hospital, availability and costs), and based on findings in literature or on expert opinion [7]. In the Netherlands, the Inspectorate of the Ministry of Health, Welfare and Sport (VWS) audits the implementation of the guidelines from the Working Party on Infection Prevention (WIP) in hospitals [7].
Isolation costs are an addition to the basic cost of care, and would not have been made if the patient was not cared for in isolation. The direct costs of isolation can roughly be divided into the following three items: (1) the costs of using additional PPE; (2) the costs of cleaning and disinfection of an isolation room, during admission and after discharge, and (3) additional personnel costs, because healthcare workers (HCWs) need time for donning and doffing of PPE [4,8]. In this study, indirect costs, which are often costs due to loss of productivity of the hospital, such as a stop on new admissions on a room or ward, or using a multiple-occupancy room for isolation of only one patient, are not included [6,9,10].
The aim of this study was to obtain detailed information on the daily direct costs generated by isolating patients with HRMOs for different types of isolation. Furthermore, to facilitate extrapolation of our results to local situations and policies, we also provide an overview of recommended use of PPE per type of isolation as described by the most common international guidelines.

isolation rooms with anteroom
Number of empty packaging material per room Type of isolation per patient per room Number of isolation days per patient per room Step 1 Convert consumption per empty packaging material to consumption per piece of PPE (e.g. per gown or surgical mask) Step 2 Counting the consumption of PPE per type of isolation (regardless of the room) Step 3 Counting the number of isolation days per type of isolation (regardless of the room) Step 4 Total consumption of PPE per type of isolation / Total number of isolation days for same isolation type Step 5 = Average use of PPE per type of isolation

Study design and setting
This observational study was performed during a nonoutbreak period from November until December 2017, and conducted at a 12-bed surgical ward at the Erasmus MC University Medical Centre (Erasmus MC), Rotterdam, the Netherlands. This ward contained solely isolation rooms with anterooms. All patients received an isolation label in their electronic patient record, stating the required type of isolation and the isolation indication.

Types of isolation
Patients could be placed into contact isolation (when identifying Enterobacterales, Stenotrophomonas maltophilia, Pseudomonas aeruginosa, Streptococcus pneumoniae, or Enterococcus faecium, with specific resistance profiles), contact-plus isolation (when identifying carbapenemaseproducing Enterobacterales) or strict isolation (when identifying meticillin-resistant Staphylococcus aureus (MRSA) or resistant Acinetobacter spp.). The indications for the different types of isolation were according to the Dutch WIP guideline for HRMOs [7]. Contact-plus isolation was initiated and implemented by the Erasmus MC for patients identified with carbapenemase-producing HRMOs and is now being used by more Dutch hospitals. We compared the Erasmus MC policy with common IPC guidelines.

Data collection
In all 12 anterooms, we placed a waste bin for the disposal of PPE packaging materials. We informed all HCWs on the ward in multiple ways about the study. Additionally, on the lid of the waste bin we attached an instruction paper with pictures of the packaging materials that had to be thrown in (Supplementary material). We collected the following packaging materials: (1) empty glove boxes, (2) empty packages of disposable gowns, (3) empty packages of surgical masks, (4) empty packages of FFP2-masks, and (5) empty packages of disposable hair caps. Furthermore, at the start of the study, all open packages in the anterooms were replaced by new boxes.
During the study period every workday at around 15.00 h the contents of all 12 waste bins were collected, the amount of packaging materials was counted, and the type of isolation per patient was noted.

Data analyses
We calculated the number of isolation days per patient, using admission and discharge dates of patients admitted to the isolation rooms. Patients without isolation, with protected isolation, or with non-HRMO indication (e.g., viral infections or other contagious diseases) were excluded from the analyses.
The numbers of used packaging material were multiplied by the number of PPE per unit of packaging material ( Figure 1, step one). For glove boxes it was multiplied by 100, for disposable gowns by 10, for surgical masks by 50 and for disposable hair caps by 150. The consumption of PPE per type of isolation, and the number of isolation days per type of isolation, were counted ( Figure 1, steps two and three). The total consumption of PPE per type of isolation was divided by the total number of isolation days for the same type of isolation ( Figure 1, step four), resulting in the average use of PPE per type of isolation per day ( Figure 1, step five).

Calculation of cost items
To calculate the average daily direct costs of isolating a patient per type of isolation, we included the following costs items: (1) additional PPE, (2) cleaning and disinfection of the isolation room, and (3) additional workload for HCWs. Throughout the manuscript, we used the exchange rate of V1 ¼ $1.1387 (as at 16 th August 2018). All mentioned prices are without 21% VAT. Costs of additional PPE were calculated by multiplying the average use per day of PPE, per type of isolation, by the manufacturers' catalogue prices.
To calculate the costs of cleaning and disinfection we used the amounts the Erasmus MC pays to an outsourced cleaning company. We took the average cleaning costs of 35 weekdays and 12 weekend days, where we made a distinction between types of isolation but also between daily cleaning and cleanings after discharge of the patient.
Time of additional workload for HCWs was based on the study by Roth et al. [4]. They calculated that the extra time a nurse needs for donning and doffing of PPE, for a patient in contact isolation while using gowns, gloves and when applicable surgical masks, was approximately 31 min per day. The average gross monthly salary of a nurse, according to the collective agreement of the University Medical Centres 2015e2017 in the Netherlands, is V2632/$2997, resulting in an average gross hourly wage of V17/$19 [11]. This means V9/$10 per day of additional personnel costs for nurses who are caring for patients in isolation. We used the average 'extra workload' of certified nurses, nursing assistants and trainee nurses combined.

Ethics
The Medical Ethics Committee of the Erasmus MC agreed to the ethical requirements of this study, and decided that the study did not require approval according to the Dutch law on Medical Research in Humans (MEC-2015-306).

Guidelines
The recommendations, of the European Society of Clinical Microbiology and Infectious Diseases (ESCMID), WHO and CDC, on the use of PPE for contact isolation are almost the same (Table I). Only the ESCMID guideline does not recommend wearing a gown in contact isolation. The guidelines differ in their recommendations on using surgical masks and hair caps in strict isolation. The ESCMID guidelines do not provide clear recommendations on the use of PPE in strict isolation.

Patient characteristics
During the 47 days of observation, a total of 44 unique patients were hospitalized on the observed ward. Of these 44  Table IV gives an overview of the average daily consumption and costs of additional PPE. The mean (median; interquartile range) consumption of PPE for gloves was 33 (34; 34), gowns 8 (8; 5), surgical masks 6 (0; 14), and hair caps 3 (0; 0). The consumption of PPE for strict isolation per day was the most expensive (V17/$20) while the consumption of PPE for contact isolation was the least expensive (V9/$10).

Cleaning and disinfection of the isolation room
Cleaning and disinfection of the isolation room included daily cleaning while the patient was present, and cleaning and disinfection after discharge. The average costs of daily cleaning were V13/$14, ranging from V10/$11 for contact isolation to V15/$17 for strict isolation. The costs of room disinfection after discharge were V18/$20 for contact isolation, V20/$23 for contact-plus, and V25/$29 for strict isolation. The costs per type of isolation were based on the costs of cleaning products and on personnel costs, which equate to V26/$29 per hour.

Costs of isolation per patient
The average daily direct cost per patient were for contact isolation V28/$31, for contact-plus isolation V36/$41 and for strict isolation V41/$47 (Table V).
The additional direct costs per patient, for the entire isolation period, were for contact isolation V111/$126 (median of four isolation days), contact-plus isolation V504/$573 (median of 14 isolation days) and strict isolation V248/$283 (median of six isolation days).

Discussion
In this observational study on a surgical ward, with solely isolation rooms with anterooms, in a tertiary-care centre, we have shown that strict isolation is the most expensive (V41/$47 per patient/day), while contact isolation is the least expensive (V28/$31 per patient/day). This difference in costs can mainly be explained by the difference in consumption of PPE and only partly by the differences in cleaning. The overview with the most common IPC guidelines shows that guidelines sometimes lack recommendations or recommendations differ per guideline, which has an impact on the costs of isolation. In this study, the costs were calculated in detail for the Erasmus MC, which enables other healthcare organizations to alter the data and calculate costs for their own organization.
The method we used to estimate the consumption of PPE, was novel and less complex and labour-intensive than already published methods [4,8,12]. When comparing the costs of PPE for patients identified with MRSA, the costs of PPE in our study (V17/$20) were slightly higher than the costs mentioned by Spence et al.($14) [12]. In the Netherlands, patients identified with MRSA are nursed in strict isolation, while in the study of Spence et al., these patients were placed in contact isolation. Using surgical masks and hair caps in strict isolation, besides using gloves and gowns, partly explains the difference in cost. However, the different methods of measuring PPE consumption could have played a role as well. Souverein et al. showed costs (e.g., V18) and consumption (e.g., 35 masks/gloves and 15 gowns per day) of PPE for strict isolation, which are more in line with our study [13]. In the study of Verlee et al. the daily costs of contact isolation ($35) are higher than in our study (V28/ $31) [8]. Verlee et al. did not include the costs of cleaning and disinfection of the isolation room, but they did report 43 min of  daily excess staff time, which might explain the difference in costs of contact isolation. These studies emphasize that it is of utmost importance that authors clearly state which PPE are used per type of isolation and which cost items are included, because this has a great influence on the costs of isolation.

Strengths and limitations
A major strength of this study is that we used an easy to apply, quick method to estimate consumption of PPE by counting empty packaging material. This is in contrast to, for example, the method used by Roth et al. where they observed 10 patients for 24 consecutive hours [4]. We believe that counting the used empty packaging material of PPE per room, instead of extracting this from hospital accounting systems or estimating it by IPC staff, as done by Murthy et al. and Spence et al., gives valuable and more detailed information [12,14]. Another strength of this study is that daily direct costs were analysed per type of isolation. This provides insights that can be used for IPC policy decision making. A third strength of this study is that the costs per type of isolation were calculated in detail and step by step. This will facilitate translation of our findings to other settings, even though the data were based on Dutch IPC guidelines and on a limited number of patients and isolation days. Unlike our study, most studies focus solely on the costs of contact isolation or on the total costs of an outbreak, instead of defining the costs for all types of isolation [9,10,12,15].
A limitation of this study is that empty packaging material was counted instead of actual used PPE. If a package was not completely used during a patient admission period, the minimum number of used PPE was 0. Even though PPE had been used during this patient period, the empty package was counted with the next patient. However, as we report on group level (i.e. type of isolation), underestimation of actual used PPE is probably small. A second limitation is that we did not measure the time for donning and doffing ourselves. Instead we used the average time for donning and doffing reported by Roth et al. [4]. Moreover, when calculating the additional time for donning and doffing, we did not take into account the different types of isolation nor the different HCWs (e.g., physicians, residents, nurses) or visitors that enter the room daily. Verlee et al. also calculated the time for donning and doffing, for contact isolation using gloves and gown, of healthcare personnel [8]. They reported 43 min of daily excess time when entering an isolation room in comparison to the 31 min used by Roth et al. [4]. A third limitation of our study is that we did not include the indirect costs of isolation. Birgand et al., Montecalvo et al. and Otter et al. did include indirect costs caused by ward or bed closure, decolonization of patients, increased length of stay and admission stop on a ward [9,10,16]. Birgand et al. found that 69% of the overall mean cost, during an HRMO outbreak, was related to loss of hospital income due to a decrease in hospital activity [9]. Moreover, Otter et al. reported V822,000 of 'opportunity costs' (i.e. staff time, bed closure and elective surgical missed revenue) in comparison with  V312,000 of 'actual expenditure' (i.e. anti-infective costs, enhanced screening, contact precautions, temporary wardbased monitors of hand and environmental practice, and environmental decontamination) [10]. These studies show that early detection and isolation of HRMO-positive patients is cost-effective, even though long-term isolation and preventive screening are also costly. The fourth limitation is that we did not collect any information on which we can estimate the care burden of the patient. The care a patient needs has a direct effect on the consumption of PPE. Our data were not corrected for this, which could have led to an under-or overestimation of the calculated consumption of PPE per type of isolation. Evans et al. showed that patients in isolation often have a higher severity of illness than nonisolated patients [17]. However, other papers, including the study of Evans et al., also show that patients in isolation tend to get less and shorter attention of HCWs when compared with non-isolated patients [17e20]. To be able to calculate the costs of isolation even more accurately in the future, it is important to also take into account the care burden of the patient.
In conclusion, the direct costs of isolating a patient for one day differs per type of isolation, with strict isolation (i.e. isolation in single-patient room and using surgical mask, gloves, gown and cap) being the most expensive. Furthermore, in our study, the costs of PPE contributed the most to the differences in costs of isolation. Insight into the direct costs of isolation is of utmost importance when developing or updating IPC policies and to be able to perform cost-effectiveness analyses where costs and effects of IPC measures are studied.