key: cord-0274728-sqrg7srd authors: Nagaraj, S.; Chandrasingh, S.; Jose, S.; B, S.; Sampath, S.; Krishna, B.; Menon, I.; Kundu, D.; Parekh, S.; Madival, D.; Nandi, V.; Ghatak, A. title: Effectiveness of a novel non-intrusive continuous-use air decontamination technology to reduce microbial contamination in clinical settings: A multi-centric study date: 2021-08-28 journal: nan DOI: 10.1101/2021.08.25.21262596 sha: 34c6f234d257860f371cdafbfb09fa9911868c74 doc_id: 274728 cord_uid: sqrg7srd Background: Despite rigorous disinfection, fumigation and air treatment, infectious microbial load has been found to circulate and survive for significant duration in health care settings. This raises significant concerns for hospital acquired infections. We have developed a novel, hybrid, trap-and-kill airborne-microbicidal technology called ZeBox which is efficient in clearing 99.999% of airborne microbial load under controlled lab conditions. In this study we evaluate the clinical performance of the ZeBox in reducing airborne and surface microbial load in two independent hospital settings. Methods: The studies were conducted in single bed and multi bed ICU of two hospitals. Airborne and surface microbial loads were collected at pre-determined sampling sites pre- and post-deployment of the ZeBox enabled device. The Normality of data distribution was determined using the Shapiro-Wilk test. Statistical significance was determined using Students T test and Mann-Whitneys U test. Pathogenic and opportunistic organisms were characterized using 16S rDNA sequencing. Furthermore, the antibiotic sensitivity of the isolated organisms was tested against current treatments of choice across major antibiotic classes. Results: Post-deployment, we found statistically significant reductions in both airborne and surface microbial load within the operating range of the ZeBox enabled technology . Across the both hospital ICUs, there was 90% reduction of airborne microbial load on average, and 75% reduction of surface microbial load on average, providing a low bioburden zone of roughly 10-15 feet diameter around the unit. These reduced microbial levels were maintained during the entire duration of device operation over several weeks. Many of the clinical isolates recovered from one of the hospitals were drug resistant, which highlighted the potential ability of ZeBox to eliminate drug-resistant microbes and thereby reduce the frequency of hospital acquired infections. Conclusions: ZeBox enabled technology can significantly reduce a broad spectrum of microbial burden in air and on surfaces in clinical settings. It can thereby serve an unmet need in reducing the incidence of hospital acquired infections. and surface microbial loads were collected at pre-determined sampling sites pre-and post-23 deployment of the ZeBox enabled device. The Normality of data distribution was determined 24 using the Shapiro-Wilk test. Statistical significance was determined using Students' T test and 25 Mann-Whitney's U test. Pathogenic and opportunistic organisms were characterized using 16S 26 rDNA sequencing. Furthermore, the antibiotic sensitivity of the isolated organisms was tested 27 against current treatments of choice across major antibiotic classes. Results: Post-deployment, we found statistically significant reductions in both airborne and 29 surface microbial load within the operating range of the ZeBox enabled technology . Across the 30 both hospital ICUs, there was 90% reduction of airborne microbial load on average, and 75% 31 reduction of surface microbial load on average, providing a low bioburden zone of roughly 10-15 32 feet diameter around the unit. These reduced microbial levels were maintained during the entire build-up of pathogens over time, particularly those that are resistant to surface or terminal 1 room disinfection and can pose a significant hazard to the next patient [13] . Under typical 2 heating, ventilation and air conditioning (HVAC) found in hospitals, Clostridium difficile 3 spores, Vancomycin resistant Enterococcus (VRE), Methicillin resistant Staphylococcus 4 aureus (MRSA) and Acinetobacter baumannii have been recovered after 4-5 months with 5 surface contamination levels exceeding the number of bacteria or virions necessary for the 6 transmission of infection [14, 15] . Causing even more concern for nosocomial spread, it has 7 been found that Pseudomonas can linger on surfaces for as long as 16 months [14] . Hospitals, dental clinics, nursing homes and long-term care facilities typically see a large 10 burden of pathogenic organisms posing a health risk to all occupants. Microbial 11 contamination in hospital wards is concentrated in hard-to-reach surfaces such as the floor 12 under beds and bed wheels as compared to higher levels of a room. This correlates both with 13 the source of infection (patients in beds) and the fact that air trapped under beds and 14 instruments is not efficiently cycled through wall mounted air purification units. There is a 15 pressing need to design microbial decontamination devices that function near microbial 16 reservoirs. 17 18 In dental clinics, aerosols generated through drills and scalers can potentially splatter or 19 aerosolise and move within the indoor environment. Body fluids or blood from patients and long-term care-home residents have shown that infections account for 27% to 63% of 1 hospitalizations in the United States [23] . Reducing the load of pathogenic organisms to below infectious level is thus crucial to 4 mitigate risk of infection, particularly in indoor spaces. The CDC recommends eliminating 5 microbes at the source as they are produced as the first line of defense against the spread of 6 infections [18] . This aspect has come into greater focus more recently with the rapid spread 7 of coronavirus disease across the globe. Indoor air decontamination is an urgent medical 8 need to maintain health and hygiene needs of occupants. 9 Currently available technologies for decontaminating room air belong to two broad 10 categories: those which merely trap suspended matter in air (inanimate dust particles along 11 with microbes) and those which are microbicidal. Each of these technologies have their 12 merits and demerits, which have been reviewed in considerable detail by others [24] [25] [26] [27] [28] [29] . 13 We believe that an ideal air decontamination technology must trap and then kill microbes 14 in situ, thus preventing any future growth and dissemination. While trap-and-kill 15 microbicidal technologies are already available-UV irradiated filters, filters made of 16 microbicidal fibers, and filters combined with plasma technology -they suffer from major 17 demerits regarding flow permeability (which determines power consumption) and 18 generation of toxic by-products during operation. We have developed a novel, hybrid, trap- 19 and-kill airborne-microbicidal technology called "ZeBox", which exploits the fact that microbicidal substrate and potentiates the substrate to instantaneously killing the trapped 1 microbes [30]. 2 In an enclosed test chamber under challenge conditions, ZeBox powered devices achieved 3 6-9-log10 reduction of a broad spectrum of microorganisms (airborne gram positive and 4 gram-negative organisms of ESKAPE group, viruses, vegetative fungi and spores) in 10 5 minutes, a performance that is at least 1000-fold superior to that reported in the literature. 6 In applications, which almost always consist of a space (enclosed or otherwise) with an 7 unceasing flux of people and patients, a continuous and rapid-action microbicidal device is 8 highly desirable. This is why the superior killing rate of ZeBox technology makes it unique 9 for continuous real-time applications. In this paper, we evaluate the clinical performance 10 of a ZeBox technology powered air decontamination device variant in reducing bacterial 11 and fungal load in air and on surfaces in two independent hospital settings. We also 12 delineate the typical pathogenic and opportunistic organisms found in these settings, to 13 characterize the risk of nosocomial transmission to patients and health care staff. 14 15 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The studies were conducted in a single bed ICU and multi bed ICU located in two 4 independent hospitals after approval from their Hospital Internal Ethics Committee. Both 5 rooms were mechanically ventilated with filtered and tempered air at 22.6±1.9°C with no 6 humidification. Housekeeping and nursing staff shared routine cleaning duties. Near-patient 7 sites were cleaned by nurses twice daily at 7 am and 7 pm using wipes (Vernacare Tuffie™ 8 wipes) and detergent (Hospec™). Terminal cleaning of the bed-space was performed 9 following discharge. Samples were collected at specific locations identified as sampling 10 sites between 2-3 pm, 7-8 hours post-cleaning of near-patient surfaces three or four times a 11 week. In the Single Bed ICU, samples were collected four times a week over 11 weeks for 12 determining baseline levels of contamination. The ZeBox powered air decontamination 13 device was deployed at the end of the 11th week and samples were collected as before for 14 another 10 weeks. 15 In the Multi bed ICU, samples were collected three times a week, over 13 weeks for 16 determining baseline levels of contamination. The ZeBox powered air decontamination 17 device was deployed at the end of the 13th week and samples were collected as before for 18 another 13 weeks. position S1 (medicine and reporting table) which was 10 feet away and position S3 (patient 1 bed rails) which was 4 feet away from the ZeBox technology powered air decontamination 2 device. 3 4 ii. Multi bed ICU 5 The room had a dimension of 30 feet x 90 feet. The sampling sites were chosen in 6 consultation with the ICU staff to ensure the deployed device did not hinder movement and 7 activities within the ICU. Sampling sites were selected such that two sites (positions 1 and 8 2) were proximal to the ZeBox technology powered air decontamination unit and served as 9 sites on which the direct effect of the device could be monitored. Two other sites (positions 10 3 and 4) were distal to the ZeBox technology powered air decontamination uni and served 11 as the control sampling sites (Figure 1b) . The deployed unit (marked by a green circle) 12 could effectively serve an area of 150sq. ft. 13 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. of 15x10 ft, samples were collected from positions S1 and S2 for quantification of total 6 bacterial and fungal population. Position S1 and S2 were 10 Feet and 6 Feet away from the 7 deployed device respectively. Sample positions S1 and S3 were 10 feet and 4 feet 8 respectively from the deployed device and used for collecting surface microbial samples. 9 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. A handheld air sampler (SAS Super ISO 100, VWR), which could sample 100 liters of air 10 per minute, was used to collect air samples. A fixed volume of air was sampled using the 11 air-sampler. Tryptic Soy Agar and Sabouraud dextrose agar plates were used to sample 12 bacteria and fungi, respectively, from the air. Plates were placed in and removed from the 13 air-sampler in an aseptic manner. Plates were incubated at 25±2 0 C (for fungal cultivation) 14 and 37±2 0 C (for bacterial cultivation) for 48 hours. After incubation, the number of 15 colonies were enumerated and converted to CFU/m 3 using statistical conversion provided 16 by the manufacturer. Control plates were used to ensure the sterility of the entire process. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint (PBS) solution. The entire 1ml solution was then plated on to Tryptic Soy Agar and 1 Sabouraud dextrose agar plates for quantification of bacteria and fungi, respectively. Plates 2 were incubated at 25±2 0 C (for fungal cultivation) and 37±2 0 C (for bacterial cultivation) for 3 48 hours. Post-incubation, the number of colonies that appeared were enumerated. Control 4 plates were used to ensure the sterility of the entire process. used as a negative control. PCR amplicons were sequenced at the The Bangalore Biocluster Information (NCBI). Sequenced genes were aligned using Clustal Omega 1 (https://www.ebi.ac.uk/Tools/msa/clustalo/) and taxonomical analysis were carried out 2 simultaneously. 3 4 E. Antibiotic sensitivity test of microbes collected from Multi Bed HICU 5 Single colony of each strain was grown in M9 medium. All test compound stocks and 6 dilutions were prepared in DMSO. Serial two-fold dilutions of antibiotics were prepared 7 separately, with concentrations ranging from 2 mg/mL to 0.015 mg/mL. To 150 µl (3-8 7×10 5 CFU/ml) of bacterial culture in 96 well microtiter plates, 3 μL compound from each 9 of the dilutions was added into respective wells to obtain final concentrations ranging from 10 40 µg/mL to 0.3 µg/mL of the test compounds. Media control, culture control and 11 appropriate reference drug controls were included. The plates were packed in gas permeable 12 polythene bags and incubated at 37 °C overnight. Growth was monitored by checking 13 absorbance at 600nm (A600). Minimum inhibitory concentration (MIC) was taken as the 14 concentration that resulted in a growth inhibition of ≥80%. 16 All data sets were tested for normal distribution using Shapiro-Wilk test (SW test), 17 following which a non-parametric test, the "Mann-Whitney's U test" (MWU test) was 18 conducted. The details are available in the Supplementary material. 19 20 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The environmental microbial load was monitored in a single bed ICU room when occupied 5 by patients. Air samples for monitoring baseline load were collected over a period of 11 6 weeks to enumerate microbial distribution at two locations within the room. This was 7 followed by ZeBox technology powered air decontamination unit deployment and sample 8 collection over a subsequent period of 10 weeks with the first sample collection within a 9 period of 3 hours after device deployment. 10 The airborne bacterial load before deployment was more or less similar at both the medicine 11 table and nurse station, and showed roughly a four-fold intra-day variation over the period 12 of 11 weeks ( Figure 2 , Table 1 ), ranging from 580-3000 CFU/m 3 (average 1168 CFU/m3) 13 at the medicine table (S1) and 80-1910 CFU/m 3 (average 1147 CFU/m3) at the nurse station 14 (S2). Similarly, the airborne fungal load before deployment showed roughly a three-fold 15 variation day to day over 11 weeks, but on some days, the CFU counts were as high as four 16 to seven-fold from the average daily counts ( Figure 3 , Table 1 ). Airborne fungal counts at 17 the medicine table (S1) ranged from 78-688 CFU/m 3 (average 157 CFU/m 3 ) and 72-698 18 CFU/m 3 (average 168 CFU/m 3 ) at the nurse station (S2) before device deployment. 19 After deployment of the decontamination device, the airborne bacterial load was reduced to Table 1 ). This accounts for a 92% reduction 22 of airborne bacterial load at the medicine table (S1) and a 91% reduction at the nurse station (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint Table 1 ). This accounts for 80% reduction of fungal load at the medicine table and a 75% 1 reduction at the nurse station. Both airborne bacterial and fungal load dropped significantly 2 within a period of 3 hours post deployment of the device. The device was in continuous 3 operation for the remaining duration of the study. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint line for both positions. The ZeBox powered air decontamination unit was deployed on day 1 72 and the first sample was taken within 3 hours of deployment. The microbial count for 2 that time point is depicted by the square. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Table 1 ). 21 This accounts for 84% reduction of airborne bacterial load at the medicine table and a 74% 22 reduction on the bed rails. 23 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 10 11 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint 1 a) Airborne microorganisms: 2 The environmental microbial load was monitored in a functional multi-bed HICU room 3 occupied by patients with regular movement of hospital personnel. Air samples were 4 collected as mentioned previously and total culturable microbial load was enumerated. 5 Baseline samples were collected over a period of thirteen weeks to understand the microbial 6 distribution at various positions in the room. This was followed by device deployment and 7 sample collection over a period of a further thirteen weeks. Depending on the position 8 sampled, the bacterial and fungal load in the air before deployment showed considerable 9 variability over time. Airborne bacterial load before device deployment ranged from 58-398 10 CFU/m 3 while the fungal load ranged from 14-130 CFU/m 3 across the four positions (Table 11 2). 12 After deployment of the decontamination device, the airborne microbial load was reduced 23 and Positions 4 (0-58 CFU/m 3 ). This accounted for 93-94% reduction of airborne fungi at 24 Positions 1 and 2, but only 51-53% reduction at Positions 3 and 4 ( Figure 5 , Table 2 ). 25 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. away from the air decontamination unit. The ZeBox powered air decontamination unit was 10 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint deployed on day 89 and the first sample was taken within 3 hours of deployment. The All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint Graphs for positions 1 (Fig. 8 a) and 3 (Fig 8 b) are shown here for comparison. Graphs 8 for position 2 and 4 can be found in the supplementary material. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint 1 Surface bacterial load before device deployment ranged from 6-620 CFU/cm 2 while the 2 fungal load ranged from 0-70 CFU/cm 2 across the four positions. After deployment of the 3 decontamination device, the microbial load was reduced to 0-180 CFU/cm 2 and 0-60 4 CFU/cm 2 for bacterial and fungal population, respectively, across the four sampling 5 positions. As before, the maximum reduction in surface bacterial load was shown at Position 6 1 (0-34 CFU/cm 2 ) and Position 2 (0-48 CFU/cm 2 ), which were 2 and 8 feet away from the 7 device, while a lowered reduction was observed at Positions 3 (2-120 CFU/cm 2 ) and 8 Positions 4 (2-180 CFU/cm 2 ), which were 24 and 26 feet away from the device. This (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint The Minimum inhibitory concentration (MIC) of each of seven antibiotics was tested 1 against each isolated organism. The seven antibiotics chosen represent current treatment 2 choices across the various classes of available antibiotics. The sensitivity of the organisms 3 to these antibiotics is tabulated in Table 3 . Among the clinical isolates characterized, several 4 strains were resistant to Ceftazidine, Azithromycin and Ampicillin. The isolates were 5 relatively more sensitive to meropenem and linezolid. Most isolates were highly sensitive 6 to Ciprofloxacin and Rifampicin. The Kytococcus and Micrococcus isolates seemed 7 resistant practically to all antibiotics tested and had a modest sensitivity towards Linezolid. 8 Brevundimonas could not be cultured for susceptibility testing. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint 1 Statistical analysis was carried out on the data set after leaving out the transition period, 2 which was one day after turning on the ZeBox powered air decontamination unit when the 3 microbial load should have settled into a new level of equilibrium. The Shapiro-Wilk test (SW 4 test) indicated that except for some data at position S1 (medicine table) We carried out two independent studies to determine the efficiency of proprietary ZeBox 21 technology powered air decontamination device in a single bed ICU and a multi bed ICU. 22 Both studies were carried out when the rooms were occupied by patients, there was expected 23 movement of hospital personnel and the hospital was fully functional. Devices underwent 24 electrical safety and emission testing as per IEC60601-1-2 standards before deployment in 25 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint data be normally distributed, we first tested the data for normality using the Shapiro-Wilk 1 test. We found that most of the datasets were not normally distributed. Therefore, to assess 2 the significance of the reduction in microbial load due to ZeBox powered air 3 decontamination unit, we conducted both a parametric test (Student's t-test) which is 4 applicable to normally distributed data, and a non-parametric test (Mann-Whitney's U test) 5 which is applicable to non-normally distributed data. Despite the different assumptions and 6 theoretical basis underlying the two tests, their conclusions were the same. Except for two 7 cases of surface microbes, one on the bed rail in the single-bed ICU and another at a location (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Our study demonstrates that the innovative ZeBox technology can provide an effective trap 5 and kill mechanism to eliminate a broad spectrum of airborne pathogens under clinical 6 conditions. This in turn prevents re-settling of bacterial and fungal microorganisms on [30] and Mycobacterium tuberculosis, bacteriophages such as MS2 phage and Phi X 174 24 (data not shown). Devices were previously shown to reduce 5log10 to 9log10 or 99.999-25 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint 99.9999999% of viable microbial load based on the starting concentration under challenge 1 conditions. In this study, we demonstrate that the ZeBox technology effectively eliminates 2 the microbial population present in normally functioning hospital environments with 3 efficiency over 95%, from the air and close to 85% from high contact surfaces like patient 4 bed rails. Reducing the environmental microbial load will reduce the occurrence of 5 nosocomial infections in healthcare environments. Although this study demonstrates the 6 device's capability in eliminating bacterial and fungal load from the environment, further 7 study is required to assess impact on viruses under clinical settings, especially respiratory 8 viruses. Nevertheless, this study successfully evaluates a novel decontamination technology 9 that can be used not only in hospitals ICUs but also in other areas such as burns units and 10 around immunocompromised patients, where the maintenance of low bioburden is critical 11 to maintaining good health and preventing difficult to treat infections. 12 13 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint The datasets supporting the conclusions of this article are included within the article. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint 3 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 24 25 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint 1 2 3 4 5 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Statistical tests were used to determine the significance of the reduction in microbial load due to 3 ZeBox. We divide the time series of measured microbial load into pre-deployment and post-4 deployment periods. After turning on ZeBox, there is a finite period of transition until the 5 microbial load settles to a new (lower) equilibrium level. The post-deployment period was 6 reckoned to begin at the end of the transition period, assumed to be 1 day after turning on ZeBox. Most statistical tests applicable to our case make two major assumptions: (1) The data in the pre-9 and post-deployment periods are statistically steady (or stationary); this implies that the statistics 10 of the microbial load within a given period does not change progressively over time, and (2) The 11 data is sampled from a normal (or, Gaussian) distribution. The first assumption is valid because 12 we have left out the transition period from consideration. Whether the second assumption is valid 13 was determined by testing the data for normality using Shapiro-Wilk test (SW test). The null 14 hypothesis for the SW test is that the dataset is sampled from a normal distribution. Table S1 and 15 Table S2 show the results of the SW test; each entry is a pair of the SW-statistic and the 16 corresponding p-value. All the statistical tests were done using the scipy-statistics package in 17 python. The cases with p-value > 0.05, marked in green, indicate that the corresponding datasets 18 are indeed sampled from a normal distribution. We see that most datasets are in fact not sampled 19 from a normal distribution. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Table S2 . SW test for multi bed data. Each entry is a pair of SW-statistic and the corresponding p-5 value. 6 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint If pre-and post-deployment datasets are both normally distributed then, to assess the 1 significance of the reduction in microbial load post-deployment of ZeBox, we may use a 2 parametric test such as the "two sample, left tailed t-test", otherwise a non-parametric test such as 3 the "Mann-Whitney's U test" (MWU test) is appropriate [1] . For the t-test, the null hypothesis is 4 that the mean microbial load in pre-and post-deployment periods are the same, and the alternative 5 hypothesis is that the ZeBox brings about a reduction in the mean microbial load (which therefore 6 requires a left-sided test). For the MWU test, the null hypothesis is that the pre-and post-7 deployment datasets are sampled from the same probability distribution, while the alternative 8 hypothesis is that they are sampled from different distributions and that the ZeBox brings about a 9 reduction in the microbial load (which therefore requires a left-sided test). 10 Results of the t-test for all the cases are shown in tables S3 and S4, and that of the MWU test in 11 tables S5 and S6. We see that the reduction brought about by deployment of ZeBox is significant Table S3 . Two sample, left-sided t-test for single bed data. Each entry is a pair of t-statistic and the 16 corresponding p-value. 17 18 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 28, 2021. ; https://doi.org/10.1101/2021.08.25.21262596 doi: medRxiv preprint The role of the 12 healthcare environment in the spread of multidrug-resistant organisms: update on current 13 best practices for containment. Therapeutic advances in infectious disease Hospital 17 respiratory protection practices in 6 U.S. states: a public health evaluation study Quantification of Mycobacterium tuberculosis and Mycobacterium avium complex in 6 human immunodeficiency virus-infected individuals with and without tuberculosis Preventing infections in nursing homes: a 11 survey of infection control practices in southeast Michigan Survival of bacterial and mold spores in air filter 16 media. Atmospheric Environment Ozone generation by indoor, electrostatic air cleaners. Aerosol 19 science and technology Review of bioaerosols in indoor environment with special 22 To test or not to test: Preliminary assessment of normality 6 when comparing two independent samples. BMC medical research methodology