Cleaning and disinfecting surfaces in hospitals and long-term care facilities for reducing hospital and facility-acquired bacterial and viral infections: A systematic review

Primary research question

How effective are bleach, ultraviolet light, hydrogen peroxide, copper surfaces, or copper treated fabrics at cleaning and disinfecting surfaces in hospitals and LTCFs and do they make a difference in reducing acquiring MDROs and HAIs? Our PICO statement was: Population: Patients in hospitals or LTCFs. Interventions: bleach, quaternary ammonium disinfectants, liquid or vapourised hydrogen peroxide, ultraviolet lighting, copper-plated surfaces or copper-treated fabrics, or disinfectant or antiseptic coated surfaces. Comparison: usual cleaning methods in the institution (or quaternary ammonium disinfectant if it is the usual method). Outcomes: rates of new MDROs and HAIs in patients. Study Design: randomized or cluster-randomized controlled trials.

Secondary research questions

(1) What are the reported MDRO and HAI rates in the region of the institution? (2) Did the study follow or test evidence-based disinfection guidelines? (3) Did the researchers perform genomic studies to track the transmission of pathogens between the community, hospitals and LTCFs? (4) Were antibiotic stewardship programmes and their effectiveness reported? and (5) Were MDRO and HAI rates reported for environmental service workers (ESWs) and other HCWs?

Data Collection

Five databases (CINAHL, Cochrane CENTRAL Register of Controlled Trials, EMBASE, Medline, and Scopus) were searched from inception to June 28, 2021. Searches combined keywords and database-specific subject headings from three concepts: (a) hospitals or acute/long term care facilities (e.g.: assisted living, long term care facilities, nursing homes) (b) disinfectants (e.g.: antisepsis, bleach, cleaning, copper plating of surfaces and copper impregnation of textiles, disinfection, decontamination, hydrogen peroxide, quaternary ammonium compounds, and ultraviolet (UV) light); and (c) randomized controlled trials. The Cochrane Collaboration’s highly sensitive search filter was used to identify relevant randomized controlled trials [6]. No language or date limits were applied. The reference lists of included studies were also searched to identify additional studies of relevance. The complete search strategy is documented in Appendix A.

Study Screening

Search results were downloaded into Covidence™ [7] for de-duplication and screening. Two authors screened abstracts and full-text papers in duplicate. Disagreements were resolved through consensus. Studies were included if they were randomized or cluster controlled trials that: (1) focused on patients in hospitals or LTCFs; (2) reported on the use of bleach, quaternary ammonium disinfectants, hydrogen peroxide, ultraviolet lights, or copper-plated surfaces/copper-treated fabrics or antiseptic coated surfaces to disinfect surfaces or reduce colonisation rates in patients; and (3) reported rates of patient infection or bacterial colonisation of surfaces.

Data Extraction and Quality Assessment

Two authors independently extracted study data and assessed the quality of all included studies. The Cochrane Risk of Bias Tool (RoB) [6] and also the Risk of Bias tool version 2 for cluster-randomized controlled trials (RoB 2) [8] were used to assess the quality of studies included in this review. All disagreements were resolved through consensus.

Data Analysis

Two authors independently analysed study data. There were insufficient studies with enough similar study arms to permit meta-analysis.

Search Results

The search identified 14 cluster -randomized controlled trials (c-RCTs) with clinical HAI and/or MDRO patient outcomes (ten with interventions focused on multiple bacterial pathogens, three on MRSA and one on C. difficile) [9-30] (Table I).

14 C-RCTs: Numbers of patients studied

Of the 14 c-RCTs, eleven reported the number of patients (total 329,254 patients), Mitchell [20] reported 3,434,439 occupied bed days but not patient numbers, Boyce also reported on 22,231 occupied bed days but not patients [14] and Ray reported basing the power computation on 1,068 cases and 1,683,928 patient days in the study facilities in 2009 but reported neither patients nor occupied bed days for the actual study period [30]. The largest studies by far were Anderson (31,226 patients exposed to MDROs and a total of 329,254 in all hospitals studied) [9-13] and Mitchell (3,534,439 bed-days) [20].

10 C-RCTs with patient outcomes and studied multiple infections

Ten of the c-RCTs studied varying interventions with outcomes of acquisition of MDROs or HAIs and three (Anderson [9-13], Mitchell [20-22], and Salgado [25]) found decreased HAIs or MDROs or both after cleaning of surfaces as a major component and one (Mody [23, 24]), with enhanced education. The largest c-RCT (Anderson [9-13] studied 31,226 patients in nine hospitals in the southeastern US who were admitted to single rooms which had been terminally disinfected after the discharge of a previous occupant with positive cultures for any of four target microorganisms (MRSA, VRE, C. difficile, or multidrug-resistant Acinetobacter) in the previous 12 months. The 11 hospitals had a total of 314,819 patients and these patients were also assessed for wider circulation of MDROs. The new admissions were randomized to rooms which had received one of four disinfection strategies. After completion of the interventions to reduce the environmental presence of the target MDROs and C. difficile, a primary composite outcome of new hospital-acquired colonisation or infection with these microorganisms was studied. The lowest rates of MDRO/C. difficile acquisition for the new admissions were after using ultra-violet light (UV) + quaternary ammonia detergent (QUAD) with 33.9 cases/10,000 exposure days which was statistically significant result (RR 0.70, 95% CI 0.50–0.98; p=0·036) compared to the standard cleaning arm. Cleaning with bleach revealed 41.6 cases/10,000 exposure days, with UV + bleach 45.6 cases/10,000 exposure days and then for the reference study arm (QUAD) 51.3 cases/10,000 patient days. For C. difficile infection alone, for bleach and for UV + bleach, there were 30.4 cases and 31.6/10,000 exposure days, respectively. Staff hand-hygiene compliance averaged 90%, room disinfection compliance 100%, and protocol compliance was QUAD 100%; and in study arms UV arm 93%, Bleach arm 96%, Bleach + UV arm 99%; and UV device use arm 92% [QUAD is also referred to as QUAT in some publications] (Table I).

The next largest c-RCT (Mitchell) was conducted in 11 acute care hospitals in Australia. The number of patients involved was not stated but the study involved 3,534,439 occupied bed days. The focus was on increasing cleaning thoroughness with the REACH bundle: using optimal cleaning agents, following manufacturers’ recommendations, defined cleaning frequency, training, audit, feedback, and a hospital-wide commitment to improve cleaning. There was a significant reduction in the primary outcomes of VRE clinical infections from 0.35/1,000 occupied bed days to 0.22 but no significant decrease in MRSA or C. difficile infections. There were significant but insufficient increases in the secondary outcome of frequent touch points cleaned in bathrooms from 55% to 76% and in bedrooms from 64% to 86% [20-22].

Salgado randomized 614 patients in ICUs in three hospitals in South Carolina either to rooms with copper surfaces on high touch surfaces (bedrails, overbed tables, IV poles, arms of the visitor’s chair, and in some the nurse call button and computers) or to rooms without copper plating. Forty-six patients (7.5%) developed HAIs and 26 (4.2%) became colonised with MRSA or VRE. For the composite primary outcome of HAI or colonisation with MRSA or VRE the rate was significantly lower in copper-plated rooms (0.071 vs. 0.123, p = 0.02). For HAI only, the incident rate was reduced from 0.081 to 0.034 (P = .013). Copper surfaces were ∼ 6 times less likely to harbour MRSA or VRE and the combined MRSA and VRE burden was 96.8% lower. Hand hygiene compliance ranged from 61% to 95% and was not significantly associated with HAIs [25].

Mody in a three-year study randomized 418 residents in 12 nursing homes in southeast Michigan who had indwelling devices either to the intervention group in which HCWs were encouraged to perform hand hygiene before and after providing any care to participants and wear gowns and gloves when performing morning and evening care, or to a usual care control group. No specific new cleaning agents or disinfection protocols were deployed in this study but rather enhanced education and precautions. Training consisted of 10 educational modules every 2-3 months over 3 years. There were no significant reductions in the intervention group for the primary outcome related to overall prevalence density of MDROs: MRSA, VRE or resistant gram-negative bacilli organisms. The secondary outcome of any device associated infection revealed significant reductions in first catheter-associated urinary tract infection occurrence (RR 0.54, 95%CI 0.30 to 0.97) and all catheter-associated infections (RR 0.69, 95%CI 0.49 to 0.99) but not for new feeding tube–associated pneumonias or skin and soft-tissue infections [23,24].

Six c-RCTs which aimed to reduce multiple MDRO acquisitions or HAI infections did not find significant reductions in their respectively designated primary outcomes. Boyce in a study of four wards on two campuses of a university-affiliated hospital in Connecticut did not report the number of patients but a total of 22,231 patient days. The wards were randomized each month during the 12-month c-RCT either to use liquid hydrogen peroxide or continue usual daily and terminal cleaning with quaternary ammonium disinfectant (QUAD). There were non-significant decreases in the primary composite outcome of patients with MRSA and VRE colonisation or infection and C. difficile infections. After peroxide cleaning the combined colonisation of room surfaces by MRSA, VRE, or hospital onset C. difficile was lower with average aerobic colony counts (ACC)/high-touch surface of 14.0 cfus and after QUAD 22.2 cfus (P = .003). The percentage of surfaces with no growth after cleaning was higher for peroxide (48%) than QUAD (35%; P < .0001). The authors proposed a cut-off of <2.5 CFUs/cm² as the definition of a clean surface, and 92.4% of surfaces were clean after peroxide and 88.4% after QUAD (P = .03) with this criterion. [14].

Chami in a cluster randomized study of 4,345 residents in 50 nursing homes in France used 30 interventions (PPE, hand hygiene, education, sterile processes, immunisation, hygiene, daily bathing, screening for incontinence, escort to bathrooms, aseptic technique, changing patient devices, preventing HAIs, and identifying infections) and for the primary outcome of the combined urinary, upper and lower respiratory and gastrointestinal infection rate found no significant differences between the intervention and control arms [15]. No specific new cleaning agents or disinfection protocols were deployed in this study.

Heredia-Rodriguez in a study at the university hospital, Valladolid, Spain, provided all cardiac surgery patients with a routine respiratory bundle of care including hand hygiene, changing ventilator circuits when soiled or at seven days, chlorhexidine 2% oral mouth wash every six hours when intubated, and semi-recumbent body position at 40 degrees, then randomized 522 to air flows exposed to UV light sterilisers and 575 to no UV light air conditioning and found no significant differences in ventilator-associated pneumonia or all infections [16].

Juthani-Mehta in 36 nursing homes in Connecticut with 834 residents randomized 434 patients to a multicomponent pneumonia prevention programme (twice daily manual tooth and gum brushing and 0.12% chlorhexidine oral rinse and upright positioning during feeding) and 400 to usual care. Adherence to the chlorhexidine protocol was 87.9% (SD 20.5%) and 100% for upright feeding. There were no significant differences in pneumonia rates over the 2.5 years of the intervention [17]. No specific new cleaning agents or disinfection protocols were deployed in this study

Lautenbach in a study of 1,021 patients in two 24-bed ICUs in Philadelphia randomized patients to use copper-treated linens (fitted sheets, sheets, pillow cover and gown) or usual linens but found no significant differences in MDRO rates. The authors commented that the study was likely underpowered [18]. Marik in a study of 1,282 patients in two ICU wards (8 beds in each) in eastern Virginia randomized the intervention group to use copper-treated linens (sheets, pillowcases, underpads, wash cloths, towels and gowns) or usual linens and there was no difference in HAIs [19].

Four C -RCTs studied single infections

Four c-RCTs focused on single pathogens. Ray studied HAI C. difficile infection in a study of 16 acute care hospitals in northeastern Ohio and reported both CDI outcomes and surface colonisation rates. He did not state the number of residents but the power computation was based on 1,068 cases and 1,683, 928 patient-days in the facilities in 2009. All hospitals used bleach wipes to daily disinfect C. difficile rooms. In the intervention arm environmental services workers (ESWs) received training on cleaning and daily individual feedback on cultures and how many fluorescent markers were removed after cleaning. There was an average increase in fluorescent marker removal over 12 months from 63% (hospitals ranged from 12% to 74%) to 69% (range 35% to 91%) and a significant decrease in C. difficile colonised rooms positive for C. difficile from 13% to 3% in the intervention group and no significant changes in the control group but there were no significant differences in the intervention or post-intervention periods in the incidence of C. difficile/10,000 patient days [30].

Three c-RCTs reported only MRSA colonisation rates. Peterson in three LTCFs in Illinois with 4,200 annual admissions randomized 12 units either to perform universal MRSA decolonisation with intranasal mupirocin and a chlorhexidine bath twice and decontamination of all flat surfaces in rooms, common rooms and equipment with bleach and enhanced cleaning of flat surfaces with bleach wipe every four months, or to usual care. New admissions were screened and decolonised. All nursing staff were trained in pathogen transmission, hand hygiene, and disinfection education. The baseline colonisation rate was 16.84% and after one year 11.61% in the intervention units (p = 0.028) and after one year 17.85% in the control units (p = 0.61) [28,29].

Baldwin in a study of 793 residents in 32 nursing homes in northern Ireland randomized 16 units to a staff infection control and education programme to reduce MRSA colonisation and 16 to usual practice. Audits were conducted at baseline, 3 and 6 months and written audit reports were sent to nursing home managers with recommendations for improvement. There was 40% attrition of residents during the study. In the intervention group over 12 months infection audit control scores increased from 56% to 84% and in the control group from 53% to 64%, but hand hygiene (66%) and equipment decontamination (75%) were still poor in the intervention group and in an ITT analysis there were no significant differences in MRSA colonisation rates between the intervention and control group after 12 months [26].

Bellini in a study of 4,750 residents in 104 nursing homes in Switzerland randomized 53 homes to the intervention group and 51 to the usual care control group. There were wide baseline ranges in MRSA rates between nursing homes: in the intervention group the nursing home average was 8.9% (range 0-38%) and in the control group the nursing home average was 8.9% (range 0-43%). In the intervention group all HCWs participated in training sessions on the concepts and practice of standard precautions that should be applied to all residents independent of their MRSA status. In the intervention group all patients received topical decolonisation of their nostrils, pharynx, skin, hair, dental prostheses and MRSA ulcers and patients’ clothes, linen, bed, bedside table, phone, walking aid, wheelchair armrests, and TV control were decolonised. The MRSA rate in the intervention arm declined from 8.9% to 5.8% (P =.003) and also in the control from 8.9% to 6.6% (P = .02) but there were no significant differences between study arms [27].

Secondary outcomes

Of the 14 included c-RCTs one reported an increase in MRSA rates in their region (eastern Switzerland) at the time of the study [27] (personal communication from Dr. Bellini). Baldwin identified a recently reported UK nursing home study with an MRSA prevalence of 17% and based his power computation on that report but did not report trends in the institutions he studied or the region [26]. Anderson did not report trends in their hospital region but identified that the addition of UV light to standard chemical disinfection in the rooms previously occupied by MDRO patients led to a significant reduction in risk for the overall group of 11 hospitals of acquiring C. difficile (RR 0.89; 95%CI 0.80 to 0.99; p = 0.031) and VRE (RR 0.56 (0.31 to 0.996; p = 0.048) [14,15]. Mitchell commented that “First, Australia has major reservoirs of C. difficile outside the hospital environment. Secondly, genetically diverse strains of C. difficile are being transmitted into hospitals and infecting patients.” [20].

Four c-RCTs referenced guidelines. Mitchell assessed their pragmatic research design against the PRagmatic-Explanatory Continuum Indicator tool (described in detail in the article [22] and appendices) and based the high touch points sampled on the CDC Environmental Cleaning Checklist. Chami in her study of 25 NHs in France presented during HCW training a”Delphi web survey of guidelines.” [15]. Mody in her study of 17 NHs in southeast Michigan reported that all Medicare-certified and Medicaid-certified NHs have an infection control program and a part-time preventionist and her Targeted Infection Prevention Intervention (TIP) included barrier precautions, surveillance and feedback for NH staff education as key infection prevention principles [23,24]. Salgado referenced the CDC Guideline for Disinfection and Sterilization in Healthcare Facilities 2008 but the study intervention was copper surfaces, a new intervention [25]. No guidelines were referenced by the other studies (Table I).

One study reported that in each of the three LTCFs the MRSA clone (pulsotype) USA !00 predominated (range 38% to 92.3%) and there were no clonal associations between the mupirocin-resistant MRSA strains [28]. Mitchell did not perform microbiological testing of the environment or whole genome sequencing because of financial constraints [20]. Anderson stated “we did not do molecular analyses to confirm that organisms included in our outcomes were related to organisms in the environment as this task was impossible given the scope of our study.” [9].

Mitchell reported that all hospitals had an antibiotic stewardship programme throughout the study period but did not report its results and found during the pre-intervention period a combined antibiotic medication monthly defined daily dose of 4,614/1000 occupied bed days and during the intervention period was 3,949/1000 [20]. Baldwin reported that the consumption of antibiotics in the previous 12 months was 12% in the control and 14% in the intervention group [26] and Bellini in the previous 30 days was 11% ± 5.6 in the control and 14% ± 8.1 in the intervention group but neither provided antibiotic stewardship data [27].

Only one C-RCT reported MDRO rates in HCWs. Baldwin reported that MRSA rates in cleaners were 6% in the intervention and 1% in the control group but did not report results for ESWs [26].

Risk of bias

Using the Cochrane Risk of Bias Tool, of the 12 c-RCTs which reported patient MDROs and or/HAIs, six were at low risk of bias for randomization (all used computer programmes) and six were unknown as they merely stated they were randomized. For allocation concealment two were at low, nine unknown and one at high risk. However, the cleaning and disinfection interventions were applied to room surfaces and not to patients by the environmental service workers employed by the hospitals and LTCFs. The thoroughness of disinfection was assessed by environmental service supervisors (ESS) employed by the hospitals and LTCFs and also by the research staff and feedback was provided by the ESSs to the environmental service workers. For blinding of participants and personnel four were at unknown and eight at high risk. For blinding of outcome assessors four were at low, four unknown and four at high risk. However, for both blinding assessments, considering that interventions were applied to all surfaces in predefined clusters using hospital and LTCF patient registries, and outcomes are laboratory-confirmed infections with MDROs or HAIs the risk of bias due to blinding was revised to low. For attrition eight were at low and four at unknown risk. Although several very large studies had a constant throughput of patients during the study period, the tracking of patients by patient bed occupancy which used the hospital admission and discharge system is likely to find all participants. All 12 studies were at low risk of selective reporting (Figures 1 and 2). The overall assessment is low risk but high for allocation concealment.

• Download figure
• Open in new tab

• Download figure
• Open in new tab

Figure 1. PRISMA Flow diagram for new systematic reviews which include searches of databases and registers only

• Download figure
• Open in new tab

Figure 2.

Cochrane Risk of Bias Graph

• Download figure
• Open in new tab

Figure 3.

Cochrane Risk of Bias Summary

Considering separately the four c-RCTs which reported patient colonisation rates Boyce [14] and Ray [30] also reported patient outcomes (reported above). For randomization three were at low and one at unknown risk, for allocation concealment four were at unknown risk, for blinding of participants and personnel one was at low risk and three at high risk, for blinding of outcome assessors one was at low, two unknown and one at high risk, and for incomplete data one was at low, two unknown and one at high risk, and for selective reporting all four were at low risk. The same comments that the interventions were provided to predetermined clusters of wards by nursing staff and/or environmental service workers employed by the hospitals and LTCFs apply also to these colonisation studies and thus the overall assessment is revised to low risk but risk was high for allocation concealment. Four studies required ongoing patient cooperation for personal decolonisation measures [15,16, 27,28]. Twelve c-RCTs reported a power computation. (Appendix B).

Six c-RCTs reported corrections for clustering. Key questions in the Cochrane RoB version 2 tool for c-RCTs are: 1. Are there baseline differences between clusters? Ten c-RCTs provided detailed tables comparing the intervention and control groups and showed no significant baseline differences; 2. Were all participants identified and recruited before randomization? In all 14 c-RCTs these processes occurred before randomization; 3. Were there any deviations from Protocols? In 10 c-RCTs the interventions were applied directly to the patients’ environment automatically, and in four the interventions were applied to the patients and followed Protocols with no reported exceptions, and 4. Were people delivering the intervention aware of the participant’s assigned intervention during the trial? Because the interventions usually lasted from one to three years it is likely that in the four where interventions were applied directly to patients by HCWs this knowledge was likely apparent to HCWs and could have been apparent to patients (Appendix B). Nevertheless, the GRADE assessment was low risk (Appendix C).

Circulation of pathogens between hospitals, nursing homes and the community

None of the 14 c-RCTs assessed the circulation of pathogens as patients transferred between hospitals, LTCFs and the community. If future studies perform these assessments asymptomatic carriers could be identified to permit better strategies to isolate patients who bring MDROs into institutions. A four-year study of whole genome sequencing of 1,250 symptomatic C. difficile cases in Oxfordshire, UK, found that 45% of the cases were genetically different from previous cases and a possible environmental source of contamination could be identified for only 2% of patients [31,32]. Bellini’s study of Swiss nursing homes found that the MRSA incidence correlated with LTCF size and inversely with staff/resident ratios and thus LTCF size, multiple patient occupancy of rooms and transfer rates to and from hospitals may affect the predominant clones and affect the difficulty of the decolonisation tasks [27].

Antibiotic stewardship

Mitchell reported that all hospitals had an antibiotic stewardship programme but did not describe the protocol or its operation for any of the hospitals. Mitchell’s was the only study which reported pre-intervention and intervention antibiotic use and for all 11 acute care hospitals. During the pre-intervention period the combined monthly defined daily dose was 4,614/1000 occupied bed days and during the intervention period was 3,949/1000. There was a wide range both pre-intervention at the 25% prescribing level (786/1000) and the 75% level (6,182/1000) and also post-intervention at the 25% level (2,090/1000) and the 75% level (5,957/1000) [20]. Antibiotic stewardship programmes and thorough supervision of their effectiveness are important in reducing MDRO colonisation. High rates of antimicrobial use were reported in Russo’s large study of 19 Australian hospitals in which 44% of patients were receiving antibiotics (excluding surgical prophylaxis) and 10% were being treated for at least one MDRO [3].

Environmental cleaning staff

Only one study reported staff MRSA colonisation rates although it is known that staff hands are often colonised [27]. The ESW staff are key to attaining thoroughness in disinfection of surfaces and several authors commented on the high turnover of ESW staff. Only Mitchells’ REACH study stated the numbers of ESWs (1,729) [21] but no c-RCT stated the number of rooms and surfaces cleaned per shift so it is unknown if the ESWs’ workload makes thorough cleaning per protocol feasible. Studies varied in the amount and frequency of feedback to cleaners. No study identified differences in effectiveness between individual cleaners, the effect of educational interventions or factors associated with superior cleaning outcomes, or understanding of the language used for training, educational materials and feedback. There are no studies of merit pay for superior cleaning outcomes. Some studies cleaned all rooms, some only known colonised rooms, the number of high touch points cleaned varied between studies, and these issues are likely to affect the overall effectiveness of ESW staff. The difficulty of cleaning LTCF rooms filled with personal possessions has not been addressed. Deterioration of surfaces making cleaning more difficult was commented on by one study [34]. The cleaning methods, diligence, duration, amount of pressure used, and type of cleaning cloths and disinfectants used will affect outcomes and need detailed comparative studies. Studies are needed to compare for individual ESWs the differences in effectiveness and the causes of heterogeneity among ESWs in relationship to their training, supervision and difficulty of the environments they are asked to disinfect. There are no economic evaluations to assess how the intensity and effectiveness of cleaning reduce total medical costs.

Mask-wearing and hand hygiene

Mask wearing and hand hygiene are key interventions to prevent person-to-person transmission of viral respiratory infections and the most comprehensive and recent systematic review is the 2020 Cochrane review by Jefferson [34] and it will need updating for studies during the COVID-19 pandemic period. In many studies the risk of bias for the RCTs and c-RCTs was high or unclear. The review included only three c-RCTs of hand hygiene in LTCFs. McConeghy’s 2017 study was assessed at unclear risk for random sequence generation and allocation concealment, and high risk for blinding of personnel and participants, outcome assessment and selective reporting [35]. Temime’s 2018 study was found to be at high risk for random sequence generation, unclear risk from allocation concealment, high risk from blinding of participants and personnel, performance assessment and incomplete data [36] while Yeung’s 2011 study was considered at unclear risk from random sequence generation, allocation concealment, selective reporting and at high risk regarding blinding of participants and personnel and outcome assessment [37]. There was only one c-RCT of hand hygiene for patients and staff which assessed bacterial contamination and there was a significant decrease in bacterial colonisation in the intervention group.

A 2020 systematic review of hand washing, distancing and mask wearing to prevent transmission of SARS-CoV-2 identified 172 observational studies in 16 countries but no RCTs and a meta-analysis of 44 non-randomized studies in health-care and non-health-care-settings (25,697 individuals) found that the odds ratio of virus transmission at distances more than one metre was 0.18 (0.09 to 0.38, moderate certainty) compared to distances < 1 metre, for mask wearing OR was 0.15 (0.07 to 0.34), but with low certainty of evidence [38].