Physical interventions to interrupt or reduce the spread of respiratory viruses

BACKGROUND
Viral epidemics or pandemics of acute respiratory infections like influenza or severe acute respiratory syndrome pose a global threat. Antiviral drugs and vaccinations may be insufficient to prevent their spread.


OBJECTIVES
To review the effectiveness of physical interventions to interrupt or reduce the spread of respiratory viruses.


SEARCH STRATEGY
We searched The Cochrane Library, the Cochrane Central Register of Controlled Trials (CENTRAL 2010, Issue 3), which includes the Acute Respiratory Infections Group's Specialised Register, MEDLINE (1966 to October 2010), OLDMEDLINE (1950 to 1965), EMBASE (1990 to October 2010), CINAHL (1982 to October 2010), LILACS (2008 to October 2010), Indian MEDLARS (2008 to October 2010) and IMSEAR (2008 to October 2010).


SELECTION CRITERIA
In this update, two review authors independently applied the inclusion criteria to all identified and retrieved articles and extracted data. We scanned 3775 titles, excluded 3560 and retrieved full papers of 215 studies, to include 66 papers of 67 studies. We included physical interventions (screening at entry ports, isolation, quarantine, social distancing, barriers, personal protection, hand hygiene) to prevent respiratory virus transmission. We included randomised controlled trials (RCTs), cohorts, case-controls, before-after and time series studies.


DATA COLLECTION AND ANALYSIS
We used a standardised form to assess trial eligibility. We assessed RCTs by randomisation method, allocation generation, concealment, blinding and follow up. We assessed non-RCTs for potential confounders and classified them as low, medium and high risk of bias.


MAIN RESULTS
We included 67 studies including randomised controlled trials and observational studies with a mixed risk of bias. A total number of participants is not included as the total would be made up of a heterogenous set of observations (participant people, observations on participants and countries (object of some studies)). The risk of bias for five RCTs and most cluster-RCTs was high. Observational studies were of mixed quality. Only case-control data were sufficiently homogeneous to allow meta-analysis. The highest quality cluster-RCTs suggest respiratory virus spread can be prevented by hygienic measures, such as handwashing, especially around younger children. Benefit from reduced transmission from children to household members is broadly supported also in other study designs where the potential for confounding is greater. Nine case-control studies suggested implementing transmission barriers, isolation and hygienic measures are effective at containing respiratory virus epidemics. Surgical masks or N95 respirators were the most consistent and comprehensive supportive measures. N95 respirators were non-inferior to simple surgical masks but more expensive, uncomfortable and irritating to skin. Adding virucidals or antiseptics to normal handwashing to decrease respiratory disease transmission remains uncertain. Global measures, such as screening at entry ports, led to a non-significant marginal delay in spread. There was limited evidence that social distancing was effective, especially if related to the risk of exposure.


AUTHORS' CONCLUSIONS
Simple and low-cost interventions would be useful for reducing transmission of epidemic respiratory viruses. Routine long-term implementation of some measures assessed might be difficult without the threat of an epidemic.

This report was prepared by the Canadian Agency for Drugs and Technologies in Health (CADTH) in partnership with the World Health Organization (WHO). The purpose of this report is to present an approach for evaluating the utility of the resources and costs of physical interventions that may be used when providing health care for epidemic-and pandemic-prone acute respiratory infections. This information was requested to help inform the revision of the evidence-based WHO interim guidelines, Infection prevention and control of epidemic-and pandemic-prone acute respiratory diseases in health care (July 2007, available at http://www.who.int/csr/resources/publications/WHO_CD_EPR_2007_6/en/index.html). These guidelines provide guidance and direction to the international community, as well as Canada.
The report contains a comprehensive review of the existing public literature, studies, materials, and other information and documentation (collectively, the source documentation) available to CADTH at the time of report preparation, and was guided by expert input and advice throughout its preparation.
The information in this report should not be used as a substitute for the application of clinical judgment in respect of the care of a particular patient or other professional judgment in any decision-making process, nor is it intended to replace professional medical advice. While CADTH has taken care in the preparation of the report to ensure that its contents are accurate, complete, and up to date, CADTH does not make any guarantee to that effect. CADTH is not responsible for any errors or omissions or injury, loss, or damage arising from or as a result of the use (or misuse) of any information contained in or implied by the information in this report.
This document may contain links to other information available on the websites of third parties on the Internet. CADTH does not have control over the content of such sites. Use of third-party sites is governed by the owners' own terms and conditions set out for such sites. CADTH does not make any guarantee with respect to any information contained on such third-party sites and CADTH is not responsible for any injury, loss, or damage suffered as a result of using such third-party sites.
While the content of this document may be used in other jurisdictions, this disclaimer and any questions or matters of any nature arising from or relating to the content or use (or misuse) of this publication will be governed by and interpreted in accordance with the laws of the Province of Ontario and the laws of Canada applicable therein, and all proceedings shall be subject to the exclusive jurisdiction of the courts of the Province of Ontario, Canada.
The authors used a novel approach for determining resource use and costing information that necessitated the adaptation of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology. Although the GRADE methodology has been increasingly used as a transparent and evidence-based approach for grading the quality of evidence and strength of recommendations, using it to assess resource use and costs is in the early stages. Nonetheless, it offers the same transparent and evidence-based approach for assessing economic studies.

Results and Conclusions
The literature search yielded 1,146 citations, which were screened against inclusion criteria based on abstracts. A total of 158 were deemed potentially relevant and retrieved for more detailed evaluation, of which 39 studies were subjected to full review. Seven studies met the inclusion criteria for the systematic review; i.e., they reported information on resource use of physical interventions or assessed the cost-effectiveness of physical interventions. The seven studies were observational in nature, had issues regarding indirectness and, in some cases, imprecision due to small sample size. In some cases where studies were based on modelling exercises, the sensitivity of results to changes in key parameters limited the confidence in study results. All of the economic studies were designed to address specific study questions, resulting in single studies being available for the assessment of physical interventions for specific respiratory viruses. Furthermore, the studies were in settings subject to specific recommendations that varied by location. Given the differences in the economic studies, the results could not be directly compared, which complicates the assessment of consistency. Consequently, the quality of the evidence in the seven studies was found to be very low, based on the use of an adapted Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology.
Based on the updated Cochrane Review, the use of physical interventions to interrupt or reduce the spread of respiratory viruses during epidemics and pandemics is effective. Studies show that the use of physical interventions increases during epidemics and pandemics. Given the general low cost of these interventions, the economic studies that were reviewed showed that use of personal protective equipment was economically attractive. These results are sensitive to assumptions about rate of transmission, facility infection rate, and compliance with interventions, with economic attractiveness increasing when transmission and fatality rates are high. Where guidelines for personal protective equipment use are not followed, cost-effectiveness could be reduced.
The generalizability of the results to different respiratory virus types and community settings requires further investigation. Additional studies are needed to inform the implications on resource use associated with physical interventions, including personal protective equipment, in interrupting or reducing the spread of various respiratory viruses.

INTRODUCTION
Viral epidemics or pandemics of acute respiratory infections like influenza or severe acute respiratory syndrome (SARS) have each posed global threats over the past decade. 1,2 The World Health Organization (WHO) interim guidelines document, Infection Prevention and Control of Epidemic-and Pandemic-Prone Acute Respiratory Diseases in Health Care, has provided infection control guidance to help prevent the transmission of acute respiratory diseases in health care, with an emphasis on acute respiratory diseases that have epidemic or pandemic potential or may constitute a public health emergency of potential concern. 3 The WHO current guidelines were published in 2007, and since then, new evidence on some specific and controversial areas has emerged, necessitating an update of the guidelines. Among the new data that will be used to update the WHO guidelines is a recently updated Cochrane Review that has found that implementation of physical interventions (screening at ports of entry, isolation, quarantine, social distancing, barriers, personal protection [e.g., wearing masks, gloves, and gowns], or hand hygiene) can be effective in containing respiratory virus epidemics or in hospital wards. 4 Given that physical interventions can be instituted rapidly, are readily available, and may be independent of any specific type of infective agent, including novel viruses, they have the potential for widespread use.
Over the last few years, there has been increasing interest in incorporating economic considerations into the development of evidence-based guidelines. The rationale is that including economic considerations allows users and policy-makers within organizations to evaluate the utility of the resources and costs with a perspective relevant to their particular setting. Resource use and economic implications are of interest to the WHO in order to gain an understanding of the amount of personal protective equipment (PPE), isolation, and other infection control measures that may be required during epidemics of acute respiratory infections. In addition, economic evaluations may be helpful for providing information on the benefits in terms of health care resources that are not used with lower rates of transmission.
As a result, this work was undertaken to determine the resource use implications associated with physical interventions to interrupt or reduce the spread of respiratory viruses.
The approach in this report for determining resource use and costing information was novel and necessitated the adaptation of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology. Although the GRADE methodology has been increasingly used as a transparent and evidence-based approach for grading the quality of evidence and strength of recommendations, using it to assess resource use and costs is in the early stages. Nonetheless, it offers the same transparent and evidence-based approach for assessing economic studies. The work related to this report has been divided into two parts: the update of the Cochrane Review 4 and the creation of evidence tables using the GRADE methodology 5 the review of literature for information on resource use associated with PPE and the development of GRADE to capture the resource use, costs, and clinical outcomes included in the studies. As there was limited published experience on the development of GRADE economic tables, an approach and methodology were established through collaboration with international experts.
The methodology and findings from the resource use work are detailed in this report.
Specifically, the research question, scope, and decisions regarding the presentation of the evidence in the GRADE tables were based on the need to inform the revision of the WHO interim guidelines and the need to assist in making recommendations on the use of PPE to reduce the transmission of respiratory viruses. In some cases, the approach may limit the usability of the report and its results for other purposes. In addition, it should be emphasized that the transmission of various respiratory pathogens differs, which is then correspondingly reflected in the resourcing and costing of specific hygienic measures and PPE usage, which in turn may limit the generalizability of the findings beyond the context in which they are presented.

RESEARCH QUESTIONS
Primary What are the resource use implications (e.g., number of units) associated with physical interventions (screening at ports of entry, isolation, quarantine, social distancing, barriers, personal protection [wearing masks, gloves, and gowns], hand hygiene) used for the interruption or reduction in the spread of respiratory viruses?

Secondary
What are the economic implications (e.g., total cost, cost-effectiveness ratios) associated with physical interventions used for the interruption or reduction in the spread of respiratory viruses?

STUDY OBJECTIVES
The following study objectives were developed: Review the economic literature related to resource implications associated with physical interventions used for interrupting or reducing spread of respiratory viruses Review the economic literature related to the cost and effectiveness of physical interventions used for interrupting or reducing spread of respiratory viruses Identify relevant studies associated with physical interventions used for interrupting or reducing spread of respiratory viruses Develop GRADE resource use tables for the identified studies.

Literature Search
An information specialist performed the literature search using a peer-reviewed search strategy. Two independent reviewers screened articles identified through the literature search, using predefined criteria. Where disagreements occurred, a third reviewer was available to provide an additional perspective to help resolve the disagreement.
The literature search focused on two categories of economic studies: 1. Economic studies (e.g., cost analyses, economic evaluations, or resource use studies) of physical interventions to control the spread of respiratory viruses 2. Burden-of-illness studies for respiratory viruses of any type, which include information on cost or resource use (utilization).

Selection Criteria
Eligible studies included economic evaluations, cost studies, utilization studies, and clinical trials (or studies examining effects). The study population included the community and hospital setting. The outcomes of interest were resource use of any of the physical interventions.

Article Selection
Two reviewers (KL and MC) independently applied the selection criteria and screened all citation titles and abstracts that were retrieved from the literature searches. The full texts of all citations and abstracts identified by the two reviewers were ordered. The reviewers then independently reviewed the full text and selected relevant studies for inclusion in this systematic review, based on a predetermined checklist of requirements. The included and excluded studies were compared and any differences were resolved by consensus. A third reviewer was available in cases where consensus could not be attained; thereafter, majority ruled. Full text articles were then reviewed.

Data Extraction
Data from each individual study were extracted by the first reviewer (MC) and reviewed by the second reviewer (KL), using a predesigned data extraction form. Any disagreements between the reviewers were resolved by consensus.

GRADE Tables
Making judgments about the quality of evidence required assessments of the validity of the results of individual studies for different outcomes. Explicit criteria were used in making these judgments. The GRADE working group has developed a standardized and transparent methodology for assessing the quality of evidence. 6 This approach has been adopted by a number of agencies and decision-making groups, including the WHO, which requested that studies in this review be presented in a GRADE format to assess the quality and reporting of the evidence.

Rating the quality of clinical evidence in economic studies
The clinical outcome measures in the economic studies were identified by the clinical reviewer (VS) in consultation with an economic reviewer (KL), and content experts (JC, CPS) conducted a check of the clinical information. The quality of the outcome measure was assessed using a standard GRADE approach as described by Guyatt et al. 6 The GRADE evidence tables for outcome measures used in the economic study were prepared using the GRADEPro program. 7 As described in the GRADE methodology, evidence derived from randomized control trials starts as highquality evidence and observational studies as low-quality evidence supporting an estimate of intervention effects (Table 1). Five factors could result in the downgrading of the evidence: risk of bias, inconsistency, indirectness, imprecision, and publication bias; however, three factors could result in upgrading the evidence: large effect, dose response, and all plausible confounders or biases would result in an underestimate of the effect size. Ultimately, the quality of evidence for each outcome falls into four categories: very low, low, moderate, and high. . It is suggested that as much as possible, the evaluation of the evidence be based on resource use estimates rather than costing information. Further, as with health outcomes, the assessment of quality should be based on the critical (or key) items of the economic study, and not on the assessment of parameters or assumptions that do not affect the results or conclusions of the study.
Reviewers (KL and VS) used their judgment for upgrading and downgrading the evidence and provided detailed reasons for doing so in the GRADE tables and the accompanying notes. The evidence was reviewed independently by the reviewers, the judgments discussed, and any conflicts resolved by consensus. Reviewers (JC and CPS) checked the information in the tables.

Study Limitations
Similar to clinical studies, non-random allocation or inadequate allocation concealment can lead to risk of bias in economic studies. Incomplete outcome data can also bias resource use estimates. If data are missing equally from treatment and control groups for similar reasons, the risk of bias may be low. Resource use should be captured over an -adequate follow-up period.‖ Adherence to the intention-to-treat principle was considered important, to maintain the prognostic balance. Further, resource use information may be collected directly or indirectly from patients. Where information is collected directly, issues of recall bias may be an issue, depending on how frequently information was collected.

Consistency of Results
Consistency should be assessed regarding variation across identified studies in terms of both magnitude and direction of the differences. Where differences exist but authors failed to provide reasons, the quality of the evidence was downgraded.

Directness of Evidence
Directness assesses the applicability of the resource use information to the setting and population for which the guideline was being developed. Where costs were included, assessment of the unit costs and whether they can be applied to the targeted setting should be provided. Further, information obtained from older studies should be assessed to ensure that the information applies to current settings. Based on GRADE guidance, it was suggested that guidelines developers will likely choose to -focus on the evidence for resource use (and costs) that is most direct.‖ Imprecision Variability in resource use between patients can be expected, but there should be sufficient power or number of events to detect differences in resources used.

Publication Bias
As with clinical studies, publication bias should be assessed for resource use studies and economic evaluations. Where only single studies are available, there is the potential for bias, as there are no other studies for comparison purposes.
Based on the information provided in the two GRADE manuscripts on economic information and resource use 8

Literature Search
The literature search yielded 1,078 citations (Search # 1: 993; Search # 2: 71; Grey: 14) and 68 citations identified in the alerts that were conducted after the literature search had been run. Of those citations, 158 (149 from literature search, three from grey literature, one from experts, five from the Public Health Agency of Canada [PHAC]) were deemed potentially relevant. A total of 39 studies were selected and reviewed, of which seven reported information on resource use of physical interventions or assessed the cost-effectiveness of physical interventions, and 32 provided economic information on specific respiratory viruses ( Figure 1).

Figure 1: Selection of Studies
Steps

SUMMARY OF KEY STUDIES
Of the studies reviewed, seven studies met the inclusion criteria; i.e., reported the cost of physical interventions or resource use. Of these studies, three reported information on resource use, either from the collection of data (Murray; 9 Macartney 10 ) or using simulation methods (Phin 11 ). The other four studies considered the cost or cost-effectiveness of physical interventions using modelling exercises. [12][13][14][15] Given the distinct nature of the research questions in the studies and fundamental differences in the biology of the transmission of the individual pathogens, the individual studies were summarized.
Data collection studies Murray (2010) 9 The study was conducted to assess the impact of the use of facial protective equipment during a pandemic (H1N1) 2009 period at the Vancouver Coastal Health (VCH) services. The VCH services 1 million people and comprises three facilities: a tertiary care hospital (644 beds) and two community care hospitals (181 beds and 254 beds). Data were collected directly from the VCH central supply department on use of surgical masks, N95 respirators, and disposable eyewear by all patients admitted to the hospital with an influenza-like illness for the period from June 28, 2009 through December 19, 2009. The PHAC recommendations (2009) for use of facial protective equipment were used as the basis for PPE use. 16 During the study period (24 weeks), 865 patients were admitted with suspected H1N1 influenza infection -149 with laboratory-confirmed infection. Mean length of stay for laboratory-confirmed infection was 8.9 ± 12.1 days from date of specimen and a mean 9.2 ± 6.8 days in ICU. Patients with suspected infection had a mean 1.8 days in isolation, while those with confirmed infection had a mean 5.4 days in isolationoverall (confirmed and unconfirmed cases), the mean number of days in isolation was 2.4. When comparing resource use to the 2009 influenza season (weeks 1 to 14) and the same period in 2008 (weeks 27 to 51), resource use increased by 79% and 130%, respectively. The authors noted a utilization ratio of 1:1 of masks to respirators per patient with laboratory-confirmed H1N1 infection during the study period. The resource use observed exceeded supply estimates from the WHO 3 by four-fold and the US by seven-fold. 17 However, WHO recommendations did differ from those of PHAC and the compliance with the recommendations is not known. When considering both confirmed and suspected cases, the authors estimated approximately 10 respirators and 10 masks per patient per day or a mean of 200 respirators and 155 surgical masks per confirmed or suspected hospitalized case. Only 14 eyewear units per suspected or confirmed patient were observed, which was below the PHAC guidance and could in part be accounted for by use of reusable eyewear. Patient outcomes were found to be similar for the area covered by VCH, compared with other jurisdictions in Canada. While there was stockpiling of facial protective equipment based on Ministry of Health-dictated supplies (10-week supply), initial shortages of facial protective equipment were encountered, which could underestimate the use under sufficient supply conditions; however, this was not felt to significantly affect the estimates. Hospitals within the VCH group all have patients at high risk of contracting tuberculosis and resource use may thus be greater than in other facilities. There was significant variation in respirator use in the three hospitals. Costs and cost-effectiveness were not considered in this study. Macartney (2000) 10 The authors sought to evaluate the clinical and cost-effectiveness of infection control in the prevention of respiratory syncytial virus (RSV) nosocomial infection (NI) in a pediatric facility. The clinical benefits to the infection control program were evaluated for four seasons prior to the intervention (1989-1992) and four seasons after the intervention (1993)(1994)(1995)(1996) in a 304-bed, pediatric hospital (Children's Hospital of Philadelphia). The control intervention included -contact precautions for all patients with symptoms of viral respiratory tract infection, consisting of hand washing before and after contact and the use of gloves and cotton cover gown by all staff for any physical interaction with a patient or the patient's 8 Physical Interventions to Interrupt or Reduce the Spread of Respiratory Viruses -Resource Use Implications environment.‖ Masks and protective eyewear were not used, and gown and glove use for visitors was not required. Resource use was determined by observing 10 isolated patients over a 24-hour period in various wards. The financial burden of RSV (beyond the cost of personal protective equipment) was estimated by comparing the hospital charges for 30 cases with 30 matched inpatients without RSV (controls). The authors estimated mean length of stay attributed to RSV NI to be 7.8 days (with sensitivity analyses ranging from 3.5 to 10.7 days). Resource use was estimated at 52 gloves per patient day and 15 gowns per patient day. Limitations with this study for the purpose of this review are as follows: the generalizability of RSV to pandemic respiratory viruses is limited; the resource use study was conducted in 10 patients; and the data were from 1996, which may no longer reflect current practice.
Costs were then calculated based on unit costs (USD 1996) by multiplying the number of predicted patient days and adjusting for compliance. The authors report that the mean cost of gloves per RSV season is $3,335 with a range of $2,223 to $4,446, based on sensitivity analysis. Similarly for gowns, the predicted mean cost per RSV season is $7,759, with a range of $5,173 to $10,345. The total cost, including personnel, materials, and RSV testing, was estimated to be $15,627, with a range of $9,418 to $24,577.

Simulation exercises
Phin (2009) 11 A simulation exercise was conducted to look at the use of PPE for a 24-hour period on a typical general medical ward, during an influenza pandemic period in the UK. The purpose of the study was to -identify operational issues and to quantify PPE usage around the provision of cohorted care to influenza patients.‖ Operational issues and PPE usage were in accordance with Department of Health, England and Health Protection Agency infection control guidance to the National Health Service (2007). 18 The ward in the study had 29 beds and a total complement of nursing staff (14 nurses, five health care assistants, four domestic staff). The simulated ward was assumed to be operating during the height of the influenza pandemic period to provide cohorted care (i.e., isolation) for patients with confirmed and suspected influenza. Staff were required to wear PPE consistent with the national guidance, and the amount of PPE use was recorded hourly. During the simulation period, the following units were used (rounded to the nearest 50): surgical masks (650), gloves (1,200 pairs), disposable aprons (750), gowns (13), FFP3 respirators (13), eye goggles (13), and visor (one), in addition to background use of PPE. A total of 167 visits were made by 115 different hospital personnel. Patients' visitors were not included in the exercise, and this could result in an underestimation of the use of PPE. Limitations are as follows: the exercise was run only for a 24-hour period; health care workers might be more proficient with PPE after continued use (i.e., it was observed that tasks and routine procedures took longer than usual) or if compliance decreases; and compliance with recommendations was not directly measured. The authors compared the results with expected use of PPEs based on the WHO guidance 3 and found that higher-level PPE (e.g., FFP3 respirators, visors, or goggles) use was less in the exercise, while basic PPE was greater. The authors stated that they felt this arose from the practicalities of dealing with a pandemic situation. Costs and costeffectiveness estimates were not considered in this study. Dan (2009) 12 The authors sought to assess the cost-effectiveness of different levels of infection control in a hospital setting in response to an epidemic respiratory virus threat, such as H1N1, over a 30-day period. The different levels of infection control prevention were provided by Singapore's Ministry of Health, 19 which correlated roughly with the WHO Pandemic Response System. 20 Costs were obtained from actual financial charges for patients treated during the 2009 H1N1 pandemic obtained from Operations and Finance Departments of hospitals in Singapore. Clinical parameters (such as influenza virulence, rate of transmission, incubation period, and second attack rate) were obtained from various sources and tested extensively in sensitivity analyses. The authors considered SARS, H1N1, and Spanish influenza separately in their analyses. They reported the number of patients infected, number of deaths, and costs.
Cost-effectiveness was reported as incremental cost per case prevented and incremental cost per death Viruses -Resource Use Implications avoided. The authors found that isolation of infected patients and selective use of PPE was associated with an attractive cost-effectiveness ratio. Results were sensitive to choices of clinical input parameters, specifically exposure rate, secondary attack rate, case-fatality ratio, and risk of transmissions from atypical cases. The authors found that the incremental cost-effectiveness ratios increased with higher levels of infection control prevention for H1N1 and Spanish flu, while the incremental cost-effectiveness ratio varied for SARS, with higher levels of infection control precautions potentially as cost-effective options. The cost-effectiveness ratios could not be replicated based on the data reported. Results were not reported in a disaggregated manner at the level of resources used.
Perloth (2010) 13 The authors developed a model to simulate the effects of three social-distancing strategies (social distancing, household quarantine, and school closure) and two antiviral medical strategies (antiviral treatment and prophylaxis), and multi-layering among the strategies to mitigate an influenza pandemic for a US community. The model focuses on the contacts between persons. The model considers the spread of influenza in a community of 10,000 people, centred on the school system. The authors state the results could be generalizable to a larger setting, assuming similar demographics, social networks, and contact rates. Information on influenza progression, resource use, costs of treatment, and health utilities was obtained from published literature. Information on medical costs was largely obtained from a published article by Molinari, 21 which evaluated the annual impact of seasonal influenza in the US based on health insurance claims. The cost per case averted was less than $35,000 (2009 USD) for the three socialdistancing strategies: $5,600 for social distancing, $15,300 for quarantine, and $32,100 for school closure. However, when considering all possible multi-layering strategies, a combination of social distancing, school closure, and antiviral treatment and prophylaxis was found to be dominant (associated with greater clinical benefits and fewer costs). The results were sensitive to the rate of infection and the case-fatality rate. The authors identified the following limitations with their study: the results may be limited to suburban communities (rather than densely populated urban centres); results for social-distancing interventions are based on the assumption that communities are able to restrict social contact; and longterm outcomes were not considered. Results were not reported in a disaggregated manner at the level of resources used. Gupta (2005) 14 The objective of this study was to assess the economic impact of widespread quarantine in Toronto (Ontario, Canada) during the SARS outbreaks in 2003. Two public health scenarios were considered in a transmission model: no significant intervention (SARS is transmitted in population), and quarantine is implemented early (where quarantine was defined as -separation and/or restriction of movement of persons,‖ and applies not only to people who are ill but those who have been exposed to the infection).
The model was populated largely using data from -other researchers, the popular press, and interviews with those involved in the Toronto outbreaks in order to make educated estimates about the unknown or uncertain variables.‖ Educated estimates applied to clinical as well as economic input parameters. The analysis was sensitive to rate of contact and transmission rates. The authors report that the total cost associated with SARS is $72 million ($48 million direct and $24 million indirect costs), while the total costs when widespread quarantine is implemented are $12.2 million ($12 million direct and $0.2 million indirect costs). Implementing quarantine during the outbreak of SARS resulted in a total cost savings of $59.8 million. Costs were reported in 2003 Canadian dollars. Results were not reported in a disaggregated manner at the level of resources used.
Putthasri (2009) 15 The authors of this study sought to evaluate the resources and capacity of the health care system in Thailand to contain an influenza pandemic, by assessing the current level of resource available (through surveys to health care institutions and providers in various provinces) and estimating likely resources required, to provide information on gaps. The information for this study was obtained largely from survey The key limitations with this study were that the scenarios were based on individual experiences rather than defined standards of guidelines, and it did not provide the results for the specific scenarios. Costs and cost-effectiveness were not considered in this study.

GRADE Resource Use Tables
The GRADE resource use tables are found in Appendix 1. These tables include the resource use for the physical interventions employed to interrupt or reduce the spread of respiratory viruses. GRADE guidance on the evaluation of economic and resource use information was used in creating the GRADE resource use tables 8  The key studies for the purpose of resource use reporting were by Murray, 9 Phin, 11 and Macartney. 10 In general, studies were subject to downgrading because of their observational nature (lack of control group). As a result, the quality of all seven studies was rated as -very low,‖ given the potential for bias in the study design. Details in each of the GRADE tables provide context for interpreting the results of the studies. The following is a brief overview of each of the GRADE tables found in Appendix 1.
Murray (2010) 9 : In general, the study was well conducted, with no serious concerns for other quality assessment aspects. A limitation would be the observational design, which limits the comparison. In terms of generalizability, resource use was based on PHAC recommendations and utilization observed in three facilities comprising the Vancouver Coast Health region. Further, the resource use information was specific to a hospital setting and for the 2009 H1N1 pandemic period (Table 1A).
Macartney (2000) 10 : The authors considered the clinical effects, resource use, and costs associated with interventions for infection control, reporting the results in a disaggregated manner. Given the design of the clinical study, with data captured pre-and post-intervention, the study was downgraded to -low‖ quality. Resource use was based only on an assessment of 10 patients over 24 hours in a pediatric facility. Issues regarding indirectness and imprecision led to a further downgrading to -very low‖ for the resource use data. In addition, the generalizability of the information to pandemic respiratory viruses is uncertain (Table 1B). Phin (2009) 11 : This study considered resource use in a simulated exercise during a 24-hour period. The observational nature of this study resulted in it being downgraded. Further, the exercise was conducted during a period that did not coincide with an influenza pandemic; this 24-hour period did not allow for participants to acclimate to the setting (potential inefficiencies with new protocols); the setting was a cohorted (or isolated) ward with 29 beds and subject to small numbers (potential for imprecision); and the Department of Health, England guidance was considered, which may not be generalizable to other jurisdictions (Table 1C).
The other four studies were based largely on modelling exercises with clinical and economic inputs obtained from a number of sources. The results of the studies were reported in terms of clinical benefits from physical interventions and total costs, and in two of the studies, cost-effectiveness ratios were Viruses -Resource Use Implications calculated. 12,13 The study by Dan 12 was downgraded primarily because of the use of observational studies (clinical) and the limited information provided on the derivation of costs. There are inherent limitations with modelling exercises, such as the need for numerous data inputs often requiring assumptions that cannot be validated, and the need for extensive sensitivity analyses. The potential lack of generalizability of the study (based largely on data specific to Singapore) is noted (Table 1D, parts i-iii). The study by Perloth 13 was based on a number of published sources (observational in nature), for clinical, cost, and quality of life information (Table 1E). Gupta 14 used information from other researchers, the popular press, and interviews, in the absence of clinical trials or published economic evaluations. The quality of the evidence was downgraded because of the extensive use of interviews, which are prone to recollection bias (Table 1F). The basis for the Putthasri 15 study was survey data that were lower in quality in the hierarchy of observational studies and were downgraded accordingly. Further, a key limitation with the study was that the results were not reported in a disaggregated manner, by severity of pandemic level (Table 1G).
Given the differences in the research questions examined by the seven included economic evaluations, the fundamental differences in the biology of transmission of the pathogens studied, and the heterogeneity of the studies, the summary of the results of the systematic review in one GRADE evidence profile was not deemed to be appropriate for the purpose of this study. While summarizing studies in one table, is the standard approach when applying GRADE, the researchers felt that the individual GRADE profiles would better address specific respiratory viruses, interventions, and study designs to ensure clarity and best meet the needs of the readership.

DISCUSSION
The work described in this report represents the second component of a two-part project. The first part was the updating of a Cochrane Review 4 on the physical interventions to interrupt or reduce the spread of respiratory viruses. This second part focuses on the economic information and, more specifically, resource use. This was of interest as Cochrane Reviews do not consider economic information.
The detailed Cochrane Review on the efficacy of physical interventions to interrupt or reduce the spread of respiratory viruses was updated. Four authors of this review (CPS, JC, VS, KL) prepared GRADE evidence tables as part of the Cochrane Review. According to the analyses using GRADE methodology, the results suggest that physical interventions to interrupt or delay the spread of virus using personal protective equipment (such as masks, gowns, gloves, or eye protection) are effective in protecting against viral transmission. The quality of evidence was very low.
The information presented in this report focuses on resource use associated with the use of physical interventions, with the inclusion of economic implications (such as costs and cost-effectiveness estimates) to supplement the findings. Three studies provide information on resource use alone, 9,11,15 which is difficult to interpret without the context of the clinical benefits, and based on the compliance with recommendations on PPE use provided by public health agencies. In general, the use of PPE increases during pandemic, and epidemic, periods. According to the Cochrane Review, simple, low-cost interventionssuch as frequent hand hygiene and, when indicated, wearing masks, gloves, and gowns and the use of these interventions in combination reduce the risk of spread of some respiratory viruses by approximately 70% to 90%, depending upon the intervention and combination of these interventions (odds ratio for infection: 0.09 to 0.45). The studies included in the review evaluated interventions for different respiratory infections (largely SARS) with different modes of transmission, different settings, and different epidemiological situations. Due to these variations, the effect size of intervention may vary depending upon the circumstances. However, given the overall (large) range of effect size, it is likely that if this data were included in an economic evaluation, these low-cost interventions would result in an 12 Physical Interventions to Interrupt or Reduce the Spread of Respiratory Viruses -Resource Use Implications attractive cost-effectiveness ratio. This was beyond the scope of the current project but represents an area of possible future research.
Of note, the pathogens considered in the Cochrane Review, as well as in the published economic evaluations, may be associated with different rates and modes of transmission. As a result, individual PPE items may differ in terms of the level of importance, depending on the pathogen. This has not been fully explored in the scope of this report.
The three economic studies that reported cost-effectiveness 10,12,13 found that the use of physical interventions was in general economically attractive, with cost-effectiveness improving with higher risk pandemics or epidemics (i.e., higher rates of transmission). Dan 12 found that when comparing Green level PPE protection to Green 0 level a in hospital, the incremental cost per death prevented was less than US$25,000. The movement to higher levels of precautions (Yellow, enforced protection in high-risk areas, and Orange, protection throughout hospital) b was found to be less cost effective, with the incremental cost-effectiveness ratio increasing as high as $827,000 when comparing Yellow with Green levels, and above $2 million when comparing Orange with Yellow, for H1N1. When considering the incremental cost per case averted, the cost-effectiveness ratios were generally less than $10,000. Given the use of antivirals in levels above Green 0, the exclusion of harms associated with antivirals may underestimate the costs, which would lead to greater cost-effectiveness estimates. Perloth 13 considered the use of various mitigation strategies in the community and found that results were dependent on the fatality rate, transmission rate, and compliance with the strategies. Of specific interest was the consideration of quarantine, social distancing, and school closure, which were found to have an incremental cost per case averted of $15,300, $5,600, and $32,100, respectively, when compared with -doing nothing.‖ The authors found that a multi-layered approach of social distancing, school closure, and antiviral therapy (treatment and prophylaxis) was the dominant strategy in their analysis. Macartney 10 reported that a hospital program to reduce the spread of RSV (targeted control program including laboratory testing, cohorting, and glove and gown use) resulted in a cost-effectiveness ratio of $1,563 per case averted. When considering the cost in terms of lost wages for caregivers, this may represent a reasonable cost from a societal perspective.
A finding from the studies by Murray 9 and Dan 12 is that PPEs may be overused and inappropriately used when faced with a new pandemic. While appropriate use of PPEs is largely seen as cost effective, where recommended use is not adhered to and overuse occurs, this could lead to PPEs not being cost effective.
Using the GRADE methodology, the economic evaluations were graded as -very low‖ quality. The components within each economic study were evaluated (clinical data, cost and/or resource use information, and quality of life information). The clinical data used in these studies tended to be observational in nature and were downgraded because of indirectness. Similarly, resource use and cost studies were observational in nature, had issues regarding indirectness, and in some cases, limitations in imprecision as a result of small sample sizes. Where studies were based on modelling exercises, the results were subject to the same limitations as the inputs, but in some cases, the sensitivity of the results to changes in key parameters affected the confidence that could be placed on the study results (imprecision). In all cases, the economic studies sought to answer specific questions, and as a result, there were only single studies available for the assessment of physical interventions for use in specific respiratory viruses, which could result in a risk of publication bias. a Green level PPE was defined as full PPE for suspected cases and all suspected cases to be isolated, tracing contacts and antiviral treatment for confirmed cases. Green 0 level (or no additional measures) was defined as triage and isolation of patients, and use of PPE as appropriate at health care institutions. b Yellow level was defined as for PPE for health care workers in high-risk contact tracing for confirmed cases, visitor restriction, restricted movement of patients and health care workers. Orange level was defined as full personal protective equipment for health care workers in high-risk contact (including N95 masks, gloves, gowns, and eye protection), visitor restriction, no interhospital movement of patients or health care workers, and prophylaxis for contacts.

Viruses -Resource Use Implications
This review is subject to a number of limitations. While numerous economic evaluations in the area of respiratory viruses have been published, the purpose of many has been to evaluate the use of antivirals in treatment or prevention. Only a few economic studies evaluating physical interventions are available, and this might largely be due to the lack of sponsorship for these studies. Of the studies published, a small number have specifically reported resource use in their publications. Where studies have considered the use of antivirals in the assessment of PPEs, the failure to consider harms (which is an emerging area of understanding) could significantly underestimate the cost-effectiveness.
Given the limited information on resource use associated with physical interventions in the interruption or reduction of the spread of respiratory viruses, information about the use of various physical interventions for specific respiratory viruses (SARS, H1N1, Spanish influenza, and influenza) is sparse. Whether results can be generalized among respiratory virus type needs to be evaluated. Similarly, because the majority of the evidence pertains to hospital populations rather than communities, assessment on how these results might be applied to the community setting requires further investigation. This review suggests that further studies on resource use and economic implications associated with PPE are needed for various respiratory viruses.
Studies measuring resource use (Murray, 9 Macartney 10 ) were based on local recommendations or guidelines, which could affect the generalizability of the results. The results of the studies are difficult to apply to other jurisdictions, as they tended to be based on specific recommendations (PHAC; Department of Health, England) and in some cases were applied in a manner specific to the interests of the jurisdiction or facility.
The economic studies evaluated different specific research questions, which may not directly inform this review. Given the differences in these studies, in most cases the results could not be directly compared. This complicates the assessment of consistency, given the uniqueness of the studies.
This study on resource use was conducted alongside the update of a Cochrane Review assessing the clinical evidence for physical interventions to interrupt or reduce the spread of respiratory viruses. The combination of the clinical findings from the Cochrane Review and the economic information gleaned from this systematic review could be considered for future work. This review found that economic evaluations exist to evaluate some physical interventions, but they are limited by the availability of clinical information to populate the input parameters (transmission rates, case fatality, and compliance). Resource use and economic implications are important considerations in understanding the potential financial burden of recommending the use of personal protective equipment, both in terms of the resources required and health care resource use that may be avoided through the reduction of infected cases and complications.

CONCLUSIONS
Based on the published economic evaluations and clinical studies, it appears that use of physical interventions to interrupt or reduce the spread of respiratory viruses increases during epidemics and pandemics. This is likely in part because local or WHO guidelines on the use of personal protective equipment and isolation were followed. Based on the updated Cochrane Review, it appears that these measures are effective at reducing the spread of viruses. Given the general low cost of these interventions, based on the economic studies reviewed, PPEs appear to be an economically attractive option for reducing the burden of respiratory viruses. These results are sensitive to assumptions regarding the rate of transmission, facility infection rate, and compliance with the interventions, with the economic attractiveness increasing when transmission is high and fatality rate is high. Where guidelines for PPE use are not followed, and inappropriate use of PPEs occurs (i.e., increased use resulting from fear of a new pandemic), this could increase the cost-effectiveness estimates.        1 Singapore Ministry of Health evaluated the impact of different alert levels. Based on hospital measures, the levels are defined as follows: Green 1 = full PPE for suspected cases, tracing contacts for confirmed cases, and antiviral treatment for confirmed cases (WHO alert level 2-3); Yellow = PPE for health care workers: for middle-risk PPE, included N95 mask, gown, and gloves (eye protection if risk of splashes); PPE including N95 mask in all patient areas, and gown and gloves, and eye protection if risk of splashes; if high-risk activity (high probability of close contact to aerosolgenerating procedures), then PPE included N95 mask in all patient areas, and gown and gloves, and eye protection if contact tracing for confirmed cases, visitor restriction, restricted movement of patients and health care workers (WHO Alert level 4). 2 Impact of an outbreak from a single index case that was not detected by hospital surveillance and was found in general ward was modelled. 3 Clinical data were generated by modelling of undetected single index case. 4 A patient with undiagnosed infection at base case with no protection measure will result in 2,580 infected patients at 30 days in this model. 5 Costs reported in 2009 USD. US$1 (2009) = US$1.02 (2011). 6 Resource use based on Disease Outbreak Response System and may not be generalizable to other jurisdictions. 7 Costs were obtained from Operations and Finance Departments from hospitals; might not be reflective of other facilities. Limited information provided on methodology; information from direct charges and assumptions regarding indirect costs. 8 Harms and attendant costs of antivirals were not included; nor was the use of antiviral prophylaxis. 9 Minimal details were provided with respect to sensitivity analyses. Viruses -Resource Use Implications    1 Resource use estimates based on a mapping exercise using information obtained from surveys from different provinces and institutional settings. 2 Simulation exercise with no control group. 3 List of 39 resources generated through previous experiences, literature searches. 4 Survey was conducted for potential resource use in different provinces of Thailand. Survey questionnaire was developed based on past experiences in Thailand and information available in literature. 5 Survey study. 6 Study was done in Thailand, which has a different health system to other countries. 7 There is potential to have another survey on the same issue. 8 Fourteen provinces participated in the survey. 9 Phase 4 refers to human-to-human transmission from case patient to caregiver. Phase 5 refers to human-to-human transmission in localized clusters. Viruses -Resource Use Implications