Conservation Practices for Personal Protective Equipment: A Systematic Review with Focus on Lower-Income Countries

During the start of the COVID-19 pandemic, shortages of personal protective equipment (PPE) necessitated unprecedented and non-validated approaches to conserve PPE at healthcare facilities, especially in high income countries where single-use disposable PPE was ubiquitous. Our team conducted a systematic literature review to evaluate historic approaches for conserving single-use PPE, expecting that lower-income countries or developing contexts may already be uniquely conserving PPE. However, of the 50 included studies, only 3 originated from middle-income countries and none originated from low-income countries. Data from the included studies suggest PPE remained effective with extended use and with multiple or repeated use in clinical settings, as long as donning and doffing were performed in a standard manner. Multiple decontamination techniques were effective in disinfecting single use PPE for repeated use. These findings can inform healthcare facilities and providers in establishing protocols for safe conservation of PPE supplies and updating existing protocols to improve sustainability and overall resilience. Future studies should evaluate conservation practices in low-resource settings during non-pandemic times to develop strategies for more sustainable and resilient healthcare worldwide.


Introduction
The emergence of SARS-CoV-2 (COVID-19) strained healthcare resources worldwide. Healthcare facilities and hospitals across the world reported severe shortages of supplies, particularly personal protective equipment (PPE) including masks and respirators, gowns, gloves, and eye protection. These shortages were driven by an unprecedented increase in consumption, compounded by factory closures along the global supply chain [1]. As a result, many hospitals and medical centers worldwide developed individual approaches to manage limited supplies based on guidance from the World Health Organization and data from previous outbreaks [2][3][4][5][6][7]. Hospitals and other health facilities enacted conservation practices such as limiting access to PPE stocks and rationing supplies. Novel, often nonvalidated, techniques emerged early in the pandemic to prolong the lifespans of single-use PPE, including prolonged use, multiple use, and reprocessing of single-use PPE [8].
Given the likelihood of continued supply disruptions in the recovery from the pandemic and the onslaught of climate change [9][10][11], it is imperative to better understand PPE conservation practices that can be safely employed in times of shortage. These practices must not reduce PPE efficacy to levels that increase risk of pathogen exposure [12]. Systematic reviews of PPE conservation that have been published since the emergence of COVID-19 have been narrowly focused on one type of PPE, mainly N95 respirators [13][14][15][16][17][18], and have not commented on practices in developing contexts. Here, we assess the body of knowledge at the start of the pandemic on various PPE conservation and reuse practices.
We hypothesized that low-and middle-income countries may have validated methods for conserving PPE due to regular shortages in supplies. Ideally, all countries should be enabled to meet their supply chain needs, but existing consumption practices from high-income countries could be improved by drawing from the experiences of countries with fewer resources, as has been suggested in the example of cataract surgeries [19][20][21]. We sought to analyze existing literature to assess effective means of conserving and reusing PPE, with implications for the ongoing pandemic and beyond. We anticipate that our analyses will help bolster healthcare system preparedness for future supply challenges, as well as support efforts to promote long-term environmental sustainability [22].

Systematic Review
On 9 June 2020, following PRISMA guidelines for systematic reviews [23], a medical librarian (TR) performed searches for studies without language or date restrictions in the LitCOVID, PubMed, Embase, HealthStar, and CINAHL databases. The search included studies describing the efficacy of methods used to conserve or reuse PPE within healthcare settings and comparisons of single-use and reusable PPE. A full description of search terms and inclusion and exclusion criteria can be found in the Supplemental Materials (SM). The citation lists of relevant systematic reviews were screened for additional studies.
Title, abstract, and full text screenings were completed by 2 researchers (GA, PL) using Covidence software [24]. The entire research team iteratively built and tested a data extraction form (see SM for further details). Data extraction from each of the included studies was carried out by four authors (GS, PS, MS, HG), with two randomized to each study; of those two, one was assigned to consolidate the two versions of extracted data into one dataset for each selected study. If needed, consolidators returned to the original study to resolve a discrepancy.

Quality Analysis
Two authors (PS, MS) independently evaluated the quality of each study to determine the validity of its recommendations. The Cochrane collaboration tool was used to assess risk of bias for randomized controlled trials [25][26][27], while the Newcastle-Ottawa quality assessment scale was used to assess risk of bias in cohort and case-control studies [28]. Discrepancies between scores were reconciled by a third author (SH). Due to the lack of a standardized quality assessment tool, the validity of studies that were lab-based, crosssectional, or descriptive were evaluated by one author (PS), who compiled the limitations listed in each study from the latter study designs as described by the study's authors.

Data Analysis
Included studies were stratified into three primary analytic domains: (1) the reuse and extended use of PPE manufactured as single-use, (2) methods for disinfecting PPE, and (3) efficacy of PPE by type and mode of pathogen transmission (droplets, aerosols, and fomites) ( Figure 1). Reuse was defined as the repeated donning and doffing of the same pieces of PPE; extended use was defined as the practice of wearing the equipment throughout a shift or prolonged period without donning or doffing. Summary statistics were used to evaluate the types of PPE, modes of pathogen transmission, and settings under study, as well as the type of study, metrics used, and outcomes evaluated. and (3) efficacy of PPE by type and mode of pathogen transmission (droplets, aerosols, and fomites) ( Figure 1). Reuse was defined as the repeated donning and doffing of the same pieces of PPE; extended use was defined as the practice of wearing the equipment throughout a shift or prolonged period without donning or doffing. Summary statistics were used to evaluate the types of PPE, modes of pathogen transmission, and settings under study, as well as the type of study, metrics used, and outcomes evaluated. Because this review sought to evaluate conservation methods for resource-limited settings, the human development index (HDI) was used to estimate the relative accessibility of each publication. In order to compare the applicability and global relevance of findings, we used the 2019 United Nations human development index to stratify study settings by relative country development level [29]. The HDI is a summary measure of achievement in health, education, and standard of living. Countries are considered to have a "very high" level of development with HDI of 0.800 or greater; "high" development with HDI of 0.700-0.799; "medium" development with HDI 0.550-0.699; and "low" with HDI < 0.550. Because this review sought to evaluate conservation methods for resource-limited settings, the human development index (HDI) was used to estimate the relative accessibility of each publication. In order to compare the applicability and global relevance of findings, we used the 2019 United Nations human development index to stratify study settings by relative country development level [29]. The HDI is a summary measure of achievement in health, education, and standard of living. Countries are considered to have a "very high" level of development with HDI of 0.800 or greater; "high" development with HDI of 0.700-0.799; "medium" development with HDI 0.550-0.699; and "low" with HDI < 0.550.

Included Studies
After removing duplicates, the initial search yielded 980 studies ( Figure 1). After screening, 50 articles met all inclusion criteria and were included for analysis. Eighteen studies assessed reuse or extended use of single-use PPE, 15 studies analyzed PPE disinfection methods, and 19 studies measured the efficacy of PPE, with 2 of the included studies meeting criteria for two categories. These studies were published in 27 different academic journals, with a plurality published by the American Journal of Infection Control (n = 10 papers) and the Journal of Occupational and Environmental Hygiene (n = 5). The publication dates ranged from 2008 to 2020, with 30% of included papers published in 2020 (prior to June) and 58% published in 2015 or later.

Geographic Distribution and Funding Sources
The 50 included studies spanned 14 countries, with some studies assessing multiple locations. The United States (45%) and China (20%) were the two most common settings. Forty-nine of the 50 studies originated in countries considered to have "high" or "very high" HDI, and 38 of these were conducted in countries ranked at a "very high" HDI. Three studies included countries with "medium" HDI, and no study was conducted in countries with "low" HDI. The average HDI of all included studies' national setting was 0.905, more than the worldwide average HDI of 0.713.
Twenty-nine of the 50 studies were federally funded; four received university funding; three credited both federal and university funding; and 14 did not specify funding sources. Of the 23 studies from the United States, 18 were federally funded and five did not specify.

PPE Type, Study Setting, and Primary Outcome
The type of PPE evaluated varied across studies. N95 or elastomeric respirators were the most common type of masks evaluated (n = 35, or 70%), followed by surgical masks (n = 15, 30%) and cloth masks (n = 4, 8%). Four studies discussed improvised masks, and one studied only melt-blown fabric (a component of masks) due to a shortage of PPE. Aside from masks, five studies examined protective eyewear, two studied gloves, and five studied gowns. Two of the gown studies evaluated gown material rather than complete suits.
Seventeen studies (34%) took place in an in situ or clinical setting, while the remaining 33 were laboratory-based studies (defined as artificially replicating conditions of disease or PPE use, rather than examining an intervention or disease transmission in a clinical setting). Half of the studies included PPE disinfection (n = 25), measuring the impact of disinfection on PPE efficacy, PPE degradation, ability to reuse, and ability to effectively disinfect against specific pathogens.
Fourteen studies (28%) focused on fit testing or filtration efficacy of filtering face piece respirators (FFRs) for aerosols, and three papers studied protection from noninfectious contaminants such as chemicals. Seventeen remaining studies examined the use of PPE to protect against one or more infectious diseases, defined as a pathogen in a clinical setting or a (potentially surrogate) virus or bacteria used in a laboratory. Twelve studies measured degradation of PPE, four focused on degradation related to reuse of single-use PPE and nine studies measured the impact of disinfection on degradation of PPE. Of the 15 studies that measured the impact of PPE disinfection on reduction in contamination, eleven measured the effect of disinfection on viruses found on PPE surfaces and four studies focused on the impact of disinfection on bacteria or bacterial spores. Among all included studies, three specifically studied SARS-CoV-2, the virus causing COVID-19 [30][31][32]. The breadth of studies dedicated exclusively to COVID-19 has notably burgeoned as the pandemic has progressed.

Reuse and Extended Use of Disposable PPE
Of the 18 studies examining reuse or extended use of single-use PPE, five focused on in-situ, clinical settings and 13 were lab-based, as shown in Table 1. Overall, studies concluded that reuse can be done safely with some carefully implemented disinfection methodologies (see Supplemental Materials, SM, for description of all disinfection methods tested in the literature). Novel PPE-surgical masks with graphene coating N/A, Self-heats with exposure to sun for disinfection and surface is hydrophobic.
Indefinite: covering masks in a layer of graphene enabled reuse of surgical masks by high-temperature self-cleaning from solar illumination. Solar illumination (temperature over 70C when graphite exposed to solar illumination)

Filtration efficacy of modified PPE not studied Yes
Many studies concluded that the extended use of eye protection and N95 respirators could be practiced safely with strict adherence to environmental and hand hygiene. For example, in a study evaluating surface contamination of N95 respirators and eye protection, no SARS-CoV-2 was detected by RT-PCR (reverse transcription polymerase chain reaction) from any surface of PPE worn by healthcare providers exiting hospital rooms of patients with COVID-19. No contamination was discovered on healthcare workers' goggles, N95 respirators, or shoes worn for extended use while interacting with a series of COVID-19 patients for up to 25 min each [30].
A cross-sectional survey of infection control coordinators at hospitals in Vietnam, Pakistan, and China (medium HDI countries potentially more prone to supply shortages), demonstrated that extended use and reuse of cloth masks, surgical masks, and respirators were common practice. Examination of sample masks from participating sites revealed high penetration of particles through reused cloth and surgical masks compared with N95 respirators. This study proposed that respirators remain the most effective option compared with surgical or cloth masks, and that respirators can be reused as long as they are not visibly soiled or damaged [33].
A cluster-randomized trial conducted by MacIntyre et al. [49] found that hospital staff who wore N95 respirators throughout work shifts had fewer cases of clinical respiratory illness compared with staff who wore N95s only for high-risk procedures or when providing direct patient care under infection isolation precautions. Studies also found that N95 respirators were still effective barriers after donning and doffing. A study based in an infectious disease hospital in Brazil found that N95 respirators could be used by healthcare workers during shifts up to 12 hours and reused for up to five days [35]. A separate study found that N95 respirators could safely undergo five uses with repeated donning and doffing while maintaining filtration factors greater than 100 (considered to be effective protection); some respirator models could undergo 20 uses without fit test failure or a drop in filtration factor [37]. In contrast, a pilot study found that tethering devices such as elastic straps from some N95 respirator models could not withstand more than three donning and doffing repetitions before breaking or losing elasticity [41].

PPE Disinfection Methods
Fifteen studies focused on the disinfection of PPE through a variety of techniques (Table 2). Five studies focused on disinfecting bacterial pathogens in lab-based settings. Of the 10 studies focused on viral pathogens, two were in situ studies and eight were labbased. In studies examining the reduction of contaminants, a 3 log 10 or 5 log 10 reduction in viable microorganisms was considered effective disinfection [50][51][52][53][54][55][56][57][58][59]. Overall, safe reuse depended on the specific mask material, with differences in contamination risk based on hydrophobicity or permeability; the type of contamination (bacterial, viral, non-diseasespecific); the mode of application of disinfectant; and a range of other factors.
Viscusi et al. [59] demonstrated that moist heat incubation and microwave-generated steam decontamination-methods that can be used in a variety of settings-did not adversely affect efficacy or comfort with N95 respirators. Additionally, studies comparing multiple methods of disinfection found that ultraviolet germicidal irradiation (UVGI), ethylene oxide, vaporized hydrogen peroxide, and bleach did not adversely affect the efficacy of N95 and elastomeric respirators [42]. However, none of these studies assessed the effect of multiple rounds of decontamination on respirator function. In a separate study, higher intensities of UVGI used for decontamination led to a loss of strength and reduced efficacy of N95 respirator material [43]. Salter et al. [44] measured the deposition of toxic residues on filtering facepiece respirators after decontamination and found that the majority of decontamination methods tested did not leave significant residues on respirators. However, decontamination with ethylene oxide left behind toxic residues, and bleach caused discoloration and a lingering odor.   While many of the decontamination methods mentioned above, such as UVGI, ethylene oxide and vaporized hydrogen peroxide, require significant infrastructure and resources, several studies found that heat sterilization, which can be performed using fewer resources, was effective and could be repeated without detrimental effect to PPE. Steam or dry heat sterilization, repeated up to five times, effectively disinfected single-use N95 respirators without degrading the pressure drop, permeability, filtration factor, or penetrance of bacterial spray [45]. A separate study by Liao et al. found that heat sterilization could be repeated up to 50 times without adversely affecting filtration efficacy of the melt-blown fabric used in N95 respirators [46].
In addition to measuring the effect of disinfection on PPE efficacy and degradation, several studies evaluated disinfection methods against specific pathogens. Many of these disinfection methods, such as steam sterilization and chlorine-based disinfectants, can be used in a variety of settings without significant investment of infrastructure. Steam sterilization, either in the form of microwave-generated steam or steam from a household rice cooker, was effective in killing bacteria (Staphylococcus aureus) and influenza virus [39,43,44]. In addition, diluted chlorine solution effectively killed bacterial spores on gown material [57].

PPE Efficacy
PPE efficacy was measured most commonly through infection rate (n = 11) or filtration factor (n = 4). All other methods were grouped into a third category, "other metric" (n = 4). The infection rate studies were all conducted in clinical settings whereas the filtration factor and "other metric" studies were conducted in lab-based settings. Overall, N95 respirators were found to be superior to surgical masks in preventing infection from viruses transmited by aerosols but not more effective than surgical masks for pathogens transmitted through bacteria or fomites. Cloth masks were generally less effective than both surgical masks and N95 respirators. Finally, the one study that compared reusable gowns to disposable gowns found disposable gowns more impermeable to respiratory fluid.

Surgical Masks and N95 Respirators
Studies comparing infection transmission using either N95 respirators or surgical masks yielded varied results, summarized in Table 3. Overall, N95 respirators were more effective than surgical masks in preventing transmission or infection with viruses transmitted via aerosols. Wang et al. [32] evaluated the effect of N95 respirator use on SARS-CoV-2 transmission and found that N95 respirators were more effective than no mask use to prevent infection from SARS-CoV-2. Surgical masks were not evaluated in this study. With respect to prevention of infection with pathogens spread primarily via droplets or fomites (such as influenza or bacteria), N95s were not consistently more effective than surgical masks. Two randomized controlled studies in China found lower rates of clinical respiratory illness among healthcare workers who wore N95 respirators during work shifts compared with workers who wore surgical masks [49,63]. However, these studies found no difference in influenza-like illness or laboratory-confirmed respiratory viruses among groups and did not take into account infection exposure outside of the work environment.   Studies evaluating the efficacy of cloth or homemade masks in preventing the spread of respiratory viruses are also shown in Table 3. In general, cloth or other homemade masks were less effective than surgical masks or N95s in blocking pathogens that are spread via aerosols. There was significant heterogeneity among studies and paucity of data related to droplet or fomite transmission. In an experimental model of respiratory virus transmission, cloth masks were half as effective as surgical masks and 50 times less effective than filtering facepiece respirators (FFP2, an equivalent to N95s) [76]. A randomized study of healthcare workers in Vietnam found that staff wearing surgical masks had lower rates of influenza-like illness and laboratory-confirmed respiratory virus compared with staff wearing cloth masks [64]. Conversely, a more recent study used aerosolized avian influenza virus and found that homemade cloth masks with internal layers of kitchen paper blocked 95% of aerosolized avian influenza virus, similar to the blocking effect of surgical masks and N95 respirators [61]. Overall, studies of novel or homemade masks yielded inferior results to N95s when evaluating aerosol transmission of pathogens. Our review found a lack of studies evaluating the effectiveness of novel PPE for protection from respiratory pathogens spread through respiratory droplets or fomites [72,74,77].

Other, Non-Mask PPE
Only one study evaluated the efficacy of reusable gowns, which could reduce reliance on disposable gowns during shortages and reduce environmental burdens [78]. In this simulation, all participants wearing reusable surgical gowns had evidence of respiratory fluid permeation onto their underlying scrubs (Table 3). After multiple washes, the reusable gowns lost their protection from permeation, while the control participant wearing a disposable gown had no permeation of respiratory fluid [38].

Quality of Included Studies
Results of our quality and bias assessment are shown in SM Figure S1 for included RCT studies and SM Table S1 for included cohort and case-control studies. Of the 50 studies reviewed, 36 did not fall into the categories of RCT, cohort, or case-control and thus were not able to be assessed for bias through either the Newcastle-Ottawa or Cochrane tools. Instead, we conducted a qualitative analysis of bias through comparison of the limitations listed in each paper. The majority of these studies took place in a lab setting and tested the efficacy of various decontamination methods.
For these 36 studies, the most commonly mentioned limitations included small sample size, evaluation of decontamination method for only one type of filtering facepiece respirator, absence of assessing fit and filtration efficiency after decontamination, lack of decontamination effect assessment on mask straps or nose pieces, and only evaluating decontamination efficacy against one microorganism. Overall concerns in many of the lab studies were that decontamination processes were not tested in a true healthcare setting, suggesting further research would be needed before implementing changes in medical facilities. Thirteen of these studies (26%) did not list limitations [31,[33][34][35]38,39,53,57,60,61,73,75,77].

Discussion
This literature review was motivated by the COVID-19 pandemic, which overwhelmed healthcare facilities, disrupted supply chains, and caused shortages of single-use PPE in the U.S. and worldwide [79]. We are also concerned about the sustainability of medical systems, including the waste generation from single-use PPE [12,80] and the impact of climate change on medical supply chains [22]. Understanding the state of the literature on the efficacy of various types of PPE and conservation practices in healthcare settings at the early stages of the COVID-19 pandemic can help contextualize how systems' responses have since developed, thus informing future supply chain management and resilience strategies. Prioritizing effective, reusable PPE practices could reduce excess waste in highly-resourced settings, allow redistribution of supplies to resource-limited settings, and help prepare for future infectious outbreaks.
We found that there are safe and effective ways to reuse N95s in circumstances when protection from aerosols is necessary. However, any reuse inevitably carries the risk of contamination or reduced efficacy. Many factors contribute to the safe reuse of respirators, such as a standardized process for donning and doffing and stable storage conditions. Improper doffing can enable the transfer of contaminants [40], and storage conditions with limited humidity prevent the survival of contaminating microorganisms [39]. When facing a low quantity of available PPE, extended use can limit waste [30,35] and can be done safely with standardized extended use practices that include proper donning and doffing [30,35,40,49]. Further research should focus on developing low-cost, widely accessible PPE for repeated use.
Some of the included studies developed effective disinfection methods that may not be financially or logistically feasible during outbreaks of infection or in low-resource settings, as they may require significant resource investment [31,57,60]. Disinfection methods that required significant infrastructure or cost include vaporized hydrogen peroxide and pulsed xenon ultraviolet light. On the other hand, effective disinfectants using common and inexpensive methods, such as microwave steam bags and steam sterilization using a kitchen rice cooker, are available to healthcare settings with limited resources [36,42,47,52,54,58,61].
Overall, included studies suggest that N95 respirators are superior to surgical masks with respect to preventing spread and acquisition of pathogens that are spread by aerosols. For pathogens that are spread through respiratory droplets and fomites, surgical masks can be as effective as N95 respirators. Several studies found surgical masks to be more effective than cloth masks. There are several potential reasons for this, including the mode of transmission studied (aerosol vs droplet) and variability in materials or fit of cloth and novel masks [61,[64][65][66]76]. This suggests that the cost of surgical masks, rather than use or reuse of cloth masks, is worth prioritizing despite limited resources. The additional cost of N95s can be reserved for aerosol-generating procedures, direct care of patients with respiratory pathogens that are transmitted via aerosols, and during seasons with high transmission of respiratory viruses.
Our findings are consistent with those of Rowan and Laffey's recent rapid review of SARS-CoV-2-related PPE use and reuse, which found varied global techniques for PPE reuse and reprocessing in the setting of the COVID-19 pandemic [8] [13][14][15][16]18,81] conclude multiple decontamination approaches are appropriate for N95s, most notably ultraviolet germicidal irradiation and vaporized hydrogen peroxide; however, all studies note a lack of strong research.
Our review spanned 14 countries and included analyses conducted both during times of disease outbreak and otherwise. It also captured multiple unconventional and innovative study methodologies. The variety of conclusions in the literature illustrates that responses to infectious disease outbreaks are highly context-and microbe-specific, requiring robust but sustainable supply chains. Standardizing the metrics of PPE efficacy could enable more accurate comparison and determination of best practices for conservation.

Limitations
Given the variety of study settings and approaches in the included literature, we were unable to quantitatively compare equivalent outcome metrics or variables across all studies. Instead, we qualitatively compared the main findings and recommendations. In addition, our review did not include publications after June 2020. Thus, we intentionally focus on the literature that existed during the early stages of the COVID-19 pandemic, when healthcare settings made emergency response decisions about PPE.
Contrary to our initial hypothesis, the average HDI of study locations included in this review was higher than the worldwide average. Only two included studies attempted to summarize the diverse field of PPE use and conservation practices in low-resource settings [33,64]. Off-label reuse practices frequently occur in settings without strict regulations [33], but the safety and efficacy of these practices have not been studied. Further research on these practices would illuminate both safer approaches to resource conservation in resource-limited settings and safe PPE conservation practices that can be employed broadly for greater environmental sustainability and resilience [19].
The methods deemed successful by this body of research cannot be generalized to lowand middle-income countries, which might have different resource and cost constraints [82]. Future literature reviews are needed to address dimensions of PPE use and reuse that were not evaluated in our systematic review, such as cost and comfort.

Conclusions
This review identified multiple methods for safe and effective reuse and disinfection of single-use PPE such as N95 respirators, which can be viable options for conserving PPE in different healthcare settings and for use by the public. It also highlights that surgical masks, which are lower in cost compared with N95 respirators, are effective in reducing the risk of infection from pathogens spread via respiratory droplets and fomites.
Future research should prioritize (i) assessing the type and effectiveness of PPE conservation practices in developing contexts, (ii) the development of low-cost PPE that can be disinfected and reused, (iii) the creation of protocols for healthcare systems to safely reuse and disinfect existing single-use PPE during times of extreme shortage, and (iv) the standardization of updated criteria for the use of surgical masks vs N95 respirators for specific infectious diseases. Such interventions would both bolster our preparedness for future supply shortages in times of healthcare overutilization, as well as reduce the environmentally harmful volume of waste generated by healthcare systems.