What is the single most important standard precaution EMTs can take to protect themselves from spreading or contracting a disease?

Cochrane Database Syst Rev. 2020; 2020(4): CD011621.

Abstract

Background

In epidemics of highly infectious diseases, such as Ebola, severe acute respiratory syndrome (SARS), or coronavirus (COVID‐19), healthcare workers (HCW) are at much greater risk of infection than the general population, due to their contact with patients' contaminated body fluids. Personal protective equipment (PPE) can reduce the risk by covering exposed body parts. It is unclear which type of PPE protects best, what is the best way to put PPE on (i.e. donning) or to remove PPE (i.e. doffing), and how to train HCWs to use PPE as instructed.

Objectives

To evaluate which type of full‐body PPE and which method of donning or doffing PPE have the least risk of contamination or infection for HCW, and which training methods increase compliance with PPE protocols.

Search methods

We searched CENTRAL, MEDLINE, Embase and CINAHL to 20 March 2020.

Selection criteria

We included all controlled studies that evaluated the effect of full‐body PPE used by HCW exposed to highly infectious diseases, on the risk of infection, contamination, or noncompliance with protocols. We also included studies that compared the effect of various ways of donning or doffing PPE, and the effects of training on the same outcomes.

Data collection and analysis

Two review authors independently selected studies, extracted data and assessed the risk of bias in included trials. We conducted random‐effects meta‐analyses were appropriate.

Main results

Earlier versions of this review were published in 2016 and 2019. In this update, we included 24 studies with 2278 participants, of which 14 were randomised controlled trials (RCT), one was a quasi‐RCT and nine had a non‐randomised design.

Eight studies compared types of PPE. Six studies evaluated adapted PPE. Eight studies compared donning and doffing processes and three studies evaluated types of training. Eighteen studies used simulated exposure with fluorescent markers or harmless microbes. In simulation studies, median contamination rates were 25% for the intervention and 67% for the control groups.

Evidence for all outcomes is of very low certainty unless otherwise stated because it is based on one or two studies, the indirectness of the evidence in simulation studies and because of risk of bias.

Types of PPE

The use of a powered, air‐purifying respirator with coverall may protect against the risk of contamination better than a N95 mask and gown (risk ratio (RR) 0.27, 95% confidence interval (CI) 0.17 to 0.43) but was more difficult to don (non‐compliance: RR 7.5, 95% CI 1.81 to 31.1). In one RCT (59 participants), people with a long gown had less contamination than those with a coverall, and coveralls were more difficult to doff (low‐certainty evidence). Gowns may protect better against contamination than aprons (small patches: mean difference (MD) −10.28, 95% CI −14.77 to −5.79). PPE made of more breathable material may lead to a similar number of spots on the trunk (MD 1.60, 95% CI −0.15 to 3.35) compared to more water‐repellent material but may have greater user satisfaction (MD −0.46, 95% CI −0.84 to −0.08, scale of 1 to 5).

Modified PPE versus standard PPE

The following modifications to PPE design may lead to less contamination compared to standard PPE: sealed gown and glove combination (RR 0.27, 95% CI 0.09 to 0.78), a better fitting gown around the neck, wrists and hands (RR 0.08, 95% CI 0.01 to 0.55), a better cover of the gown‐wrist interface (RR 0.45, 95% CI 0.26 to 0.78, low‐certainty evidence), added tabs to grab to facilitate doffing of masks (RR 0.33, 95% CI 0.14 to 0.80) or gloves (RR 0.22, 95% CI 0.15 to 0.31).

Donning and doffing

Using Centers for Disease Control and Prevention (CDC) recommendations for doffing may lead to less contamination compared to no guidance (small patches: MD −5.44, 95% CI −7.43 to −3.45). One‐step removal of gloves and gown may lead to less bacterial contamination (RR 0.20, 95% CI 0.05 to 0.77) but not to less fluorescent contamination (RR 0.98, 95% CI 0.75 to 1.28) than separate removal. Double‐gloving may lead to less viral or bacterial contamination compared to single gloving (RR 0.34, 95% CI 0.17 to 0.66) but not to less fluorescent contamination (RR 0.98, 95% CI 0.75 to 1.28). Additional spoken instruction may lead to fewer errors in doffing (MD −0.9, 95% CI −1.4 to −0.4) and to fewer contamination spots (MD −5, 95% CI −8.08 to −1.92). Extra sanitation of gloves before doffing with quaternary ammonium or bleach may decrease contamination, but not alcohol‐based hand rub.

Training

The use of additional computer simulation may lead to fewer errors in doffing (MD −1.2, 95% CI −1.6 to −0.7). A video lecture on donning PPE may lead to better skills scores (MD 30.70, 95% CI 20.14 to 41.26) than a traditional lecture. Face‐to‐face instruction may reduce noncompliance with doffing guidance more (odds ratio 0.45, 95% CI 0.21 to 0.98) than providing folders or videos only.

Authors' conclusions

We found low‐ to very low‐certainty evidence that covering more parts of the body leads to better protection but usually comes at the cost of more difficult donning or doffing and less user comfort, and may therefore even lead to more contamination. More breathable types of PPE may lead to similar contamination but may have greater user satisfaction. Modifications to PPE design, such as tabs to grab, may decrease the risk of contamination. For donning and doffing procedures, following CDC doffing guidance, a one‐step glove and gown removal, double‐gloving, spoken instructions during doffing, and using glove disinfection may reduce contamination and increase compliance. Face‐to‐face training in PPE use may reduce errors more than folder‐based training.

We still need RCTs of training with long‐term follow‐up. We need simulation studies with more participants to find out which combinations of PPE and which doffing procedure protects best. Consensus on simulation of exposure and assessment of outcome is urgently needed. We also need more real‐life evidence. Therefore, the use of PPE of HCW exposed to highly infectious diseases should be registered and the HCW should be prospectively followed for their risk of infection.

Plain language summary

Protective clothes and equipment for healthcare workers to prevent them catching coronavirus and other highly infectious diseases

Background

Healthcare workers treating patients with infections such as coronavirus (COVID‐19) are at risk of infection themselves. Healthcare workers use personal protective equipment (PPE) to shield themselves from droplets from coughs, sneezes or other body fluids from infected patients and contaminated surfaces that might infect them. PPE may include aprons, gowns or coveralls (a one‐piece suit), gloves, masks and breathing equipment (respirators), and goggles. PPE must be put on correctly; it may be uncomfortable to wear, and healthcare workers may contaminate themselves when they remove it. Some PPE has been adapted, for example, by adding tabs to grab to make it easier to remove. Guidance on the correct procedure for putting on and removing PPE is available from organisations such as the Centers for Disease Control and Prevention (CDC) in the USA.

This is the 2020 update of a review first published in 2016 and previously updated in 2019.

What did we want to find out?

We wanted to know:

what type of PPE or combination of PPE gives healthcare workers the best protection;

whether modifying PPE for easier removal is effective;

whether following guidance on removing PPE reduced contamination;

whether training reduced contamination.

What did we find?

We found 24 relevant studies with 2278 participants that evaluated types of PPE, modified PPE, procedures for putting on and removing PPE, and types of training. Eighteen of the studies did not assess healthcare workers who were treating infected patients but simulated the effect of exposure to infection using fluorescent markers or harmless viruses or bacteria. Most of the studies were small, and only one or two studies addressed each of our questions.

Types of PPE

Covering more of the body leads to better protection. However, as this is usually associated with increased difficulty in putting on and removing PPE, and the PPE is less comfortable, it may lead to more contamination. Coveralls are the most difficult PPE to remove but may offer the best protection, followed by long gowns, gowns and aprons. Respirators worn with coveralls may protect better than a mask worn with a gown, but are more difficult to put on. More breathable types of PPE may lead to similar levels of contamination but be more comfortable. Contamination was common in half the studies despite improved PPE.

Modified PPE

Gowns that have gloves attached at the cuff, so that gloves and gown are removed together and cover the wrist area, and gowns that are modified to fit tightly at the neck may reduce contamination. Also, adding tabs to gloves and face masks may lead to less contamination. However, one study did not find fewer errors in putting on or removing modified gowns.

Guidance on PPE use

Following CDC guidance for apron or gown removal, or any instructions for removing PPE compared to an individual’s own preferences may reduce self‐contamination. Removing gown and gloves in one step, using two pairs of gloves, and cleaning gloves with bleach or disinfectant (but not alcohol) may also reduce contamination.

User training

Face‐to‐face training, computer simulation and video training led to fewer errors in PPE removal than training delivered as written material only or a traditional lecture.

Certainty of the evidence

Our certainty (confidence) in the evidence is limited because the studies simulated infection (i.e. it was not real), and they had a small number of participants.

What do we still need to find out?

There were no studies that investigated goggles or face shields. We are unclear about the best way to remove PPE after use and the best type of training in the long term.

Hospitals need to organise more studies, and researchers need to agree on the best way to simulate exposure to a virus.

In future, simulation studies need to have at least 60 participants each, and use exposure to a harmless virus to assess which type and combination of PPE is most protective.

It would be helpful if hospitals could register and record the type of PPE used by their workers to provide urgently needed, real‐life information.

Search date

This review includes evidence published up to 20 March 2020.

Summary of findings

Background

Description of the condition

Over 59 million people are employed in the healthcare sector worldwide (WHO 2006). Some of these healthcare workers (HCW) are at risk of developing life‐threatening infectious diseases due to contact with patients’ blood or body fluids such as mucus, vomit or exhaled droplets. The risk of infection and its consequences vary, but it is well recognised as an occupational risk (Heptonstall 2010; Sepkowitz 2005). Especially during epidemics, these risks become more visible as the infection rate among HCW is higher than in the general population. Another risk of HCW infection is that infected HCWs will infect patients or that they will act as a vector for the transfer of the disease between patients. In addition, during epidemics, infected HCW will further diminish the capacity of an already overburdened healthcare system.

The 2013 to 2015 Ebola Virus Disease (EVD) epidemic put HCW at high risk of a disease with a very high fatality rate in the epidemic areas (Ebola 2014). According to the World Health Organization (WHO), healthcare workers were between 21 and 32 times more likely to be infected with Ebola than people in the general adult population (Forrester 2014; WHO 2015a). According to the statistics from the 2013‐2015 West Africa EVD epidemic, there were 1049 registered cases of infected HCW with 535 deaths (Kilmarx 2014; WHO 2015b).

Just a decade earlier during the 2002 to 2003 Severe Acute Respiratory Syndrome (SARS) epidemic, 20% of all patients were healthcare workers of whom about 10% lost their lives (WHO 2003).

During the COVID‐19 pandemic, HCW are at higher risk of infection than the general population, just as during other epidemics. Experts strongly urge the use of proper personal protective equipment (PPE) for the HCWs' and patients' safety (Adams 2020; Chang 2020). In a Chinese case‐series of 138 consecutive patients that were hospitalised in Wuhan, China during the month of January 2020, 30% were HCW, which is considerably higher than expected (Wang 2020). Remuzzi 2020 reports that in Lombardy, Italy as of 12 March 2020, 20% of HCW at intensive care units became infected, while Giwa 2020 estimates that at least 10% of HCW in Italy will become infected in spite of using PPE.

HCW may become infected through various routes of transmission, depending on the pathogen. Infection can occur through splashes and droplets of contaminated body fluids on non‐intact skin, or via needle‐stick injuries through intact skin. Infection can also occur when splashes or droplets of contaminated body fluids land on the mucous membranes in the eyes, mouth or nose, or when the same mucous membranes come into contact with contaminated skin, such as when rubbing the eyes with a hand carrying pathogens after touching a patient or contaminated surface (Siegel 2019). For EVD, contact transmission is the main route of transmission. For SARS, the highest risk of infection was due to inhalation of aerosols, but the disease was also transmitted through droplet and contact infection. For COVID‐19 the main route of exposure is through droplet transmission and contact transmission but other transmission routes are also possible (Chang 2020; Otter 2016; Peng 2020).

Here, we focus on highly infectious diseases, which means that contamination with infectious material can readily lead to clinical disease. We also focus on those infections that have serious consequences, such as a high case fatality rate, because the motivation of HCW to protect themselves will be different in situations where the risk is low and the consequences are not serious. The term 'high consequence pathogen' is also used but the list of what constitutes a high consequence pathogen varies from country to country. The European Network for Infectious Diseases defines highly infectious disease as an infectious disease easily transmitted from person to person, causing life‐threatening disease, presenting a serious hazard in healthcare settings and in the community, and requiring specific control measures (Brouqui 2009).

Description of the intervention

In the occupational health field, the 'hierarchy of controls' is best practice. This means that measures with a general effect such as control of exposure should have priority over more individual control measures such as PPE. Exposure of HCW can be best controlled by organisational measures that minimise the exposure to contaminated body fluids or infected patients. The most important preventive measure is the proper organisation of the hospital or healthcare unit to avoid unnecessary contact. Once this has been implemented, the main strategy for reducing physical exposure to highly infectious diseases is through PPE. Both in the European Union (EU) and in the USA, it is mandatory for employers to protect their workers against blood‐borne pathogens and other infections at work (OSHA 2012; EU 2010).

Coveralls, gowns, hoods, masks, goggles and face shields, among others, are used to prevent skin and mucous membranes from becoming contaminated and respirators are used to prevent inhalation. Depending on the transmission route and the specifics of the infection, different types of PPE are recommended. PPE in health care are usually considered as part of what is called transmission‐based precautions. Standard precautions or universal precautions are based on the principle that all blood, body fluids, secretions, excretions except sweat, non‐intact skin, and mucous membranes may contain transmissible infectious agents. Depending on anticipated exposure, hand hygiene and the use of PPE such as gloves, gowns, masks, eye protection (i.e. goggles or face shields) should be implemented. When the route(s) of transmission is (are) not completely interrupted using standard precautions alone, there are three categories that elaborate the precautions to be taken: contact precautions, droplet precautions, and airborne precautions (Siegel 2019).These precautions contain a number of measures including appropriate PPE to prevent the specific modes of transmission.

PPE will only be effective if the equipment can form a barrier between the HCW and the contaminated body fluids. Therefore, standards have been developed that, when complied with, ensure that PPE is of sufficient quality to protect against biohazards (Mäkelä 2014; NIOSH 2014). Even though the biohazard symbol (Figure 1), is widely used to indicate the presence of biohazards, it is not a label for protective clothing. For biohazards, these standards are based on laboratory tests that evaluate to what extent the fabric and the seams of protective clothing are leak‐tight, that is, are they impermeable for liquids, viruses, or both at certain pressure levels. The standards in the EU and the USA are different. PPE should contain a label that specifically indicates the standards against which it has been tested.

International symbol indicating biohazards

Technical standards for PPE

Technical standards for PPE are complicated and the categorisation is confusing. In the EU, there is standard EN 14126 for clothing, specifically coveralls that protect workers against biological hazards from micro‐organisms. Clothing compliant with the standard EN 14126 is further classified according to routes of contamination and the circumstances in which contamination may occur (pressurized contaminated liquid, mechanical contact with substances containing contaminated liquid, contaminated liquid aerosols, contaminated solid particles) based on ISO 2004a and ISO 2004b test methods. There is a separate standard for surgical gowns, EN 13795, but this standard is specifically designed to protect the patient.

In the USA, ANSI/AAMI PB70 2012 standard classifies surgical and isolation gowns according to their liquid barrier performance with four levels of protection, with level 4 offering the most protection against viral and liquid penetration but level 1 offering only minimal water resistance. There are several differences between ANSI/AAMI PB70 2012 and EN 13795 surgical gown classifications. Because the test methods and performance requirements cannot be compared directly, it is difficult to assign equivalency between surgical gowns classified according to ANSI/AAMI PB70 2012 and EN 13795. There is also US standard NFPA 1999 which was specifically developed to address a range of different protective clothing items worn by emergency medical service first responders, and also applies to medical first receivers. NFPA 1999 lists many performance requirements for protective clothing used by emergency medical personnel, including (but not limited to) viral penetration resistance, tensile strength, liquid integrity, and seam strength.

To summarise, the qualities of protective clothing certified by different standards are not fully comparable and complex. Nonetheless, they all aim to ensure that protective clothing is of a quality that prohibits water and blood‐like fluids with virus particles, applied under a specified amount of pressure, from passing through. In addition, some standards have requirements that the whole piece of clothing, including the seams, must be non‐permeable to liquids (NFPA 1999).

Clothing that is manufactured according to the standards mentioned above, at the appropriate level of protection, is impermeable to body fluids and viruses and will technically prevent skin contamination. However, this review does not deal with the technical physical standards of equipment, but rather whether and how its use in practice will prevent contamination and infection.

Guidelines for choosing proper PPE

In 2014, the WHO developed a guideline for infection prevention and control of epidemic‐ and pandemic‐prone acute respiratory infections in health care. The guideline strongly recommends using appropriate PPE as determined by risk assessment (according to the procedure and suspected pathogen). Appropriate PPE when providing care to patients presenting with acute respiratory infection (ARI) syndromes may include a combination of: medical mask (surgical or procedure mask); gloves; long‐sleeved gowns; and eye protection (goggles or face shields). For aerosol‐generating procedures (AGPs) this combination including a surgical or a procedural mask or a particulate respirator is conditionally recommended. If splashing with blood or other body fluids is anticipated and gowns are not fluid‐resistant, a waterproof apron should be worn over the gown (WHO 2014).

For COVID‐19, recommendations for PPE are gloves, masks, goggles or face shields, and long‐sleeved gowns (WHO 2020a; WHO 2020b) with N95 respirators recommended over masks for AGPs, consistent with the WHO 2014 guideline. Masks are further described as medical mask (flat, pleated or cup‐shaped, affixed to head with a strap). Otherwise there are no quality criteria provided for the PPE parts. This is especially worrying because isolation gowns can have very different qualities, of which the end users are usually not aware (Kilinc‐Balci 2016). Most isolation gown models also leave the neck unprotected, which could be a source of contamination (Zamora 2006). Centers for Disease Control and Prevention (CDC) recommends that non‐sterile, disposable patient isolation gowns, which are used for routine patient care in healthcare settings, are appropriate for use by HCW when caring for patients with suspected or confirmed COVID‐19. Current US guidelines do not require use of gowns that conform to any standards (CDC 2020a). If there is a medium to high risk of contamination, CDC recommends isolation gowns that claim moderate to high barrier protection (ANSI/AAMI PB70 2012 level 3 or 4; CDC 2020b). For a proper overview of requirements for and use of isolation gowns see Kilinc‐Balci 2015 and Kilinc‐Balci 2016.

During the EVD epidemic, several guidelines became available for choosing proper PPE (Australian NHMRC 2010; CDC 2014; ECDC 2014; WHO 2016). Even though all guidelines propose using similar protective clothing, there are differences. For example, ECDC 2014 proposes taping gloves, boot covers and goggles onto the coveralls to prevent leaving any openings but the other guidelines do not recommend this. Most guidelines have recently been updated. There are also recommendations for the technical quality of the PPE to be used with Ebola. For gowns, WHO 2016 currently recommends EN 13795 high‐performance surgical gowns or ANSI/AAMI PB70 2012 level 3 (option 1), or level 4 (option 2), or equivalent. As the first option for coveralls, WHO currently recommends protection equivalent to EN 14126, level 3 protection against blood level 2 against viruses.

Overprotection can be a problem. Some propose using three layers of gloves, because according to their experience, this is best practice (Lowe 2014). However, it may make work more difficult, and eventually lead to an increased rather than a decreased risk of infection, especially during doffing (i.e. removing the PPE). For example, the combined use of several respirators probably does not lead to more protection, but considerably increases the burden on the worker (Roberge 2008a; Roberge 2008b).

Donning and doffing of PPE

Despite using proper PPE, probably the biggest risk of infection is associated with self‐contamination by HCW inappropriately removing the PPE (Fischer 2014). Some types of PPE make donning and doffing more difficult, thereby increasing the risk of contamination (Zamora 2006). There is evidence that when doffing PPE, the use of a double pair of gloves decreases the risk of contamination (Casanova 2012). How contamination of PPE occurs has also been clearly illustrated with a simulation study about cleaning up vomit (Makison 2014). The results of such simulation studies should increase HCW's confidence in executing the donning and doffing procedures correctly, and thus can also be an incentive for their uptake and compliance with the guidelines. Therefore, specific guidance has been developed for donning and doffing PPE (CDC 2014; WHO 2016).

Compliance with guidance on correct PPE use in health care is historically poor. HCW sometimes distrust infection control, and using PPE is stressful (Zelnick 2013). For respiratory protection such as masks and respirators, compliance has been reported to be around 50% on many occasions (Nichol 2008). Due to lack of proper fitting and incorrect use, real field conditions almost never match laboratory standards (Coia 2013; Howie 2005). Also, reports of hand hygiene show that there is still much room for improvement, and guidelines recommend education and training in combination with other implementation measures (WHO 2009). From reports of HCW, it is clear that most appropriate PPE is not user‐friendly in tropical conditions. It prevents heat loss through sweating because it is not made of breathable material. A common reason for a breach in the barrier of the PPE is the worker sweating and then instinctively wiping their face (Cherrie 2006).

In this review, we only concentrated on PPE for highly infectious diseases that have serious consequences for health, such as EVD and COVID‐19. We excluded other highly infectious, but less serious viral infections, such as norovirus, as we expected the effect of PPE to be different. We included SARS as it was highly infectious to HCW, sometimes fatal, and had similar recommendations on PPE use and training to COVID‐19.

We did not specifically study the effects of hand hygiene or of respiratory protection to prevent transmission through inhalation. Hand hygiene is also crucial in preventing skin contamination, but this has already been covered in another review (Gould 2010). The protective effect of different types of respiratory protection, and effects of interventions to increase their uptake are covered in two other reviews (Jefferson 2011; Luong Thanh 2016).

How the intervention might work

First, HCW, their supervisors, or occupational health professionals should choose the proper type of PPE, as indicated in the guidance described above. Then, the HCW needs to know how to don and doff PPE according to the guidelines provided. Next, the HCW needs to comply with established procedures for correctly using, donning and doffing PPE. Education and training are used to increase compliance. The emphasis in teaching the correct use of PPE is on doing everything slowly and carefully to minimise the risk of making a mistake. Often an assistant or buddy, sometimes coupled with a mirror, is used while donning PPE, while a hygienist supervises doffing.

Compliance can be increased by personal supervision and instruction, checklists, audits of performance, by providing feedback, and by allowing sufficient time for donning and doffing. Education and training on uptake and compliance with PPE should have an effect in both the short term and the long term (Northington 2007; Ward 2011). Education and training can be seen as one method to increase compliance (Gershon 2009; Hon 2008). Compliance with PPE can also be improved by providing sufficient, comfortable, well‐fitting, and more user‐ and patient‐friendly PPE. Compliance with guidelines has been studied for hand hygiene. There is some evidence that multifaceted interventions and staff involvement are important, but altogether, there is little evidence that allows firm conclusions (Gould 2010).

Why it is important to do this review

From studies conducted during the SARS epidemic and the EVD epidemic it has become clear that the use of gloves, gowns and masks each help to reduce the infection rate in HCW (Appendix 1, Verbeek 2016a). More consistent use of gloves, gowns, masks and goggles was each related to fewer infections among HCW. Also, theoretically, protecting the skin and the mucous membranes of the mouth nose and eyes will prevent transmission. We have therefore little doubt that in a technical sense PPE will help and that the minimum amount of PPE needed is gloves, gown, and mouth, nose and eye protection, as recommended by WHO and CDC. The guidance does not, however, indicate which type or quality‐level of PPE is most protective. In this review, we concentrate on finding out which PPE protects best by only including studies that compare one type of PPE against an alternative type of PPE, such as gowns against coveralls or goggles against face shields only when used as part of full PPE. We do not include studies that compare the use of PPE against no PPE, or studies comparing one type of PPE to another when not used as part of a set of full‐body PPE.

There is still uncertainty about the optimal type, composition, amount, and ways of using full‐body PPE to prevent skin and mucous membrane contamination of HCW while treating patients infected with highly infectious diseases. This is also reflected in the different ways guidelines for PPE are implemented in Europe (De Iaco 2012), and acknowledged in current WHO guidelines regarding EVD (WHO 2016). WHO realises that a safer, more comfortable and culturally appropriate protective system commensurate with the risk is needed and has provided guidance for industry, health workers, engineers, innovators, medical and scientific researchers, and others to re‐think, energise, and innovate for a better PPE system for the HCW responding to Ebola virus outbreaks in tropical climates (WHO 2018).

Since full‐body protection has mainly evolved as a direct result of experiences gained from the recent outbreaks of deadly viruses, there are still many types of PPE available with varying types of components. The comparative effectiveness of one type against another is still unknown. Regarding the equipment, there is uncertainty whether face shields protect as well as goggles, especially when goggles are combined with a hood. There is uncertainty whether and when double or triple gloves would be more protective than single gloves. Regarding suits, it is unclear if gowns are as protective as coveralls, and how breathable and impermeable for liquids or viruses they should be. Some argue that using more breathable material would decrease the risk of contamination (Kuklane 2015).

When it comes to donning and doffing procedures for EVD, there is uncertainty about the effect of integrity checks of gloves and other equipment, and whether gloves should be changed when highly contaminated. With doffing especially, it is unclear if this should be done in pairs with a helper buddy removing part of the PPE, or if this can be done alone. Another element of the doffing procedure that is uncertain is if spraying with a disinfectant such as chlorine spray is more protective than not using spray. It is not clear which disinfectant is the best antiviral: chlorine solution or alcohol gel, and at which concentration.

Also, for COVID‐19, different procedures for donning and doffing PPE are recommended. Giwa 2020 proposes a specific procedure of doffing PPE, but the procedure is not consistent with the procedures proposed by CDC (CDC 2020c). Others, including the CDC, have proposed that gown and gloves should be doffed in a one‐step procedure (Osei‐Bonsu 2019), to minimise self‐contamination.

It is also unclear what are the best ways to train HCW and how to best maintain the skills needed for proper use of PPE.

This review is a timely update of the Verbeek 2019 review, the results of which indicated that more research is still needed to answer the review's questions.

Objectives

To evaluate which type of full‐body PPE and which method of donning or doffing PPE have the least risk of contamination or infection for HCW, and which training methods increase compliance with PPE protocols. In particular, we evaluated the effect of:

• different types of PPE on contamination and infection rates or on compliance (one type or component of full‐body protection PPE versus another);

• different donning or doffing procedures on contamination and infection rates or on compliance (one procedure for donning and doffing full‐body PPE versus another); and

• different types of education and training aiming to improve compliance with guidelines for full‐body PPE on compliance, contamination and infection rates, (one type of training versus another).

Methods

Criteria for considering studies for this review

Types of studies

Since the circumstances for evaluation studies are difficult during epidemics, we anticipated including a broad range of study designs.

We included any prospective or retrospective controlled field study. Field study here refers to a study that tests interventions with healthcare staff in a real‐life exposure situation. This also includes case‐control studies that compare the use of interventions retrospectively between cases that have become infected and comparable controls that did not get infected.

We also included randomised as well as non‐randomised prospective controlled studies that simulated exposure to contaminated body fluids with the use of marker chemicals or harmless viruses or bacteria.

We excluded studies without a comparison group, but did not exclude studies on the basis of type of comparison group.

Types of participants

For simulation studies, we included any type of participants (volunteers or HCW) using PPE designed for EVD or comparable highly infectious diseases with serious consequences.

For field studies, we included studies only if they were conducted with HCW or ancillary staff exposed to body fluids from patients in the form of splashes, droplets, or aerosols contaminated with particles of highly infectious diseases that have serious consequences for health such as EVD, SARS, or COVID‐19.

We excluded studies conducted with laboratory staff because the preventive measures in labs are more detailed and easier to comply with.

Types of interventions

1. We included studies that evaluated the effectiveness of different types of full‐body protection (PPE), or comparing different types, compositions, or amounts of the following PPE components:

• body protection such as gowns, coveralls, or hazardous materials (hazmat) suits;

• eye and face protection such as glasses, goggles, face shields or visors, or masks or hoods that cover the entire head;

• hand protection: gloves; and

• foot protection: overshoes or boots.

We defined PPE as any of the equipment listed above that is designed or intended to protect healthcare staff from contamination with infected patients' body fluids.

2. We included studies that evaluated the effectiveness of different PPE parts or different procedures or protocols for donning and doffing of the PPE.

For example, extra assistance during donning and doffing, extra disinfection, or the use of extra gloves to prevent contamination in comparison to standard protocols.

3. We included studies that evaluated the effectiveness of training to increase compliance with existing guidance on the selection or use of PPE, including but not limited to:

• education (courses);

• practical training;

• information only (such as posters, guideline leaflets, etc.);

• audit and feedback; or

• monetary or organisational incentives.

Types of outcome measures

Primary outcomes

We included all studies that had measured the effectiveness of interventions as:

• contamination of skin or clothing, measured with any type of test material to visualise contamination (e.g. stains made visible with UV light) or harmless viruses or bacteria;

• infection with EVD, another viral haemorrhagic fever, or comparable highly infectious disease with serious consequences such as SARS, or COVID‐19;

• compliance with guidance on selection of type and use of PPE measured, for example, with an observation checklist.

Secondary outcomes

• User‐reported assessment of comfort and convenience

• Costs or resource use

• Time to don and doff the PPE

The secondary outcomes were not a criterion for including studies in this review.

Search methods for identification of studies

Electronic searches

We conducted a systematic literature search to identify all published and unpublished trials that could be considered eligible for inclusion in this review. We adapted the search strategy we developed for Medline through PubMed (see Appendix 2) for use in the other electronic databases. The literature search identified potential studies in all languages. 

We searched the following electronic databases from inception to the dates presented underneath for identifying potential studies (search dates provided below). We searched with different interfaces for the various updates. The searches are listed in the appendices for all interfaces. For the 2020 update we did not search OSH‐Update because the earlier search yielded so little.

• Cochrane Central Register of Controlled Trials (CENTRAL; 2020, Issue 3) via Wiley Online Library (Appendix 3);

• NIOSHTIC (OSH‐UPDATE) (Appendix 10) to 31 December 2018;

• NIOSHTIC‐2 (OSH‐UPDATE) to 31 December 2018;

• HSELINE (OSH‐UPDATE) to 31 December 2018;

• CISDOC (OSH‐UPDATE) to 31 December 2018;

We also conducted a search of ClinicalTrials.gov (www.ClinicalTrials.gov), and the WHO trials portal (www.who.int/ictrp/en/), which includes the Pan African Registry for potential studies on EVD for the 2016 and 2019 updates. For the 2020 update we searched the WHO trials portal for COVID 19/SARS‐CoV‐2. We searched all databases from their inception to the present for the first versions of the review. We searched from the earliest date of search to the present for updates of the review. We did not impose a restriction on language of publication.

Searching other resources

We checked reference lists of all primary studies and reviewed articles for additional references. For the 2016 version of the review, we contacted non‐governmental organisations involved in medical relief operations in the high‐risk EVD areas to identify additional unpublished materials on protection against EVD (Médécins Sans Frontières (MSF) and Save the Children). We also used Twitter to ask for unpublished reports from people in the field. Evidence Aid helped in locating relevant organisations and in asking them for unpublished reports. We also contacted DuPont, and 3M, PPE manufacturers, to request unpublished studies.

In addition, we used Google to find any unpublished or grey literature on our question that may not be available from the sources listed above by using the following terms: 'personal protective equipment ebola'. For the March 2020 update we conducted a search of Google Scholar using the search phrase ('SARS CoV 2' OR 'COVID' AND 'protective equipment' AND 'healthcare worker').

Data collection and analysis

Selection of studies

Pairs of review authors (JV, RS, BR, ET, BB, CT, SI, JR) independently screened titles and abstracts of all systematic search results to identify studies for inclusion. The same review authors coded them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. We retrieved the full‐text study reports/publication and pairs of review authors (JV, ET, BR, RS, BB, CT, SI, JR) independently screened the full text, identified studies for inclusion, and identified and recorded reasons for exclusion of the ineligible studies. We used the computer programme Covidence for the selection of references and full‐text studies. We resolved any disagreement through discussion, except in two cases where a third‐person assessment (SI or CT) was needed. We identified and excluded duplicates and collated multiple reports of the same study so that each study rather than each report is the unit of interest in the review. We recorded the selection process and completed a PRISMA flow diagram (Moher 2009), for the search for our original review (Figure 2), our updated review (Figure 3) and this update (Figure 4). We also completed a 'Characteristics of excluded studies' table.

PRISMA study flow diagram for search up to January 2016

PRISMA study flow diagram for search between 2016 and 2018

Study flow diagram for 2020 April update

Data extraction and management

We used Covidence for extracting study characteristics and outcome data. Two review authors (JV, BR, BB, ET, CT, RS, SI, JR) independently extracted the following study characteristics from included studies.

• Methods: study design, total duration of study, study location, study setting, withdrawals, and date of study

• Participants: number, mean age or age range, sex, severity of condition, diagnostic criteria if applicable, inclusion criteria, and exclusion criteria

• Interventions: description of intervention, comparison, duration, intensity, content of both intervention and control condition, and co‐interventions

• Outcomes: description of primary and secondary outcomes specified and collected, and at which time points reported

• Notes: funding for trial, and notable conflicts of interest of trial authors, country where trial was conducted

Pairs of review authors (JV, BR, CT, SI, JR, ME, RS) independently extracted outcome data from included studies. We noted in the 'Characteristics of included studies' table if outcome data were not reported in a usable way. We resolved disagreements by consensus so there was no need to involve a third review author. One review author (JV or BR) transferred the data into Review Manager 5 (Review Manager 2014). We double‐checked that data had been entered correctly by comparing the data presented in the systematic review with the study reports. A second review author (CT or JV) spot‐checked study characteristics for accuracy against the trial report.

Assessment of risk of bias in included studies

Pairs of two review authors (JV, BR, CT, SI, JR, ME, RS) independently assessed risk of bias for each randomised study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). We resolved any disagreements by discussion so there was no need to involve another review author. We assessed the risk of bias according to the following domains in all RCTs.

1. Random sequence generation

2. Allocation concealment

3. Blinding of participants and personnel

4. Blinding of outcome assessment

5. Incomplete outcome data

6. Selective outcome reporting, and

7. Other bias

We rated each potential source of bias as high, low, or unclear and provided a quote from the study report or study author together with a justification for our judgment in the 'Risk of bias' table. We summarised the 'Risk of bias' judgements across different studies for each of the domains listed. For compliance, we considered blinding to PPE type significant for the outcome assessor only. Where information on risk of bias relates to unpublished data or correspondence with a study author, we noted this in the 'Risk of bias' table.

We considered randomised studies to have a low overall risk of bias when we judged random sequence generation and blinded outcome assessment to have a low risk of bias and none of the other domains to have a high risk of bias.

We used the domains blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other bias for all non‐randomised studies. Instead of the domains random sequence generation and allocation concealment, we used the following items as suggested in the ROBINS‐I tool (Sterne 2016), for the assessment of risk of bias in non‐randomised intervention studies.

• Bias due to confounding. We made an overall assessment of risk of bias based on the following questions if the signalling question, 'Is confounding of the effect of intervention unlikely in this study?' was answered with no.

  • Did the authors use an appropriate analysis method that adjusted for all the critically important confounding domains?

  • Were confounding domains that were adjusted for measured validly and reliably by the variables available in this study? For this review question, we considered baseline differences between compared groups in the following factors significant: prior experience with PPE, healthcare qualification, or education of HCW, age and sex, ambient temperatures, and stressful activities.

• Bias due to selection of participants into the study. We made an overall assessment of this risk of bias based on the following questions if the signalling questions, 'Was selection into the study unrelated to intervention or unrelated to outcome?, and 'Do start of follow‐up and start of intervention coincide for most participants?' were answered with no.

  • Were adjustment techniques used that are likely to correct for the presence of selection biases?

  • For case‐control studies: were the controls sampled from the population that gave rise to the cases, or using another method that avoids selection bias?

We considered the domains of confounding and selection of participants to yield high, low, or unclear risk of bias. For a non‐randomised study as a whole, we considered the study to have a low risk of bias if all domains received a judgment of low risk of bias comparable to an RCT. This means receiving a low 'Risk of bias' judgment on the two domains listed above as well as domains three to seven in the previous section.

When considering treatment effects, we took into account the risk of bias for the studies that contributed to that outcome.

We judged studies to have a low overall risk of bias if we judged them to have a low risk of bias in the following domains: both random allocation and allocation concealment, or both confounding and selection bias, and incomplete outcome data and selective reporting. We considered the blinding of participants and outcome assessors less important because the outcomes were objective or we could not imagine that participants would have an interest in a certain type of attire and outcome.

Assessment of bias in conducting the systematic review

We conducted the review according to the published protocol (Verbeek 2015), and where there were deviations from it, we reported these in the 'Differences between protocol and review' section of the systematic review.

Measures of treatment effect

We entered the outcome data for each study into the data tables in Review Manager 2014 to calculate the treatment effects. We used risk ratios (RRs) for dichotomous outcomes, and mean differences (MDs) or standardised mean differences (SMDs) for continuous outcomes. When studies reported only effect estimates and their 95% confidence intervals or standard errors, we entered these data into Review Manager 2014 using the generic inverse variance method. When study authors used multivariate analyses, we used the most adjusted OR (odds ratios) or RRs. We ensured that higher scores for continuous outcomes had the same meaning for the particular outcome, explained the direction and reported where the directions were reversed, if this was necessary. If, in future updates of this review, we come across studies reporting results that we cannot enter in either way, we will describe them in the 'Characteristics of included studies' table, or we will enter the data into additional tables. For cohort studies that compare an exposed to a non‐exposed population we intended to report both the RR for the intervention versus the control at baseline and at follow‐up for dichotomous outcomes to indicate the change brought about by the intervention but we did not find any such studies.

Unit of analysis issues

If in future updates of this review we come across studies that employ a cluster‐randomised design and that report sufficient data to be included in the meta‐analysis but do not make an allowance for the design effect, we will calculate the design effect based on a fairly large assumed intra‐cluster correlation of 0.10. We based this assumption of 0.10 being a realistic estimate by analogy on studies about implementation research (Campbell 2001). We will follow the methods stated in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) for the calculations.

We intended to take the paired nature of the cross‐over design in the included studies into account in our data analysis. However, the included studies did not present sufficient data to do so and the results presented here are based on the unpaired test that is implemented in Review Manager 2014 which resulted in wider confidence intervals than with the use of a paired t‐test.

Dealing with missing data

We contacted investigators to verify key study characteristics and obtain missing numerical outcome data where possible (e.g. when a study was identified as abstract only). If in future updates of this review we come across studies where this is not possible, and the missing data are thought to introduce serious bias, we will explore the impact of including such studies in the overall assessment of results by a sensitivity analysis.

Similarly, if in future updates of this review we come across studies where numerical outcome data are missing, such as SDs or correlation coefficients and they cannot be obtained from the authors, we will calculate them from other available statistics such as P values, according to the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017).

Assessment of heterogeneity

We assessed the clinical homogeneity of the results of included studies based on similarity of population, intervention, outcome and follow‐up. We considered populations as similar when they were HCWs directly engaged in patient treatment (nurses, doctors, paramedics) versus those who were not involved in patient therapy directly (cleaning and transport staff).

We considered interventions as similar when they fell into one of the intervention categories as stated in Types of interventions.

We considered any assessment of contamination of the skin or mucous membranes as similar enough to combine.

We considered the following follow‐up times as similar: from immediately following a procedure up until the end of the work shift (short‐term), and any time after the incubation time (long‐term).

If in future updates of this review we come across studies with results that we can pool with meta‐analysis, we will use the I² statistic (Higgins 2003), to measure heterogeneity among the trials in each analysis. Where we identify substantial heterogeneity, we will report it and explore possible causes by prespecified subgroup analysis. We will regard an I² value above 50% as substantial heterogeneity (Deeks 2017).

Assessment of reporting biases

For a future update, if we are able to pool more than five trials in any single meta‐analysis, we will create and examine a funnel plot to explore possible small study biases.

Data synthesis

In future updates of this review we will pool data from studies we judge to be clinically homogeneous using Review Manager software (RevMan Web 2019). If more than one study provides usable data in any single comparison, we will perform meta‐analysis. We will use a random‐effects model when I² is above 40%; otherwise we will use a fixed‐effect model. When I² is higher than 75% we will not pool results of studies in meta‐analysis. We will include a 95% confidence interval (CI) for all estimates (Deeks 2017).

We will describe the results in the case of skewed data reported as medians and interquartile ranges.

Where multiple trial arms are reported in a single trial, we will include only the relevant arms. If two comparisons are combined in the same meta‐analysis, we will halve the control group to avoid double‐counting.

Subgroup analysis and investigation of heterogeneity

If future updates of this review find a sufficient number of studies, we will carry out the following subgroup analyses:

• high income versus low and middle‐income countries; and

• PPE that is certified for biological hazards versus PPE that does not have such a certification.

We will also use our primary outcomes in subgroup analyses, and we will use the Chi² test, as implemented in RevMan Web 2019, to test for subgroup interactions. At this time, we have not identified enough studies to allow for such a subgroup analysis.

Sensitivity analysis

If future updates of this review find a sufficient number of studies, we will perform sensitivity analyses defined a priori to assess the robustness of our conclusions. This involves including only studies we judge to have a low risk of bias. At this time we have not identified enough studies to allow such a sensitivity analysis.

Reaching conclusions

We based our conclusions only on findings from the quantitative or narrative synthesis of included studies that we judged to have the lowest risk of bias. Consequently, we used findings from non‐randomised studies when we did not find evidence from randomised studies. We avoided making recommendations for practice based on more than just the evidence, such as values and available resources. Our implications for research suggest priorities for future research and outline what the remaining uncertainties are in the area.

Summary of findings and assessment of the certainty of the evidence

Studies used numerous comparisons to measure the effect of PPE and we limited the 'Summary of findings' tables to the findings of the comparisons we judged most useful. We created a series of 'Summary of findings' tables to present the primary outcomes for different types of PPE (one type versus another) and donning or doffing procedures (one procedure versus another). We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the certainty of a body of evidence as it related to the studies that contributed results data for the prespecified outcomes. We used methods and recommendations described in Section 8.5 (Higgins 2017), and Chapter 12 (Schünemann 2017), of the Cochrane Handbook for Systematic Reviews of Interventions, using GRADEpro GDT software. We justified all decisions to down‐ or upgrade the certainty of evidence using footnotes and we made comments to aid reader's understanding of the review where necessary. With non‐randomised studies, we started at low‐certainty evidence and with randomised studies at high‐certainty evidence. In future updates of this review, if the outcomes are measured in many different ways, we will prioritise the reporting of outcomes as follows: infection rates, contamination rates and compliance rates.

Results

Description of studies

Results of the search

The search to January 2016 resulted in 10,268 references for screening (see Figure 2). From these references we selected 205 articles for full‐text assessment. Through checking the references of included articles we found 18 additional articles. We found another five articles by using Google, and we found one more through contacting NGOs (Tomas 2015). Our contacts with the manufacturers did not yield any responses or data. Most of the studies that we located outside our electronic searches were studies of PPE use during the SARS epidemic that did not make reference to any type of PPE in the title or abstract. For the same reason we did not locate Nyenswah 2015 because there was no reference to PPE. By using Google search, we found one additional article (Bell 2015), that was not indexed in any of the databases that we searched. Based on a request of one of the peer referees we also searched the African Index Medicus, which yielded 24 references but no new studies to include. Contacting PPE manufacturers did not lead to any responses. This added up to 205 papers that we checked full‐text for inclusion. Of these, we excluded 196. This resulted in nine included studies.

We updated the searches in Embase up to 22 May 2018, in Medline through PubMed up to 15 July 2018, in CINAHL up to 31 July 2018, in OSH‐update on 31 December 2018, and in CENTRAL up to 18 June 2019. We did not have access to Embase after May 2018 and used Scopus to update the Embase search up to 18 June 2019. This yielded 1698 new references after de‐duplication. We assessed 68 articles in full‐text and subsequently we excluded 58 articles. This resulted in 10 new studies that fulfilled our inclusion criteria (see Figure 3) of which we could include eight in the review and two were awaiting assessment.

For the 2020 update we reran the searches including the search word 'decontamination' and PPE as a MeSH term in Medline. We did not update the OSHupdate search because this yielded so little for the previous version. We also searched African Index Medicus but it did not add any new articles. Altogether we retrieved 3792 references through database searching and 17 additional records through searching Google Scholar. We removed 1760 duplicates (see Figure 4). Thus, we screened 2049 records, which led to 65 full‐text assessments. Of these, we excluded 58 records, mainly because the studies did not have a comparison or were already included in the review. The selection process finally resulted in seven new studies included in the review which includes the two studies awaiting assessment in the previous version of this review (Andonian 2019; Chughtai 2018; Drews 2019; Hajar 2019; Kpadeh Rogers 2019; Osei‐Bonsu 2019; Suen 2018).

Included studies

We contacted Bell 2015; Casalino 2015; Casanova 2016; Curtis 2018; Drews 2019; Hall 2018; Suen 2018 and we got additional information from all but Casanova 2016. We entered this information in the 'Characteristics of included studies' table.

Study types

We included 24 studies in total. Twenty‐two were simulation studies, of which 18 simulated exposure to contaminated body fluids and measured contamination outcomes, and four studies provided alternative PPE or procedures and measured compliance with donning and doffing procedures.

Of these simulation studies 14 were randomised trials (seven with parallel groups (Andonian 2019; Bell 2015; Curtis 2018; Hung 2015; Osei‐Bonsu 2019; Tomas 2016; Wong 2004), seven had a cross‐over design (Chughtai 2018; Hajar 2019; Guo 2014; Mana 2018; Strauch 2016; Suen 2018; Zamora 2006)), and one was a quasi‐RCT (Gleser 2018).

There were seven non‐randomised controlled studies (five with a cross‐over design ((Buianov 2004; Casanova 2012; Drews 2019; Kpadeh Rogers 2019; Hall 2018) and two with parallel groups (Casalino 2015; Casanova 2016)).

In addition, we found two retrospective cohort studies. One study evaluated the effect of PPE training on SARS infection rates and noncompliance with the doffing protocol (Shigayeva 2007). In this study, the authors located all HCW that had been exposed to SARS patients and assessed, by questionnaire, compliance with PPE guidelines and PPE doffing guidelines. Houlihan 2017 evaluated the risk of EVD infection according to donning and doffing practices and the use of disinfectant in HCW that had been deployed in West Africa during the EVD epidemic.

Participants

In the simulation studies, researchers included 816 intervention and 367 control participants, when we take into account that studies used a cross‐over design and thus all participants were intervention participants. In the cohort studies, there were 863 intervention and 232 control participants. Altogether there were 2278 participants.

The participants in all studies were healthcare workers with a mixture of occupations, but mainly physicians, nurses and respiratory technicians. One study included medical students during their internships (Casalino 2015). No studies included other healthcare staff such as people working in emergency services or cleaning staff.

In the two retrospective cohort studies, exposure of participants was to the SARS epidemic in one study (Shigayeva 2007), and to the EVD epidemic in another study (Houlihan 2017).

In the simulation studies, 12 studies simulated exposure using a fluorescent agent, three studies exposed participants to a harmless virus or microbes, and another three studies used both ways of exposure simulation. Studies used a wide range of different fluorescent agents and a range of exposure methods that varied from rubbing 0.5 mL of fluorescent agent over the gloved hands to throwing 100 mL of fluorescent agent onto the torso of the gown (see Table 17). The situation was similar in the studies that used viruses or bacteria to simulate exposure. Four studies simulated donning and doffing to assess compliance with guidance (Casalino 2015; Curtis 2018; Drews 2019; Hung 2015).

1

Exposure and outcome in simulation studies

Countries

Twelve studies were performed in the USA, four in China and Hong Kong, two in Canada, two in the UK, one each in Australia, Germany and Russia, and one was performed in three countries at the same time: France, Mexico and Peru (Casalino 2015). One study in Canada was performed during the SARS epidemic and one study in the UK was among HCW that had returned from the West‐African EVD epidemic.

Time period

All studies were conducted after the year 2000, with six before, and 18 after 2015.

Interventions and comparisons

Of the 24 included studies, 17 studies evaluated an intervention and a control condition. Four studies (Buianov 2004; Guo 2014; Houlihan 2017; Shigayeva 2007), evaluated two interventions. One study compared three types of PPE (Suen 2018), one study five types (Hall 2018), and one study 10 types (Chughtai 2018).

Fourteen studies compared one type of PPE to one or more other types. Eight studies compared two or more different ways of donning and doffing. One of these studies named the intervention 'enforced training' but we categorised it under different ways of doffing because it entailed giving instructions during the donning and doffing process versus not giving instructions (Casalino 2015). Three studies evaluated the effect of training.

Comparison of different types or parts of full‐body PPE

Fourteen simulation studies compared different types or parts of full‐body PPE outfits or compared an adapted design versus a standard design PPE, but all in a different way. Only a couple of studies were similar enough to allow us to combine their results. None of the included studies used a standardised classification of the properties of the PPE that protect against viral penetration such as the EN 14126.

Two simulation studies compared different types of masks or respirators as part of full‐body PPE. Buianov 2004 compared two different types of powered, air‐purifying respirator (PAPR) that were especially developed for this project in Russia to protect healthcare personnel against Ebola and similar viruses. Buianov 2004 also compared the effect of different airflow rates that varied from 50 L to 300 L per minute. The intervention participants were required to carry out a step test that lasted for four hours. The study authors did not describe the equipment they tested in sufficient detail for us to be able to judge their technical qualities. Zamora 2006 compared PPE combined with a PAPR in use at the study hospital with PPE without a PAPR according to CDC recommendations to prevent respiratory infection at the time of the study, the so‐called Enhanced Respiratory and Contact Precautions (E‐RCP).

Six simulation studies compared different types of gowns and protective clothing. Wong 2004 compared four types of PPE according to their material properties. First, they tested the material according to the American Association of Textile Chemists and Colorists' standards 22 and 127. We excluded the surgical‐gowns‐only category since it had no water repellency and insufficient viral barrier properties. Type A had good water repellency and water penetration resistance, but at the cost of poor air permeability. Type B had good water repellency and good air permeability, but poor water penetration resistance. Type C was the surgical gown with both poor water repellency and water penetration resistance. Type D, Barrierman, was made of Tyvek and had good water repellency, poor air permeability and fair water resistance. Bell 2015 compared commercially available PPE, compliant with CDC recommendations, with locally available clothing, such as rain coats that were thought to be as protective as the commercially available ones. Guo 2014 compared three types of PPE: a disposable water‐resistant, non‐woven gown, a reusable, woven, cotton gown, and a disposable non‐woven plastic apron. The second one was a cotton, water permeable gown, like a surgical gown. We left this arm out of the analysis because surgical gowns alone are not used for EVD. The study authors tested the fabrics for water repellency and liquid penetration according to the American Association of Textile Chemists and Colorists' standard 22. The gown and the apron received ratings of 4 and 5 respectively on a scale of 0 to 5 for water repellency. One simulation study compared different full‐body PPE ensembles. Hall 2018 compared five different PPE ensembles used in EVD surge units in hospitals, which all met the guidance of the Advisory Committee on Dangerous Pathogens endorsed by Public Health England (PHE). Three ensembles used gowns while two ensembles used coveralls. Some PPE ensembles were comprised of gowns with surgical caps and other ensembles of coveralls with hoods. Some PPE comprised boots only and others boot covers. Some taped the second pair of gloves whereas others did not. Suen 2018 compared three types of PPE, which differed with respect to the use of a waterproof gown, isolation gown, or coverall. Chughtai 2018 compared 10 different outfits that complied with guidance given by WHO or in specific countries, including the guidance for donning and doffing.

Modifications to existing PPE

Strauch 2016 compared a N95 filtering face piece respirator (FFR) mask to a modified FFR mask with tabs placed on the elastic band as a doffing aid. The study authors reported having evaluated contamination of the hands and head in two different trials but they reported their results in the same article.

Tomas 2016 compared a standard gown to a prototype seamless PPE that consisted of a polyethylene gown with nitrile gloves attached by a contact bond adhesive to enable the removal of the gown and gloves at the same time. Mana 2018 compared a standard polyethylene gown to a modified gown with a double elastic neck closure for easier removal, more gown coverage on the palm of the hand and smaller thumb holes and elastic wrist bands to create a snugger fit. Hajar 2019 also evaluated a gown with improved glove gown interface.

One simulation study compared different types of gloves. Gleser 2018 compared a modified glove with a small tab near the thumb to aid in glove removal without contamination to standard medical examination gloves. Both types of gloves were made of the same material from the same company. The study authors did not provide any more information.

Studies comparing different types of eye protection or footwear are missing.

Contamination rates are not only determined by the type of PPE but also by the donning and doffing procedures. All studies had a priori determined donning and doffing procedures. It should be noted that these studies evaluated the totality of the type of PPE inclusive of the donning and doffing procedure. We have described the procedures in the 'Characteristics of included studies' table.

Donning or doffing procedures (one procedure for donning or doffing versus another)

Eight studies compared different donning or doffing procedures.

Extra gloves

Casanova 2012 compared the effect of wearing two pairs of gloves with wearing one pair of gloves on contamination rates. We classified the study under methods of doffing because the intention of the double‐gloving was to decrease contamination during doffing. Doffing was done as per CDC recommendation, which describes how to do both single‐gloving and double‐gloving. Osei‐Bonsu 2019 also compared the CDC procedure for doffing with doffing with double gloves.

Structured procedures versus individual ways of donning and doffing

One simulation study compared individual's own versus recommended procedures. Guo 2014 compared the effect of doffing a gown or an apron according to an individual's own views versus the procedure recommended by CDC in the USA in 2007. Participants were given the following instructions: "Gown front and sleeves are contaminated! Unfasten neck, then waist ties. Remove gown using a peeling motion; pull gown from each shoulder toward the same hand. Gown will turn inside out. Hold removed gown away from body, roll into a bundle and discard into waste or linen receptacle".

Alternative procedures versus CDC procedure

One study (Osei‐Bonsu 2019) compared the CDC procedure for doffing with a one‐step procedure in which gloves are doffed at the same time as the gown.

Extra instruction

Two simulation studies compared the effect of extra assistance during donning or doffing versus no instructions. Casalino 2015 compared standard (unassisted) donning or doffing procedure to reinforced (extra assistance) procedures. The reinforcement consisted of an instructor saying out loud the next step of donning or doffing. The study authors used the reinforcement with both basic PPE (impermeable apron without a hood) and enhanced PPE (full‐body suit and hood). Andonian 2019 compared training in teamwork to conventional donning and doffing.

Disinfection procedures

Four simulation studies, and one field study, compared donning or doffing procedures with extra disinfection during the process. Casanova 2016 compared the self‐contamination of skin with two surrogate viruses when either an alcohol‐based hand rub or hypochlorite solution was used for the glove hygiene step of a PPE doffing protocol. Houlihan 2017 intended to compare the PPE removal with and without chlorine spray and also with and without assistance but there was collinearity between these variables and being in clinical work or in laboratory work. All those that were in clinical work reported having used chlorine spray and assistance whereas those in laboratory work did not. Therefore we could not analyse these data. Kpadeh Rogers 2019 compared the effect of alcohol‐based hand rub, quaternary ammonium or bleach to no glove disinfection. Osei‐Bonsu 2019 compared the recommended CDC procedure to the same procedure plus extra hand hygiene with alcohol‐based hand rub.

Type of training or education (one type of training or education versus another)

Three studies evaluated different training methods for donning and doffing procedures.

Hung 2015, a simulation study, compared a conventional training session for donning and doffing procedures to a procedure in which the conventional session was complemented with a computer simulation later.

Shigayeva 2007, a field study, evaluated the effect of active and passive training versus no training on compliance rates. We defined active training as training that involved any group or face‐to‐face interaction. We defined passive training as watching a video or receiving written instructions. This allowed us to make an indirect comparison between the effect of active and passive training. We calculated the effect of active training compared to passive training by subtracting the OR for passive training from the OR for active training, as outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We calculated the variance of this indirect comparison by summing the variances of both direct comparisons. Then we calculated the standard error by taking the square root of the combined variance. We used this as input for the generic inverse variance method in Review Manager 2014.

Curtis 2018, a simulation study, compared a video‐based learning session on instructions for PPE use for patient decontamination as part of a disaster medicine training to a traditional lecture before participating in a practical exercise.

Outcomes

Infection rates

One study (Houlihan 2017), evaluated the effect of interventions on infection rates. The study authors measured the level of immunoglobulin G (IgG) specific for EVD in an oral fluid sample to assess if there had been undetected infections in HCW exposed to EVD.

Contamination outcomes

Simulation studies measured contamination either as the proportion of people contaminated, as the number of contaminated spots, or as the area of the body contaminated in studies using a fluorescent marker (see Table 17). Study authors measured contamination with the help of a UV lamp (when using fluorescent marker), or by directly measuring viral or microbe presence or viral or microbial load (when using a non‐pathogenic virus or microbes). However, across studies, different body locations were contaminated and also different body locations were measured for the contamination outcome. In the control groups there was a median of 67% of participants contaminated and across intervention groups this was 25%. There were two studies in which there were participants that had zero contamination with a specific PPE outfit (Chughtai 2018; Hall 2018).

Compliance with guidance: noncompliance rates with donning and doffing procedures

Ten studies evaluated the effect of interventions on noncompliance (Casalino 2015; Casanova 2012; Curtis 2018; Drews 2019; Hajar 2019; Hung 2015; Shigayeva 2007; Suen 2018; Zamora 2006)

Four contamination simulation studies (Casanova 2012; Drews 2019; Hajar 2019; Zamora 2006), measured non‐compliance as the number of participants that did not follow the correct order of the protocol, omitted elements, or did not use the correct equipment.

Shigayeva 2007 measured noncompliance in their training study as the number of violations against protocol as recorded from interviews. There were two different compliance outcomes. One was called consistent adherence and was calculated as the proportion of exposure episodes with full compliance with PPE. The other one was called unsafe doffing, measured if one or more of the elements of the doffing procedure were violated. We recalculated outcomes in such a way that they represented the frequency of noncompliance.

Hung 2015 measured compliance as a total score on a 16‐item checklist for donning and 20‐item checklist for doffing. To get results comparable to the other studies we subtracted the mean compliance values from the maximum score and used these as noncompliance values.

Casalino 2015 measured noncompliance as the number of errors per person for donning and for doffing and the number of people with one or more errors as measured by the specialist trainer or instructor, who also gave the spoken instructions in case of reinforcement. The study authors also measured critical errors, which were those where there was contact between skin and potentially contaminated PPE, but we did not consider this a valid measure of contamination and disregarded it. We took measurement of the errors at the last training session as the effect of the intervention. We disregarded the error measurements at earlier training sessions.

Suen 2018 measured non‐compliance as the average of the percentage errors of all items of a checklist.

Curtis 2018 measured compliance as the percentage of the maximum attainable score that an external evaluator gave on a practical skills test for both donning and doffing PPE.

Secondary and other relevant outcomes

No studies reported on costs or other economic outcomes such as resource use.

Wong 2004 and Lai 2011a measured time, and Wong 2004 and Drews 2019 measured satisfaction. Buianov 2004 measured heart rate and body temperature. We chose to report the results of this outcome as well, as we identified it as an additional outcome that appeared relevant to the questions being addressed.

Excluded studies

Description of case series or outbreak

One reason for excluding important studies was that the researchers only described a case‐series of HCW cases' use of PPE for EVD (Muyembe‐Tamfum 1999), Marburg Haemorraghic Fever infection (MHF) (Borchert 2007; Colebunders 2004; Jeffs 2007; Kerstiens 1999), Congo Crimean Haemorraghic Fever (CCHF) (Gozel 2013), or for SARS (Christian 2004; Ho 2003; Ofner 2003; Ofner‐Agostini 2006). None of these studies described the use of PPE by the cases in such detail that they could be replicated. In combination with the lack of a control condition, it is difficult to conclude how much PPE, or the lack thereof, contributed to the infection. The only different study of a series of cases during an outbreak was the study by Dunn 2015 that contained proper descriptions of PPE.

Studies that evaluated only one type of PPE and not part of full‐body PPE

Ogendo 2008 measured eye protection only. Bearman 2007 measured universal glove use only. Chughtai 2013, Lindsley 2012 and Lindsley 2014 measured masks or face shields only. Even though these studies yield valuable information, it is unclear how well the results also cover the use of these items as part of full‐body protection and therefore we excluded these studies.

Participants not exposed to highly infectious diseases with serious consequences

Many studies evaluated PPE use for diseases other than EVD and related haemorraghic fevers, such as HIV or other nosocomial infections that were not considered highly infectious or having serious consequences, or both, and we excluded these studies (Anderson 2017; Bischoff 2019; Malik 2006; Makovicka 2018; Ransjo 1979; Sorensen 2008). In another study participants were not HCW (Kahveci 2019).

Training or simulation studies without a control group

There were a number of studies that evaluated training but that did not use a control group. This makes it difficult to draw inferences about the effect of one type of training compared to another (Abrahamson 2006; Beam 2014; Hon 2008; Northington 2007; Tomas 2015).

Inconsistent use of PPE during the SARS epidemic

After intensive discussion, we excluded 11 studies that measured the use of PPE (mask, gloves, gowns, goggles) during the SARS outbreak and related that to the risk of SARS infection. One line of thinking was that these studies did not fulfil the inclusion criteria because the comparison here was not clearly one type of PPE versus another type of PPE. Another line of thinking was that the studies compared different types of PPE composition and thus would fulfil the inclusion criteria. We finally decided to deal with these studies in the discussion section only (Ho 2004; Lau 2004; Le 2004; Liu 2009; Loeb 2004; Nishiura 2005; Park 2004; Pei 2006; Scales 2003; Seto 2003; Teleman 2004).

Risk of bias in included studies

See Figure 5 for an overview of our judgment of the risk of bias per study. Figure 6 gives an overview of risk of bias per domain. Since the figures contain the 'Risk of bias' assessments for both randomised and non‐randomised studies, not all cells are applicable to both study types and those that are not applicable remain empty.

'Risk of bias' summary: review authors' judgements about each 'Risk of bias' item for each included study

'Risk of bias' graph: review authors' judgements about each 'Risk of bias' item presented as percentages across all included studies

Allocation

Allocation was random in 14 studies but only five of them stated adequately what method they had used for generating the random sequence and where thus rated as at low risk of bias for random sequence generation. Five studies reported an appropriate method (Osei‐Bonsu 2019; Suen 2018; Wong 2004; Zamora 2006), and for one we received additional information from the study authors (Mana 2018). One used alternation and we rated it as having a high risk of bias (Gleser 2018). The other studies were rated at unclear risk of bias

Allocation concealment was unclear in all but two of the randomised studies (Mana 2018; Osei‐Bonsu 2019). We judged only these two studies to have a low risk of selection bias.

Blinding

In the simulation studies, the participants could not be blinded for the type of attire they were wearing or the type of donning or doffing procedure they were following. It is unclear if they could have contaminated themselves more with attire that they thought was not good, or they did not like, but for the majority of the studies we considered this unlikely and assessed the risk of performance bias to be low. For one study, Casalino 2015, we rated the risk of performance bias as high because the instructors who provided the intervention were very much aware if instruction was given or not and they were the also the assessors. We also rated the risk of performance bias as high for Drews 2019 and Hajar 2019 because the outcomes were subjective and the participants unblinded. We judged the risk of performance bias as low in 15 studies.

For the non‐randomised SARS study (Shigayeva 2007), we considered the risk of performance bias low because the study was retrospective and the participants did not know they were part of a study.

The risk of detection bias was unclear in most studies, as they did not report whether outcome assessors were blinded. We considered the risk to be high in one study (Casalino 2015), as providers of the intervention were also the assessors of compliance, and in a second study (Shigayeva 2007), because the intervention and the outcome were assessed with the same questionnaire at the same time. We judged the risk to be low in four studies because the study authors stated that assessors were blind to group status (Curtis 2018; Hung 2015; Mana 2018; Zamora 2006). We judged the risk of detection bias to be low for Houlihan 2017 because they used antibodies against Ebola, an objective outcome, which would not be affected by assessors' knowledge of treatment. All in all, we judged the risk of detection bias as low in eight studies.

Incomplete outcome data

We judged the risk of attrition bias to be low in 14 studies and unclear in 10 studies. All but two studies were short‐term experiments and therefore most had a complete follow‐up of all participants.

Selective reporting

It was difficult for us to judge selective reporting because none of the included studies had published a protocol. We judged seven studies (Andonian 2019; Casalino 2015; Casanova 2016; Chughtai 2018; Guo 2014; Kpadeh Rogers 2019; Suen 2018), to have a low risk of reporting bias as the study authors appeared to have reported all relevant data as specified in their articles' methods. We judged Bell 2015 to be at high risk of reporting bias because they did not report outcomes separately for the intervention and the control. We also judged Hung 2015 to have a high risk of reporting bias as the study authors did not fully report the results of the computer usability questionnaire. In addition, Gleser 2018 and Osei‐Bonsu 2019 did not fully report all results. In total we judged four studies to be at high risk of reporting bias.

Other potential sources of bias

We did not consider that any of the included studies were at risk of other sources of bias except for Gleser 2018, where we considered that there was a substantial financial conflict of interest because the first author was also the director of the company that produced the gloves that were part of the intervention.

Bias due to selection of participants into the study (non‐randomised studies)

We judged there to be a low risk of bias due to selection of participants into the study for five non‐randomised studies (Buianov 2004; Casalino 2015; Casanova 2012; Hall 2018; Shigayeva 2007), and unclear for one study (Casanova 2016). We considered the risk of selection bias to be high in two studies. Houlihan 2017, because they recruited participants based on snowball sampling, and Kpadeh Rogers 2019, where different HCW performed tests with different bacteria.

Overall risk of bias per study

We judged none of the included studies to be at low risk of bias overall. According to our judgment they were all at either unclear (N = 15) or at high risk of bias (N = 9).

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8; Table 9; Table 10; Table 11; Table 12; Table 13; Table 14; Table 15; Table 16

Summary of findings 1

Personal protective equipment (PPE) types: powered, air‐purifying respirator (PAPR) plus coverall versus N95 mask plus gown

Summary of findings 2

Personal protective equipment (PPE) types: more protective versus less protective

Summary of findings 3

Personal protective equipment (PPE) types: gowns versus aprons

Summary of findings 4

Personal protective equipment (PPE) types: different types of PPE attire

Summary of findings 5

Modified personal protective equipment (PPE): sealed gown‐glove interface versus standard gown

Summary of findings 6

Modified personal protective equipment (PPE): gown ‐ easy to doff compared to standard gown

Summary of findings 7

Modified personal protective equipment (PPE): gown with gown‐glove improvement compared to standard gown and gloves

Summary of findings 8

Modified personal protective equipment (PPE): gloves with tab versus standard gloves

Summary of findings 9

Modified personal protective equipment (PPE): mask plus tabs versus standard masks

Summary of findings 10

Procedures: doffing according to Centers for Disease Control and Prevention method versus individual doffing

Summary of findings 11

Procedures: single‐step doffing compared to Centers for Disease Control and Prevention standard

Summary of findings 12

Procedures: doffing with double gloves compared to doffing with single gloves

Summary of findings 13

Procedures: donning and doffing with instructions compared to without instruction

Summary of findings 14

Procedures: doffing with extra sanitation of gloves compared to standard no sanitation

Summary of findings 15

Procedures: doffing with hypochlorite versus doffing with alcohol‐based glove sanitiser

Summary of findings 16

Teaching: video‐based learning versus traditional lecture

1. Different types of PPE compared

1a Different types of mouth and nose protection

1.1 Powered air‐purifying respirator (PAPR) versus PPE for enhanced respiratory and contact precautions (E‐RCP)

Outcome: contamination with fluorescent marker

Zamora 2006 found that the PAPR system in use in their hospital led to less contamination than using the E‐RCP system (RR 0.27, 95% CI 0.17 to 0.43; Analysis 1.1). Other ways of measuring contamination also led to less contamination with the PAPR system: contamination more than 1 cm (RR 0.21, 95% CI 0.12 to 0.36). The total contaminated area was also less with a mean difference of −81.10 cm² (95% CI −96.07 to −66.13). This was mainly due to a lack of protection of the neck in the E‐RCP system.

Analysis

Comparison 1: PAPR versus E‐RCP Attire, Outcome 1: Any contamination

Outcomes: compliance with guidance ‐ donning and doffing noncompliance

Noncompliance with donning guidelines occurred more with the PAPR system as this consists of more elements (RR 7.50, 95% CI 1.81 to 31.10; Analysis 1.4; Zamora 2006). Noncompliance with doffing guidelines was more frequent with the E‐RCP system, but this was not statistically significant (RR 0.50; 95% CI 0.20 to 1.23; Analysis 1.5).

Analysis

Comparison 1: PAPR versus E‐RCP Attire, Outcome 4: Donning noncompliance

Analysis

Comparison 1: PAPR versus E‐RCP Attire, Outcome 5: Doffing noncompliance

Outcomes: donning and doffing time

The donning (MD = 259 seconds) and doffing time (MD = 337 seconds) were considerably longer with the PAPR system (Analysis 1.6; Analysis 1.7; Zamora 2006).

Analysis

Comparison 1: PAPR versus E‐RCP Attire, Outcome 6: Donning time

Analysis

Comparison 1: PAPR versus E‐RCP Attire, Outcome 7: Doffing time

1.2 One type of PAPR versus another and different airflow rates

Outcome: contamination with microbial aerosol

Buianov 2004 found that the suit that had the hood attached to the suit (СКБ‐I) had a lower 'contamination penetration rate' than the suits that had separate hoods and coveralls with a percentage of 8.10‐8 for the suit and 2.10‐1 for the coveralls. However, we could not understand the meaning of the penetration rate and we decided that we would not use these results for our conclusions (their results are not shown in data tables).

Outcomes: heart rate and body temperature

Buianov 2004 also found that contamination stopped beyond the 250 L/minute airflow rates. Body temperature and heart rates were also lower at these airflow rates.

1b Different types of body protection

1.3 Four types of PPE versus another

Wong 2004 compared four types of PPE according to their material properties. Type A had good water repellency and water penetration resistance but at the cost of poor air permeability. Type B had good water repellency and good air permeability but poor water penetration resistance. Type C was the surgical gown with both poor water repellency and water penetration resistance. Type D, Barrierman, was made of Tyvek and had good water repellency, poor air permeability, and fair water resistance.

Outcomes: contamination, user‐reported assessment of comfort and convenience ‐ usability, donning and doffing times

There were no considerable differences in contamination (Analysis 2.1) between Type A and Type B for face, neck, trunk, foot, or hand, but Type B scored about 10% higher on usability (MD −0.46, 95% CI −0.84 to −0.08; Analysis 2.2); this was due especially to better breathability of the fabric. There were no considerable differences in donning and doffing times (Analysis 2.3; Analysis 2.4).

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 1: A vs B Contamination, mean number of spots

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 2: A vs B Usability score (1‐5)

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 3: A vs B Donning time

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 4: A vs B Doffing time

There were considerable differences in contamination of the foot (MD −4.1 spots, 95% CI −6.94 to −1.26) and the hand (MD −12.76 spots, 95% CI −21.62 to −3.9) between Type A and Type D (Analysis 2.5). Donning (MD 33 seconds, Analysis 2.7) and doffing (MD 17 seconds, Analysis 2.8) times were also much worse for Type D. Usability was rated as not considerably differently (MD 0.25, 95% CI −0.12 to 0.62; Analysis 2.6).

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 5: A vs D Contamination, mean number of spots

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 6: A vs D Usability score (1‐5)

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 7: A vs D Donning time

Analysis

Comparison 2: Four types of PPE attire compared, Outcome 8: A vs D Doffing time

It was unclear how many participants had no contamination. On average, all types of PPE had some contamination.

1.4 Formal PPE versus locally available PPE

Outcome: contamination with fluorescent marker

Bell 2015 compared contamination in four participants with formal PPE with four participants with locally available protective gear, such as raincoats. They found contamination in one participant in both study arms. The study was so small that it is difficult to draw conclusions (Analysis 3.1).

Analysis

Comparison 3: Formal versus local available attire, Outcome 1: Contamination

1.5 Gown versus apron

Outcome: contamination with fluorescent marker

Guo 2014 compared a gown with an apron and found that the gown left less contamination than an apron, regardless of the way of doffing (Analysis 4.1; Analysis 4.2).

Analysis

Comparison 4: Gown versus apron, Outcome 1: Contamination with marker; individual doffing

Analysis

Comparison 4: Gown versus apron, Outcome 2: Contamination with marker; CDC doffing

1.6 Five types of PPE attire compared

Outcome: contamination with fluorescent marker

Hall 2018 compared post‐doffing contamination of five types of PPE ensembles used in different hospital wards across the UK. No analysis of contamination rates of the different suits was available since the authors reported the data on contamination sites only and not according to type of attire. They argued that the contamination rates were too low to provide a valid comparison.

1.7 Three different types of PPE attire compared

Outcome: contamination with fluorescent marker

Suen 2018 measured small and large patches of contamination in three different ensembles with PPE 1, a surgical gown used with EVD with a hood covering the neck, PPE 2, a coverall also used for EVD, and PPE 3, an isolation gown. They reported the median number of patches across 10 body sites and four environmental contamination sites. The median number of contaminations for small patches was respectively 5, 7 and 7 and for large patches it was 39, 43 and 47. These differences were reported as being statistically significantly different but there were insufficient data to check this. This would mean that a long gown protects better than a coverall and that the commonly used isolation gown protects least. According to the study authors, the reduced protection for the isolation gown is especially due to the lack of coverage of the neck, "which resulted in many small or extra‐large patches in the anterior and posterior neck region after spraying of the fluorescent solution onto the face shield and anterior surfaces of the gown".

Outcomes: compliance with guidance ‐ donning and doffing noncompliance

Suen 2018 also measured compliance and reported the average percentage of errors across the items measured. For PPE 1, PPE 2 and PPE 3, the averages for donning were 6.1, 6.0 and 3.7 and for doffing 3.0, 9.5 and 3.5. This seems to give an indication that coveralls are more difficult to doff.

Outcomes: time for donning and doffing

Suen 2018 also measured the time needed to don and doff the PPE (Analysis 5.1; Analysis 5.2). PPE 3, the isolation gown, was quickest to don and doff, while the coverall doffing took significantly longer, with on average more than 10 minutes for doffing. The attire with the long surgical gown took twice as long as the isolation gown to put on and was also slower to doff because more PPE items were used. We were not able to conduct a proper paired analysis because of the lack of detail in the study report. We analysed the trial as if it were a two‐group parallel trial, which leads to too wide confidence intervals.

Analysis

Comparison 5: Three types of PPE compared, Outcome 1: Time for donning

Analysis

Comparison 5: Three types of PPE compared, Outcome 2: Time for doffing

1.8 Ten different types of PPE ensembles compared

Outcome: contamination with fluorescent marker

Chughtai 2018 evaluated 10 different PPE ensembles recommended for use with EVD by global and national authorities. Six of these used coveralls and four used gowns. There were also differences in the use of a PAPR or a respiratory mask. Each ensemble was tested in total three times by part of 10 volunteers. There were only four ensembles that led to contamination: the ensemble recommended by WHO, North Carolina authorities, CDC and Health Canada. The first three consist of coveralls and the last one is a gown.

Outcome: user satisfaction

Chughtai 2018 also asked users to rate the ease of donning and doffing. The ECDC coverall and protocol was rated highest for ease of donning and doffing. Since there were only three ratings per ensemble, this has only a limited meaning.

2. Modifications versus standard gear

2.1 Sealed gown‐glove interface versus traditional gown‐glove interface

Outcome: contamination

Tomas 2016 found that participants doffing a gown that had continuous coverage of skin from arm to hand (sealed suit) were less likely to contaminate themselves with fluorescent lotion than those doffing traditional PPE of gown and glove (RR 0.27; 95% CI 0.09 to 0.78; Analysis 6.1). The study authors obtained similar results when they used MS2 bacteriophage as the contaminate (RR 0.68; 95% CI 0.47 to 0.98; Analysis 6.2).

Analysis

Comparison 6: Gown sealed gloves versus standard gown, Outcome 1: Contamination fluorescent lotion

Analysis

Comparison 6: Gown sealed gloves versus standard gown, Outcome 2: Contamination MS2

2.2 Easy‐doffing gown versus traditional gown

Outcome: contamination

Mana 2018 compared a gown with modified neck and wrist design to facilitate doffing with a traditional gown and found fewer people with contamination with both fluorescent marker (RR 0.08, 95% CI 0.01 to 0.55; Analysis 7.1) and with harmless virus (RR 0.53, 95% CI 0.29 to 0.94; Analysis 7.2). Even though we received additional information from the study authors we were unable to conduct a proper paired analysis.

Analysis

Comparison 7: Gown easy to doff versus standard gown, Outcome 1: Contamination with fluorescent marker

Analysis

Comparison 7: Gown easy to doff versus standard gown, Outcome 2: Contamination with bacteriophage

2.3. Modified gown‐glove interface versus standard gown‐glove interface

Outcome: contamination

Hajar 2019 modified the gown‐glove interface with more overlap between gown and glove. They evaluated this in two different groups. In one they compared the modified gown to a standard gown and in the other they added extra education to both intervention and control group. This led to considerably less contamination (RR 0.45, 95% CI 0.26 to 0.78, Analysis 8.1) in the meta‐analysis of the two trials. We could not take into account that the trials had a cross‐over design but analysed these as if they were parallel trials with twice the number of participants. This may have led to a slight overestimation of the precision.

Analysis

Comparison 8: Gown with gown‐glove improvement vs standard gown‐gloves, Outcome 1: People with contamination

2.4. Modified‐inside gown versus standard gown

Outcome non‐compliance: errors during donning, doffing, performance

Drews 2019 redesigned the gown based on observed errors during doffing, donning and performing tasks. They found a similar number of people with errors while donning (RR 0.93, 95% CI 0.50 to 1.72; Analysis 9.1), while performing tasks (MD −0.30, 95% CI −0.67 to 0.07; Analysis 9.2) and while doffing (RR 0.81, 95% CI 0.33 to 2.00; Analysis 9.3).

Analysis

Comparison 9: Gown with marked inside versus standard gown, Outcome 1: Noncompliance donning: people with errors

Analysis

Comparison 9: Gown with marked inside versus standard gown, Outcome 2: Noncompliance: errors during performance

Analysis

Comparison 9: Gown with marked inside versus standard gown, Outcome 3: Noncompliance doffing: people with errors

2.5 Gloves with tabs versus gloves without tabs

Outcome: contamination

Gleser 2018 found a decrease in people with contamination when doffing gloves with tab near thumb and wrist compared to standard gloves (RR 0.22, 95% CI 0.15 to 0.31; Analysis 10.1).

Analysis

Comparison 10: Gloves with tab versus standard gloves, Outcome 1: Any contamination of hands

2.6 Masks with tabs versus masks without tabs

Outcome: contamination

Strauch 2016 found that contamination from hands to the head was less when the participant doffed a mask with tabs on the strap engineered as a doffing aid compared to a mask without tabs (RR 0.33, 95% CI 0.14 to 0.80; Analysis 11.1). There was no difference in contamination rates when participants doffed a contaminated mask that either had or did not have tabs (RR 0.96; 95% CI 0.83 to 1.12; Analysis 11.2).

Analysis

Comparison 11: Mask with tabs versus no mask tabs, Outcome 1: Contamination of head from hands

Analysis

Comparison 11: Mask with tabs versus no mask tabs, Outcome 2: Contamination of hands from mask

3. Changes in donning or doffing procedures

3.1 Double‐gloving versus single‐gloving

Outcome: contamination with MS2 virus

Both Casanova 2012 and Osei‐Bonsu 2019 found that contamination with the use of double gloves was less than with single gloves. We felt that the studies were comparable even though the first used harmless virus and the second harmless bacteria as the simulated exposure. When all contaminated sites were taken together the RR was 0.34 (95% CI 0.17 to 0.66; Analysis 12.1). For the specific body parts the reduction was less clear (Analysis 12.1). Also when measured with fluorescent marker, there was no difference between double‐ and single‐gloving (RR 0.98, 95% CI 0.75 to 1.28; Analysis 12.4).

Analysis

Comparison 12: Doffing with double gloves versus doffing with single gloves, Outcome 1: Contamination: virus detected

Analysis

Comparison 12: Doffing with double gloves versus doffing with single gloves, Outcome 4: Contamination with fluorescent

All participants had some level of contamination. Measured as the quantity of virus found, the hands were less contaminated after degloving when participants used double gloves but due to missing data we could not test this.

Outcome: compliance with guidance ‐ compliance errors

No more errors in compliance occurred with the donning or doffing protocol for double‐gloving compared to single gloving (RR 1.08, 95% CI 0.70 to 1.67; Analysis 12.3).

Analysis

Comparison 12: Doffing with double gloves versus doffing with single gloves, Outcome 3: Non‐compliance: any error

3.2 CDC‐recommended procedure versus individual doffing

Outcome: contamination

Guo 2014 found that the CDC's recommended way of doffing a gown or an apron led to a different decrease in contamination compared to individually chosen doffing. When doffing the gown, there were 5.4 fewer smaller contamination patches (95% CI −7.4 to −3.4) and 5.2 fewer stains in the environment (95% CI −7.3 to −3.3), but no difference in small contamination patches on the hands, shoes or underwear. With doffing the apron, there were fewer smaller stains, stains on the hands, shoes, and environment, but more large stains and a similar number of stains on the underwear (Analysis 13.1; Analysis 13.2).

Analysis

Comparison 13: CDC versus individual doffing, Outcome 1: Gown: contamination with fluor marker

Analysis

Comparison 13: CDC versus individual doffing, Outcome 2: Apron: contamination with fluor marker

3.3 CDC‐recommended procedure versus single step

Outcome: contamination

Osei‐Bonsu 2019 evaluated doffing gown and gloves in a single step versus the standard gloves first procedure and found no difference in contamination with fluorescent marker (RR 0.98, 95% CI 0.75 to 1.28; Analysis 14.1) but with bacterial contamination there was a considerable difference (RR 0.20, 95% CI 0.05 to 0.77; Analysis 14.2). It is unclear what would cause this difference in effect between the two outcome measures. We would be inclined to assume that the bacterial simulation is more realistic than the fluorescent powder.

Analysis

Comparison 14: Single‐step doffing vs CDC standard, Outcome 1: Fluorescent contamination

Analysis

Comparison 14: Single‐step doffing vs CDC standard, Outcome 2: Bacterial contamination

3.4 Doffing with extra disinfection of gloves

a. Alcohol‐based sanitation of gloves versus no extra glove sanitation

Outcome: bacterial contamination

Osei‐Bonsu 2019 compared alcohol‐based glove sanitation versus no glove sanitation and found no considerable reduction in the number of people contaminated (RR 0.75, 95% CI 0.39 to 1.45; Analysis 15.1). Kpadeh Rogers 2019 found a non‐significant reduction in bacterial contamination from a median 2.4 colony‐forming units (CFUs) to 2.2 CFUs for both bacteria used when alcohol‐based hand rub was used versus no extra sanitation of gloves.

Analysis

Comparison 15: Doffing with extra sanitation of gloves versus standard no sanitation, Outcome 1: Bacterial contamination

b. Quaternary ammonium versus no extra glove sanitation

Outcome: bacterial contamination

Kpadeh Rogers 2019 found a significant reduction in bacterial contamination from a median 2.4 CFUs to 0 CFUs for both bacteria used for simulating exposure when quaternary ammonium‐based hand rub was used versus no extra sanitation of gloves.

c. Bleach versus no extra glove sanitation

Outcome: bacterial contamination

Kpadeh Rogers 2019 found a significant reduction in bacterial contamination from a median 2.4 CFUs to 0 CFUs for both bacteria used for simulating exposure when bleach‐based hand rub was used versus no extra sanitation of gloves.

d. Hypochlorite sanitation versus alcohol‐based sanitation

Outcome: viral contamination

Casanova 2016 found non‐significantly greater self‐contamination of bacteriophage MS2 to the hands, face or scrubs when hypochlorite solution was used for the glove sanitising step of the doffing protocol compared to the use of an alcohol‐based hand rub (RR 4.00, 95% CI 0.47 to 34.24; Analysis 18.1). The study authors did not detect contamination of bacteriophage Ph6 when using either alcohol‐based hand rub or the hypochlorite solution (Analysis 18.2).

Analysis

Comparison 18: Doffing with hypochlorite versus doffing with alcohol‐based glove sanitiser, Outcome 1: Contamination MS2

Analysis

Comparison 18: Doffing with hypochlorite versus doffing with alcohol‐based glove sanitiser, Outcome 2: Contamination Ph6

e. Chlorine spray versus no spray

Houlihan 2017 compared the risk of HCW contracting Ebola when either using or not using a chlorine spray during the doffing of PPE. However, there was no variation in the use of chlorine spray among clinical workers. The use only varied between clinical and laboratory workers. Since it is not possible to disentangle risk of exposure and the use of hypochlorite solution, no conclusions can be drawn from this study with regard to PPE.

3.5 Additional spoken personal instructions versus no such instructions

3.5.1. Outcome: compliance with guidance ‐ noncompliance

Casalino 2015 found that there were substantially less noncompliance (people with one or more errors) after additional spoken instruction compared to no instructions with (RR 0.31, 95% CI 0.11 to 0.93) and also that the mean number of errors fell by on average almost one (MD −0.89, 95% CI −1.36 to −0.41) in the group with spoken instructions (Analysis 16.1; Analysis 16.2).

Analysis

Comparison 16: Donning and doffing with instructions versus without instructions, Outcome 1: People with one or more errors

Analysis

Comparison 16: Donning and doffing with instructions versus without instructions, Outcome 2: Non‐compliance: mean errors

Andonian 2019 organised team work between the person with PPE and doffing assistants who guided the donning and doffing process and found a decrease in the number of sites contaminated with either fluorescent marker or particles (MD −5.00, 95% CI −8.08 to −1.92). We assumed that the median reported by the study authors would be roughly equal to the mean and the interquartile range equalled, 1.35 SD.

3.5.2. Outcome: infection rate

One study compared infection rates between people who had instructions while donning and doffing versus rates in those without instructions. Due to the fact that the exposure was also different between these two groups, we were unable to draw conclusions about the protective effect of instructions (Houlihan 2017).

4. Training and instructions

4a. Training and instruction for proper and complete PPE use

4a.1 Active training versus passive training

4a.1.1 Outcome: compliance with guidance ‐ noncompliance with PPE guidance

Shigayeva 2007 defined consistent adherence as always wearing gloves, gown, mask, and eye protection. We transformed this to inconsistent use as being noncompliant with the guidance. The study found that active training led to less noncompliance than no training (OR 0.37, 95% CI 0.2 to 0.58). For passive training, they found a lower risk of noncompliance compared to no training (OR 0.58, 95% CI 0.33 to 1.00). For the indirect comparison, active versus passive training, the OR was 0.63 (95% CI 0.31 to 1.30; Analysis 20.1).

Analysis

Comparison 20: Computer simulation versus no simulation, Outcome 1: Number of errors while donning

4b. Training and instruction for PPE donning and doffing

4b.1. Active versus passive instruction

4b.1.2. Outcome: compliance with guidance ‐ noncompliance with doffing procedures

Shigayeva 2007 found no considerable effect of active (OR 0.70, 95% CI 0.45 to 1.11) or passive training (OR 1.56, 95% CI 0.83 to 2.94) compared to no training on the number of errors in compliance with the doffing protocol. For the indirect comparison, active versus passive training, the OR was 0.45 (95% CI 0.21 to 0.98; Analysis 19.1).

Analysis

Comparison 19: Active training in PPE doffing versus passive training, Outcome 1: Noncompliance doffing protocol

4b.2. Additional computer simulation versus no additional computer simulation

4b.2.1. Outcome: compliance with guidance ‐ noncompliance

Even though the number of errors was already low, Hung 2015 found that adding computer simulation reduced the number of errors with on average half an error for donning (MD, −0.52, 95% CI −0.90 to −0.14; Analysis 20.1) and with more than one error for doffing (MD −1.16, 95% CI −1.63 to −0.69; Analysis 20.2).

Analysis

Comparison 20: Computer simulation versus no simulation, Outcome 2: Number of errors while doffing

4b.3 Video‐based learning versus traditional learning

Curtis 2018 compared skills in donning PPE when taught with a video‐based learning method versus a traditional lecture. Those that participated in the video learning had a higher mean score on the post‐exam than those who attended a traditional lecture. (MD 30.7, 95% CI 20.14 to 41.26; Analysis 21.1).

Analysis

Comparison 21: Video‐based learning versus traditional lecture, Outcome 1: Skills in PPE donning

5. Subgroup and sensitivity analysis

We planned a subgroup analysis of studies conducted in high‐ versus low‐ and middle‐income countries. However, there were not enough studies for such a subgroup analysis to be meaningful.

We also planned a sensitivity analysis including only studies we judged to have a low risk of bias. As none of the included studies fulfilled this criterion, we could not perform this analysis.

6. Certainty of the evidence

We judged if there was a reason to downgrade the certainty of the evidence for each domain of GRADE. Since we judged all studies to have a high or unclear risk of bias, we downgraded the evidence for all comparisons by one level. We considered simulation studies to be indirect evidence, and downgraded the evidence yielded by these studies by one level as well. In addition, when there was only one small study, we downgraded because of imprecision. All in all, the certainty of the evidence is low to very low for all comparisons. For the non‐randomised studies, there were no reason to upgrade the certainty of the evidence.

Discussion

Summary of main results

Almost all findings are based on one or at most two small simulation studies. Therefore, we judged the certainty of the evidence as very low or low.

One type of PPE compared to another

One study found less contamination when a PAPR with hood and coverall was used compared to a gown and a N95 mask but there were more errors in donning with the PAPR (Table 1).

Three studies compared different types of body protection. One study found that more protective gear protected slightly better but was more uncomfortable because of lack of breathability (Table 2). Another study found gowns to be better than aprons (Table 3). The third study did not provide data.

Three studies compared more recently proposed PPE ensembles according to different guidelines. One study found too few contamination events to draw conclusions. Another study found that long gowns protected better than a coverall or isolation gown and the coverall was difficult to doff (Table 4).

Modifications versus standard attire

Three studies compared changes to gowns especially related to improved doffing and changed glove‐gown interface and found considerably less contamination (Table 5; Table 6; Table 7). One study modified the inside of the gown and the closure system but found no difference in errors with donning or doffing or during performance.

Two studies evaluated the effect of tabs to improve ease of donning and found less contamination with tabs on masks or gloves (Table 8; Table 9).

One type of donning or doffing procedure compared to another

There are eight studies that compared donning and doffing procedures.

Following CDC recommendations for doffing gowns and aprons compared to individually chosen ways may decrease the risk of contamination (Table 10). Doffing of gloves and gown in one step may also decrease the risk of contamination (Table 11).

For doffing, there is very low‐certainty evidence that double‐gloving as part of full‐body PPE may reduce the risk of contamination and reduce the viral load on the hands without increasing the frequency of noncompliance with the doffing protocol (Table 12). Instructions during doffing may increase compliance (Table 13). Adding extra steps to the process in the form of glove disinfection may not be effective for alcohol‐based rub but may decrease viral and bacterial contamination when quaternary ammonium or bleach is used (Table 14). There is no difference in contamination between using alcohol‐based hand rub during doffing and using chlorine based disinfection (Table 15).

One type of training versus another

Three studies compared training models. There is very low‐certainty evidence from one SARS‐related study and two simulation studies that more active training in PPE use decreases noncompliance with donning and doffing guidance more than passive training. The active training used in the studies was video or computer simulation or face‐to‐face training compared to lectures (passive) only (Table 16).

We found no audit reports or other unpublished reports or data from our contact efforts to manufacturers and other organisations.

Overall completeness and applicability of evidence

Most studies provided sparse descriptions of the level of chemical protection (ISO 2013), or viral protection (EN 14126; ISO 2004a), of the PPE they used, or the outfits used varied so much in their components that it was impossible to make uniform comparisons.

For some PPE parts such as face shields and goggles, we found no studies that compared the two. There is, however, evidence from studies with viruses that do not have serious consequences and from simulation studies with manikins that each protects compared to no intervention (Agah 1987; Lindsley 2014). In a thorough overview of face shields for infection prevention, Roberge 2016 concludes that even though face shields can considerably reduce droplet contamination of the face, more research is needed into their efficacy. Other technical laboratory studies without involvement of humans also support the findings of this review. Kahveci 2019 found that double gloving can reduce contamination by reducing the fluid leakages through the glove‐gown interface.

Doffing procedures are fairly easy to evaluate in simulation studies. We found several studies that confirmed that it is important to follow procedures. However, all studies were small and only the comparisons about double‐gloving disinfection procedures and spoken instructions had more than one study. It seems that it would not be difficult to perform more and better simulation studies to find out how important these procedures are.

Because studies seem feasible and because we searched exhaustively, there must be other reasons why there is so little evidence available with infection rates as an outcome. One of these is probably the highly politicised context in which such a study has to be performed during an epidemic. However, retrospective cohort and case‐control studies are possible as has been shown during the SARS epidemic. The studies conducted after the SARS epidemic show that the consistent use of PPE rather than type of PPE was most important (see Appendix 1). At the start of the epidemic, SARS patients were not appropriately diagnosed, and the importance of PPE was not immediately clear. PPE compliance was higher in the later stages, and infections occurred less frequently (Nishiura 2005). SARS also affected comparatively higher‐income countries such as China, Hong Kong and Canada. The experiences from retrospective studies during Ebola epidemics are similar. During the 1995 Ebola epidemic in Kikwit in the Democratic Republic of the Congo, a study also reported that once PPE and other control measures were used, there were very few HCW infections (Kerstiens 1999). Dunn 2015 is a case study from the Ebola epidemic that also provided systematic information on the use of PPE and infection rates. We reanalysed the excluded study by Dunn 2015 as a cohort study of exposed HCW (Verbeek 2016a). The risk ratio of contracting Ebola infection for HCW using gloves only versus those not using PPE was 0.16 (95% CI 0.04 to 0.71) indicating that using gloves already provides a lot of protection. For using gloves or a gown or more compared to no PPE, the RR was 0.03 (95% CI 0.00 to 0.57; Verbeek 2016a). This is very similar to the findings of the SARS studies mentioned above. It is also, to a certain extent, reassuring for those situations during an epidemic or in low‐ and middle‐ income countries, when sufficient PPE is not available (see Levy 2015), that some PPE decreases the risk of infection considerably. In this version of the review we were able to include one retrospective cohort study from the 2015 West Africa Ebola epidemic. Unfortunately, the information on PPE was not detailed enough to be able to draw conclusions.

While the included studies show that more active training prevented errors, it is not clear how long the effects of training last. Northington 2007 showed that at six months after training, only 14% of participants were able to correctly don and doff PPE. It is unclear from the included studies, if fit‐testing of masks is part of training. This is a commonly accepted prerequisite for proper functioning of respiratory protection.

We included only one study conducted in a low‐ and middle‐income country. Since most serious haemorrhagic fever epidemics occur in some parts of Africa, this is a serious disadvantage of the current evidence. However, in such resource‐poor settings, appropriate research is the lowest priority for the local decision makers. Consequently, the initiative has to come from WHO and international organisations that work in these epidemics.

Quality of the evidence

We rated the certainty of the evidence as very low or low for all comparisons, mainly because our conclusions are based on single studies or two small studies and all the included studies had a high or unclear risk of bias. The retrospective cohort studies have a high risk of recall bias because participants had to recall their use of PPE after the epidemic occurred. The simulation studies had small sample sizes or very few events across compared groups.

One of the major problems is that most of the studies did not indicate if the PPE that they used complied with one or more of the international standards for protective clothing and whether they used whether they used protective clothing that is constructed with viral resistant fabrics and seams. The lack of attention to the designation of PPE as being protective for viruses is also problematic in practice.Also the lack of description of the PPE significantly reduces the ability make clear conclusions.

The many different labels and standards that are in use to designate protection make it almost impossible for a HCW in practice to make the right choice. For EVD, it was especially problematic because HCW needed the highest standard of protection. The confusing language of infection control has also been reported for isolation practices in general. This is why Landers 2010 called for the adoption of internationally accepted and standardised category terms for isolation precautions. Others have tried to improve the standardisation by providing HCW with a summary card of the various types of precautions that have to be taken and indicated that this increased the implementation of precautionary measures (Russell 2015).

In simulation studies, it is not clear how well the exposure represents real life exposure. Some studies used 'high volume exposure to simulate splash' (Bell 2015), whereas other studies only used a powdered fluorescent marker spread in the room (Beam 2011). It is also not clear how well the fluorescent marker can indicate that there is no viral contamination. Casanova 2008 showed that in spite of no fluorescent marker being detected, there could still be viral contamination with bacteriophage MS2. On the other hand, Osei‐Bonsu 2019 did not find a difference in effect with fluorescent marker as the outcome but did find a difference with bacterial exposure.

Only one of the case studies that we collected (Dunn 2015), properly described the use of PPE. Better description would enable better analysis.

Potential biases in the review process

We excluded all studies that evaluated only one piece of PPE, such as goggles or masks. However, none of these excluded studies would have answered the questions that in our current review remained unanswered. From Casanova 2012, it became clear that using double gloves as part of full‐body PPE is important, because it facilitates the removal of the other pieces of PPE without contaminating the hands. This shows that it is important to consider the effect of one piece of PPE as part of full‐body PPE. In addition, seldom is there only one clear transmission route. Even with SARS, which, as a respiratory infection, was spread by droplets and aerosols, consistent use of other pieces of PPE besides respiratory protection was still important to prevent contact transmission. Therefore, we think that our strict inclusion criteria did not bias the results of our review.

We assumed that adherence to PPE use and training would work in a similar way between SARS, EVD, and simulation studies. However, there is an important difference. At the start of the SARS epidemic, the causal virus and its transmission were unclear and workers were probably not instructed well enough to protect themselves. On the other hand, it has been known for years that EVD is a highly contagious disease with a very high fatality rate. Thus, compliance and effectiveness of training concerning EVD might be higher than we concluded from the SARS study. In the SARS studies that we excluded, there was high heterogeneity in the effects of consistently wearing PPE that we could not explain. The heterogeneity in effect is also underpinned by studies that did not find any SARS infections in spite of imperfect protection with PPE. This means that at best the effectiveness of PPE is not fully understood.

Twelve of the included simulation studies are cross‐over studies. But the authors of only four studies analysed the data with tests that took into account the paired nature of the data: Zamora 2006 used the Mailand‐Gart test; Guo 2014 used repeated measures; and Casanova 2012 and Strauch 2016 the paired t‐test but the methods used in Mana 2018 were unclear. We could not use the results of these tests in our analyses in Review Manager 2014, which resulted in wider confidence intervals than using a paired analysis. There were insufficient data in the studies to properly adjust for the cross‐over effect in our analyses. However, all results that were reported as being statistically significant were also statistically significant in our analyses. Therefore, we think that this has not biased our results.

With the simulation studies the way exposure was simulated is an important element to consider. This varied highly between the studies. However, most studies used a worst case scenario, spraying fluorescent marker over large parts of the body but some studies applied only small amounts. Hall 2018 used a sophisticated manikin with an internal mechanism simulating exposure as described by Poller 2018. Future studies urgently need consensus from experts in the field on how exposure can be best simulated. This is best possible under the auspices of WHO or other internationally recognised bodies.

With the included non‐randomised studies, we assessed risk of bias with a hybrid version of the Cochrane 'Risk of bias tool' (Higgins 2017) and the recently developed ROBINS‐I tool (Sterne 2016). This might not have been the optimal way to assess risk of bias. However, we believe that the limitations of the available studies are profound and a more rigorous 'Risk of bias' assessment could not have lowered (or improved) our confidence in the evidence any further.

Agreements and disagreements with other studies or reviews

We found two other reviews that have evaluated the effect of PPE for highly infectious diseases with serious consequences in HCW: Hersi 2015 and Fischer 2015. Hersi 2015 was commissioned by WHO to underpin the PPE guidelines issued for HCW exposed to EVD. The authors originally included only controlled studies of interventions to protect HCW against EVD and similar haemorrhagic fever infections with infection rates as outcomes. During the review process the authors decided to also include case studies and case series but they were not able to draw conclusions from these studies because the PPE use was not well described. Fischer 2015 took a more pragmatic but unsystematic approach and included all articles pertaining to filovirus transmission and PPE and in addition articles that evaluated donning and doffing strategies. They conclude that there is a lack of evidence but that simulation studies could provide evidence for guidelines.

Heat stress and breathability is an important issue in PPE especially for Ebola. Kuklane 2015 argued that using other materials would substantially reduce the heat stress but these come at a tenfold higher price. Other researchers that have looked into this problem have found inconsistent results. Coca 2015 found that PPE on manikins led to a critical body core temperature of 38.4ºC in one hour. On the other hand, Grélot 2015 found that HCW caring for Ebola patients had only a 0.46ºC rise in core body temperature after being at work for one hour. Of the 25 workers studied, only four reached a core body temperature over 38.5ºC.

An independent panel of experts that evaluated the Ebola response concluded, among many other things, that a coordinated research effort is needed to build a better global system for infectious disease outbreak and response (Moon 2015). Their recommendation is that research funders should establish a worldwide research and development financing facility for outbreak‐relevant drugs, vaccines, diagnostics, and non‐pharmaceutical supplies (such as PPE). This is very much in line with what we experienced and found in this review.

Missair 2014 reviewed implications of EVD patient management for anaesthetists based on a literature review of all types of studies on EVD. This is why their inclusion criteria were very broad and non‐specific. Finally the authors relied on PPE guidelines as provided by WHO and MSF to make recommendations with no evidence of their comparability. This makes their results difficult to compare to ours.

Moore 2005 reviewed all measures to prevent healthcare workers from SARS and other respiratory pathogens in a narrative format, from 168 publications. They concluded that a positive safety climate is the most important factor for adherence to universal precautions. They recommend using adequate PPE, but they do not define 'adequate'. Their inclusion criteria were much broader than ours and their results are difficult to compare with ours. The same research group formulated valuable advice about research gaps based on this review but focused only on respiratory protection (Yassi 2005). They corroborate the findings of Jefferson 2011, that N95 respirators may not be superior, citing the early containment of the SARS epidemic without these in Hanoi.

The Cochrane Review by Jefferson 2008, updated in Jefferson 2011, evaluated the effect of physical interventions to interrupt the spread of respiratory viruses for all patient and staff populations. Even though they only included studies on respiratory infections and any type of protection for any person at risk, 10 studies in their review are about SARS and protecting HCW. The authors did not conduct a subgroup or additional analysis of these HCW studies because the infection risk for HCW is substantially different from the populations they protect. The Jefferson 2011 results are not applicable to HCW.

Authors' conclusions

Implications for practice

In addition to other infection control measures, consistent use of full‐body personal protective equipment (PPE) can diminish the risk of infection for healthcare workers (HCW). EN (European) and ISO (international) standards for protective clothing and fabric permeability for viruses are helpful to determine which PPE should technically protect sufficiently against highly infectious diseases. However, the risk of contamination depends on more than just these technical factors. In simulation studies, contamination happened in almost all intervention and control arms.

For choosing between PPE types, there is very low‐certainty evidence, based on single‐exposure simulation studies. Covering more parts of the body leads to better protection but usually comes at the cost of more difficult donning (putting on) or doffing (taking off) and user comfort, and may therefore even lead to more contamination. A powered, air‐purifying respirator (PAPR) with a hood may protect better than an N95 mask with a gown but is more difficult to don. A long gown may be the best compromise between protection and ease of doffing. Coveralls may be more difficult to doff. A more breathable fabric may still lead to similar levels of contamination protection to less breathable fabric, and may be preferred by users.

For changes to PPE, there is low‐ to very low‐certainty evidence that adding tabs to gloves or masks or closer fit of gowns at the neck or the wrist may decrease contamination, even though one study could not show a decrease in donning or doffing errors.

For different procedures of donning and doffing, there is very low‐certainty evidence that double gloves, as part of PPE and following Centers for Disease Control and Prevention (CDC) guidelines, and providing users with help or spoken instructions during donning and doffing may reduce the risk of contamination. Extra disinfection of gloves with bleach or quaternary ammonium may decrease hand contamination but not alcohol‐based hand rub.

For various training procedures there is very low‐certainty evidence that more active training (including video or computer simulation or spoken instructions) may increase compliance with instructions compared to passive training (lectures or no added instructions). No studies compared methods to retain PPE skills needed for proper donning and doffing in the long term.

The certainty of the evidence is low to very low for all comparisons because conclusions are based on one or two small studies and a high or unclear risk of bias in studies, indirectness of evidence, and small numbers of participants. This means that we are uncertain about the estimates of effects and it is therefore possible that the true effects may be substantially different from the ones reported in this review.

Implications for research

We concur with the World Health Organization (WHO) that there is a need to carry out a re‐evaluation of how PPE is standardised, designed, and tested (WHO 2018). What is missing is a harmonised set of PPE standards and a unified design for PPE to be used when taking care of patients with highly infectious diseases. This holds for PPE as used for preventing contact transmission as well as other ways of transmission. There is, for example, no unified technical standard for isolation gowns. There is also a need for a more transparent and uniform labelling of infection control measures, such as droplet precautions, and the protection level of PPE for HCW. We believe that this is an important prerequisite for the universal implementation of infection control measures for HCW.

Simulation studies are a feasible and relatively simple way to compare different types of PPE and to find out which protects best against contamination. It is a prerequisite for a reliable answer that methods of simulation studies are standardised in terms of exposure and outcome measurement. We recommend developing a core outcome set (COS) in this field that would provide critical outcomes measures to enable better comparisons and synthesis across trials. Viral marker bacteriophage MS2 seems to be the most sensitive marker and we would advocate using this. Studies should have sufficient power. A sample size of 62 would be needed to be able to detect a relatively large risk ratio of 0.5 with a large control group rate of contamination of 0.7, assuming α = 0.05 and β = 0.80. In addition, it would help evidence synthesis if study authors would better adhere to the appropriate reporting guidelines (Cheng 2016).

To find out how PPE behaves under real exposure, we need prospective follow‐up of HCW involved in the treatment of patients with highly infectious diseases, with careful registration of PPE, donning and doffing and risk of infection. Here, the effect sizes would be smaller and thus the sample size should be bigger than 60.

In addition, case‐control studies comparing PPE use among infected HCW and matched healthy controls, using rigorous collection of exposure data, can provide information about the effects of PPE on the risk of infection. The sample sizes should be much bigger than the current case studies because we would like to detect small but important differences in effect between various combinations of PPE such as gowns versus coveralls. There is a need for collaboration between organisations serving epidemic areas to carry out this important research in circumstances with limited resources, and during the throes of an outbreak.

We also need more randomised controlled studies of the effects of one type of training versus another, to find out which training works best, especially at long‐term follow‐up of one year or more. Here also, the effect size seems to be quite large and thus a sample size of around 60 seems to provide adequate power.

Feedback

Unified design for PPE, September 2019

Summary

We noted the timely and welcome update of the above review by Dr Verbeek and his team. As stated in the introduction, in epidemics of highly infectious diseases such as Ebola Virus Disease (EVD), healthcare workers (HCW) are at much greater risk of infection than the general population. Sadly the review comes at a time when this once again is being proved, with recent (20th July, 2019) data from the Democratic Republic of Congo (DRC) recording that since the beginning of the epidemic, the cumulative number of cases has been 2564 (2470 confirmed and 94 probable) with 1728 deaths (1634 confirmed and 94 probable cases). Of those, the cumulative number of confirmed/ probable cases among health workers is 138 (5% of all confirmed/probable cases) including 41 deaths (1). This comes shortly after the World Health Organization declared EVD in DRC a Public Health Emergency of International Concern (2). With the United Nations also recognising the seriousness of the emergency, by activating the Humanitarian System‐wide Scale‐Up to support the EVD response, this increases the possibility of HCW from around the globe being called upon to provide practical support in country, or travellers to affected countries returning with infection, and with it the need for personal protection from exposure to patients’ contaminated body fluids.

We were pleased that data from our recent research (3) was included in the review. In our study, we compared five PPE ensembles used in different high consequence infectious disease (HCID) units around the UK for examination of a ‘suspected case’, using a medical training manikin to expose HCW wearing the PPE to four different body fluid simulants, each tagged with different colour fluorochromes, and UV light to visualise any cross‐contamination during dry doffing. We note and accept the conclusions of the Review that “what is missing is a harmonised set of PPE standards and a unified design for PPE to be used when taking care of patients with highly infectious diseases”, also that the quality of the evidence was low because conclusions were based on single studies or on small numbers of participants.

While resources did not allow us to address the ‘small numbers’ issue, we have addressed the ‘unified design for PPE’ in a paper which was published after the Review literature search cut‐off date. In this follow‐up work, we presented the outcome of the initial research to the HCID units and reached a consensus on a unified PPE ensemble for examination of a suspected HCID case. Again, using HCW volunteers, we tested the unified PPE ensemble with fluorochromes as before, the result being no cross contamination events from 20 volunteers (4). In subsequent HCW training for one HCID unit, a further 40 challenges using 35 volunteers tested the PPE ensemble with only one cross contamination event through a known deviation from the doffing protocol (unpublished data).Therefore, there were 60 challenges with 54 volunteers with one breach. Public Health England plan in the near future to make written and video guidance available to demonstrate safe use of this unified PPE ensemble, and similar guidance is already available through Health Protection Scotland (5).

While more is needed, we believe this adds to the body of evidence required to ensure HCW can conduct the important business of patient care with confidence that they will be protected from potential infection.

Brian Crook(a), Anne Tunbridge(b), Bozena Poller(b), Samantha Hall(a), Cariad Evans(c) on behalf of the High Consequence Infectious Diseases Project Working Group UK

(a) Health and Safety Executive, Harpur Hill, Buxton SK17 9JN UK, (b) Sheffield Teaching Hospitals NHS Foundation Trust, Department of Infectious Diseases, Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF, UK, (c) Sheffield Teaching Hospitals NHS Foundation Trust, Department of Virology, Northern General Hospital, Herries Road, Sheffield, S5 7AU, UK

References

1. DRC Ministry of Health. Epidemiological situation report, Fri 20 Jul 2019. Available at: https://mailchi.mp/sante.gouv.cd/ebola_kivu_20juil19?e=77c16511ad

2. WHO news release 17th July, 2019. Available at: https://www.who.int/news-room/detail/17-07-2019-ebola-outbreak-in-the-democratic-republic-of-the-congo-declared-a-public-health-emergency-of-international-concern

3. Hall S, Poller B, Bailey C, Gregory S, Clark R, Roberts P, Tunbridge A, Poran V, Evans C, Crook B. Use of ultraviolet‐fluorescence‐based simulation in evaluation of personal protective equipment worn for first assessment and care of a patient with suspected high‐consequence infectious disease. J Hosp Infect 2018;99: 218‐228

4. Poller B, Tunbridge A, Hall S, Beadsworth M, Jacobs M, Peters E, Schmid ML, Sykes A, Poran V, Gent N, Evans C, Crook B on behalf of the High Consequence Infectious Diseases Project Working Group. A unified personal protective equipment ensemble for clinical response to possible high consequence infectious diseases: A consensus document on behalf of Public Health England and the Health and Safety Executive. Journal of Infection 2018;77:496–502

5. NHS Education for Scotland. Viral Haemorrhagic Fever‐ The correct order for donning and the safe order for removal and disposal of Personal Protection Equipment. Available at:https://www.nes.scot.nhs.uk/education-and-training/by-theme-initiative/public-health/health-protection/travel-and-international-health/viral-haemorrhagic-fever.aspx

Reply

Thank you for the comments and for supporting the conclusions of our review. It is great to see that the use of PPE for highly infectious diseases is becoming standardised in the UK. Compared to the current diversity in outfits this is certainly an improvement.

We believe that controlled studies form the best evidence in showing the protective capabilities of PPE against highly infectious diseases. We have no doubts that PPE helps in preventing infection. The question remains what the best possible PPE is. Given that infections still occur among health care workers and that users are not very satisfied with the PPE ensembles currently in use, improvement is still possible. Therefore, we included only controlled studies that compared newly designed PPE with existing PPE. The 20 test of one type of PPE by 17 volunteers in the Poller 2018 study were an uncontrolled experiment. Unfortunately, the paper did not provide data on the test of volunteers but only reported that there were no contaminations. Without knowing further details of this study, for example how many times the volunteers tested the new PPE ensemble, it is difficult to judge the significance of this result.

We also noticed that the agreed PPE ensemble currently does not include tags on gloves and masks or a sealed gown‐glove combination. These are both aspects that are supported by some evidence in our updated Cochrane review, meaning that these may prevent contamination more than conventional PPE. Therefore, we think that the agreed PPE ensemble could still be improved. We also hope that the newly agreed ensemble, and any further improvements upon it, will be tested against the currently used ones in a sufficiently large randomised experiment of simulated exposure.

Contributors

Jos H Verbeek, Blair Rajamaki, Sharea Ijaz, Christina Tikka, Jani H Ruotsalainen, Riitta Sauni

Certainty of the evidence, November 2019

Summary

This is a large scale and important review. On the next update, the review may benefit from up‐to‐date application of GRADE. Authors should use the term 'certainty' rather than 'quality' of evidence. At present, the term 'certainty' features in GRADE tables, but 'quality' throughout the text. Although the authors GRADE all comparisons as very low certainty, in the abstract the authors present findings using the term 'may', for example: "may protect better". The accepted plain language for very low certainty evidence is 'we do not know', and the review may therefore over‐represent the certainty of evidence. The authors should consider how best to ensure the very low certainty of evidence is adequately reflected for each result presented.

Paul Hine, Honorary research fellow, Cochrane Infectious Diseases Group

Reply

Thank you very much for your comments on our review and pointing out the inconsistency in using quality and certainty of the evidence. We will repair this throughout the review with the next update.

We don't think that the phrase 'may improve' instead of ‘we don't know’ over‐represents the certainty of the evidence. At the beginning of the abstract we state: 'Evidence for all outcomes is based on single studies and is very low quality'. Recent GRADE guidance says that very low certainty evidence can be reported as 'may improve but the evidence is very uncertain'.1 This is also the guidance in the latest version of the Cochrane Handbook (Table 15.6.b). We will add the additional "but the evidence is very uncertain" to the phrase 'may improve' in the review update in line with the most recent GRADE guidance.

1Santesso N, Glenton C, Dahm P, Garner P, Akl E, Alper B, Brignardello‐Petersen R, Carrasco‐Labra A, De Beer H, Hultcrantz M, Kuijpers T, Meerpohl J, Morgan R, Mustafa R, Skoetz N, Sultan S, Wiysonge C, Guyatt G, Schünemann HJ, for the GRADE Working Group, GRADE guidelines 26: Informative statements to communicate the findings of systematic reviews of interventions, Journal of Clinical Epidemiology (2019)

Contributors

Jos H Verbeek, Blair Rajamaki, Sharea Ijaz, Christina Tikka, Jani H Ruotsalainen, Riitta Sauni

Comparisons, April 2020

Summary

The comparison of PAPR and gown, vs N95 and gown is a poor comparison. And should not be used as a fair comparison and will likely give HCW the wrong impression. PAPR/CAPR and gown vs N95 and face shield with gown should be the equal comparison arms. Please consider this as this will give people the wrong messaging as we know face shields are an important component to PPE and we should do a real comparison as to determine whether one protection is better than another.

Michael Bolaris

Reply

Thank you for comments. We were already quite strict with including only studies that intended to evaluate full‐body protection. We don't have control over the interventions and controls that have been evaluated in studies. This specific combination of intervention and control was evaluated in one trial as a result of the problems encountered with aerosol generating procedures during the SARS epidemic. We think that the results are still very useful for the current COVID‐19 pandemic. The difference in contamination with the two types of PPE was especially caused by contamination of the neck in the PPE that consisted of mask and gown but where the gown did not cover the neck. We doubt that a face‐shield would have made much difference here because it does not protect the neck. The current WHO COVIID‐19 guidelines suggest the use of a gown with face mask and goggles as minimum PPE. This PPE ensemble leaves a part of the body uncovered and the trial mentioned above shows that this can easily lead to contamination

Contributors

Jos H Verbeek, Blair Rajamaki, Sharea Ijaz, Riitta Sauni, Elaine Toomey, Bronagh Blackwood, Christina Tikka, Jani H Ruotsalainen, F Selcen Kilinc Balci

What's new

History

Protocol first published: Issue 4, 2015
Review first published: Issue 4, 2016

Notes

Disclaimer. The findings and conclusions in this Cochrane systematic review are those of the authors and do not necessarily represent the official position of the National Institute for Occupational Safety and Health, Centers for Disease Control and Prevention. Mention of product names does not imply endorsement

Acknowledgements

We thank the Cochrane Editorial Unit for providing financial support to undertake this review. For this 2020 update, we thank Ruth Foxlee for updating the searches, Bozena Poller and Nick Gent for peer referee comments, Euan S Shearer for consumer comments, Toby Lasserson, Liz Bickerdike, Anne‐Marie Sephani, Robin Featherstone, and Helen Wakeford for editorial input, and Denise Mitchell for copy‐editing the review.

We thank Kaisa Neuvonen, Erja Mäkelä and Raluca Mihalache and Michael Edmond for their contributions as authors to the previous version of this review (Verbeek 2016b). We thank Toby Lasserson, Hannah Ryan, Darrel Singh, Fiona Smaill, Mauriccio Ferri, Annalee Yassi, Nuala Livingstone and Julian Higgins for their comments on the text for the first publication of this review (Verbeek 2016b). We extend a special thankyou to Consol Serra for her editorial work. We thank Claire Allen from Evidence Aid for her help in trying to locate unpublished reports. We thank Alexey Pristupa for assessing the studies written in Russian.

Sharea Ijaz's time for this update was supported by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care West (CLAHRC West) at University Hospitals Bristol NHS Foundation Trust.

Appendices

Appendix 1. Effects of wearing personal protective equipment (PPE) consistently on the risk of SARS infection

Wearing PPE consistently versus wearing PPE inconsistently

During and just after the SARS epidemic a number of studies evaluated the impact of the use of PPE on SARS infection rates. Six of these studies were case‐control studies and five were retrospective cohort studies. Since information in these studies was collected in the same retrospective way by questionnaires and/or interviews we combined the results of these studies.

There were two studies (Le 2004; Park 2004), one in a single hospital in Vietnam and the other in multiple hospitals in the USA, that reported no cases in spite of sufficient exposure to SARS patients. The Vietnamese study claimed that this was because of the almost universal use of N95 masks later during the epidemic. The US study could not find an explanation because the use of PPE was not optimal in many cases. We could find no reasons to explain this result because these studies were similar to the other studies included. Also, in another hospital near the one in the Vietnamese study, SARS cases did occur among healthcare workers but this was more at the beginning of the epidemic and it was unclear how well PPE had been used (Reynolds 2006).

1 Consistent mask use versus inconsistent use

We were able to combine six studies (Liu 2009; Loeb 2004; Nishiura 2005; Scales 2003; Seto 2003; Teleman 2004), in a meta‐analysis that showed a beneficial effect of consistent mask use as part of PPE both in a fixed‐effect (OR 0.28, 95% CI 0.17 to 0.46, I² = 42%) and in a random‐effects meta‐analysis model (OR 0.27, 95% CI 0.13 to 0.53).

2 Consistent gown/suit use versus inconsistent use

Four studies (Loeb 2004; Nishiura 2005; Pei 2006; Teleman 2004), could be combined and showed that consistent gown use had a preventive effect on SARS infection both in a fixed‐ and random‐effects analysis (OR 0.22, 95% 0.10 to 0.50, I² = 53%). The data in Teleman 2004 were reported as OR 0.5, 95% CI 0.4 to 6.9 P = 0.6). However, this is an apparent mistake as the confidence interval does not fit with the OR nor with the P value. We corrected this to OR 0.5, 95% CI 0.04 to 6.9 which makes the results consistent.

3 Consistent glove use versus inconsistent use

Also consistent glove use in six studies (Loeb 2004; Nishiura 2005; Pei 2006; Scales 2003; Seto 2003; Teleman 2004), led to a decrease in the risk of SARS infection both in fixed‐effect meta‐analysis (OR 0.54 95% CI 0.33 to 0.89, I² = 0%) and in a random‐effects analysis (OR 0.53, 95% CI 0.28 to 1.01) but this was not statistically significant.

4 Consistent use of more than one PPE part versus inconsistent use

Ho 2004, Lau 2004, and Scales 2003 measured consistent use of more than one PPE part compared to no use at all. The combination of more than one PPE had a similar effect on SARS infection risk but this was not statistically significant, neither in the fixed‐effect analysis (OR 0.36, 95% CI 0.09 to 1.39, I² = 35%) nor in the random‐effects analysis (OR 0.37, 95% CI 0.07 to 1.98).

Appendix 2. MEDLINE search strategy 15 July 2019

#1

"Protective Clothing"[Mesh] OR gown*[tw] OR coverall*[tw] OR "protective layer"[tw] OR "protective layers"[tw] OR "surgical toga"[tw] OR apron*[tw] OR "smock"[tw] OR "smocks"[tw] OR "hazmat suit"[tw] OR (hazmat[tw] AND suit[tw]) OR "Gloves, Protective"[Mesh] OR "glove"[tw] OR "gloves"[tw] OR "Respiratory Protective Devices"[Mesh] OR "Masks"[Mesh] OR "mask"[tw] OR "masks"[tw] OR "air‐purifying respirator"[tw] OR "PAPR"[tw] OR "enhanced respiratory and contact precautions" OR "E‐RCP"[tw] OR "respiratory protection"[tw] OR "transparent panel"[tw] OR "surgical mask"[tw] OR "surgical masks"[tw] OR "filtering face piece"[tw] OR "filtering facepiece"[tw] OR "Eye Protective Devices"[Mesh] OR goggle*[tw] OR "visor"[tw] OR "facial protection equipment"[tw] OR "safety glass"[tw] OR "safety glasses"[tw] OR "safety spectacles"[tw] OR "personal protective equipment"[tw] OR "PPE"[tw] OR "protective equipment"[tw] OR overshoe*[tw] OR "shoe cover"[tw] OR "shoe covers"[tw] OR "rubber boot"[tw] OR "rubber boots"[tw] OR "head cover"[tw] OR "head covering"[tw] OR "face shield"[tw] OR "face shields"[tw] OR "surgical hood"[tw] OR "hood"[tw] OR "Equipment Contamination/prevention and control"[Mesh] OR "Infection Control"[Mesh] OR "infection control"[tiab] OR "gloving"[tw] OR "donning"[tw] OR "doffing"[tw]

#2

"Communicable Diseases"[Mesh] OR "infectious disease"[tiab] OR "infectious diseases"[tiab] OR "Disease Transmission, Infectious"[Mesh] OR "disease transmission"[tw] OR "Infectious Disease Transmission, Patient‐to‐Professional"[Mesh] OR "infection control precautions"[tw] OR "human‐to‐human transmission"[tw] OR "parenteral transmission"[tw] OR "Virus Diseases/prevention and control"[Mesh] OR "viral disease"[tw] OR "viral diseases"[tw] OR "Bacterial Infections/prevention and control"[Mesh] OR "bacterial infection"[tw] OR "filovirus"[tw] OR "Ebolavirus"[Mesh] OR "Hemorrhagic Fever, Ebola"[Mesh] OR "Ebola"[tw] OR "Marburg virus"[tw] OR "Lassa virus"[tw] OR "haemorrhagic fever"[tw] OR "HIV Infections/prevention and control"[Mesh] OR "HIV"[ti] OR "hiv infection"[tiab] OR "hiv transmission"[tw] OR "Influenza, Human/prevention and control"[Mesh] OR "SARS Virus"[Mesh] OR "Severe Acute Respiratory Syndrome Virus"[tw] OR "SARS"[tw] OR "MERS"[tw] OR "respiratory infection"[tw] OR "Influenza, Human/prevention and control"[Mesh] OR "influenza"[tiab] OR "Tuberculosis/prevention and control"[Mesh] OR "tuberculosis"[tiab] OR "Hepatitis A"[Mesh] OR "hepatitis a"[ti] OR "Hepatitis B/prevention and control"[Mesh] OR "hepatitis b"[ti] OR "Hepatitis C/transmission"[Mesh] OR "hepatitis c"[ti] OR "bioterrorism"[tw] OR "aerosol‐generating procedure"[tw] OR "Cross Infection"[Mesh] OR "bacterial contamination"[tw] OR "microbial contamination"[tw] OR "self‐contamination"[tw] OR "decontamination"[tw] OR "surface decontamination"[tw] OR "skin decontamination"[tw]

#3

"Health Personnel"[Mesh] OR "Personnel, Hospital"[Mesh] OR "health care worker"[tw] OR "health care workers"[tw] OR "health care personnel"[tw] OR "health personnel"[tw] OR "health‐personnel"[tw] OR "health provider"[tw] OR "health providers"[tw] OR "health care provider"[tw] OR "health care providers"[tw] OR "medical staff"[tw] OR "medical personnel"[tw]OR "medical professional"[tw] OR "medical worker"[tw] OR "medical workers"[tw] OR "dental personnel"[tw] OR "dental staff"[tw] OR "Dentists"[Mesh] OR "dentist"[tw] OR "dentists"[tw] OR "dental assistant"[tw] OR "dental assistants"[tw] OR "Dental Assistants"[Mesh] OR "nursing staff"[tw] OR "Nurses"[Mesh] OR "nurse"[tw] OR "nurses"[tw] OR "nursing assistant"[tw] OR "nursing assistants"[tw] OR "Nurses' Aides"[Mesh] OR "Nurse Midwives"[Mesh] OR "midwife"[tw] OR "midwives"[tw] OR "military‐medical personnel"[tw] OR "Physicians"[Mesh] OR "physician"[tw] OR "physicians"[tw] OR "emergency medical services"[tw] OR "Emergency Medical Services"[MeSH] OR "transporting patients"[tw] OR "patient transport"[tw] OR "Ambulances"[Mesh] OR "Allied Health Personnel"[Mesh] OR paramedic[tw] OR paramedics[tw] OR paramedical personnel[tw] OR "Burial"[Mesh] OR burial staff OR cleaning workers[tw] OR cleaner work OR cleaner[tw] OR cleaners[tw]

#4

(#1 AND #2 AND #3)

Appendix 3. CENTRAL search strategy 20 March 2020

#1 MeSH descriptor: [Personal Protective Equipment] explode all trees

#2 MeSH descriptor: [Protective Clothing] explode all trees

#3 MeSH descriptor: [Respiratory Protective Devices] explode all trees

#4 MeSH descriptor: [Masks] explode all trees

#5 MeSH descriptor: [Eye Protective Devices] explode all trees

#6 MeSH descriptor: [Equipment Contamination] explode all trees

#7 MeSH descriptor: [Infection Control] explode all trees and with qualifier(s): [methods ‐ MT]

#8 (glove* or gloving):ti,ab,kw

#9 (gown* or coverall* or (protective NEXT layer*) or (surgical NEXT toga*) or apron* or smock* or (hazmat NEXT suit*)):ti,ab,kw

#10 (mask* or (air NEXT purifying NEXT respirator*) or PAPR or "enhanced respiratory and contact precautions" or ERCP or "respiratory protection" or (transparent NEXT panel*) or (filtering NEXT face NEXT piece*) or (filtering NEXT facepiece*)):ti,ab,kw

#11 (goggle* or visor* or (safety NEXT glass*) or "safety spectacles" or overshoe* or (shoe NEXT cover*) or (rubber NEXT boot*) or (head NEXT cover*) or (face NEXT shield*) or hood* or "protective equipment" or PPE or donning or doffing):ti,ab,kw

#12 "infection control":ti,ab,kw

#13 {or #1‐#12}

#14 MeSH descriptor: [Health Personnel] explode all trees

#15 MeSH descriptor: [Personnel, Hospital] explode all trees

#16 ((health NEXT care NEXT worker*) or (healthcare NEXT worker*) or "health care personnel" or "healthcare personnel" or "health personnel" or (health NEXT provider*) or (health NEXT care NEXT provider*) or "medical staff" or "medical personnel" or (medical NEXT professional*) or (medical NEXT worker*) or "military‐medical personnel" or "military medical personnel"):ti,ab,kw

#17 MeSH descriptor: [Dentists] explode all trees

#18 MeSH descriptor: [Dental Assistants] explode all trees

#19 ("dental personnel" or "dental staff" or dentist* or (dental NEXT assistant*)):ti,ab,kw

#20 MeSH descriptor: [Nurses] explode all trees

#21 MeSH descriptor: [Nursing Assistants] explode all trees

#22 MeSH descriptor: [Nurse Midwives] explode all trees

#23 MeSH descriptor: [Nursing Staff] explode all trees

#24 (nurse or nurses or nursing or midwife OR midwives):ti,ab,kw

#25 MeSH descriptor: [Physicians] explode all trees

#26 physician*:ti,ab,kw

#27 MeSH descriptor: [Emergency Medical Services] explode all trees

#28 MeSH descriptor: [Ambulances] explode all trees

#29 ("emergency medical services" or "transporting patients" or "patient transport" or paramedic* or (ambulance NEXT worker*)):ti,ab,kw

#30 MeSH descriptor: [Allied Health Personnel] explode all trees

#31 MeSH descriptor: [Burial] explode all trees

#32 "burial staff":ti,ab,kw

#33 ("cleaning workers" or cleaner* or janitor*):ti,ab,kw

#34 {or #14‐#33}

#35 MeSH descriptor: [Communicable Diseases] explode all trees

#36 MeSH descriptor: [Disease Transmission, Infectious] explode all trees

#37 MeSH descriptor: [Virus Diseases] explode all trees and with qualifier(s): [prevention & control ‐ PC, transmission ‐ TM]

#38 MeSH descriptor: [Bacterial Infections] explode all trees and with qualifier(s): [prevention & control ‐ PC, transmission ‐ TM]

#39 MeSH descriptor: [Ebolavirus] explode all trees

#40 MeSH descriptor: [Hemorrhagic Fever, Ebola] explode all trees

#41 MeSH descriptor: [Marburg Virus Disease] explode all trees

#42 MeSH descriptor: [Lassa virus] explode all trees

#43 MeSH descriptor: [Influenza, Human] explode all trees and with qualifier(s): [prevention & control ‐ PC, transmission ‐ TM]

#44 MeSH descriptor: [SARS Virus] explode all trees

#45 MeSH descriptor: [Severe Acute Respiratory Syndrome] explode all trees

#46 MeSH descriptor: [Middle East Respiratory Syndrome Coronavirus] explode all trees

#47 MeSH descriptor: [HIV Infections] explode all trees and with qualifier(s): [prevention & control ‐ PC, transmission ‐ TM]

#48 MeSH descriptor: [Tuberculosis] explode all trees and with qualifier(s): [prevention & control ‐ PC]

#49 MeSH descriptor: [Hepatitis A] explode all trees and with qualifier(s): [prevention & control ‐ PC, transmission ‐ TM]

#50 MeSH descriptor: [Hepatitis B] explode all trees and with qualifier(s): [prevention & control ‐ PC, transmission ‐ TM]

#51 MeSH descriptor: [Cross Infection] explode all trees

#52 ((infectious NEXT disease*) or "disease transmission" or "infection control precautions" or "human‐to‐human transmission" or "human transmission" or "parenteral transmission"):ti,ab,kw

#53 ((viral NEXT disease*) or (bacterial NEXT infection*) or filovirus or ebola or "Marburg virus" or "Lassa virus" or "haemorrhagic fever" or "hemorrhagic fever" or (HIV NEAR/3 infection*) or "Severe Acute Respiratory Syndrome Virus" or SARS or "Middle East Respiratory Syndrome" or MERS or coronavirus* or (corona NEXT virus*) or COVID or "COVID 19" or “severe acute respiratory syndrome coronavirus 2” or "SARS CoV 2" or (SARS NEXT CoV*)):ti,ab,kw

#54 ("surface decontamination" or "skin decontamination" or "self contamination" or self‐contamination):ti,ab,kw

#55 {or #35‐#54}

Appendix 4. Medline OVID search strategy 20 March 2020

1 exp Personal Protective Equipment/

2 exp Protective Clothing/

3 exp Respiratory Protective Devices/

4 exp Masks/

5 exp Eye Protective Devices/

6 exp Equipment Contamination/

7 exp Infection Control/mt [Methods]

8 (glove* or gloving).ti,ab.

9 (gown* or coverall* or protective layer* or surgical toga* or apron* or smock* or hazmat suit*).ti,ab.

10 (mask or masks or air purifying respirator* or PAPR or "enhanced respiratory and contact precautions" or ERCP or "respiratory protection" or transparent panel* or filtering face piece* or filtering facepiece*).ti,ab.

11 (goggle* or visor* or facial protection equipment or safety glass* or safety spectacles or overshoe* or shoe cover* or rubber boot* or head cover* or face shield* or hood* or protective equipment or PPE or donning or doffing).ti,ab.

12 infection control.ti.

13 or/1‐12

14 exp Health Personnel/

15 exp Personnel, Hospital/

16 (health care worker* or healthcare worker* or health care personnel or health personnel or health care provider* or health provider* or medical staff or medical personnel or medical professional* or medical worker* or military medical personnel).ti,ab.

17 exp Dentists/

18 exp Dental Assistants/

19 (dental personnel or dental staff or dentist* or dental assistant*).ti,ab.

20 exp Nurses/

21 exp Nursing Assistants/

22 exp Nurse Midwives/

23 exp Nursing Staff/

24 (nurse or nurses or nursing or midwife or midwives).ti,ab.

25 exp Physicians/

26 physician*.ti,ab.

27 exp Emergency Medical Services/

28 exp Ambulances/

29 (emergency medical services or transporting patients or patient transport or paramedic* or ambulance worker*).ti,ab.

30 exp Allied Health Personnel/

31 exp Burial/

32 burial staff.ti,ab.

33 (cleaning worker* or cleaner* or janitor*).ti,ab.

34 or/14‐33

35 exp Communicable Diseases/

36 exp Disease Transmission, Infectious/

37 exp Virus Diseases/

38 exp Bacterial Infections/

39 exp Ebolavirus/

40 exp Hemorrhagic Fever, Ebola/

41 exp Marburg Virus Disease/

42 exp Lassa virus/

43 exp Influenza, Human/

44 exp SARS Virus/

45 exp Severe Acute Respiratory Syndrome/

46 exp Middle East Respiratory Syndrome Coronavirus/

47 exp HIV Infections/pc, tm [Prevention & Control, Transmission]

48 exp Tuberculosis/pc, tm [Prevention & Control, Transmission]

49 exp Hepatitis A/pc, tm [Prevention & Control, Transmission]

50 exp Hepatitis B/pc, tm [Prevention & Control, Transmission]

51 exp Cross Infection/

52 (infectious disease* or disease transmission or infection control precautions or (human* adj3 transmission) or parenteral transmission).ti,ab.

53 (viral disease* or viral infection* or bacterial infection* or filovirus or ebola* or Marburg virus or Lassa virus or h?emorrhagic fever or (HIV adj3 infection*) or Severe Acute Respiratory Syndrome Virus or SARS or Middle East Respiratory Syndrome or MERS or coronavirus* or corona virus* or COVID or severe acute respiratory syndrome coronavirus or SARS CoV 2 or SARS‐CoV‐2).ti,ab.

54 (skin decontamination or surface decontamination or self contamination).ti,ab.

55 or/35‐54

56 13 and 34 and 55

57 exp animals/ not humans.sh.

Appendix 5. Embase OVID search strategy 20 March 2020

1 exp protective equipment/

2 exp protective clothing/

3 exp mask/

4 exp eye protective device/

5 exp medical device contamination/

6 infection control/pc [Prevention]

7 (glove* or gloving).ti,ab.

8 (gown* or coverall* or protective layer* or surgical toga* or apron* or smock* or hazmat suit*).ti,ab.

9 (mask or masks or air purifying respirator* or PAPR or "enhanced respiratory and contact precautions" or ERCP or "respiratory protection" or transparent panel* or filtering face piece* or filtering facepiece*).ti,ab.

10 (goggle* or visor* or facial protection equipment or safety glass* or safety spectacles or overshoe* or shoe cover* or rubber boot* or head cover* or face shield* or hood* or protective equipment or PPE or donning or doffing).ti,ab.

11 infection control.ti.

12 or/1‐11

13 exp health care personnel/

14 exp hospital personnel/

15 (health care worker* or healthcare worker* or health care personnel or health personnel or health care provider* or health provider* or medical staff or medical personnel or medical professional* or medical worker* or military medical personnel).ti,ab.

16 exp dentist/

17 exp dental assistant/

18 (dental personnel or dental staff or dentist* or dental assistant*).ti,ab.

19 exp nurse/

20 exp nursing assistant/

21 exp nurse midwife/

22 exp nursing staff/

23 (nurse or nurses or nursing or midwife or midwives).ti,ab.

24 exp physician/

25 physician*.ti,ab.

26 exp emergency health service/

27 exp ambulance/

28 (emergency medical services or transporting patients or patient transport or paramedic* or ambulance worker*).ti,ab.

29 exp paramedical personnel/

30 exp burial/

31 burial staff.ti,ab.

32 (cleaning worker* or cleaner* or janitor*).ti,ab.

33 or/13‐32

34 exp communicable disease/

35 exp disease transmission/

36 exp virus infection/

37 exp bacterial infection/

38 exp ebolavirus/

39 exp Ebola hemorrhagic fever/

40 exp Marburg hemorrhagic fever/

41 exp Lassa virus/

42 exp filovirus infection/

43 exp influenza/

44 exp SARS coronavirus/

45 exp severe acute respiratory syndrome/

46 exp Middle East respiratory syndrome coronavirus/

47 exp Human immunodeficiency virus infection/pc [Prevention]

48 exp tuberculosis/pc [Prevention]

49 exp hepatitis/pc [Prevention]

50 exp cross infection/

51 (infectious disease* or disease transmission or infection control precautions or (human* adj3 transmission) or parenteral transmission).ti,ab.

52 (viral disease* or viral infection* or bacterial infection* or filovirus or ebola* or Marburg virus or Lassa virus or h?emorrhagic fever or (HIV adj3 infection*) or Severe Acute Respiratory Syndrome Virus or SARS or Middle East Respiratory Syndrome or MERS or coronavirus* or corona virus* or COVID or severe acute respiratory syndrome coronavirus or SARS CoV 2 or SARS‐CoV‐2).ti,ab.

53 (skin decontamination or surface decontamination or self contamination).ti,ab.

54 or/34‐53

55 12 and 33 and 54

56 exp experimental organism/

57 animal tissue/

58 exp animal disease/

59 exp carnivore disease/

60 exp bird/

61 exp experimental animal welfare/

62 exp animal husbandry/

63 animal behavior/

64 exp animal cell culture/

65 exp mammalian disease/

66 exp mammal/

67 exp marine species/

68 nonhuman/

69 animal.hw.

70 or/56‐69

71 70 not human/

72 55 not 71

Appendix 6. Scopus search strategy 18 June 2019

#1

"protective clothing" OR gown* OR coverall* OR "protective layer" OR "protective layers" OR "surgical toga" OR apron* OR smock OR smocks OR "hazmat suit" OR glove OR gloves OR "respiratory protective devices" OR mask OR "air‐purifying respirator" OR "PAPR" OR "enhanced respiratory and contact precautions" OR "E‐RCP" OR "respiratory protection" OR "transparent panel" OR "surgical mask" OR "surgical masks" OR "filtering face piece" OR "filtering facepiece" OR "eye protective device" OR goggle* OR visor OR "facial protection equipment" OR "safety glass" OR "safety glasses" OR "safety spectacles" OR "personal protective equipment" OR "PPE" OR "protective equipment" OR overshoe* OR "shoe cover" OR "shoe covers" OR "rubber boot" OR "rubber boots" OR "head cover" OR "head covering" OR "face shield" OR "face shields" OR "surgical hood" OR hood OR gloving OR donning OR doffing)

#2

"health care personnel" OR "hospital personnel" OR "health care worker" OR "health care workers" OR "health care personnel" OR "health personnel" OR "health‐personnel" OR "health provider" OR "health providers" OR "health care provider" OR "health care providers" OR "medical staff" OR "medical personnel" OR "medical professional" OR "medical worker" OR "medical workers" OR "dental personnel" OR "dental staff" OR "dentist" OR "dentists" OR "dental assistant" OR "dental assistants" OR "nursing staff" OR "nurse" OR "nurses" OR "nursing assistant" OR "nursing assistants" OR "midwife" OR "midwives" OR "military‐medical personnel" OR "physician" OR "physicians" OR "emergency medical services" OR "transporting patients" OR "patient transport" OR "ambulance" OR "paramedical personnel" OR paramedic OR paramedics OR "burial staff" OR "cleaning workers" OR cleaner OR cleaners

#3

"virus infection" OR "viral disease" OR "filovirus" OR "ebola" OR "marburg virus" OR "lassa virus" OR "haemorrhagic fever" OR "Severe Acute Respiratory Syndrome Virus" OR "SARS" OR "MERS" OR "bioterrorism" OR "bacterial contamination" OR "microbial contamination" OR "self‐contamination" OR "decontamination" OR "surface decontamination" OR "skin decontamination"

#4

LIMIT‐TO ( PUBYEAR, 2018 )

#5

#1 AND #2 AND #3 AND #4

Appendix 7. Embase search strategy embase.com 15 July 2016

#7

#6 NOT [medline]/lim) (646)

#6

#5 AND [embase]/lim (2,227)

#5

#4 AND [humans]/lim (5,270)

#4

#1 AND #2 AND #3 (5,675)

#3

'communicable disease'/de OR "infectious disease":ab,ti OR 'disease transmission'/de OR "disease transmission" OR "infection control precautions" OR "human‐to‐human transmission" OR "parenteral transmission" OR 'virus infection'/de OR "viral disease":ab,ti OR 'bacterial infection'/de OR "bacterial infection":ab,ti OR "filovirus" OR 'ebola virus'/de OR 'hemorrhagic fever ebola'/de OR "ebola" OR "marburg virus" OR "lassa virus" OR "haemorrhagic fever" OR 'sars coronavirus'/de OR "Severe Acute Respiratory Syndrome Virus" OR "SARS" OR "MERS" OR "bioterrorism" OR 'cross infection'/de OR "bacterial contamination" OR "microbial contamination" OR "self‐contamination" OR "decontamination" OR "surface decontamination" OR "skin decontamination" (323,524)

#2

'health care personnel'/de OR 'hospital personnel'/de OR "health care worker" OR "health care workers" OR "health care personnel" OR "health personnel" OR "health‐personnel" OR "health provider" OR "health providers" OR "health care provider" OR "health care providers" OR "medical staff" OR "medical personnel" OR "medical professional" OR "medical worker" OR "medical workers" OR “dental personnel” OR “dental staff” OR "dentist" OR "dentists" OR "dental assistant" OR "dental assistants" OR "nursing staff" OR 'nurses'/de OR "nurse" OR "nurses" OR "nursing assistant" OR "nursing assistants" OR 'nursing assistant'/de OR 'nurse midwife'/de OR "midwife" OR "midwives" OR "military‐medical personnel" OR 'physician'/de OR "physician" OR "physicians" OR "emergency medical services" OR “transporting patients” OR “patient transport” OR 'ambulance'/de OR 'paramedical personnel'/de OR "paramedical personnel" OR paramedic OR paramedics OR 'posthumous care'/de OR "burial staff" OR "cleaning workers" OR "cleaner work" OR cleaner OR cleaners (1,287,399)

#1

'protective clothing'/de OR gown* OR coverall* OR "protective layer" OR "protective layers" OR "surgical toga" OR apron* OR smock OR smocks OR "hazmat suit" OR (hazmat AND suit) OR glove OR gloves OR 'respiratory protective devices'/de OR 'mask'/de OR mask OR "air‐purifying respirator" OR "PAPR" OR "enhanced respiratory and contact precautions" OR "E‐RCP" OR “respiratory protection” OR "transparent panel" OR "surgical mask" OR "surgical masks" OR "filtering face piece" OR "filtering facepiece" OR 'eye protective device'/de OR goggle* OR visor OR "facial protection equipment" OR "safety glass" OR "safety glasses" OR "safety spectacles" OR "personal protective equipment" OR "PPE" OR "protective equipment" OR overshoe* OR "shoe cover" OR "shoe covers" OR "rubber boot" OR "rubber boots" OR "head cover" OR "head covering" OR "face shield" OR "face shields" OR "surgical hood" OR hood OR 'medical device contamination'/de OR 'infection control'/de OR 'infection control':ab,ti OR gloving OR donning OR doffing (160,118)

Appendix 8. CINAHL EBSCO search strategy 20 March 2020

S51 S10 AND S32 AND S50

S50 S33 OR S34 OR S35 OR S36 OR S37 OR S38 OR S39 OR S40 OR S41 OR S42 OR S43 OR S44 OR S45 OR S46 OR S47 OR S48 OR S49

S49 TI ("surface decontamination" OR "skin decontamination") OR AB ("surface decontamination" OR "skin decontamination")

S48 AB (“viral disease” OR “viral diseases” OR “viral infection” OR “viral infections” OR “bacterial infection” OR “bacterial infections” OR filovirus OR ebola* OR “Marburg virus” OR “Lassa virus” OR “haemorrhagic fever” OR “hemorrhagic fever” OR “HIV infection” OR “HIV infections” OR “Severe Acute Respiratory Syndrome Virus” OR SARS OR “Middle East Respiratory Syndrome” OR MERS OR coronavirus* OR “corona virus” OR “corona viruses” OR COVID OR “severe acute respiratory syndrome coronavirus” OR “SARS CoV 2” OR “SARS‐CoV‐2”)

S47 TI (“viral disease” OR “viral diseases” OR “viral infection” OR “viral infections” OR “bacterial infection” OR “bacterial infections” OR filovirus OR ebola* OR “Marburg virus” OR “Lassa virus” OR “haemorrhagic fever” OR “hemorrhagic fever” OR “HIV infection” OR “HIV infections” OR “Severe Acute Respiratory Syndrome Virus” OR SARS OR “Middle East Respiratory Syndrome” OR MERS OR coronavirus* OR “corona virus” OR “corona viruses” OR COVID OR “severe acute respiratory syndrome coronavirus” OR “SARS CoV 2” OR “SARS‐CoV‐2”)

S46 TI (“infectious disease” OR “infectious diseases” OR “disease transmission” OR “infection control precautions” OR “human‐to‐human transmission” OR “parenteral transmission”) OR AB (“infectious disease” OR “infectious diseases” OR “disease transmission” OR “infection control precautions” OR “human‐to‐human transmission” OR “parenteral transmission”)

S45 (MH "Cross Infection+")

S44 (MH "Hepatitis B+/PC/TM")

S43 (MH "Hepatitis A/PC/TM")

S42 (MH "Tuberculosis+/PC/TM")

S41 (MH "HIV Infections+/TM/PC")

S40 (MH "Middle East Respiratory Syndrome") OR (MH "Middle East Respiratory Syndrome Coronavirus")

S39 (MH "SARS Virus") OR (MH "Severe Acute Respiratory Syndrome")

S38 (MH "Influenza, Human+")

S37 (MH "Hemorrhagic Fever, Ebola") OR (MH "Ebola Virus")

S36 (MH "Bacterial Infections+")

S35 (MH "Virus Diseases+")

S34 (MH "Disease Transmission, Patient‐to‐Professional") OR (MH "Disease Transmission, Professional‐to‐Patient")

S33 (MH "Communicable Diseases+")

S32 S11 OR S12 OR S13 OR S14 OR S15 OR S16 OR S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31

S31 TI ("cleaning worker" OR "cleaning workers" OR cleaner* or janitor*) OR AB ("cleaning worker" OR "cleaning workers" OR cleaner* or janitor*)

S30 (TI burial) or (AB burial)

S29 (MH "Allied Health Personnel+")

S28 TI (“emergency medical services” OR “transporting patients” OR “patient transport” OR paramedic* OR “ambulance worker” OR “ambulance workers”) OR AB (“emergency medical services” OR “transporting patients” OR “patient transport” OR paramedic* OR “ambulance worker” OR “ambulance workers”)

S27 (MH "Ambulances")

S26 (MH "Emergency Medical Services+")

S25 (TI physician*) OR (AB physician*)

S24 (MH "Physicians+")

S23 AB (nurse OR nurses OR nursing OR midwife OR midwives)

S22 TI (nurse OR nurses OR nursing OR midwife OR midwives)

S21 (MH "Nurse Midwives")

S20 (MH "Nursing Assistants")

S19 (MH "Nurses+")

S18 AB ("dental personnel" OR "dental staff" OR dentist* OR "dental assistant" OR "dental assistants")

S17 TI ("dental personnel" OR "dental staff" OR dentist* OR "dental assistant" OR "dental assistants")

S16 (MH "Dental Assistants")

S15 (MH "Dentists+")

S14 AB (“health care worker” OR “health care workers” OR “healthcare worker” OR “healthcare workers” OR “health care personnel” OR “health personnel” OR “health care provider” OR “health care providers” OR “health provider” OR “health providers” OR “medical staff” OR “medical personnel” OR “medical professional” OR “medical professionals” OR “medical worker” OR “medical workers” OR "military medical personnel")

S13 TI (“health care worker” OR “health care workers” OR “healthcare worker” OR “healthcare workers” OR “health care personnel” OR “health personnel” OR “health care provider” OR “health care providers” OR “health provider” OR “health providers” OR “medical staff” OR “medical personnel” OR “medical professional” OR “medical professionals” OR “medical worker” OR “medical workers” OR "military medical personnel")

S12 (MH "Personnel, Health Facility+")

S11 (MH "Health Personnel+")

S10 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7 OR S8 OR S9

S9 AB (glove* OR gloving OR gown* OR coverall* OR "protective layer" OR "protective layers" OR "surgical toga" OR apron* OR smock* OR "hazmat suit" OR “hazmat suits” OR mask* OR "air‐purifying respirator" OR PAPR OR "enhanced respiratory and contact precautions" OR E‐RCP OR ECPR OR "respiratory protection" OR "transparent panel" OR "surgical mask" OR "surgical masks" OR "filtering face piece" OR "filtering facepiece" OR goggle* OR visor* OR "facial protection equipment" OR "safety glass" OR "safety glasses" OR "safety spectacles" OR "personal protective equipment" OR PPE OR "protective equipment" OR overshoe* OR "shoe cover" OR "shoe covers" OR "rubber boot" OR "rubber boots" OR "head cover" OR "head covering" OR "face shield" OR "face shields" OR "surgical hood" OR "hood" OR "infection control OR donning OR doffing)

S8 TI (glove* OR gloving OR gown* OR coverall* OR "protective layer" OR "protective layers" OR "surgical toga" OR apron* OR smock* OR "hazmat suit" OR “hazmat suits” OR mask* OR "air‐purifying respirator" OR PAPR OR "enhanced respiratory and contact precautions" OR E‐RCP OR ECPR OR "respiratory protection" OR "transparent panel" OR "surgical mask" OR "surgical masks" OR "filtering face piece" OR "filtering facepiece" OR goggle* OR visor* OR "facial protection equipment" OR "safety glass" OR "safety glasses" OR "safety spectacles" OR "personal protective equipment" OR PPE OR "protective equipment" OR overshoe* OR "shoe cover" OR "shoe covers" OR "rubber boot" OR "rubber boots" OR "head cover" OR "head covering" OR "face shield" OR "face shields" OR "surgical hood" OR "hood" OR "infection control OR donning OR doffing)

S7 (MH "Infection Control+/PC")

S6 (MH "Equipment Contamination/PC")

S5 (MH "Respiratory Protective Devices")

S4 (MH "Gloves")

S3 (MH "Eye Protective Devices")

S2 (MH "Masks")

S1 (MH "Protective Clothing+")

Appendix 9. CINAHL search strategy 31 July 2018

S5 S4 MEDLINE records excluded (878)

S4 (S1 AND S2 AND S3) (2,584)

S3

(MH "Communicable Diseases") OR (TI "infectious disease") OR (AB "infectious disease") OR (MH "Disease Transmission) OR TX "disease transmission" OR (MH "Disease Transmission, Patient‐to‐Professional") OR TX "infection control precautions" OR TX "human‐to‐human transmission" OR TX "parenteral transmission" OR (MH "Virus Diseases/PC") OR TX "viral disease" OR TX "viral diseases" OR TX "bacterial infection" OR (MH "Bacterial infection/PC") OR TX "filovirus" OR TX "ebolavirus" OR (MH "Hemorrhagic Fever, Ebola") OR TX "ebola" OR TX "marburg virus" OR TX "lassa virus" OR TX "haemorrhagic fever" OR (MH "SARS Virus") OR TX "severe acute respiratory syndrome virus" OR TX "SARS" OR TX "MERS" OR TX "respiratory infection" OR TX "bioterrorism" OR TX "aerosol‐generating procedure" OR (MH "Cross Infection") OR TX "bacterial contamination" OR TX "microbial contamination" OR TX "self‐contamination" OR TX "decontamination" OR TX "surface decontamination" OR TX "skin decontamination" (37,937)

S2

(MH Protective Clothing) OR TX gown* OR TX coverall* OR TX "protective layer" OR TX "protective layers" OR TX "surgical toga" OR TX apron* OR TX "smock" OR TX "smocks" OR TX "hazmat suit" OR TX (hazmat AND suit) OR (MH "gloves protective") OR TX glove OR TX gloves OR (MH "Respiratory Protective Devices") OR (MH "Masks") OR TX mask OR TX masks OR TX "air‐purifying respirator" OR TX "PAPR" OR TX "enhanced respiratory and contact precautions" OR TX "E‐RCP" OR TX "respiratory protection" OR TX "transparent panel" OR TX "surgical mask" OR TX "surgical masks" OR TX "filtering face piece" OR TX "filtering facepiece" OR (MH "Eye Protective Devices") OR TX goggle* OR TX "visor" OR TX "facial protection equipment" OR TX "safety glass" OR TX "safety glasses" OR TX "safety spectacles" OR TX "personal protective equipment" OR TX "PPE" OR TX "protective equipment" OR TX overshoe* OR TX "shoe cover" OR TX "shoe covers" OR TX "rubber boot" OR TX "rubber boots" OR TX "head cover" OR TX "head covering" OR TX "face shield" OR TX "face shields" OR TX "surgical hood" OR TX "hood" OR (MH "Equipment Contamination/PC") OR (MH "Infection Control") OR (TI "infection control") OR (AB "infection control") OR TX "gloving" OR TX "donning" OR TX “doffing” (28,554)

S1

(MH "Health Personnel") OR TX health care workers OR TX health care personnel OR TX health personnel OR TX health‐personnel OR TX health providers OR TX health care providers OR TX medical staff OR TX medical personnel OR TX medical professional OR TX medical workers OR TX dental personnel OR TX dental staff OR (MH "Dentists") OR TX dentist OR TX dental assistant OR TX nursing staff OR (MH "Nurses") OR TX nurse OR TX nursing assistant OR (MH "Allied Health Personnel" OR (MH "Midwives") OR TX nurse midwife OR TX nurse midwives OR TX military‐medical personnel OR (MH “Physicians") OR TX physician OR TX emergency medical services OR (MH “Emergency Medical Services”) OR TX transporting patients OR TX patient transport OR (MH "Ambulance") OR (MH "Allied Health Personnel") OR TX paramedic OR TX paramedical personnel OR (MH "Burial") OR TX burial staff OR TX cleaning worker OR TX cleaner work OR TX cleaner OR TX cleaners (498,394)

Appendix 10. OSH‐update search strategy

Notes

Edited (no change to conclusions)

Data and analyses

Comparison 1

PAPR versus E‐RCP Attire

Analysis

Comparison 1: PAPR versus E‐RCP Attire, Outcome 2: Contamination > 1 cm

Analysis

Comparison 1: PAPR versus E‐RCP Attire, Outcome 3: Contamination area

Comparison 2

Four types of PPE attire compared

Comparison 3

Formal versus local available attire

Comparison 4

Gown versus apron

Comparison 5

Three types of PPE compared

Comparison 6

Gown sealed gloves versus standard gown

Comparison 7

Gown easy to doff versus standard gown

Comparison 8

Gown with gown‐glove improvement vs standard gown‐gloves

Comparison 9

Gown with marked inside versus standard gown

Comparison 10

Gloves with tab versus standard gloves

Comparison 11

Mask with tabs versus no mask tabs

Comparison 12

Doffing with double gloves versus doffing with single gloves

Analysis

Comparison 12: Doffing with double gloves versus doffing with single gloves, Outcome 2: Contamination: virus quantity

Comparison 13

CDC versus individual doffing

Comparison 14

Single‐step doffing vs CDC standard

Comparison 15

Doffing with extra sanitation of gloves versus standard no sanitation

Comparison 16

Donning and doffing with instructions versus without instructions

Analysis

Comparison 16: Donning and doffing with instructions versus without instructions, Outcome 3: Fluorescence contamination

Comparison 17

Active training in PPE use versus passive training

Analysis

Comparison 17: Active training in PPE use versus passive training, Outcome 1: Noncompliance with PPE

Comparison 18

Doffing with hypochlorite versus doffing with alcohol‐based glove sanitiser

Comparison 19

Active training in PPE doffing versus passive training

Comparison 20

Computer simulation versus no simulation

Comparison 21

Video‐based learning versus traditional lecture

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Characteristics of excluded studies [ordered by study ID]

Characteristics of ongoing studies [ordered by study ID]