Introduction
To protect healthcare personnel (HCPs) caring for patients with communicable diseases, protocols have been established to mitigate the risk of transmission. 1 Central to these protocols is personal protective equipment (PPE). The PPE used to minimize exposure to high-consequence infectious diseases (HCIDs), such as Ebola virus disease, utilizes layers of barrier precautions including fluid-resistant coveralls, impervious aprons or gowns, fluid-resistant footwear, powered air-purifying respirators (PAPRs), and gloves. Protocols that outline proper donning and doffing of the PPE are fundamental to mitigating self-contamination for HCPs and preventing the transmission of contaminants outside patient rooms. Reference Casanova, Erukunuakpor and Kraft2 However, PPE protocols are often based on manufacturer recommendations of individual products and infection prevention best practices. Accordingly, ensembles of PPE and their corresponding protocols usually have not been empirically validated.
Previous studies have demonstrated that adherence to PPE doffing protocols is challenging and variable among HCPs Reference Casanova, Erukunuakpor and Kraft2 ; previous studies have quantitatively examined and discovered high rates of deviations from established protocols. Reference Kwon, Burnham and Reske3 Doffing protocols would optimally be safe even considering this high underlying variability. Moreover, rigorous risk assessments of PPE ensembles should consider this factor in evaluating PPE safety.
This study explores contamination risks of an established doffing protocol. To validate this protocol’s efficacy, we applied a quantitative analysis of the PPE ensemble to test for self-contamination. This investigation consisted of two phases: (1) examination of the original PPE ensemble and its doffing protocol with fluorescence and bacteriophages; and (2) determination if an amended protocol decreased self-contamination.
This study highlights the need for infection prevention protocols, such as high-containment PPE doffing, to be evaluated in a quantitative, experimental fashion. We present results from both phases of our investigation below.
Methods
Investigational procedures
Movie analysis
Doffing protocols were captured on video. Each movie was analyzed by authors BAB, KBB, APS, and JSM who viewed and stopped the movie and recorded comments when potential hazards were observed.
For each doffing trial, footage was taken from four camera angles using camcorders (Canon, Japan, Model #HF R80) supplied by the Healthcare and Human Factors Lab at Emory. Comments were made for possible contamination during the Phase 1 phage trial and deviations from the protocol in the Phase 2 phage trial (Supplementary Table 1).
Fluorescent testing
To verify the contact we observed in the movie described above, we used a 0.5% fluorescein solution (Millipore Sigma, USA, Catalogue #F6377) and Glo Germ powder (Glo Germ, USA, #GGP10) to visualize contamination and identify the steps where spread of contaminant may occur. To conduct this test, a trained HCP donned a complete high-containment PPE ensemble and was then sprayed with the fluorescein solution. The solution was sprayed to coat surfaces of the PPE that may be exposed to contact with patients (patient-facing surfaces). The HCP then doffed according to the protocol. Pictures were taken from multiple angles to record how fluorescein transfer; ultraviolet flood lights (Onforu, China, #UFLAU004102) were used to emphasize fluorescence. A separate test was performed for aerosolization risk where the heavy-loading filter on top of the PAPR hood was laden with Glo Germ powder. After the filter was removed, pictures were taken with ultraviolet to capture results of the experiment.
Phage testing
Using a previously validated procedure, Reference Burke, Berryhill and Garcia4 we sprayed with small spray bottles (Bürkle, Germany, #10216-888) with high densities (approximately 108 phages per mL) of three genetically distinct λ phages on HCP volunteers to reflect densities of pathogens found in patient samples (Supplementary Table 2). 2 mL of λ phages were sprayed onto three locations, and each site was sprayed with a different variant of λ: one marked with a kanamycin resistance gene was sprayed on the wrists; one marked with a chloramphenicol resistance gene was sprayed on the back of the hood; and one lacking an antibiotic marker at the critical triangle, described below. HCPs then doffed, after which their hands, forearms, and PAPR were swabbed and scrubs were collected and tested for bacteriophage presence and identification.
Materials and technical methods
Strains
Escherichia coli strain C was acquired from Marie-Agnès Petit from INRAe, France. Bacteriophages λ (λTemp), λChl, and λKan was obtained from Maroš Pleška at The Rockefeller University.
Bacteriophage lysate preparation and distribution
Each type of λ phage lysate was inoculated, shaken, centrifuged, and filtered per the methods in Burke et al. Reference Burke, Berryhill and Garcia4 to create high-titer lysates (PFU/mL between 1×108 and 1×109). These lysates were stored in spray bottles, transported, and primed per the methods in Burke et al. Reference Burke, Berryhill and Garcia4 Immediately after priming, each lysate was sprayed with one pump from a distance of 10 cm onto the target sites for initial contamination. The spray dried clear and was unidentifiable to the naked eye. Contamination occurred no earlier than five minutes before the start of the doffing.
Bacteriophage recovery
Immediately after the doffing procedure, skin was sampled by applying a saline wipe (Hygea, USA, #C22370) around the hands and a wipe around the forearms; these wipes were stored in conical tubes (Corning, USA, #352070). Disposable scrubs were then stored in a Whirl-Pak (Nasco, USA, #B01542). Four sites of interest were swabbed with self-contained saline swabs (Hardy Diagnostic, USA, #SRK35) using a progressive back-and-forth motion until the entire surface appeared damp. To liberate phage from the saline wipe, the wipe was squeezed to remove excess liquid and the extracted solution was tested. To recover phage from the scrubs, 300 mL of deionized water was added to the bags that contained the scrubs and shaken vigorously to ensure scrubs were fully saturated. Excess liquid was poured into a conical tube for testing. To recover phage from the swabs, the saline containers were vortexed vigorously.
After bacteriophage recovery, all surfaces with possible phages were sprayed with 70% ethanol (Decon Labs, USA, #2716) and wiped with Sani-Cloth Disposable Wipes (Professional Disposables International, Inc., USA, #Q55172).
Bacteriophage identification and quantification
Phage identification was performed by polymerase chain reaction (PCR), using the methods and materials used by Burke et al. Reference Burke, Berryhill and Garcia4 Band sizes of 800bp were called λTemp, 1500bp called λChl, and 1900bp called λKan. The PCR was performed with an O’Gene Ruler DNA Ladder (Thermo Fisher Scientific, USA, #SM1563).
The serum resistance lipoprotein (bor) gene (Gene ID: 2703532, NCBI) of the λ phages was amplified by PCR using the following primers designed in PrimerBLAST (NCBI): Forward (borRG1Fw) 5’-GCTCTGCGTGATGATGTTGC-3’ and Reverse (borRG1Rv) 5’-GCAGAGAAGTTCCCCGTCAG-3’. Using the double-layer soft agar method. Reference Kropinski, Mazzocco, Waddell, Lingohr and Johnson5 LB soft agar overlays containing 0.1 mL of a turbid E. coli C overnight were prepared and allowed to harden. 0.01 mL of serially diluted saline recovery solution was spotted on the overlay at four densities. These plates were grown overnight at 37°C, and plaques were enumerated the next day.
If samples were determined to be PCR positive but negative via spot testing, 100 µL of sample were cultured with 1×107 CFU/mL log-phase E. coli C in 10 mL of LB broth. These cultures were grown with shaking for 6 hours, centrifuged, and filtered through a 0.22 μm filter to generate boosted lysates. 300 µL of these lysates were plated on E. coli C lawns to determine viable bacteriophage presence.
Process documentation (videography and still photography)
For the fluorescein and Glo Germ experiments, pictures were taken with an iPhone under ultraviolet illumination in a dark room; footage was recorded with one camcorder. During doffing trials, pictures and footage were recorded under standard room lighting.
Results
Initial protocol analysis
Analysis of initial doffing movie
The investigation of the initial protocol (detailed in Supplementary Tables 3, 4, and 5) began with a movie analysis. Table 1 notes observations of potential sources of contamination to the HCP during the doffing protocol displayed in Movie 1. Steps can be correlated to the steps of the original protocol in Supplementary Table 3.
Table 1. Events of concern noted during annotation of the personal protective equipment doffing movie
* This step in the movie deviated from the written protocol.
Evaluating contamination via fluorescence testing
In Figure 1 we present several pictures of testing with fluorescence that illustrate the concerns raised by the movie. We note that the protocol as performed in our trials, including this one, follows the written protocol and differs slightly from Movie 1; Movie 2 accurately portrays the written doffing protocol. There are six pairs of pictures: The left (L) panes show areas of concern, and the right (R) panes show the spread of fluorescence from those events.
Figure 1. Fluorescent visualization of areas and actions of concern during doffing. Experimental results of doffing with fluorescent markers present for specific actions of concern or highlighting areas of concern, as found during review of the doffing protocol movie. Left (L): Before; shows areas that may be of concern. Right (R): After; shows potential concerns by transfer of fluorescence. (A) The original personal protective equipment (PPE) ensemble was in natural light, both with the apron and with the apron removed. (B) Patient-facing surfaces of the PPE not covered by the apron could become contaminated. (C) Critical Triangle area of the powered air-purifying respirator (PAPR). (D) Interaction of the coverall sleeves with the inside of the PAPR hood. (E) Transfer of contamination from the PAPR hood to the forearm that can occur during doffing. (F) Aerosolization of fine powder trapped on the heavy-loading filter.
The fluorescein reveals the materials in the complete PPE ensemble that are patient-facing. 1AL shows a complete ensemble, and 1AR shows the ensemble without the apron. 1BL shows the ensemble under blacklight, and 1BR shows contamination not covered by the apron. 1CL and 1CR highlight the “Critical Triangle,” which includes parts of the shoulder, the side of the abdomen, and the arm.
The movement of the fluorescein demonstrated in Figure 1 reveals how contamination can move from the outside of the PPE ensemble to an intermediate location, then ultimately to the HCP. These contamination pathways are demonstrated in 1D, 1E, and 1F. 1DL shows how contamination may reach the arm or Critical Triangle of the coveralls. From here, contamination could transfer to the underside of the shroud (shown in 1DR) which could then move to scrubs. 1E shows a second pathway, where contamination on a patient-facing shoulder (1EL) transfers to arms when reaching up to roll up the PAPR hood, with that contamination demonstrated in 1ER. 1F demonstrates how aerosolized pathogens land on skin, scrubs, and footwear.
Using these results, three initial locations on the PPE were determined to pose a high contamination risk. These locations are (i) the PAPR hood Critical Triangle, located to the left and right of the apron and near shoulders, (ii) the wrist/lower forearm area of the protective coverall, and (iii) the back of the PAPR hood near the filter and shoulders.
Doffing in the presence of a bacteriophage proxy
To more accurately mirror pathogenic contamination, we inoculated three genetically marked variants of λ on the three sites above to determine both the origin and final location of each virus. Presented in Figure 2 are the results of doffing performed by four HCPs with varying heights and body types and varying experience in performing the protocol.
Figure 2. Phage recovery after doffing personal protective equipment. Experimental results of doffing protocols performed by four healthcare personnel with three bacteriophages initially inoculated on the powered air-purifying respirator (PAPR) Critical Triangle, coverall cuffs, and the back of the PAPR hood. Numbers inside each square represent the number of PFU/mL recovered from that location.
Figure 2 demonstrates that phages moved to the four locations we had hypothesized could become contaminated. Moreover, we found these phages to be viable and present at high densities. All four HCPs demonstrated contamination.
These four HCPs were recorded from multiple angles while performing the doffing procedure. We present in Table 2 behaviors noted during our analysis of the movies that would increase the risk of contamination.
Table 2. Concerns noted while reviewing movie of four healthcare personnel doffing
* Critical Triangle = The side of the PAPR hood, coveralls, and arm, which may be exposed and facilitate contamination.
Updated protocol analysis
Changes to the protocol
Our analysis and experiments of the first protocol revealed insufficiencies that led to contamination of the HCP. We aimed to eliminate viable phage recovery by limiting the observed contamination pathways. Accordingly, we altered the protocol in both equipment and doffing steps (Movie 3; Supplementary Table 3; Supplementary Table 4; Supplementary Table 5). Below is a table detailing changes made to the PPE ensemble and procedure (Table 3).
Table 3. Updates to the protocol
Six amendments were made to the protocol. Adjustments were made based largely upon concerns raised by the phase one analysis, but amendments were also incorporated for ease of doffing. Of the six changes, three were changes in equipment; one was an additional step made for added equipment; and two were reordered steps.
Phage testing with updated protocol
We next evaluated the updated protocol with the phage testing described previously. Nine HCPs doffed using the updated protocol. The results of these doffing trials are presented in Figure 3.
Figure 3. Bacteriophage recovery after doffing personal protective equipment with the altered protocols. Experimental results of the altered doffing protocols performed by nine healthcare personnel (HCPs) with three bacteriophages initially inoculated on the powered air-purifying respirator (PAPR) Critical Triangle, coverall cuffs, and the back of the PAPR hood. An X denotes that the phage DNA from the origin location was found at that sampled location at the end of doffing via PCR. To test for viable phages below the limit of detection (1×102 PFU/mL) samples were incubated with a susceptible bacteria host and no viable phages were recovered from any HCP.
Following the updated protocol, no viable phages were recovered. Phage DNA was found via PCR from several locations (indicated by an X) but viable phages were unable to be recovered from these PCR-positive samples even after providing a bacterial host. This indicates that the phages moved during doffing, but these phages were likely inactivated by the alcohol-based sanitizer during hand hygiene. Even if a contamination pathway was not eliminated, the updated protocol limited those pathways to contamination on gloves where sanitation could deactivate the bacteriophages.
Deviations from the doffing protocol by HCPs could contribute to variability in the results shown in Figure 3. We analyzed footage of each HCP doffing and noted deviations from the protocol which may lead to the spread of bacteriophages (Supplementary Table 1). Several HCPs deviated from the procedure. However, these deviations did not increase contamination per the results in Figure 3.
Discussion
PPE forms the cornerstone of safety for HCPs, but for HCPs working with HCIDs, satisfactory high-containment PPE is especially important. Reference MacIntyre, Chughtai, Seale, Richards and Davidson6 Hundreds of HCPs experienced near-miss events, infections, or death from Ebola virus disease. Reference World Health7,Reference Houlihan, McGowan and Dicks8 Although individual pieces of equipment receive National Institute for Occupational Safety & Health approval, PPE ensembles and their doffing protocols do not. Indeed, Koh et al. wrote over twenty years ago that PPE needed to be evaluated for efficacy against infection from severe acute respiratory syndrome—this is a problem that has needed addressal for decades, Reference Koh, Lim and Chia9 and later, the same call to action was issued for empirical review of Ebola PPE and ensembles. Reference Sprecher, Caluwaerts and Draper10
This PPE ensemble had not been assessed by empirical means. Our goal was to evaluate the ensemble and doffing protocol for possible contamination pathways and offer interventions to mitigate potential contamination. Even a slight contamination of an HCID could be a threat to HCPs (Supplementary Table 2). Thus, the aim of our interventions was to prevent viable self-contamination. We note that although other methods exist for analyzing the antiviral and disinfection qualities of PPE and its ensembles, Reference Xue, Coleman and Duncan11,Reference Pfuderer, Stark and Grass12 we elected to focus exclusively on how this ensemble performed in the transfer of pathogens.
Our first phase of this study began with an examination of a movie depicting the original protocol. The original protocol followed infection prevention best practices and was designed with disease containment in mind, but an in-depth evaluation revealed potential contamination pathways. We found from the fluorescence testing that contamination moved through the pathways we had hypothesized onto scrubs and skin.
Fluorescence experiments, however, carry limitations. Reference Hall, Poller and Bailey13 Fluorescence does not reflect sanitation measures and can be visually tracked by participants. Contamination experiments with phages resolve both failings. The viruses are visually undetectable and can be inactivated via alcohol-based sanitation but pose no appreciable risk. Reference Burke, Berryhill and Garcia4,Reference Burke, Berryhill and Garcia14 With the original doffing protocol, all participants had at least one contamination with at least 1800 virions present—an amount far greater than the minimum infective dose of many HCIDs (Supplementary Table 2).
We next offered an assortment of interventions. Modifications were made not only to reduce contamination by contact but also to make doffing easier. Reducing discomfort for HCPs may reduce deviations from a protocol, reducing contamination. The result of these changes manifested in the phage experiment with the updated protocol. In the second phage experiment, we did not recover a viable population of phage on any of the nine HCPs. Through PCR, we found phage DNA in several locations, indicating that the phages were inactivated by the use of alcohol-based sanitizer during the doffing. Phages that may have contaminated several locations were routed through pathways that included successful hand sanitation. Moreover, updates to the protocol eliminated intermediate contamination locations present in the original doffing protocol, which would have re-contaminated the HCP at later doffing steps. These results were observed despite deviations from the protocol by the HCPs during their doffings. The protocol, built to include redundancies and reduce events of contamination, allowed for small deviations without self-contamination.
This study does contain limitations. The original contamination was deliberately placed according to the fluorescent test with the intention of revealing contamination pathways. Thus, we cannot wholly capture contamination that would occur in a clinical setting – instead, we show how specific contamination can be tracked and eliminated through specific procedures. Further studies are needed to capture how contamination may move throughout a clinical environment, on PPE and otherwise. We further note that phages are only proxies. Using HCIDs for studies such as this is not ethical, but accordingly, we are closely approximating how they would function in a clinical setting through phages.
With the initial PPE ensemble and doffing protocol, contamination occurred that would have endangered the individual HCP and the community at large had it occurred with a dangerous pathogen. Through modifications of both protocol and equipment, the doffing protocol was successfully improved from initially incurring dangerous amounts of contamination to eliminating viable contaminants in all cases. These tests did not pose a great financial burden. Excluding PPE costs, each trial cost less than $45 USD, and our interventions were modest. Based on our results, validation of other healthcare PPE protocols by quantitative methods such as those we employed here is both logistically feasible and informative. No hospital procedure is designed for failure, but with empirical validation, those procedures can ensure they provide necessary protection.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2024.84
Data availability statement
All data generated for this manuscript are available in the manuscript or its supplementary material.
Acknowledgments
We thank the other members of the Levin Lab, particularly Joshua Manuel and Bruce Levin, for their advice and assistance. We also Jessica Carag and Darryl Grant for their help with revising this manuscript.
Author contributions
Conceptualization: CSK, JSM, BAB.
Methodology: JBV, JM, JT, JSM, KBB, BAB.
Investigation: JSM, APS, KBB, BAB.
Visualization: APS, KBB, BAB.
Funding Acquisition: CSK, JMM.
Project Administration: CSK, JBV.
Supervision: JSM, BAB.
Writing – Initial Draft: CSK, JT, JSM, APS, KBB, BAB.
Writing – Review & Editing: CSK, JMM, JBV, JM, JT, JSM, APS, KBB, BAB.
Financial support
This work was supported by funds from the US National Institute of General Medical Sciences [grant number R35GM136407 to Bruce R. Levin]; the US National Institute of Allergy and Infectious Diseases [grant number U19AI158080-02 to Bruce R. Levin]; and the Centers for Disease Control and Prevention [grant number NU38CK000481 to Joel M. Mumma]. The funding sources had no role in the design of this study and will not have any role during its execution, analysis, interpretation of the data, or drafting of this report. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Centers for Disease Control and Prevention.
Competing interests
The authors have no potential conflicts of interest to declare.
Open access
