See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/41668533

Particulate face masks for protection against airborne pathogens - one size
does not fit all: an observational study

Article  in  Critical care and resuscitation: journal of the Australasian Academy of Critical Care Medicine · March 2010
Source: PubMed

CITATIONS READS

9 990

4 authors, including:

Jane H Thomas Joshua S Davis

Royal Darwin Hospital Menzies School of Health Research
13 PUBLICATIONS   273 CITATIONS    121 PUBLICATIONS   3,284 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Improving Aboriginal Kidney Transplantation: "I-AKT" View project

Antimicrobial Stewardship in Primary Care View project

All content following this page was uploaded by Joshua S Davis on 17 January 2014.

The user has requested enhancement of the downloaded file.

 OR I G I N A L A R T I C L E S

Particulate face masks for protection against airborne

pathogens — one size does not fit all: an observational study
Susan Winter, Jane H Thomas,
Dianne P Stephens and Joshua S Davis

In the past decade, the world has seen epidemics of ABSTRACT

severe acute respiratory distress syndrome,1 multidrug-
Crit Caretuberculosis,
resistant Resusc ISSN: 21441-2772
highly 1pathogenic
March avian Objective: To determine the proportion of hospital staff
2010 12
3 1 24-27
influenza and pandemic influenza H1N1 2009.4 Trans- who pass fit tests with each of three commonly used
©Crit Care Resusc 2010
mission of these diseases from patients to health care
www.jficm.anzca.edu.au/aaccm/journal/publi- particulate face masks, and factors influencing preference
cations.htm
workers has been described,5 in some cases resulting in and fit test results.
Original articles
death.6 To avoid such transmission, the use of personal Design: Observational study.
protective equipment (PPE) is recommended,7 the most Setting and participants: 50 healthy hospital staff
important aspect of which is a P2 respirator (known in volunteers in an 18-bed general intensive care unit in
the United States as an N95 respirator). This is often an Australian teaching hospital.
referred to in a clinical context as a particulate face mask. Interventions: Participants were administered a
These masks filter out airborne infectious particles. questionnaire about mask use and their preferred mask
Guidelines suggest their use for health care workers and underwent qualitative fit-testing with each of three
caring for patients with pandemic influenza8 or sus- different particulate masks: Kimberly–Clark Tecnol
pected or confirmed tuberculosis.9 FluidShield N95 particulate filter respirator (KC), 3M Flat
Although a number of previous studies have evaluated Fold 9320 particulate respirator and 3M 8822 particulate
characteristics and fit testing of P2 masks,10-12 they have not respirator with exhalation valve. Participants who failed fit-
included the types of mask commonly stockpiled in Aus- testing were trained in correct mask donning, and fit-
tralia.13 Australian guidelines recommend that “P2 (N95) testing was repeated.
respirators should be used within the context of a respir- Main outcome measures: Proportion of participants who
atory protection program that includes fit testing, fit check- passed the fit test for each mask and the effect of training.
ing and training”.14 Qualitative fit tests rate masks as either Results: The proportion of participants who passed a fit
pass or fail, based on the wearer detecting a test substance test was low for all three masks tested (KC, 16%; flat fold,
through taste or smell. These tests are generally easy and 28%; and valved, 34%). Rates improved after training: the
relatively cheap to perform and do not require specialised first mask tested fitted in 18% of participants pre-training
personnel or equipment, but give only a subjective measure and 40% post-training (P = 0.02). None of the masks fitted
of the quality of seal of the face mask.15 In contrast, for 28% of participants. There were no significant
quantitative tests are costly, but provide an objective meas- predictors of fit-test results.
ure of the quality of the mask’s seal. There is some evidence
Conclusions: A large proportion of individuals failed a fit
that the results correlate well between qualitative and
test with any given mask, and we were not able to identify
quantitative tests.16
any factors that predicted mask fit in individuals. Training on
We evaluated three P2 masks that are commonly avail-
mask use improved the rates of adequate fit. Hospitals
able in Australia and were available at our hospital. For each
should carry a range of P2 masks, and should conduct
of these masks, we aimed to determine the proportion of
systematic P2 mask training and fit-testing programs for all
health care workers who achieved a good fit with the mask,
staff potentially exposed to airborne pathogens.
factors that affected fitting, the effect of training, and
acceptability of the mask. Crit Care Resusc 2010; 12: 24–27

Methods hospital. Exclusion criteria were the presence of facial hair

The study was approved by our institutional human or pregnancy.
research ethics committee. Participants were volunteers After providing written informed consent, participants
drawn from staff involved in clinical care of patients in the filled in a questionnaire and underwent fit testing with each
18-bed general intensive care unit of an Australian teaching of the three masks, in random order, without training.

24 Critical Care and Resuscitation • Volume 12 Number 1 • March 2010

 OR I G I N A L A R T I C L E S

Those who failed a fit test with any particular mask received manoeuvres (such as talking and walking) over at least 3
training on correct fitting, and the fit test was repeated. minutes. A test was deemed a pass if no taste was detected
Training was conducted using posters and DVDs provided by the end of the above protocol (mask effective), and a fail if
by the masks’ manufacturers. the substance was tasted at any time (mask ineffective).
The three masks evaluated were the Kimberly–Clark Continuous variables were compared using the Student t
Tecnol FluidShield N95 particulate filter respirator (KC; test, and proportions were compared using χ2 tests. A
Kimberly–Clark, Sydney, NSW), the 3M Flat Fold 9320 P < 0.05 was considered significant. All statistical analyses
particulate respirator (3M flat fold; 3M, Sydney, NSW), and were performed using Intercooled Stata version 10 (Stata-
the 3M 8822 particulate respirator with exhalation valve corp, Tex, USA).
(3M valved; 3M, Sydney, NSW).
Fit testing was performed using 3M Qualitative Taste Fit
Testing Kits F10 and F30 (3M, Sydney, NSW), according to the Results
manufacturer’s instructions. Participants underwent a sensi- Fifty volunteers participated in the study (Table 1). Of these,
tivity test while wearing a test hood but no mask, to 46 (92%) had previously used the KC mask; one of these
determine their ability to taste saccharin. If saccharin could had also used the 3M flat fold, and another of these, the
not be tasted, then denatonium benzoate was used instead. 3M valved mask. Seven (14%) had been previously trained
Participants then rinsed their mouths with water, waited 10 in the use of any P2 mask.
minutes, and donned the test hood over a mask. A test
solution of saccharin (830 mg/mL in water) or denatonium Preferred masks
benzoate (1.688 mg/mL in isotonic saline) was aerosolised The 3M valved was the most preferred mask before testing,
into the hood. The participants performed a series of dynamic and the 3M flat fold afterwards. The preference for the KC
dropped significantly after testing (Table 2).

Fit testing
Table 1. Occupation and age of participants The 50 participants underwent a total of 150 fit tests, one
⭐ 29 30–39 40–49 ⭓ 50 test each for each of the three types of mask. Considering all
years years years years Total fit tests together (pre- and post-training), 14 participants
Doctor 4 3 3 3 10 (28%) found that none of the masks fitted. Of the 36 who
Nurse 15 5 3 2 25 passed a fit test with any mask, 18 passed with only one type
Medical student 4 1 0 1 6 of mask, eight with two types, and 10 with all three types.
Other* 1 5 2 1 9 All three masks had low rates of fitting, with the KC
having the lowest at 16% (Table 3). Training significantly
* Included patient care assistants, physiotherapists, pharmacists and
laboratory staff. improved the rate of successful fit tests for both the 3M
masks but not the KC mask (Table 3).

Table 2. Preferred masks before and after testing and reasons for stated preference
Before testing (n = 50) After testing (n = 50)
KC 3M flat fold 3M valved KC 3M flat fold 3M valved
No. who nominated mask as preferred option 21 (42%) 6 (12%) 22 (44%) 5 (10%) 24 (48%) 20 (40%)
Reasons for preference*
Familiarity 21 0 3 2 2 0
Shape 1 4 6 1 3 7
Best predicted fit 0 5 13 1 9 10
Brand 0 0 0 0 0 1
†
Extra features 0 0 7 0 0 2
Good fit – – – 5 13 15
Comfortable to wear – – – 3 10 9
KC = Kimberly–Clark Tecnol FluidShield N95 particulate filter respirator. 3M flat fold = 3M Flat Fold 9320 particulate respirator. 3M valved = 3M 8822
particulate respirator with exhalation valve.
* Reasons were not mutually exclusive: participants could tick as many features as they considered relevant. † All cited the exhalation valve.

Critical Care and Resuscitation • Volume 12 Number 1 • March 2010 25

 OR I G I N A L A R T I C L E S

The considerable failure rate with all three masks sug-

Table 3. Number of participants (%) who passed a gests that hospitals should stock more than three types of
fit test with each mask, before and after training P2 mask. Although our data did not show this, it is likely
that a significant minority of individuals will not be fitted by
Total (pre− +
Pre-training post-training) commonly used masks because of facial characteristics,
(n = 50) (n = 50) P such as face size and shape.12,17 We could find no factors
KC 8 (16%) 14 (28%) 0.15 that would allow individuals to select an appropriate mask
3M flat fold 14 (28%) 24 (48%) 0.04 without undergoing fit testing. It was particularly worrying
3M valved 17 (34%) 27 (54%) 0.04 that the KC mask, which is the type generally used in our
First mask tested 9 (18%) 20 (40%) 0.02
hospital, fitted the fewest participants, with or without
training. Many participants initially chose this mask as their
KC = Kimberly–Clark Tecnol FluidShield N95 particulate filter respirator.
3M flat fold = 3M Flat Fold 9320 particulate respirator.
preferred one, citing familiarity, but changed their minds
3M valved = 3M 8822 particulate respirator with exhalation valve. when a fit was not achieved.
This study has several limitations. Qualitative fit testing is
a participative process, relying on an individual’s percep-
To evaluate the effect of training overall and to eliminate tion of taste. As each individual was aware which mask
the effect of learning contamination and mask type, we was being tested, it is conceivable that their attitude
considered only the first mask fitted for each participant. towards the mask may have influenced their taste thresh-
The first mask fitted was equally distributed among the old and thus the result of the test. However, qualitative
three mask types. The proportion of participants who testing is widely available and more practical for small to
achieved a fit with the first mask increased significantly medium-sized hospitals than quantitative testing, which
from 9/50 (18%) before training to 20/50 (40%) after requires a large initial financial outlay. Thus, the way in
training (P = 0.02). which we performed our study provides good generalisa-
bility to many Australian health care settings. Testing the
Predictors of fit test results three masks sequentially may have contaminated the
On univariate logistic regression analysis, none of the results of later tests because of a learning effect. However,
variables tested — sex, age, head circumference, occupa- this is unlikely to be the case, as we endeavoured to
tion and order of mask testing — were significant predic- exclude this bias by testing the masks in random order in
tors of failure to fit any of the three masks. Moreover, there each individual.
was no difference in the mean head circumference (SD) of In conclusion, we found that a large proportion of health
those who did not achieve a fit with any mask (56.3 [2.4] care workers failed a fit test with any given mask, but there
cm) compared with those who achieved a fit with at least were no factors that predicted mask fit in individuals.
one mask (56.6 [1.7] cm) (P = 0.60). Training on using a mask improved the rates of adequate
fit. These findings strongly support the recommendation
that all acute hospitals should undertake P2 mask fit-testing
Discussion and training programs for all clinical staff. Given that,
We found unexpectedly high failure rates for three common despite a choice of three masks, a fit was not achieved in
P2 masks. We also found that a large proportion (28%) of all individuals, a variety of masks (at least five types) should
health care workers were unable to pass a fit test with any be made available to individuals who work in hospitals
of the three masks, even after training. It is concerning that where they may be exposed to airborne transmission of
almost a third of our sample, all of whom work in a clinical pathogens.
setting and are required to wear these masks, could not fit
any of the three masks easily accessible in Australia. This, Author details
along with the fact that there were no significant predictors
Susan Winter, Staff Intensivist1
of fit-test failure, reinforces the need for hospitals to
Jane H Thomas, Research Nurse1
conduct systematic fit-testing programs for all staff who Dianne P Stephens, Director of Intensive Care1
may be exposed to airborne pathogens. Furthermore, while Joshua S Davis, Infectious Disease Consultant2,3
training had a less marked effect than one might predict, it 1 Intensive Care Department, Royal Darwin Hospital, Darwin, NT.
did significantly improve the chance of a successful fit test 2 Department of Medicine, Darwin Hospital, Darwin, NT.
with a given mask. Thus, training of relevant staff in the use 3 Menzies School of Health Research and Charles Darwin University,
of P2 masks should be a routine part of hospital infection Darwin, NT.
control and prevention activities. Correspondence: [email protected]

26 Critical Care and Resuscitation • Volume 12 Number 1 • March 2010

 OR I G I N A L A R T I C L E S

References 10 Lee SA, Grinshpin SA, Reponen T. Mask performance offered by P2

masks and surgical mask human subject evaluation with NaCl
1 Smith RD. Responding to global infectious disease outbreaks: aerosol representing bacterial and viral particle size range. Ann
lessons from SARS on the role of risk perception, communication Occup Hyg 2008; 52: 177-85.
and management. Soc Sci Med 2006; 63: 3113-23. 11 Weber A, Witteke K, Marchoni R, et al. Aerosol penetration and
2 Loudon RG, Spohn SK. Cough frequency and infectivity in patients leakage characteristic of masks used in the health care industry. Am
with pulmonary tuberculosis. Am Rev Respir Dis 1969; 99: 109-11. J Infect Control 1993; 21: 167-73.
3 Stephenson I; Zambon M. The epidemiology of influenza. Occup 12 Lawrence RB, Duling MG, Calvert CA, Coffey CC. Comparison of
Med (Lond) 2002; 52: 241-7. performance of three different types of respiratory protection
4 Zimmer SM, Burke DS. Historical perspective — emergence of devices. J Occup Environ Hyg 2006; 3: 465-74.
influenza A (H1N1) viruses. N Engl J Med 2009; 361: 279-85. Epub 13 Coffey CC, Lawrence RB, Campbell DL, et al. Fitting characteristics
2009 Jun 29. of eighteen P2 filtering-face piece masks. J Occup Environ Hyg
5 Apisarnthanarak A, Puthavathana P, Kitphati R, et al. Outbreaks of 2004; 1: 262-71.
influenza A among nonvaccinated healthcare workers: implications 14 National Influenza Pandemic Action Committee. Interim infection
for resource-limited settings. Infect Control Hosp Epidemiol 2008; control guidelines for pandemic influenza in healthcare and com-
29: 77-80. munity settings. Annex to Australian health management plan for
6 Cooper BS, Fang LQ, Zhou JP, et al. Transmission of SARS in three pandemic influenza, June 2006. http://www.flupandemic.gov.au/
Chinese hospitals. Trop Med Int Health 2009; 14 Suppl 1: 71-8. internet/panflu/publishing.nsf/Content/7587228D3932D8D1CA
7 Centers for Disease Control and Prevention [US]. Update on influ- 2573D4001EB02F/$File/pandemic-infec-gl.pdf (accessed Jan 2010).
enza A (H5N1) and SARS L interim recommendations for enhanced 15 Marsh JL. Evaluation of saccharin qualitative fitting test for masks.
US surveillance, testing and infection controls. Atlanta: CDC, 2004. Am Ind Hyg Assoc J 1984; 45: 371-6.
8 Australian Government Department of Health and Ageing. Austral- 16 American National Standards Institute. American national standard
ian health management plan for pandemic influenza, May 2006. for respirator fit testing methods. ANSI Z88.10. New York: ANSI,
http://www.health.gov.au/internet/panflu/publishing.nsf/Content/ 2001.
ahmppi-1 (accessed Jan 2010). 17 Jensen PA, Lambert LA, Piademarco M, Ridzon R. Guidelines for
9 Centers for Disease Control and Prevention. Guidelines for prevent- preventing the transmission of Mycobacterium tuberculosis in
ing the transmission of Mycobacterium tuberculosis in health-care health care settings. Atlanta: Division of Tuberculosis Elimination
settings, 2005. MMWR Morb Mortal Wkly Rep 2005; 54 (RR17): 1. National Centre for HIV, STD and TB Prevention, 2005. ❏

Critical Care and Resuscitation • Volume 12 Number 1 • March 2010 27

View publication stats