• health-care workers;
  • infection control;
  • respirator


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results and Discussion
  6. Conclusion
  7. References

The instant measurements provided by the Portacount fit-test instrument have been used as the gold standard in predicting the protection of an N95 respirator in a laboratory environment. The conventional Portacount fit-test method, however, cannot deliver real-time measurements of face-seal leakage when the N95 respirator is in use in clinical settings. This research was divided into two stages. Stage 1 involved developing and validating a new quantitative fit-test method called the Personal Respiratory Sampling Test (PRST). In Stage 2, PRST was evaluated in use during nursing activities in clinical settings. Eighty-four participants were divided randomly into four groups and were tested while performing bedside nursing procedures. In Stage 1, a new PRST method was successfully devised and validated. Results of Stage 2 showed that the new PRST method could detect different concentrations and different particle sizes inside the respirator while the wearer performed different nursing activities. This new fit-test method, PRST, can detect face seal leakage of an N95 respirator being worn while the wearer performs clinical activities. Thus, PRST can help ensure that the N95 respirator actually fulfils its function of protecting health-care workers from airborne pathogens.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results and Discussion
  6. Conclusion
  7. References

In 1995, the U.S. Centers for Disease Control and Prevention (CDC)'s National Institute for Occupational Safety and Health enacted the Guidelines for the Prevention of Transmission of Mycobacterium Tuberculosis requiring health-care workers to use particulate air-purifying respirators. CDC's Guidelines, issued 1 December 2010, recommend that health-care workers conducting the highest exposure risk activities (i.e. aerosol-generating procedures) should wear only fit-tested N95 respirators to reduce the risk of infection.[1]

The World Health Organization (WHO) recommends use of the N95 respirator by any health-care worker performing high-risk aerosol-generating procedures (e.g. bronchoscopy, or any procedure involving aspiration of the respiratory tract).[2] The respirator can provide protection only when the face seal fits tightly. If the respirator does not provide a tight seal between the edge of a respirator and the wearer's face, some particles will enter through the gap.[3] Fit testing of respirators has been recommended, legislated and implemented in the USA and Canada.[4] In the USA, Occupational Safety and Health Administration (OSHA) regulations state that, for his or her own safety, no one should wear an N95 without being fit-tested.[5] In Hong Kong, a fit test is done when a health-care worker joins the workforce (i.e. hospital), when any registered user has facial contour changes or when a particular respirator model is discontinued by the manufacturer. However, there is no recommended standard fit-test method suggested by WHO or CDC to check the tightness of the seal before use in clinical settings.

Although no particular method is recommended for evaluating the tightness of the face seal, the method almost universally used involves calculation of a ‘fit factor’ which represents the performance of the respirator. The fit factor is calculated from measurements made while the wearer performs certain exercises in the laboratory environment, and is expressed as the ratio of the mean concentration of ambient particles outside the respirator to the concentration of particles inside the respirator.[6] A fit factor greater than or equal to 100 is taken as an indication that the respirator fits the wearer well.[7]

The calculation of fit factor depends on measuring the number of particulates in the ambient air. This is feasible in a controlled environment, such as a laboratory, but is not feasible in a ward setting.[8] But ironically a ward setting is where the measurement most needs to be made if the respirator is to effectively protect the wearer.

The number of air particles inside the respirator is affected by: (i) particles penetrating directly through filter fibre; and (ii) face-seal leakage. The quantitative fit-test measures total aerosol penetration, that is, leakage that has occurred through the filter medium and through the face-seal. No method clearly differentiates between the two pathways under actual breathing conditions.[9]

Particle concentration by size is a key predictor of the ability of a particulate exposure to cause adverse health effects.[10] Some adverse health effects, particularly those related to cardiovascular function, appear to be related to short-term changes in particle concentration by size.[11, 12] However, the conventional Portacount fit test does not measure particle concentration by size in real time in clinical settings; it only measures particle concentration inside and outside the respirator during fit-test exercise in the laboratory environment. Therefore, ensuring that a wearer's respirator fits properly throughout his/her work requires an instrument that can measure particle concentration in real time in a clinical setting.

In the conventional Portacount fit-test method, leakage is measured during a sampling period of 30 s; this figure is used to calculate a mean leakage for the period of the entire fit test. An instantaneous leak or relatively high leakage during a single exercise can be offset by longer periods of lower leakage in that single exercise, resulting in an overall passing fit factor in a single exercise.[13] Studies show that health-care workers are increasingly becoming exposed to, and contracting, Pulmonary tuberculosis (TB) in the workplace, particularly multiple-drug-resistant TB.[14]

The Portacount fit-test method has four drawbacks: (i) it cannot detect face-seal leakage in real time; (ii) it cannot identify aerosol particles by size; (iii) it can underestimate or overestimate protection of N95 respirator by comparing fit factors of individual exercise; and (iv) it cannot be used with certain clinical movements.

The purpose of this study was to develop a method to evaluate N95 respiratory protection in real time, in a clinical setting. The study is significant because it can identify particle size distribution in determining the leakage of the N95 respirator in a real-time manner. It is necessary to perform fit test while the wearer performs actual clinical nursing practices to get an accurate measure of how well an N95 respirator fits, which is how effectively it protects the wearer.

Therefore, two questions guided this study. In Stage 1: Was there any difference between conventional Portacount fit-test method and investigator-developed fit-test method Personal Respiratory Sampling Test (PRST) in measuring air particles? In Stage 2: How well did the investigator-developed fit-test method PRST evaluate the fitness of N95 respirator performance as compared with that of conventional Portacount fit-test method? The research hypothesis was that the investigator-developed fit-test method PRST had better predictive power than the conventional Portacount fit-test method in detecting face-seal N95 respirator leakage.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results and Discussion
  6. Conclusion
  7. References

This research was divided into two stages. Stage 1 involved developing and validating a new fit-test method called the PRST. In Stage 2, PRST was evaluated in use during nursing activities in clinical settings.

Stage 1

The experiment was set up in an enclosed environment with an average air speed of 0.1775 m/s, using a low velocity flow analyser type 54N50, Dantec Type 54R10 (Dantec Dynamics, Skovlunde, Denmark) with a low velocity transducer, at a temperature of 20 °C and relative humidity 77% (HoBo data logger). The aerosol was measured by two real-time portable aerosol spectrometers, namely the PAS 1.109 (Grimm Technologies, Inc., Dorfstrasse, Germany). The PAS 1.109 measures particulates in the size range of 0.25–32 μm in 31 size channels. Stainless steel tubes (4mmOD × 3mmID) provided by the manufacturer were used as the inlets for the PASs. A particle generator Model 8026 (TSI Incorporated, Shoreview, MN, USA) was used to spray a suspension of polydisperse into a box chamber placed in one end of the chamber, and it was then mixed with a fan. The reservoir jar of particle generator was filled with clean tap water with one salt tablet added. The output adjustment screw was set to maximize the aerosol flow. The flow rate inside the box was maintained at 0.17 m/s, and was measured by a flow analyser (TSI Incorporated). Particles were measured in the chamber in real time with two PAS 1.109 instruments. PAS 1.109 was set to report size distribution every minute. The instruments were used to measure the aerosol particles in the chamber three times: at 5 min, 15 min and 30 min, to estimate measurement precision.

Stage 2

A method for evaluating the protection afforded by a respirator, similar to Lee et al.[15] and named PRST, was set up (Fig. 1). The set-up consisted of five components: (i) sampling probe (Adaptor Kit 8025-N95, TSI Inc., St Paul, MN, USA); (ii) two Portable Aerosol Spectrometers 1.109 (Grimm Technologies, Inc.) for 31 size channels; (iii) 60 cm cylindrical plastic tubes; (iv) N95 respirators (Model 1860, 1860s, 1862, 3M); and (v) one backpack. Our equipment and test differed from Lee et al.'s study[15] in three aspects. First, the cylindrical plastic tube was directly attached (with adhesive) to leak locations on the respirator rather than to the helmet. When attached to the helmet, the plastic tube hindered performance of nursing activities. Second, we used different leak locations. In this study, potential leakage was measured on the right and left sides of the bridge of the nose and at the chin (Fig. 1), instead of at the nose, left cheek and chin. Myers et al.'s study[16] showed that 85% of leakage occurs at the nose, 20% leakage at the chin and 19% leakage at the cheek. Studies have showed that in using respirators, there is a higher chance of air leakage around the nose.[17] Therefore, in this study, both sides of the nose were monitored. Third, we tracked particles 0.3 μm in size because they are the most penetrating. Because of these three characteristics of our study, we believed that PRST would be able to evaluate the fitness of the N95 respirator more meaningfully than previous methods.


Figure 1. Schematic representation of the PRST. image, sites testing for leakage; image, portable Aerosol Spectrometer (PAS006, PAS007). PRST, Personal Respiratory Sampling Test.

Download figure to PowerPoint

Two Portable Aerosol Spectrometers 1.109 (Grimm Technologies, Inc.) were used for continuous measurement of the size distribution of various aerosol particles. These two spectrometers were put into the backpack. A sampling probe (Adaptor Kit 8025-N95, TSI Inc.) punctured the respirator and was secured with a push nut on from the other side of the respirator. One 60 cm-long cylindrical plastic tube was used to connect the exposed end of the sampling probe to the spectrometer for measuring size distribution of aerosol particles in the respirator, and one 60 cm-long cylindrical plastic tube was anchored to the outer surface of the respirator near the nose region to measure ambient concentration.

The pilot study provided a set of data for statistical analysis from which the sample size required for the main study could be estimated, that is, four in each group when the effect size was 0.86, with an achieved power of 62% and alpha was 0.05. Sample size for the main study was then obtained from the power table.[18] The minimum required sample size was 18 in each group when the effect size was 0.40 and power was 80%.

Participants were recruited through convenience sampling. We put a poster to recruit participants in a local university. Participants were students 18 years of age or older, and all were Year 1 nursing students. To minimize the confounding effects resulting from prior clinical experience and training in how to wear an N95 respirator, the exclusion criteria were that a participant had learned how to perform fit test and fit check (a self-check the wearer performance to detect air leaks of the respirator when donning a respirator every time) before. Their lack of experience was important because experience can have a significant impact on fit-test results.[19] Participants who were pregnant, or who had been diagnosed with respiratory problem or back injury, were excluded. The purposes and procedures of the study were explained to all participants. Written consent was obtained from each participant before the researcher recorded any personal information. The ethics was approved by the Human Subjects Ethics Sub-committee of the Hong Kong Polytechnic University.

Eighty-four eligible participants ranging in age from 18 to 21 years old were recruited for the study. They were divided randomly into four groups (Groups A to D), each group with 21 participants. Both Groups A and B participants received training on how to perform fit check. Whereas Group A participants performed the Portacount fit test, Group B participants did not.

Participants in Groups C and D did not receive training on how to perform fit check; therefore, they were not required to perform the fit-check demonstration. Whereas Group C participants were asked to perform the Portacount fit test, Group D participants did not. Researcher assessed all their performance in doing their fit checks by using a fit-check checklist developed by the researcher, based on the CDC guidelines. Participants in Groups A and C were told about the recommended type of N95 respirators after performing the fit test (Fig. 2).


Figure 2. Flowchart of the study procedure. PRST, Personal Respiratory Sampling Test.

Download figure to PowerPoint

In Session 1, participants were required to perform the quantitative fit test conducted with a Portacount Plus machine. In Session 2, participants performed fit testing by the PRST in performing bedside nursing procedures in a nursing laboratory. The laboratory test simulated the physical work that workers experience in clinical settings. The PRST was measured by two portable aerosol spectrometers at the same time; one (PAS1.109-007) measured the ambient air particle concentration outside the respirator, whereas the other (PAS1.109-006) measured the air particle concentration inside the respirator. The tests were conducted while the participants were wearing N95 respirators in doing bedside nursing procedures. Nursing procedures were suction, napkin changing (moving exercise) and normal breathing (non-moving exercise). The measurements were conducted continuously for 15 min, with particle concentration measurements averaged over 1 min.

Participants were required to close their mouths and breathe through their noses during the experiment to minimize water vapour generated from their mouth entering the sampling system. This prevented the mixing of particles that have leaked in around the face-seal of the facepiece with those that have been generated by the wearers. The plastic tubing was changed after each participant to prevent water content generated by the participant's exhalation from entering the instrument. Upon completion, participants, still wearing their N95 respirators, were asked to breathe normally.

Statistical analysis

Statistical Package for Social Sciences Version 17.0 (SPSS 17.0; IBM, New York, USA) was used for the statistical analysis of particle size distribution. The demographic distribution of all participants was examined, and all numeric data were presented as means and standard deviations (mean ± SD). One-way analysis of variance (ANOVA) was performed to compare the performance of the respirator in screening out particles of different sizes. The concentrations of particles of different sizes showed the performance of the respirator.

Results and Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results and Discussion
  6. Conclusion
  7. References

Stage 1: validating PRST

To evaluate the reliability of portable aerosol spectrometers in measuring ambient particle number concentration and particle size, an inter-instrument reliability test was performed in the laboratory. Two sample t-tests were used to evaluate and compare the two spectrometers. In the 5 min intervals, there was no significant difference in the particle number concentration measured in > 0.25–> 0.40 μm and > 0.58–> 3.0 μm size ranges. Thus, the particle size distributions measured by the two instruments were essentially identical in > 0.25–> 0.40 μm and > 0.58–> 3.0 μm size ranges.

Correlation analysis was used to examine the relationships between spectrometers and particle generator, and the relationship between the two spectrometers. Results show strong positive correlation between spectrometer 1 and spectrometer 2 (r = 0.714, P = 0.00) and P < 0.001. This indicates that the relationship between spectrometer 1 and spectrometer 2 was statistically significant (Table 1). The intraclass correlation coefficient (ICC) indicates reliability of a single trial of multiple instruments, and values greater than 0.8 are considered highly reliable. The one-way ICC between the two spectrometers was 0.83. This showed that the two spectrometers measured ambient particle concentration consistently.

Table 1. Correlations of Spectrometer 1 and particle generator; Spectrometer 2 and particle generator; Spectrometer 1 and Spectrometer 2 (N = 156)
VariableSpectrometer 1Spectrometer 2Particle generator
  1. ** P < 0.001.

Spectrometer 11.00  
Spectrometer 20.71**1.00 
Particle generator0.72**0.71**1.00

Stage 2: performance of PRST

The results of mean fit factor by group was plotted to present a graphical view of the performance of each group (Fig. 3). Group B (no fit test and trained fit check) had the highest mean fit factor among the four groups, and Group D (no fit test and untrained fit check) had the lowest mean fit factor. The fit factor of Groups A (Portacount fit test and trained fit check) and C (Portacount fit test and untrained fit check) were similar, whereas the fit factor of Groups B and D were in the middle.


Figure 3. Performing nursing procedures among four groups (values are mean and standard deviation). image, diff_NB; image, diff_suction; image, diff_napkin; image, diff_rNB.

Download figure to PowerPoint

High concentration of particles > 4.0 μm inside the respirator were found more frequently in Group D overall and more frequently for all groups during moving exercise, especially during heavy working condition (napkin changing). The findings corresponded to results reported by Grinshpun et al.[9] ANOVA was performed to compare the concentrations of different particle sizes inside the respirator while the wearer performed different bedside nursing procedures. A fit factor of less than 10 was found more frequently in participants performing moving exercise with heavy working condition (i.e. changing napkin). During the experiment, participants performed the nursing procedures one after the other, without manipulation of the respirator. Results showed that air leakage occurred during these procedures as well as during non-moving exercise (normal breathing) after heavy working condition (changing napkin). Even so, the fit factor was still very low, less than 10 as compared with other exercises. There were significance differences in particle counts while participants were doing different nursing procedures of suction of particle size > 4.0 μm between groups, P = 0.04, F-ratio 2.83. There were no significant differences in concentrations of particle sizes of 0.3 and 1 μm during other bedside nursing procedures. Results indicated that during moving exercise, more large particles (i.e. greater than penetration size, e.g. > 4.0 μm) leaked into the N95 respirator.

Kruskal–Wallis test was used to compare the conventional Portacount fit test and the PRST in identifying N95 respiratory protection in terms of fit-tested respirator model. There was no significant difference in the respiratory protection of N95 respirator between conventional Portacount fit test (H(3) = 2.87, P = 0.41) with a mean rank of 37.67 for Group A, 45.29 for Group B, 41.52 for Group C and 45.43 for Group D, and the PRST method in terms of fit-tested N95 respirator type among the four groups (H(3) = 5.31, P = 0.15), with a mean rank of 40.00 for Group A, 48.00 for Group B, 44.00 for Group C and 38.00 for Group D. These results indicate that the PRST method and conventional Portacount fit test are equally reliable to evaluate respiratory performance of N95 respirator.

No significant differences were found in demographic characteristics among the four groups (Table 2). Movement increases leakage; this assumption is in agreement with earlier results presented by Chen et al.[20, 21] which show that the fraction of particles leaking through the face-seal decreases when the constant flow rate through the respirator increases from 5 to 95 L/min. More exercise produced higher face-seal leakage-to-filter ratios than non-moving ones. The results of this study are consistent with those of Grinshpun et al.[9] Lee et al.[15] showed that particle size affects penetration through filter materials and face-seal leaks. Other factors like higher breathing rate might increase penetration and influence face-seal fit.[22]

Table 2. Demographic distribution (N = 84)
Demographic dataGroup A n = 21 n (%)Group B n = 21 n (%)Group C n = 21 n (%)Group D n = 21 n (%)Chi-squareP* (< 0.05)
  1. *Significant when P < 0.05.  Mean (SD).  ANOVA.

SexMale12 (57.14)14 (66.67)6 (28.57)9 (42.86)70.07
Female9 (42.86)7 (33.33)15 (71.43)12 (57.14)  
Age 19.33 (0.66)19.57 (1.08)19.1 (0.83)19.29 (0.96)10.39
Height 166.57 (8.17)168.95 (7.86)165.76 (6.18)165.81 (9.44)0.740.53
Weight 56.45 (9.52)60.17 (9.60)56.11 (8.10)54.05 (9.46)1.610.19
Place of birthHong Kong16 (76.19)13 (61.9)19 (90.5)17 (81.0)6.620.36
Mainland China5 (23.8)7 (33.3)2 (9.5)3 (14.3)  
Macau01 (1.2)01 (1.2)  
Programme studiedGeneral19 (90.1)19 (90.1)21 (100)19 (90.1)8.760.46
Mental2 (9.5)2 (9.5)02 (9.5)  

More exercise produced higher face-seal leakage to filter ratios than non-moving ones. This study confirms these results. Our study results suggest that the PRST can be used to measure respirator fitness in clinical settings. The PRST can detect particle concentration and particle size inside the respirator during moving and non-moving exercises. We found no significant difference between conventional Portacount fit-test method and PRST in assessing the fitness of N95 respirator. Results show that the conventional fit test and PRST are equally reliable in evaluating respirator fitness; however, the conventional fit test cannot be used on the job. This is crucial, because face-seal leakage is most likely to occur spontaneously and unexpectedly during work. Thus, all N95 respirator wearers should be trained in performing PRST so that they can use it whenever and as often as they need to.

Implications and limitations

In the current practice, health-care workers only pay attention to the overall fit factor of fit-test result. Although the overall fit factor is passed (i.e. greater than 100), the fit factor during a specific individual exercise might fail (i.e. fall below 100). When this happens, the respirator is leaking and the health-care worker is exposed to infectious agent risk during work even though he or she is wearing a respirator that passed the fit test. If the fit-test result indicates that the health-care worker has respirator leakage when performing napkin changing, we can recommend that he or she adjusts the N95 respirator before and after performing this procedure in a ward. That is, once a specific person knows the respirator leaks when he/she does a procedure, every time he/she does that same procedure, he/she should check and adjust the respirator. Therefore, through PRST, health-care workers can fully understand the strengths and weaknesses of the N95 respirator when they perform bedside nursing procedures. Through PRST, each health-care worker can acquire specific individual knowledge about when and where his/her respirator is likely to leak so that he/she can refit the respirator as appropriate during his/her clinical work.

A limitation of this study was that all the participants recruited in this study were between 18 and 20 years of age. Hence, the results might not be generalizable to health-care workers of other age groups, particularly different age groups with different facial contours. Participants were required to wear the spectrometer of 5 kg in weight, and to close their mouths and breathe through their noses during the experiment to minimize water vapour generated from their mouth entering the sampling system. This prevented the mixing of particles that have leaked in around the face-seal of the face-piece with those that have been generated by the wearers. These experiments under such requirements would not totally reflect the real working situation.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results and Discussion
  6. Conclusion
  7. References

Current methods for respirator fit testing cannot truly estimate the protection of N95 respirator in clinical settings while wearer/health-care workers are performing their duties. Active work can cause the respirator to shift, thus, wearers need some way to test and check whether their respirator fits properly, and whether it continues to protect them from airborne pathogens.

A self-developed method to evaluate protection provided by N95 respirator, named PRST, was devised and tested in this study. This new fit-test method can detect face-seal leakage. In particular, it can identify different particles of sizes ranging from 0.3 to 4 μm inside the respirator and ambient environment in real time every 1 min. If implemented, the PRST can help wearers evaluate the actual performance of their respirators in clinical settings.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results and Discussion
  6. Conclusion
  7. References
  • 1
    Centers for Disease Control and Prevention. Respirator trusted-source information page. 2010, December 1. Available from URL: Accessed 29 December 2011.
  • 2
    WHO. Clinical management of human infection with pandemic (H1N1) 2009: Revised guidelines. November 2009. Available from URL: Accessed 15 March 2011.
  • 3
    Borwegen B. Airborne infections and respirators: How prepared is your workplace? The American Journal of Nursing 2006; 106: 3334.
  • 4
    Clayton M, Vaughan N. Fit for purpose? The role of fit testing in respiratory protection. The Annals of Occupational Hygiene 2005; 49: 545548.
  • 5
    Rebmann T. Protect yourself, protect your patients: The essentials of PPE. Nursing made incredibly easy 2007; Jan/Feb: 3039.
  • 6
    Clifton CD, Murphy RW, Van Ert MD. A comparison of controlled negative and aerosol quantitative respirator fit test systems by using human subjects. American Industrial Hygiene Association Journal 1993; 54: 1015.
  • 7
    Dosman J, Senthilselvan A, Kirychuk S et al. Positive human health effects of wearing a respirator in a Swine Barn. Chest 2000; 118: 852860.
  • 8
    Coffery CC, Lawrence R, Zhuang A, Campbell DL, Jensen PA, Myers WR. Comparison of five methods for fit-testing N95 filtering-facepiece respirators. Applied Occupational and Environmental Hygiene 2002; 17: 723730.
  • 9
    Grinshpun SA, Haruta H, Eninger RM, Reponen T, McKay RT, Lee SA. Performance of an N95 filtering facepiece particulate respirator and a surgical mask during human breathing: Two pathways for particle penetration. Journal of Occupational and Environmental Hygiene 2009; 6: 593603.
  • 10
    Schlesinger RB, Kunzli NG, Hidy M, Gotschi T, Jerrett M. The health coherence of toxicological and epidemiological inferences. Inhalation Toxicology 2006; 18: 95125.
  • 11
    Peters AD, Dockery W, Muller JE, Murray MD, Mittleman MD. Increased particulate air pollution and the triggering of myocardial infarction. Circulation 2001; 103: 28102815.
  • 12
    Wellenius GA, Schwartz J, Mittleman MA. Air pollution and hospital admissions for ischemic and hemorrhagic stroke among medicare beneficiaries. Stroke; a Journal of Cerebral Circulation 2005; 36: 25492553.
  • 13
    Janssen LL, Luinenburg MD, Mullins HE, Nelson TJ. Comparison of three commercially available fit-test methods. AIHA Journal: A Journal for the Science of Occupational and Environmental Health and Safety 2002; 63: 762767.
  • 14
    Radonovich LJ, Hodgson MJ, Cohen HJ. Do respirators protect health-care workers from airborne infectious diseases? Respiratory Care 2008; 53: 16601663.
  • 15
    Lee SA, Grinshpun SA, Adhikari A et al. Laboratory and field evaluation of a new personal sampling system for assessing the protection provided by the N95 filtering facepiece respirators against particles. The Annals of Occupational Hygiene 2005; 49: 245257.
  • 16
    Myers WR, Allender JR, Iskander W, Stanley C. Causes of in-facepiece sampling bias—1. Half-facepiece respirators. The Annals of Occupational Hygiene 1988; 32: 345359.
  • 17
    Holton PM, Tackett DL, Willeke K. Particle size-dependent leakage and losses of aerosols in respirators. American Industrial Hygiene Association Journal 1987; 48: 848854.
  • 18
    Cohen J. Statistical Power Analysis for the Behavioral Sciences, 2nd edn. Hillsdale, NJ, USA: Lawrence Erlbaum Associates, 1988; 315.
  • 19
    Johnston AR, Myers WR, Colton CE, Birkner JS, Campbell CE. Review of respirator performance testing in the workplace issues and concerns. Industrial Hygiene Association Journal 1992; 53: 705712.
  • 20
    Chen CC, Ruuskanen J, Pilacinski W, Willeke K. Filter and leak penetration characteristics of a dust and mist filtering facepiece. American Industrial Hygiene Association Journal 1990; 51: 632639.
  • 21
    Chen CC, Willeke K. Characteristics of face seal leakage in filtering facepieces. American Industrial Hygiene Association Journal 1992; 53: 533539.
  • 22
    Reponen T, Lee SA, Grinshpun SA, Johnson E, McKay R. Effect of fit testing on the protection offered by N95 filtering facepiece respirators against fine particles in a laboratory setting. The Annals of Occupational Hygiene 2011; 55: 264271.