Abstract
N95 respirators are the core equipment used by healthcare workers to prevent the spread of respiratory diseases. The protective effect of N95 against infection spread depends on the fit of the N95 to the wearer, which is related to the wearer’s facial dimensions. The purpose of this cross-sectional study was to assess the relationship between the fit of three types of N95 and facial dimensions. A total of 305 healthcare workers from ten hospitals in Beijing were recruited for this study. Facial dimensions of workers were measured using Intel RealSense Depth Camera D435. Fit testing was conducted on three types of N95 using the TSI-8038 Porta Count Pro + Respirator Fit Tester. Possible associations between the fit test results and facial dimension data were examined. A Porta Count reading of 100 was used as the criterion for an acceptable fit. The fit of the folding respirators was positively correlated with nose length (r = 0.13, p = 0.02), nose height (r = 0.14, p = 0.02), and face width (r = 0.12, p = 0.03), whereas that of flat respirators was correlated with nose width (r = 0.16, p < 0.01), chin length (r = 0.18, p < 0.01), and pro-face width (r = 0.13, p = 0.02), and that of arched respirators was correlated with the nose length (r = 0.13, p = 0.03). The fit of N95 for wearers depends on their facial features. The results of this study can provide advice for medical workers to choose the appropriate N95. Medical staff should fully consider their facial dimensions when choosing an appropriate N95 to improve the protective efficacy of respirators and to reduce the risk of infection by respiratory diseases.
Introduction
In recent years, there have been several outbreaks of infectious respiratory diseases, such as severe acute respiratory syndrome (SARS), the H1N1 influenza pandemic, avian influenza A, and COVID-19. These outbreaks have highlighted the importance of respiratory protection in the healthcare setting.
N95 respirators have been recommended as important protective equipment by the Centers for Disease Control and Prevention (CDC) and The World Health Organization (WHO) for healthcare workers in hospitals to prevent the transmission of airborne infectious diseases (Beneviat et al. Citation2023). They are also a type of respiratory protection recommended by the National Institute for Occupational Safety and Health (NIOSH) in the United States for personal protection against respiratory infectious diseases and are widely used in the prevention and control of airborne diseases (Derrick et al. Citation2005). However, the protective efficacy of N95 in preventing the spread of airborne infectious diseases is related to the filtering efficiency of respirator materials and the fit to the wearer (Rebmann et al. Citation2013). The facial suitability of choosing an N95 respirator has not been given a great amount of attention in the healthcare industry (Lam et al. Citation2011). Studies have shown that properly wearing an N95 does not guarantee absolute protection, and a risk of leakage exists (CDC Citation1998). There was an average ambient aerosol penetration of 33% in poorly fitting respirators and 4% in well-fitting respirators. Gaps between the respirator and the wearer’s face can lead to aerosol penetration, which is termed leakage. Airborne contaminants may leak into the respirator, resulting in ineffective or compromised protection. Respirators, including N95s, rely on structural features attached to the face to isolate ambient air from clean air inside the respirator. Therefore, adaptability between the respirator and the user’s face is important. Studies have found that facial dimensions have different effects on the facial fitness of various types of N95s. Face length, face width, and nose height may affect the N95’s fit (Niu et al. Citation2023).
Quantitative fit testing (QNFT) is recognized by the United States Occupational Safety and Health Administration (OSHA) as the gold standard for determining fit and is widely used to determine whether a respirator’s fit is appropriate (Prince et al. Citation2021). The QNFT measures the amount of leakage into the respirator to assess the adequacy of the fit. Factors affecting the fit of N95 include respirator material, respirator type, and dimensions of the wearer’s face. Some researchers have studied the relationship between facial dimension parameters and respirator suitability and found that those who wore arched N95 respirators and passed the fitness test had wider faces, wider jaws, and longer faces than those who did not (Cheng et al. Citation2012). In some studies, the methods used to determine human facial dimensions were manual measurements using a ruler, tape measure, and other tools (Trehan et al. Citation2021). This manual measurement method required the participation of trained professionals and direct contact with participants, which made the measurement process cumbersome. These problems can be avoided by using computer-vision technology to measure the features of human faces. Currently, error tolerances in 3D face scanning are within 1 mm, and the facial features obtained by automatic calculation are comparable to those obtained by manual measurements. Therefore, the use of computer-vision technology to obtain 3D facial measurements provides the advantage of being more accurate and more stable in the measurement of 3D facial features (Sreenath et al. Citation2001; Yang et al. Citation2007; Fakherpour et al. Citation2021).
This study aimed to investigate the relationship between the fit factor (FF) of wearers wearing three different types of N95s and facial dimension data to guide medical workers in choosing the appropriate N95 respirator to reduce the risk of respiratory infections (Derrick and Gomersall Citation2005). Moreover, this study uses computer-vision technology to measure participants’ facial dimensions.
Methods
Design and participants
This study used a descriptive, cross-sectional, prospective design. A total of 305 participants were randomly selected from eligible medical staff. Inclusion criteria included individuals who were in good health and able to wear N95 for more than 4 h. Those who could not tolerate wearing a mask for more than 4 h were excluded. The subjects were selected from 10 hospitals of different levels and professional types in Beijing and included 143 males and 162 females. Participants were invited to participate in the experiment through the staff of the Hospital Infection Management Office, and the subjects who participated in the experiment received a monetary honorarium of about 200 yuan (approximately $28).
Setting
All tests in this project were conducted in an air-conditioned room with a humidity of approximately 75% and temperature set at 23 °C to reduce the variation of suspended particle and dust concentrations in the environment. Subjects from each hospital completed respirator fit tests and face scans in eligible rooms in their hospitals. Fit tests and face scans were done at different times in the same room. The room size was required to be between 20 and 30 square meters and the height between 3 meters and 3.2 meters. One person was tested at a time.
Data collection
A standardized protocol and guidelines for performing fit tests and collecting facial dimension information were initially presented to the investigators and participants. QNFT, which is a reliable method for fit testing, and recommended by the OSHA regulation 29 CFR 1910.134 Respiratory Protection Appendix A, was used to evaluate whether a respirator fits a wearer. The QNFT results with three respirators, facial dimension data, including face length and face width, and demographic data, including age and sex, of the participants were recorded on a data sheet. The data entry process adopted double entry to ensure the authenticity, integrity, and reliability of the data.
Measuring facial dimensions
An Intel RealSense D435 depth camera (Intel Corporation, Santa Clara, California, USA.) was used to collect red, green, blue, and depth (RGB-D) images, and face geometry data were computed from the RGB-D images using computer-vision technology. Specifically, three RGB-D images were used to reconstruct the 3D face model of a human subject, and face geometry data were computed based on the 3D face model. In this study, facial geometry data included face height, face width, nose height, mouth width, and other facial parameters, which affected the fit of the N95. lists the definitions of the face parameters.
Table 1. The definition of face parameters.
Fit testing
A QNFT device was used to measure the fit of the N95 respirator to the wearer. The TSI-8038 Porta Count Pro+ (TSI, Shoreview, MN, USA) respirator fit tester system was used in this study. This device uses a miniature continuous-flow condensation nucleus counter to count particles in the air from 0.02 to 1.00 mm in diameter (Han et al. Citation2022). Participants wore three different types of N95 (3M9132: folding, NTPN95N: flat, WN-N95: arched) (Images and characteristics of each respirator are presented in Supplementary Figure S1) and performed six exercises in sequence for approximately 20 min as guided by the device. These six exercises imitated the frequency of required movements of the head and face of medical staff during the diagnosis and treatment of patients ( shows the sequence of exercises performed during fit testing). The TSI-8038 Porta Count Pro + respirator fit tester system provided an individual FF (range: 0–200) for each exercise, and the overall FF, which ranged from 0 to 200, was provided after all actions were completed. An overall FF ≥ 100 indicates a “pass” rating, which means the respirator is correctly fitted to the wearer and is operating as intended, otherwise, a “fail” rating is indicated, implying that the respirator fit was not acceptable. A daily check was performed to ensure that the machine and system performed as per the standard procedure before testing.
Table 2. Sequence of exercises performed in fit testing.
Ethical considerations
Ethical approval was obtained from the Beijing Municipal Health Commission and Ethics Committee of Beijing You’an Hospital, Capital Medical University (Commission’s approval number: LL-2021-135-K). All participants signed an informed consent form before participating in the study. The informed consent form included the project introduction, purpose of the research, benefits, and possible damages to the participants, right to confidentiality, research process, approval signatures, right to withdraw, and consent statement. A trained researcher distributed consent forms to all the participants and provided a detailed explanation.
Data analysis
All data analyses were conducted using SAS software (version 9.1; SAS Inc., Cary, NC, USA). An FF of 100 was used as the boundary value to determine whether a respirator “passed” or “failed” the fit test according to the OSHA QNFT protocol. An independent sample t-test was used to examine if there was any significant difference in face dimensions between the “pass” and “fail” groups. Spearman’s rank correlation analysis was used to further study the correlation between the facial dimensions and the FF value of each respirator. The Spearman’s rank correlation analysis method has the advantage that it does not require the distribution of original variables, is suitable for data that do not conform to a bivariate normal distribution, and is unsuitable for product-difference correlation analyses. The level of significance was set at p < 0.05.
Results
Of the 305 participants, 143 were male and 162 were female. All participants were from Beijing and worked in 10 different hospitals. The average age was 37.32 ± 10.16 for males and 36.61 ± 10.74 for females; average height was 174.21 ± 6.13 cm for males and 161.72 ± 4.68 cm for females, and average weight was 74.86 ± 12.06 kg for males and 59.29 ± 10.57 kg for females. The mean BMI was 24.63 ± 3.42 for males and 22.65 ± 3.85 for females (see ). The occupational distribution between participants was 41 physicians, 85 nurses, and 179 logistics and management workers.
Table 3. Basic characteristics of participants.
The respirators selected for the study were unisex and sex did not significantly affect the pass rate of the tests. There were significant differences in some facial dimensions between the “pass” and “fail” groups when the participants wearing a specific respirator were tested. The participants wore folding respirators for the quantitative fitness test, and a t-test was performed to compare the facial dimension parameters of the “pass” and “fail” groups. The results revealed that the average nose length (p < 0.01) and nose height (p = 0.05) of the “pass” group were greater than those of the “fail” group, and the difference was statistically significant. Whereas t-test results showed that the chin-length (p < 0.01) and nose width (p = 0.04) of those who passed the test were greater than those who failed for the flat respirator, and the differences were statistically significant. For arched respirators, t-test results showed that the average values of several facial parameters in the passing group were greater than those in the failing group; however, there were no statistically significant differences in all indexes involved in the comparison. shows a summary of the t-test results.
Table 4. Summary of comparison results of facial dimensions between the "pass" and "fail" groups (independent sample t-test).
Although the facial dimensions were normally distributed, the FF values of the three types of respirators did not follow a normal distribution. Therefore, Spearman’s rank correlation analysis was used for correlation analysis of these two groups of data. Spearman’s rank correlation analysis of FF values and nose length and height showed that nose length (p = 0.02) and nose height (p = 0.02) were positively correlated with the fitness factors for the folding respirator. Although the t-test revealed no statistical difference in face width between the “pass” and “fail” groups, Spearman’s rank correlation analysis showed a significant positive correlation between face width and the fit of folding respirators (p = 0.03). The Spearman’s rank correlation analysis also revealed that chin length (p < 0.01), nose width (p < 0.01), and profile face length (p = 0.02) were significantly positively correlated with the FF value of the flat respirator. While t-test result shows although the average profile face length of the “pass” group was higher than that of the “fail” group, there was no statistically significant difference. The results of Spearman’s rank correlation analysis indicated nose length was positively correlated with the FF value of the arch-type respirator (p = 0.03). shows a summary of Spearman’s rank correlation analysis.
Table 5. Summary of Spearman’s rank correlation coefficients of FF values and face dimensions.
Discussion
The results indicate that the fit of a respirator to the wearer’s face is related to the dimensional parameters of the face and that this effect varies according to the design type of the respirator (folding, flat, or arched). The fit of different design types of respirators is affected differently by the different facial dimensions. In this study, two different analysis methods were used to analyze the relationship between facial dimension and respirator fit, and the analysis results were not completely consistent. This may have to do with the fact that there is a certain internal correlation between the facial dimensions themselves.
The fit of the folding respirator is related to the nose length, nose height, and face width of the wearer. This may be because the taller and longer nose fits the nose clip of the folding respirator better and supports the respirator so that it does not collapse easily, while the wider face allows the respirator to fit more closely to the sides of the face. When designing and manufacturing this type of respirator, an improvement in the plasticity of the nose clip should be considered, as well as the manufacture of several models, of various sizes to adapt to people with small noses or small faces.
Chin length, nose width, and profile face length positively affect the fit of the flat respirators. The flat respirator does not have a hard nose clip and is worn and shaped by adjusting the elastic band on the upper and lower edges of the respirator. When tightening the elastic, a wider nose and longer chin make the respirator fit better on the face, and a longer side face increases the overall tension of the respirator's strap and makes the respirator fit better on the face. People with narrow noses and short chins are advised to prioritize other types of respirators.
Arched respirators had the least influence, and only nose length was found to be a possible influence. The winning arched respirator selected in this study has a relatively large cup shape and a circle of elastic material inside the edge of the respirator, which may be the reason why this respirator has good adaptability. Other respirator manufacturers can also make similar improvements.
This study has its own unique advantages. Based on previous studies, this study combined computer-vision technology with a respirator confirmation quantitative test to study the influence of facial features on N95, which is a major step in the study of the performance of N95. The results of this study not only provide a reference for medical workers to choose proper medical protective respirators but also provide a reference for respirator manufacturers to improve respirators.
Limitations
A limitation of this study is that the population selection was limited to medical personnel in Beijing, and the sample size and sample coverage were limited. Subjective comfort of wearing a medical protective respirator was not included in this study. As independent individuals, medical staff cannot avoid the influence of comfort when choosing medical protective respirators. These limitations could be addressed in future research. Another limitation of the current study was that only one respirator brand/model was used for each design type (foldable, flat, or arched). The study aimed to investigate the effects of different types of N95 on facial dimensions. By choosing several brands and models of respirators for each design type, more comprehensive results can be obtained. Nevertheless, the results have reference value for studying the influence of facial dimensions on the performance of different types of respirators. Also, further studies with a larger sample size may lead to a clearer conclusion regarding the relationship between arched respirators and the facial parameters of wearers. The authors plan to expand the scope of the study and increase the sample size for a more detailed and in-depth study.
Conclusion
N95s are an important tool to prevent the spread of respiratory diseases, and it is necessary for medical workers to choose N95s suitable for their individual face shape and to wear the respirator correctly to maximize the fit of the respirator and reduce occupational exposure. It is well known that the use of an N95 is superior to surgical masks in preventing respiratory infections; however, only when a good fit is achieved and the respirator is worn correctly, can respirators play a protective role in the health and safety of healthcare workers (Qian et al. Citation1998; Yu et al. Citation2012).
The fit of N95 respirators is related to the facial dimensions, and the fit of different types of respirators assessed in this study was affected by different facial dimensions. It is recommended that medical staff carry out a respirator fit test before using an N95. If possible, it is recommended that medical institutions provide multiple types of respirators, so that medical staff can choose more suitable masks according to their facial dimensions, such as face length, face width, nose height, and nose length. For example, medical staff with a long and high nose can prioritize folding respirators, whereas medical staff with a wide nose and a long chin can prioritize flat respirators. By using appropriately fitting respirators, medical staff can better protect themselves against exposure to airborne pathogens that can be caused by inappropriate respirator fit.
This study shows that the fit of different types of respirators is affected by facial dimensions, and the collection of facial dimension information from a representative target population is the basis for improving the fit of respirators. China has not established a representative population facial dimension database, and existing facial dimension parameters for the design and manufacture of N95 respirators are insufficient, resulting in the generally poor fitting respirators in China and in other Asian countries (Seo et al. Citation2020). Therefore, it is necessary to carry out a nationwide collection of human head and face size data, extract the factors related to the fit of respirators, and provide data support for improving respirator design and manufacturing for improved fit amongst Asian populations.
Supplemental Material
Download MS Word (674.7 KB)Acknowledgments
The authors thank all the medical staff who participated in this research and all members of the project team who helped carry out this study. We thank all participants and investigators for their contribution to the study, Zaofang Yan and Lei Li, for assistance with data collection, and hospitals participating in the study for support on room usage, specific equipment, and consumables.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Additional information
Funding
References
- Beneviat AR, Waldhoff SC, Vick DJ, Kerr BJ, Jr. 2023. A review of 2020 state and DC face mask guidelines for U.S. healthcare workers during the Covid-19 pandemic. Hosp Top. 101(2):146–155. doi:10.1080/00185868.2021.1997130.
- Centers for Disease Control and Prevention (CDC). 1998. Laboratory performance evaluation of N95 filtering facepiece respirators, 1996. MMWR Morb Mortal Wkly Rep. 47(48):1045–1049.
- Cheng WJ, Yu YY, Yu D, Ping J, Chen WH. 2012. [The relationship between fit characteristics of two kinds of self-inhalation air-purifying dust respirators and head-face dimensions of subjects]. Zhonghua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi. 30(5):352–355.
- Derrick JL, Chan YF, Gomersall CD, Lui SF. 2005. Predictive value of the user seal check in determining half-face respirator fit. J Hosp Infect. 59(2):152–155. doi:10.1016/j.jhin.2004.09.009.
- Derrick JL, Gomersall CD. 2005. Protecting healthcare staff from severe acute respiratory syndrome: filtration capacity of multiple surgical masks. J Hosp Infect. 59(4):365–368. doi:10.1016/j.jhin.2004.10.013.
- Fakherpour A, Jahangiri M, Seif M, Charkhand H, Abbaspour S, Floyd EL. 2021. Quantitative fit testing of filtering face-piece respirators during the COVID-19 pandemic reveals anthropometric deficits in most respirators available in Iran. J Environ Health Sci Eng. 19(1):805–817. doi:10.1007/s40201-021-00648-3.
- Han DH, Seo H, Kang BK, Jang H, Kim H, Shim S. 2022. Comparisons of fit factors between two quantitative fit testers (PortaCount vs. MT). Saf Health Work. 13(4):500–506. doi:10.1016/j.shaw.2022.10.001.
- Lam SC, Lee JK, Yau SY, Charm CY. 2011. Sensitivity and specificity of the user-seal-check in determining the fit of N95 respirators. J Hosp Infect. 77(3):252–256. doi:10.1016/j.jhin.2010.09.034.
- Niu X, Koehler RH, Yermakov M, Grinshpun SA. 2023. Assessing the fit of N95 filtering facepiece respirators fitted with an ear loop strap system: a pilot study. Ann Work Expo Health. 67(1):50–58. doi:10.1093/annweh/wxac051.
- Prince SE, Chen H, Tong H, Berntsen J, Masood S, Zeman KL, Clapp PW, Bennett WD, Samet JM. 2021. Assessing the effect of beard hair lengths on face masks used as personal protective equipment during the COVID-19 pandemic. J Expo Sci Environ Epidemiol. 31(6):953–960. doi:10.1038/s41370-021-00337-1.
- Qian Y, Willeke K, Grinshpun SA, Donnelly J, Coffey CC. 1998. Performance of N95 respirators: filtration efficiency for airborne microbial and inert particles. Am Ind Hyg Assoc J. 59(2):128–132. doi:10.1080/15428119891010389.
- Rebmann T, Carrico R, Wang J. 2013. Physiologic and other effects and compliance with long-term respirator use among medical intensive care unit nurses. Am J Infect Control. 41(12):1218–1223. doi:10.1016/j.ajic.2013.02.017.
- Seo H, Kim JI, Kim H. 2020. Development of Korean head forms for respirator performance testing. Saf Health Work. 11(1):71–79. doi:10.1016/j.shaw.2019.12.008.
- Sreenath A, Weed J, Church T. 2001. A modified protocol for quantitative fit testing using the PortaCount. Appl Occup Environ Hyg. 16(10):979–988. doi:10.1080/104732201300367245.
- Trehan RS, McDonnell EP, McCoy JV, Ohman-Strickland PA, Donovan C, Quinoa TR, Morrison DS. 2021. Comparing the quantitative fit-testing results of half-mask respirators with various skin barriers in a crossover study design: a pilot study. J Hosp Infect. 111:125–131. doi:10.1016/j.jhin.2021.02.010.
- Yang L, Shen H, Wu G. 2007. Racial differences in respirator fit testing: a pilot study of whether American fit panels are representative of Chinese faces. Ann Occup Hyg. 51(4):415–421. doi:10.1093/annhyg/mem005.
- Yu Y, Benson S, Cheng W, Hsiao J, Liu Y, Zhuang Z, Chen W. 2012. Digital 3-d headforms representative of Chinese workers. Ann Occup Hyg. 56(1):113–122. doi:10.1093/annhyg/mer074.