Abstract
Improving the fit of a half-mask respirator can be achieved by developing a design, fit, and sizing strategy to fit the faces of the general population or a specific group such as race, age group, or occupation. The purpose of this study was to define respirator fit based on the body product relationship and to develop a new set of facial landmarks and measurements for half-mask respirator design. 3D scan data and quantitative fit factor scores from 47 healthcare workers and 9 researchers in healthcare-related fields were utilized to investigate the relationship of new anthropometry measurements to respirator fit. A mask fit association model was validated through logistic regression. The respirator fit prediction model incorporating highly correlated face measurements opens the possibility of developing a system for judging respirator fit success and failure based on face dimensions; it can be integrated with automated measuring technologies and machine learning.
Providing adequate respirator fit for individuals continues to be a major concern, especially in occupational settings. Improving the fit of a half-mask respirator can be achieved by developing a design, fit, and sizing strategy to fit the faces of the general population or a specific group such as race, age group, or occupation. The purpose of this study was to define respirator fit based on the body-product relationship and to develop a new set of facial landmarks and measurements for half-mask respirator design. 3D scan data and quantitative fit factor scores from 47 healthcare workers and 9 researchers in healthcare-related fields were utilized to investigate the relationship of new anthropometry measurements to respirator fit. PCA results revealed the limitations of existing studies on describing mask fit with a small number of facial dimensions. A mask fit prediction model was validated through logistic regression. The respirator fit prediction model incorporating highly correlated face measurements opens the possibility of developing a system for judging respirator fit success and failure based on face dimensions; it can be integrated with automated measuring technologies and machine learning. The proposed method can be applied to a range of respirators and allow simple fit testing procedures for occupational workers.
Providing properly fitting PPE to workers is imperative, yet complex in industrial settings. This research presents a model for innovative respirator design and future research. The respirator fit prediction model unlocks the possibility of developing a system for judging respirator fit success and failure based on face dimensions; it can be integrated with automated measuring technologies and machine learning. The proposed method can be applied to a range of respirators. It can be utilized with technologies such as photo-based 3D face shape estimation and app-based face scanning with an in-depth camera to develop advanced algorithms that enable the occupational workers to capture their own faces and select suitable respirators. While this prediction model and future recommender systems do not substitute the need for proper respirator fit testing in industrial settings, it does create a simplified process for workers and employers to select and stock the best fitting respirators.
