Determination of Environmental Factors Affecting Airborne Bioaerosols in Hospitals Using Nonlinear Models

Abstract

Introduction: Environmental and structural factors have a significant effect on the concentration of bioaerosols in the indoor air; however, this issue has not received enough attention in research. . Therefore, this study aimed to determine the distribution of bioaerosols and environmental factors affecting the concentration of bioaerosols by using an artificial neural network.Methods: In this study, the concentration of bioaerosols was determined using the inactive method for one hour in the human respiratory tract. The teypticas soy agar and sabouroud dextrose agar were used to culture bacteria and fungi; then, they were incubated at 15, 25, and 37° C. The independent variable includes three categories of structural variables (number of ward beds, room dimensions, and window dimensions), hospital management and management variables (ward air change per hour), and analytical variables (incubation temperature and type of culture medium). The effect of these variables was modeled by an artificial neural network with Levenberg–Marquardt algorithm.Results: This study showed that the distribution of bioaerosols in dialysis and surgery wards was higher than in other wards. Also, the artificial neural network can predict the effect of environmental and structural factors of the hospital on bioaerosols, especially fungi because the maximum and minimum concentrations of bacteria and fungi in the actual data were similar to the data predicted by the artificial neural network. In addition, the correlation coefficient for the fungi was 0.95, and the correlation rate between the actual data and the data predicted by the artificial neural network was higher than 0.92.Conclusion: ANN could be a desirable and reliable method for detecting bioaerosol in hospitals based on the environmental and structural parameters.

Keywords

• airborne
• bacteria
• fungi
• neural network
• environmental factors