TIME SERIES ANALYSIS ON THE INCIDENCE OF INFLUENZA IN ELEMENTARY SCHOOLS DURING WINTER SEASON USING TEMPERATURE AND HUMIDITY

Study on the relationship between indoor air environment and the incidence of influenza in schools

Abstract

 A deteriorated indoor climate is commonly related to increases in sick building syndrome symptoms, respiratory illnesses, sick leave, reduced comfort and losses in productivity. In schools, sick leave might to be a significant index of productivity. Elementary schools are thought to be amplifiers for infections in influenza in communities. In School Health and Safety Act, children are told to stay home in case of an epidemic of influenza. Therefore the indoor air environment should be controlled for avoiding an epidemic of influenza. The humidity ratio had been reported to be a critical factor for infection of influenza. However the relationship between the really measured indoor thermal factors and the incidence of influenza is rarely investigated. Then the room temperature and relative humidity are continuously measured in twelve classrooms in a public elementary school in Tokyo. And the period of analysis was set as 3 winter seasons from 2005/2006 to 2007/2008.
 The average values of the room temperature per every winter season in all classrooms were met the School Environmental Sanitary Standard. On the other hand, the average values of indoor relative humidity were not met the standard in some classrooms and lower than the standard value.
 The high correlation coefficients were not obtained in the relationship between the average values of indoor thermal factors, i.e., air temperature, relative humidity, and humidity ratio per every winter season in each classroom, and the number of absence on account of flu. Although the tendency was found that the lower the humidity, the greater the number of absence on account of flu, the correlation coefficient was low and not significant.
 Focusing on the indoor climatic data when the absence of flu was occurred, the highest correlation coefficient was observed at 0.984 in the relationship between the daily average indoor temperature of two days before the absence and the probability of flu. The second highest correlation coefficient was observed in the relationship between the daily average indoor room relative humidity of one day before the absence and the probability of flu.
 The significant high correlation was seen in the relationship between the probability of flu and the daily average outdoor air temperature of the absence day obtained from the meteorological data. The significant high correlation was also found in the relationship between the probability of flu and the daily average outdoor humidity ratio of one day before the absence.
 A multiple regression analysis was conducted with the probability of flu as a objective variable. The explanatory variables were indoor and outdoor climatic factors. As a result of the analysis, the room temperature of two days before the absence was merely a significant factor (p<0.05). A time-series analysis for indoor/outdoor climate especially a few days before the absence could be useful for understanding the infection of flu.