An understanding of interplay between asthma, criteria air pollutants, and meteorological factors is essential to predicting human health and reducing the capital costs on controlling asthma prevalence. A systematic relationship between asthma and variables related to air quality and weather in a location will help mitigate or reduce impact of asthma over the people in it. The impact of industries, transportation, and weather on health is complex and needs continuous monitoring for controlling health disorders. The present study examines Poisson Regression Model of the daily data of asthma admissions, Particulate Matter (PM 2.5 ), Ozone (O 3 ), Nitrogen dioxide (NO 2 ), temperature, and humidity in selective locations of Mississippi coastal region of Gulf of Mexico for the period 2003 to 2011. The study region consists of three locations, namely Gulfport, Pascagoula, and Waveland because of the extent of data availability. Overall, the results indicate a negative correlation of asthma with temperature and the effect was statistically significant (p < 0.05) in all the regions. The correlation of other variables was not consistent uniformly, and their influence on asthma was not statistically significant except in few cases.
Keywords:Poisson Regression Model; Pearson correlation; Asthma; Criteria Pollutants; time series
Materials and MethodsConsider a random variable Y t , representing the time series count data Y 1 ,Y 2 , …., Y T and X i representing the regression covariate variables, then Y t Poisson distributed over X with a mean value of µ is given by,
Y t~P oisson (μ t )The log linear regression model of Poisson distribution in terms of regression coefficient (β) is given by, log(μ t ) = β 0 +β 1 X 1 +β 2 X 2 +...In the present situation, Y t is the daily counts of asthma admissions at time t and X i the independent variable (for the air pollution and meteorological parameters); the parameter β i represents regression coefficient and is a measure of association between an independent variable, X i (for the air pollution and meteorological parameters), and the risk of the outcome Y for the asthma admission. The log linear relationship is given by, log(μ t )=β 0 +β 1 O 3 +β 2 NO 2 +β 3 PM 2.5 +β 4
temperature+β 5 HumidityThe regression coefficient β i also signifies forecasting proportional change in the value of Y t for a given a unit change in X t , Using SPSS statistical package, the Poisson Regression model is applied to the asthma data without time lag, and analyzed the associations by considering the causes variables individually. This was done to utilize the maximum number of valid datasets. log(μ t ) = β 0 +β 1 O 3 log(μ t )= β 0 + β 1 NO