Bayesian and geostatistical analysis of the effect of air pollution on asthma hospitalisation in Perth

Asthma is a respiratory disease which carries a risk of hospitalisation in severe cases. Numerous studies have shown that air pollution is one of the factors that exacerbates asthma symptoms and while population increases in asthma hospitalisation have been shown for Perth, the nature of exposure and risk across areas within the metropolitan area have not been determined. This study investigates the effect of air pollution concentrations on asthma hospitalisation through building a mathematical model which calculates the risk of asthma hospitalisation for Perth metropolitan areas. The data used in this study are: averaged emission inventory data (NO, CO and PM10) in 2006 and records of asthma hospitalisation for 75 postcodes in the Perth metropolitan region. A time series analysis of the data shows that the seasonality for NO2 and CO coincides with the health data seasonality for all years and all postcode groups. The study makes use of geostatistical and statistical techniques within the hierarchical stages of modelling procedure. The air pollutant concentration in Perth was estimated with the Gaussian Plume Model (GPM) and the lognormal kriging. The risk of asthma hospitalisation was calculated using Bayesian Hierarchical Model (BHM). The outputs from the air pollution model and asthma risk model were incorporated in a Generalised Linear Model (GLM), where the risk of asthma hospitalisation was regressed against the air pollution concentration estimate. The results of GLM analysis showed that the influence of air pollutant on asthma hospitalisation risk varied over time and space. The comparisons of the yearly GLM output suggested that the asthma hospitalisation risk has decreased over the study period, although few acute rises in the risk were observed in some years. This study has demonstrated how environmental data and health data could be integrated in a series of mathematical modelling procedures for predicting possible health hazards at a postcode level.

Grantor dc:publisher

Edith Cowan University, Research Online, Perth, Western Australia

Year dc:date

2010

Author dc:creator

• Yano, Yuichi

• respiratory
• smog
• prediction
• forecast
• modelling
• statistics
• health
• Respiratory Tract Diseases
• Statistics and Probability

Repository record dc:identifier

https://ro.ecu.edu.au/theses/1837

OAI identifier oai:identifier

oai:ro.ecu.edu.au:theses-2838