A Weather-Based Predictive Model for Barmah Forest Virus Disease in Brisbane, Australia

Abstract

Listen

Translate article

P-505 Introduction: Barmah Forest virus (BFV) disease is one of the most common vector-borne diseases in Australia. The study examined the impact of weather variability and tides on the incidence of Barmah Forest virus and developed a weather-based predictive model for Barmah Forest virus disease in Brisbane, Australia. Methods: We obtained data on the monthly Barmah Forest virus cases, weather variables (eg: mean minimum and maximum temperature, total rainfall and mean relative humidity) and mean high tide, and population size in Brisbane between January 1992 and December 2001 from the Queensland Department of Health, Australian Bureau of Meteorology and Queensland Transport, and Australian Bureau of Statistics, respectively. Results: The results of the correlations show that Barmah Forest virus incidence was statistically significantly associated with minimum temperature (rs= 0.18, P = 0.04) and relative humidity (rs= 0.26, P = 0.00). A series of seasonal auto-regressive integrated moving average (SARIMA) models were used to determine the effect of weather variables on Barmah Forest virus incidence. We have used two models, a model (Model I) excluding the weather variables and the other (Model II) including weather variables for prediction of the incidence of Barmah Forest virus disease. Modelling outcomes were compared for both the models. The model (Model II) with the input of minimum temperature with a moving average of 0–3 months was statistically significantly and positively associated with the incidence of Barmah Forest virus (Regression coefficient = 0.057, P < 0.001) and therefore was a better model for predicting Barmah Forest virus incidence (Model I: Akaike's Information Criterion = 63.99, Root Mean Square error = 2.5; Model II: Akaike's Information Criterion = 40.47, Root Mean Square error = 1.3). However, no significant association was observed for other weather variables. The validation of the model II showed it could be used as a potential tool in forecasting. Discussion and conclusions: These findings may be used to develop effective control and prevention programs towards Barmah Forest virus disease.