Predictive modelling of Ross River virus using climate data in the Darling Downs.

Area of Science:

• Environmental Science
• Epidemiology
• Public Health

Background:

• Ross River virus (RRV) is Australia's most prevalent mosquito-borne illness, causing significant health and economic burdens.
• Current lack of effective treatments or vaccines for RRV disease necessitates improved predictive strategies.
• The complex ecology of RRV involves diverse reservoirs, vectors, environments, and climates across Australia.

Purpose of the Study:

• To develop and validate a predictive model for monthly Ross River virus case numbers and outbreaks.
• To assess the utility of readily available climate data in forecasting RRV activity.
• To inform public health interventions through timely outbreak predictions.

Main Methods:

• Utilized negative binomial regression to build a predictive model for RRV.
• Trained the model using human RRV notifications and climate data from July 2001 to June 2014 in Queensland, Australia.
• Validated model predictions with data from July 2014 to June 2019.

Main Results:

• The model demonstrated moderate effectiveness in predicting monthly RRV case numbers (Pearson's r = 0.427).
• The model achieved 65% accuracy in predicting RRV outbreaks, with 59% sensitivity and 73% specificity.
• Climate data proved valuable for forecasting RRV case numbers and outbreaks.

Conclusions:

• Readily accessible climate data can be effectively used for timely prediction of Ross River virus outbreaks.
• The developed model offers a practical tool for public health authorities to anticipate and manage RRV disease.
• Further research can refine these models to enhance prediction accuracy and geographic scope.