Prediction of Ross River Virus Incidence Using Mosquito Data in Three Cities of Queensland, Australia

Wei Qian; Elvina Viennet; Kathryn Glass; David Harley; Cameron Hurst

doi:10.3390/biology12111429

Prediction of Ross River Virus Incidence Using Mosquito Data in Three Cities of Queensland, Australia

Biology (Basel). 2023 Nov 13;12(11):1429. doi: 10.3390/biology12111429.

Authors

Wei Qian^{1

2}, Elvina Viennet^{3

4}, Kathryn Glass⁵, David Harley², Cameron Hurst^{6

7}

Affiliations

¹ School of Public Health, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China.
² UQ Centre for Clinical Research, The University of Queensland, Herston, QLD 4029, Australia.
³ Strategy and Growth, The Australian Red Cross Lifeblood, Kelvin Grove, QLD 4059, Australia.
⁴ School of Biomedical Sciences, Queensland University of Technology, Kelvin Grove, QLD 4059, Australia.
⁵ Research School of Population Health, Australian National University, Acton, ACT 0200, Australia.
⁶ Molly Wardaguga Research Centre, Charles Darwin University, Brisbane, QLD 4001, Australia.
⁷ Department of Statistics, QIMR Berghofer Medical Research Institute, Brisbane, QLD 4006, Australia.

Abstract

Ross River virus (RRV) is the most common mosquito-borne disease in Australia, with Queensland recording high incidence rates (with an annual average incidence rate of 0.05% over the last 20 years). Accurate prediction of RRV incidence is critical for disease management and control. Many factors, including mosquito abundance, climate, weather, geographical factors, and socio-economic indices, can influence the RRV transmission cycle and thus have potential utility as predictors of RRV incidence. We collected mosquito data from the city councils of Brisbane, Redlands, and Mackay in Queensland, together with other meteorological and geographical data. Predictors were selected to build negative binomial generalised linear models for prediction. The models demonstrated excellent performance in Brisbane and Redlands but were less satisfactory in Mackay. Mosquito abundance was selected in the Brisbane model and can improve the predictive performance. Sufficient sample sizes of continuous mosquito data and RRV cases were essential for accurate and effective prediction, highlighting the importance of routine vector surveillance for disease management and control. Our results are consistent with variation in transmission cycles across different cities, and our study demonstrates the usefulness of mosquito surveillance data for predicting RRV incidence within small geographical areas.

Keywords: Ross River virus; exposures; lagged effects; mosquitoes; prediction; surveillance.

Grants and funding

na/This work was supported by the University of Queensland Research Training Scholar-ship and Frank Clair Scholarship. The funders had no role in study design, data collec-tion and analysis, decision to publish, or preparation of the manuscript.