A localised risk model for liver fluke infection

Area of Science:

• Veterinary Epidemiology
• Parasitology
• Predictive Modeling

Background:

• Liver fluke infection in cattle is often sub-clinical, making accurate risk assessment difficult.
• The parasite's complex life cycle is sensitive to climate and requires intermediate snail hosts.
• Targeted risk assessments are crucial for effective disease management and producer awareness.

Purpose of the Study:

• To develop a predictive modeling tool for forecasting location-specific liver fluke infection risk in cattle.
• To incorporate changing climate conditions and spatio-temporal data into the risk assessment model.
• To create localized risk profiles for Australian postcodes to inform cattle producers.

Main Methods:

• Utilized processor data from over 4 million cattle (2016-2020) with observed liver fluke infection status.
• Developed a spatio-temporal model incorporating location-specific and time-varying covariates.
• Employed Random Forest (RF) and Generalised Additive Models (GAM) for prediction, with RF showing slight preference.

Main Results:

• The spatio-temporal model successfully predicted liver fluke infection risk.
• Localized risk profiles were generated for Australian postcodes.
• Forecasts for 2022 indicated seasonal variations, highlighting high-risk regions.

Conclusions:

• The developed predictive model can forecast liver fluke infection risk in cattle based on climate and location.
• The model's output aids in identifying high-risk areas, particularly wetter regions.
• This tool supports informed, targeted management decisions for cattle producers to mitigate infection risks.