Exploring the longitudinal dynamics of herd BVD antibody test results using model-based clustering

Area of Science:

• Veterinary Epidemiology
• Infectious Disease Dynamics
• Animal Health Management

Background:

• Bovine Viral Diarrhoea (BVD) control and eradication are challenged by difficulties in accurately determining herd infection status.
• Current methods for assessing herd status may not fully capture the dynamic nature of Bovine Viral Diarrhoea virus (BVDV) antibody responses.
• Longitudinal analysis of antibody test results offers a potential complementary approach to improve herd classification.

Purpose of the Study:

• To explore novel methods for classifying cattle herds based on Bovine Viral Diarrhoea virus (BVDV) antibody test result dynamics.
• To investigate the epidemiological similarities between farms by clustering their antibody percentage positivity (PP) trajectories.
• To assess the utility of trend, magnitude, and shape analyses of antibody PP data for herd classification.

Main Methods:

• Analysis of longitudinal BVDV antibody test data (PP values) from 15,500 bovines across 39 herds (2007-2010).
• Application of Gaussian mixture models for clustering herds based on the magnitude and shape of antibody PP trajectories.
• Investigation of epidemiological similarities within identified farm clusters.

Main Results:

• Development of methods to classify herds based on antibody PP trajectories, capturing herd exposure dynamics.
• Identification of epidemiologically meaningful clusters of farms using magnitude and shape clustering approaches.
• Moderate concordance observed between clustering based on antibody PP trajectory shape and magnitude.

Conclusions:

• Classifying herds by the dynamics of BVDV antibody test results provides a complementary approach to current methods.
• These novel clustering methods show potential for enhancing control strategies for BVD and other livestock diseases.
• The approach offers improved stability in herd status classification compared to traditional year-to-year assessments.