Using Data Collected for Production or Economic Purposes to Research Production Animal Welfare: An Epidemiological Approach
Publication Type: |
Journal Article |
Year of Publication: |
2009 |
Authors: |
Cate Dewey, Charles Haley, Tina Widowski, Robert Friendship, Janet Sunstrum, Karen Richardson |
Publication/Journal: |
Journal of Applied Animal Welfare Science |
Keywords: |
animal welfare, data collection, epidemiology, industry, methodology |
ISBN: |
1088-8705 |
Abstract:
Epidemiologists use the analyses of large data sets collected for production or economic purposes to research production nonhuman animal welfare issues in the commercial setting. This approach is particularly useful if the welfare issue is rare or hard to reproduce. However, to ensure the information is accurate, it is essential to carefully validate these data. The study used economic data to research in-transit deaths of finishing pigs. The most appropriate model to fit the distribution of the outcome must be selected. A negative binomial model fit these data because the prevalence was low and most lots of pigs had no deaths. The study used hierarchical dummy variables to identify thresholds of temperature and humidity above which in-transit losses increased. Multiple variable modeling provides the foundation for the strength of epidemiological research. The model identifies the association between each factor and the outcome after controlling for the other factors in the model. The study evaluated confounding and interaction. Bias may be introduced when data are limited to one farm system, one abattoir, or one season. Census data enable us to understand the entire industry.