Can an animal welfare risk assessment tool identify livestock at risk of poor welfare outcomes?

If livestock at risk of poor welfare could be identified using a risk assessment tool, more targeted response strategies could be developed by enforcement agencies to facilitate early intervention, prompt welfare improvement and a decrease in reoffending. This study aimed to test the ability of an Animal Welfare Risk Assessment Tool (AWRAT) to identify livestock at risk of poor welfare in extensive farming systems in Australia. Following farm visits for welfare- and non-welfare-related reasons, participants completed a single welfare rating (WR) and an assessment using the AWRAT for the farm just visited. A novel algorithm was developed to generate an AWRAT-Risk Rating (AWRAT-RR) based on the AWRAT assessment. Using linear regression, the relationship between the AWRAT-RR and the WR was tested. The AWRAT was good at identifying farms with poor livestock welfare based on this preliminary testing. As the AWRAT relies upon observation, the intra- and inter-observer agreement were compared in an observation study. This included rating a set of photographs of farm features, on two occasions. Intra-observer reliability was good, with 83% of Intra-class Correlation Coefficients (ICCs) for observers ≥ 0.8. Inter-observer reliability was moderate with an ICC of 0.67. The AWRAT provides a structured framework to improve consistency in livestock welfare assessments. Further research is necessary to determine the AWRAT's ability to identify livestock at risk of poor welfare by studying animal welfare incidents and reoffending over time.