Temporal dynamics and intra-farm variability of animal welfare indicators in fattening pigs

The Welfare Quality^® (WQ^®) protocols are broadly implemented for monitoring farm animal welfare and certification purposes. However, they represent a single snapshot in time, demand a significant amount of time for execution, appraise only a subset of the animals, and to achieve high repeatability among assessors, they must be continuously trained. In growing pigs, the dynamic nature of physiological and behavioural changes during the production cycle may affect the consistency of monitoring over time and thus the reliability of the method. Hence, the aim of the present study was to evaluate the temporal variability of principles, criteria and measures of the WQ^® protocol in growing pigs, and the intra-farm (between buildings) variability of selected measures. The full WQ^® protocol was applied every two weeks during the fattening period in two separate buildings of the same farm by the same two trained assessors (seven and eight visits respectively). A descriptive temporal analysis was conducted, and the relation of measures was explored using Spearman's correlation. Intra-farm variability of selected measures was assessed for each visit using a permutation test for independence. Certain variability in some measures, criteria and principles was observed depending on the timing of the assessment during the growing period. The "good housing" principle was the most fluctuating variable over time, due to the substantial downward trend of the scores of both the "comfort around resting" and "ease of movement" criteria. Although the "good health" principle was the most stable in terms of absolute score, the categorization changed in one assessment in one of the studied buildings. Both "good feeding" and "appropriate behaviour" remained stable in their categorization. More variability was observed at the criteria and measure levels. In terms of intra-farm variability, measures from four conditions were significantly different, in at least one visit, being the "fear of humans" measure the most frequently variable. This study highlighted the possible differences that may arise in WQ^® scoring depending on the time of the assessment during a growing period. The temporal dynamics and intra-farm variability of this protocol outcomes warrant further investigation to improve the reliability of welfare monitoring protocols.