The correlation of Qualitative Behavior Assessments with Welfare Quality® protocol outcomes in on-farm welfare assessment of dairy cattle

The effort to develop methods for assessing animal welfare at farm level has grown dramatically since the end of the 1990s, culminating in the protocols developed by the European-wide project Welfare Quality (R) (WQ). However, these protocols are time consuming and lack transparency in how scores are aggregated into welfare outcomes. The current study investigates the potential of Qualitative Behavior Assessment (QBA), a much less time-consuming approach, to be used as a stand-alone integrative screening tool for identifying farms with compromised welfare before applying the full WQ protocol. QBA is a 'whole-animal' approach asking human observers to summarize animals' expressive demeanor and its context into descriptors such as relaxed, anxious, content or frustrated -terms which given their emotional connotation appear to have direct relevance to animal welfare. Two trained QBA-assessors, and one trained Welfare Quality (R) assessor visited 43 Danish dairy cattle farms at different times, the former focusing on QBA and the latter making a full WQ protocol assessment. The QBA scores were analyzed using Principal Component Analysis (correlation matrix, no rotation), and WQ protocol data were analyzed and integrated according to the WQ protocol. The resulting QBA and WQ protocol outcomes were correlated using non-parametric methods (Spearman Rank and Kendall W). Highly significant inter-observer agreement was found between the two QBA-assessors (P<0.0001). QBA scores showed some weak correlations to WQ measures but no meaningful pattern of relationship between these measures emerged. The present study does not support the application of QBA as a stand-alone welfare assessment tool capable of predicting the outcome of the larger WQ protocol. (C) 2012 Elsevier B.V. All rights reserved.