On-farm qualitative behaviour assessment in sheep: Repeated measurements across time, and association with physical indicators of flock health and welfare

Abstract

Qualitative Behavioural Assessment (QBA) is a ‘whole-animal’ methodology that assesses the expressive qualities of animal behaviour using terms such as ‘tense’, ‘relaxed’, ‘anxious’, and ‘content’. The reliability and validity of QBA as an indicator for on-farm welfare assessment in pigs, cattle, poultry and sheep has been examined in a number of ways. However, the use of QBA on farms over longer periods of time has not yet been examined. The aim of this study was to investigate whether and how on-farm QBA of sheep varies over the different seasons of the year, and whether it is associated with physical measures of sheep health and welfare such as lameness. A trained assessor visited each of 12 farms six times within a one year period at two month intervals, and made group level assessments of approximately 100 sheep selected ad hoc (assuming homogeneity within the flock). The sheep flocks were assessed with a list of twelve QBA descriptive terms previously developed for sheep. Following QBA, the same sheep were also assessed with seven physical indicators of health and welfare (‘dull physical demeanour’, lameness, breech and abdominal soiling, pruritis, wool loss, and coughing). QBA scores from all visits were analysed together, and also in combination with the physical measures, with Principal Component Analysis (PCA – correlation matrix, no rotation). The effect of visit on PCA flock scores was analysed with random-effects multiple linear regression models. The association between PCA flock scores and physical measures was investigated using Spearman rank correlation (rS), and the correlation of flock rankings across visits was examined with Kendall Coefficient of Concordance. PCA distinguished two main dimensions of sheep expression: PC1 (47% variation) ranging from content/relaxed/thriving to distressed/dull/dejected (summarised as ‘mood’) and PC2 (21%), which ranged from anxious/agitated/responsive to relaxed/dejected/dull (summarised as ‘responsiveness’). No significant effect of visit on PC1 scores was found (p=0.155), and PC1 flock scores correlated at W=0.84 (p<0.001) across the 6 visits, indicating high consistency of characterisations of individual flock mood over the year. However there was an effect of visit on PC2 scores (p<0.001), and PC2 flock scores were correlated at W=0.60 (p<0.001) across visits, indicating that the presence of young lambs may have had a consistently relaxing effect on flocks. There was also an effect of visit period on lameness (p=0.025), and on breech (p<0.001) and abdominal (p=0.0048) soiling. With the exception of lameness and breech and abdominal soiling, the physical indicators were observed at a low prevalence (<2%) across the study farms. The highest lameness levels were observed during the winter period (mean 17.86%, 95% CI 7.83–27.90) whilst breech soiling was highest in spring (mean 23.83%, 95% CI 11.86–35.81). An effect of farm type was found on lameness scores (p=0.0176) and an effect of flock size on abdominal soiling scores (p=0.025). PC1 ‘mood’ scores were negatively correlated to the proportion of lame sheep (n=72; rS=−0.72, p<0.001), and to the proportion of animals with dull physical demeanour (rS=−0.70, p<0.001), while PC2 ‘responsiveness’ scores showed a weak correlation with breech soiling (rS=0.42, p<0.001). In summary, these results suggest that QBA has the potential to serve as a sensitive, meaningful indicator for on-farm welfare assessment in sheep.