Qualitative behaviour assessment captures the effects of acute pain on emotional expressivity in Angus beef calves

Pain assessment methods in farm animals often fail to capture the range of emotional expression, and reliable and valid methods to assess pain-related emotional states remain limited, particularly for on-farm applications. This study aimed to apply Qualitative Behavioural Assessment (QBA) to evaluate acute pain-related emotional states in Angus beef calves and assess its inter- and intra-observer reliability and construct and criterion validities. Sixty-nine calves, 6–8 weeks old, were randomly assigned to either surgical castration (CAST, n = 34) or sham castration (SHAM, n = 35) groups. Calves were individually video recorded in a holding pen before and after castration. A panel of ten cattle-experienced individuals viewed 20 videos (1-min long) displaying 10 CAST and 10 SHAM calves and, through discussion, generated a fixed-term list of 20 descriptors with short definitions. Ten different QBA-trained observers scored 40 videos (20 CAST, 20 SHAM) twice, 3 weeks apart, using the fixed-term list. Principal component analysis (PCA-1; no rotation) was conducted for both sessions to extract principal component (PC) values. Intra-observer reliability (Kendall coefficient, W), using PC values of PCA-1, ranged from strong to very strong on PC1 (W = 0.55–0.90), PC2 (W = 0.53–0.90), and PC4 (W = 0.53–0.88), and from moderate to very strong on PC3 (W = 0.37–0.83) and PC5 (W = 0.42–0.80). Inter-observer reliability ranged from strong to very strong for PC1 (W = 0.76, 0.57 for sessions 1 and 2, respectively), and from moderate to weak (W ' 0.50) for both sessions on other PCs. PCA-2 was performed (Varimax rotation) using the first assessment by 8 observers with intra-observer reliability ≥ 0.60 on at least 3 PCs from PCA-1. PCA-2 yielded 4 PCs, which explained 64.69% of the variance. Considering all terms with loadings of ' -0.5 or loadings ' 0.5, PC1 (32%) was labelled as ‘emotional arousal’, PC2 (14.3%) as ‘positive mood’, PC3 (11.2%) as ‘negative mood’, and PC4 (7.3%) as ‘attentiveness’. Construct validity was evaluated using a generalized mixed effects model for PC values from PCA-2. The CAST calves had lower values for emotional arousal and positive mood, and higher values for negative mood, compared to SHAM calves (P ' 0.01). Criterion validity was evaluated using Spearman correlations (r _s) between PC values and other pain measures, including the calf grimace scale (CGS), activity, locomotion, the activity-locomotion score, and miscellaneous behaviours from the UNESP-Botucatu cattle pain scale. Emotional arousal (PC1) had moderate negative correlation with most pain indicators (r _s= −0.47 to −0.58, P ' 0.01). Negative mood (PC3) had moderate to strong positive correlations (r _s = 0.54–0.77, P ' 0.01) with all indicators. Other PCs had no correlations (P ' 0.05), except PC2 with the CGS front view, r _s = -0.42, (P = 0.01). In conclusion, QBA is a valid and reliable tool for assessing acute pain-related emotional states in calves.