Refining and Validating an on-site Canine Welfare Assessment Tool Developed for use in Commercial Breeding Kennels.
Accurate assessments of behavior and welfare are needed to evaluate the state of domestic dogs maintained in commercial breeding (CB) and other types of kennels. Field assessments of dogs’ states of being must be reliable, valid and efficient. However, observer subjectivity and situational variation in dogs’ responses pose a challenge to incorporating behavioral metrics into welfare assessment tools. The published Field Instantaneous Dog Observation (FIDO) tool, designed to capture the immediately observable physical and behavioral status of dogs in kennels, was thus evaluated on its reliability and validity. Specifically, the main goals were to determine 1) reliability of the behavioral scoring when used by novice raters, 2) whether and to what extent dogs’ behavioral responses to stranger-approach changed during a 30-second observation period, and 3) the predictive power of the FIDO scoring on behavioral responses of dogs placed within a standardized arena with a stranger. Behavioral responses to stranger-approach were organized into three categories: red, indicating a fearful response to approach, green, indicating an affiliative or neutral response, and yellow, indicating an ambivalent response. In study one, behavior assessment was conducted by two novice raters with 50 dogs housed at two US shelters. A stranger approached the home pen of each dog in a non-threatening manner, stood quietly, extended a hand to the dog and scored the response while the test was video-recorded. Intra-rater reliability was assessed by comparing each rater’s live observation scores with their scores of the same dogs using video recordings. Inter-rater agreement between scores from video recordings was also calculated. In study two, 81 commercial breeding dogs maintained at four USDA-licensed CB facilities in the US were approached by one observer and scored once every five seconds for 30 seconds. Of the 81 subject dogs, 56 met the criteria for full sampling. In study three, 40 pairs of dogs were scored using the FIDO tool in their home pens at four USDA-licensed CB facilities and assessed on responses to an unfamiliar observer in a field test conducted in an outdoor arena. Behavioral measures from the field test such as latencies to approach the stranger and durations in different areas of the arena were captured from video recordings and subjected to a principal components analysis [PCA] for reduction of variables. A multivariate multiple regression analysis was thereafter used, with principal component scores obtained from the PCA as outcome variables, and FIDO scores as predictor variables, accounting for other factors such as breed and group composition. Results from study one indicated that raters showed almost perfect agreement between their own scores of live and video-recorded shelter dog responses (kappa = 0.83, 0.89) and between each other’s video-recorded scores (kappa = 0.87), indicating high intra- and inter-rater reliability. Results from study two indicated that over a 30-second time frame with five-second increments, 91% of the dogs showed no change in their behavioral response to approach. This suggests that the first five seconds of scoring provide a reliable time point for assessing behavior using the FIDO tool and indicates no benefit to extending the FIDO scoring period to gauge dogs’ immediate responses to stranger-approach. Results from study three showed that dogs scored as red, as determined by the FIDO scoring, also showed higher scores associated with avoidance, indicated by greater time spent away and more time taken to approach the unfamiliar person in the field test (p = 0.039). FIDO scoring was, however, not significantly predictive of other behavioral responses such as interaction with the unfamiliar person. How dogs were housed was also significantly associated with incidences of escape attempts from the arena; dogs paired as a mixed sex (i.e. male-female) showed lower scores on incidences of escape attempts than dogs paired as a same sex (i.e. female-female) (p = 0.003). Taken together, these results suggest that the FIDO tool can be used by individuals without much expertise in canine behavior, and also attest to the practicality of the tool via a reliable five-second approach. Further, results also indicate that the FIDO scoring can help to identify fearful dogs in need of greater socialization towards unfamiliar people. Future research would, however, be valuable in validating behavioral scores obtained using the tool against long-term indicators of overall welfare.