Hirsch, E. N., Andersson, M., Loberg, J. et al. 2021. Development of existing scoring systems to assess behavioural coping in shelter cats. Applied Animal Behaviour Science 234, 105208.
Assessing how cats cope with the housing and husbandry at shelters is an important part of maintaining good animal welfare. There are non-invasive methods to assess how cats cope with their environment. The aim of this study was to investigate the reliability of the behaviours used in an extended Stress Assessment protocol for cats to detect stress. Looking at which behaviours are salient and possible to observe accurately, and which correlate with time until adoption. The study was carried out at a non-governmental medium sized cat rescue shelter in Midwestern USA. The shelter had a no-kill policy with screening of cats before intake from county shelters. The observed cats were either group-housed in five rooms (n = 70) or singly housed in double cages (n = 13). Observations were carried out during both morning and afternoon sessions, during which two 1-min observations recorded if cats performed any of 85 behavioural elements (BEs). Time at shelter and if cats were declawed or not were collected from shelter records after the observations. Statistical analysis of the BEs that best predicted the total time at shelter was calculated using the Survival Analysis based on the Cox proportional hazards regression model using a stepwise regression analysis separately for each scoring. The median time at shelter for group-housed cats was 26 days (IQR = 6–54) and for single-housed cats 29 days (IQR = 7–97). In total, 24 % of the BEs (20 of 85) were never recorded, however there were significantly more BEs recorded in group-housed cats (63 BEs) than in single-housed cats (49 BEs, p < 0.05). The survival analysis found 16 unique BEs to predict “Short time at shelter” (14 BEs in group-housed, two in single-housed), 14 were positively correlated meaning that they increased the chance of early adoption and two were negative meaning that they decreased the chance of early adoption. The survival analysis also calculated “Long time at shelter” and found 14 unique BEs where 12 BEs were in group-housed cats and three BEs were in single-housed cats. Seven of these were positively correlated meaning that they decreased the chance of early adoption, whereas seven were negatively correlated meaning that they increased the chance of early adoption. The conclusion is that the extended Stress Assessment could be used to detect BEs indicating stress of cats at shelters, and that there are BEs that can predict shorter time at the shelter. However, further investigations could help reduce the number of BEs needed.