Surveillance of cat shelters by official veterinarians in Austria is based on legal requirements and does not directly assess cat welfare. The aim of this study was to develop animal-based parameters that are feasible to measure in a surveillance setting, stable over time, and reproducible by different raters. We assessed physical condition (body condition, eye and nose discharge, and coat and skin) and behavior of the cats (aggressive interactions, play behavior, behavior toward humans) and also collected data about housing conditions, shelter management, and the shelter cat population. The validity of the animal-based parameters was assessed by examining relationships with housing conditions. Thirty animal shelters housing a median of 63 cats each were studied. Cats were kept mostly in groups (mean size, 7). Adult cats housed in the shelter for at least 4 weeks (24 +/- 18 per shelter) were included. Test-retest reliability of the animal-based parameters was assessed by visiting 10 shelters twice approximately 2 months apart. To assess interrater reliability (IRR), 36 pens were evaluated by 2 raters. Two parameters were stable over time with acceptable IRR. The proportion of very thin cats (body condition score of 1/5) and the proportion of cats with a poor coat condition correlated highly with no significant difference between visits (%catsverythin: r(s) = 0.84, P = 0.002, Wilcoxon test: P = 0.225; % catspoorcoat: r(s) = 0.86, P = 0.002, Wilcoxon test: P = 0.128). IRR was moderate for % catsverythin (r(s) = 0.63, P < 0.001) and good for % catspoorcoat (r(s) = 0.87, P < 0.001). The physical condition of the cats was found to correlate with housing conditions; for example, an increased proportion of very thin cats correlated with a higher proportion of pens with less than 1 lying area per cat (r(s) = 0.54, P < 0.05) and with a lower proportion of pens with hiding places for all cats (r(s) = -0.54, P < 0.05). Poorer scores for coat condition correlated with longer mean length of stay in the shelter (r(s) = 0.46, P < 0.05), fewer cat toilets per cat (r(s) = -0.37, P < 0.05), and a more unpleasant odor (r(s) = 0.57, P < 0.05). Our results show that simple animal-based parameters can give an indication of how well cats cope with their environment and suggest that the use of these parameters, in addition to the assessment of conformance with legal requirements for shelter surveillance, could be beneficial for cat welfare. (C) 2015 Elsevier Inc. All rights reserved.