Inter-evaluator bias and applicability of feline body condition score from visual assessment

Emily C. Graff¹Christopher R Lea¹Diane Delmain¹Erin D Chamorro¹Xiaolei Ma¹Jingyi Zheng¹Yue Zhang¹Emily Brinker²Kenzii Kittell¹Mackenzie Hicks¹Casey Pfister¹Heather Hamilton¹Qinghong Li³Douglas Martin¹Xu Wang^1,4*

• ¹Auburn University, Auburn, United States
• ²Tufts University, Medford, Massachusetts, United States
• ³Purina Institute (United States), St Louis, Missouri, United States
• ⁴College of Veterinary Medicine, Auburn University, Auburn, Alabama, United States

Body Condition Score (BCS) is an effective tool for assessing body weight and fat mass, as well as diagnosing obesity and abnormal weight loss. A method for visual assessment of BCS in cats would be useful to expand access for feline health and research. The goal of this study is to determine whether BCS can be accurately assessed solely from photographs of cats, and to measure inter-evaluator bias in visually assessed BCS. To do this, a set of online-sourced cat images was administered as a quiz to nine evaluators. Inter-evaluator bias was relatively low (mean±SE=0.35±0.03) with ~ 50% complete agreement. To validate the results, a BCS was clinically assessed during routine wellness exams for 38 cats, enrolled, through palpation by one evaluator and visual assessment by all nine evaluators using photographs collected at the exam. The visual assessment of BCS deviated from the clinically assessed BCS by 0.61±0.04, which was slightly higher than the deviation observed in the online-sourced image set. In both scenarios, the majority voting among all evaluators achieved the highest accuracy, demonstrating its effectiveness in reducing evaluator bias. Inter-evaluator bias caused a 15.5% misclassification between ideal and overweight BCS but 1.8% between ideal and obese, indicating minimal bias in diagnosing feline obesity. The ability to accurately assess BCS through photographic evaluation will enhance remote consultations in telemedicine and support large-scale epidemiological studies. This study has developed a method for evaluating and minimizing inter-evaluator bias in BCS assessments across diverse practitioners and settings, thereby improving consistency and comparability and improving our understanding and application of BCS as a tool for feline health.