Buczinski S, Faure C, Jolivet S, Abdallah A
a Clinique Ambulatoire Bovine, Département des Sciences Cliniques , Faculté de Médecine Vétérinaire, Université de Montréal , CP 5000, St-Hyacinthe, Québec , Canada J2S 7C6.
b Ecole Nationale Vétérinaire de Toulouse (ENVT) , Université de Toulouse, Institut National Polytechnique (INP) , Toulouse , 31076 , France.
N Z Vet J. 2016 Jul;64(4):243-7. doi: 10.1080/00480169.2016.1153439. Epub 2016 Mar 2.
To determine inter-observer agreement for a clinical scoring system for the detection of bovine respiratory disease complex in calves, and the impact of classification of calves as sick or healthy based on different cut-off values.
Two third-year veterinary students (Observer 1 and 2) and one post-graduate student (Observer 3) received 4 hours of training on scoring dairy calves for signs of respiratory disease, including rectal temperature, cough, eye and nasal discharge, and ear position. Observers 1 and 2 scored 40 pre-weaning dairy calves 24 hours apart (80 observations) over three visits to a calf-rearing facility, and Observers 1, 2 and 3 scored 20 calves on one visit. Inter-observer agreement was assessed using percentage of agreement (PA) and Kappa statistics for individual clinical signs, comparing Observers 1 and 2. Agreement between the three observers for total clinical score was assessed using cut-off values of ≥4, ≥5 and ≥6 to indicate unhealthy calves.
Inter-observer PA for rectal temperature was 0.68, for cough 0.78, for nasal discharge 0.62, for eye discharge 0.63, and for ear position 0.85. Kappa values for all clinical signs indicated slight to fair agreement (<0.4), except temperature that had moderate agreement (0.6). The Fleiss' Kappa for total score, using cut-offs of ≥4, ≥5 and ≥6 to indicate unhealthy calves, was 0.35, 0.06 and 0.13, respectively, indicating slight to fair agreement.
There was important inter-observer discrepancies in scoring clinical signs of respiratory disease, using relatively inexperienced observers. These disagreements may ultimately mean increased false negative or false positive diagnoses and incorrect treatment of cases. Visual assessment of clinical signs associated with bovine respiratory disease needs to be thoroughly validated when disease monitoring is based on the use of a clinical scoring system.
确定用于检测犊牛牛呼吸道疾病综合征的临床评分系统的观察者间一致性,以及基于不同临界值将犊牛分类为患病或健康的影响。
两名三年级兽医学生(观察者1和2)和一名研究生(观察者3)接受了4小时关于对乳用犊牛呼吸道疾病体征进行评分的培训,包括直肠温度、咳嗽、眼鼻分泌物及耳部位置。观察者1和2在三次访问犊牛饲养设施期间,相隔24小时对40头断奶前乳用犊牛进行评分(80次观察),观察者1、2和3在一次访问中对20头犊牛进行评分。通过一致性百分比(PA)和卡方统计量评估个体临床体征的观察者间一致性,比较观察者1和2。使用≥4、≥5和≥6的临界值来表明不健康犊牛,评估三名观察者在总临床评分上的一致性。
直肠温度的观察者间PA为0.68,咳嗽为0.78,鼻分泌物为0.62,眼分泌物为0.63,耳部位置为0.85。所有临床体征的卡方值表明一致性为轻微至中等(<0.4),除温度为中等一致性(0.6)。使用≥4、≥5和≥6的临界值来表明不健康犊牛时,总评分的Fleiss卡方值分别为0.35、0.06和0.13,表明一致性为轻微至中等。
使用经验相对不足的观察者时,在对呼吸道疾病临床体征进行评分方面存在重要的观察者间差异。这些分歧最终可能意味着假阴性或假阳性诊断增加以及病例治疗错误。当基于临床评分系统进行疾病监测时,与牛呼吸道疾病相关的临床体征的视觉评估需要进行全面验证。