Wilson D J, González R N, Hertl J, Schulte H F, Bennett G J, Schukken Y H, Gröhn Y T
Department of Population Medicine and Diagnostic Sciences, Cornell University, Ithaca, NY 14850, USA.
J Dairy Sci. 2004 Jul;87(7):2073-84. doi: 10.3168/jds.S0022-0302(04)70025-9.
The objective of this study was to estimate the milk production losses associated with clinical mastitis using mixed linear models and correlation structures that have not been available previously. Data used included computer-recorded daily milk yields and detailed and accurate recordings of clinical mastitis cases. Two commercial Holstein dairy farms in New York State participated in the study, one with 650 lactating cows and another that began the study with 830 lactating cows and increased to 1120 cows by the end of the study. Cows on both farms were housed in free stall barns and milked 3 times daily in milking parlors. Electrical conductivity was used as a diagnostic aid for clinical mastitis on both farms. Date of clinical onset was recorded for every episode of clinical mastitis as well as for 8 other diseases defined using standardized case definitions (dystocia, milk fever, retained placenta, metritis, ketosis, displaced abomasum, lameness, and cystic ovarian disease) during the study period of October 1, 1999 to July 31, 2001. The mixed linear model for explaining variation in the outcome variable daily milk yield relative to non-mastitic herdmates found the terms for all 9 diseases studied, including clinical mastitis, significant. The model with an autoregressive correlation structure was preferred based on -2 * log likelihood, Akaike's information criterion, and Bayesian information criterion as well as savings in degrees of freedom. Separate analyses were run for first lactation cows and for second-plus lactation cows because their lactation curves were shaped differently. Adjusting for the effects of the other 8 diseases, milk production loss from clinical mastitis during the whole lactation was estimated as approximately 598 kg for second-plus lactation cows. However, cows that contracted mastitis had a daily production advantage of 2.6 kg over their herdmates until they contracted the disease. When compared with this potentially higher milk production, the total loss from clinical mastitis was estimated as 1181 kg.
本研究的目的是使用先前未曾有过的混合线性模型和相关结构，估算与临床乳腺炎相关的产奶量损失。所使用的数据包括计算机记录的每日产奶量以及临床乳腺炎病例的详细准确记录。纽约州的两个商业荷斯坦奶牛场参与了该研究，一个有650头泌乳奶牛，另一个在研究开始时有830头泌乳奶牛，到研究结束时增加到1120头。两个农场的奶牛都饲养在自由栏舍中，每天在挤奶厅挤奶3次。两个农场都将电导率用作临床乳腺炎的诊断辅助手段。在1999年10月1日至2001年7月31日的研究期间，记录了每例临床乳腺炎发作的临床发病日期以及使用标准化病例定义确定的其他8种疾病（难产、产乳热、胎盘滞留、子宫炎、酮病、真胃移位、跛行和卵巢囊肿病）的发病日期。用于解释相对于非患乳腺炎同群奶牛而言结果变量每日产奶量变化的混合线性模型发现，所研究的所有9种疾病（包括临床乳腺炎）的相关项均具有显著性。基于-2 *对数似然值、赤池信息准则、贝叶斯信息准则以及自由度节省情况，具有自回归相关结构的模型更受青睐。对初产奶牛和经产奶牛分别进行了分析，因为它们的泌乳曲线形状不同。在调整了其他8种疾病的影响后，经产奶牛整个泌乳期因临床乳腺炎造成的产奶量损失估计约为598千克。然而，患乳腺炎的奶牛在患病前比同群奶牛每日产奶量高2.6千克。与这种可能更高的产奶量相比，临床乳腺炎造成的总损失估计为1181千克。
