China Animal Husbandry and Veterinary Medicine ›› 2023, Vol. 50 ›› Issue (8): 3267-3274.doi: 10.16431/j.cnki.1671-7236.2023.08.024

• Preventive Veterinary Medicine • Previous Articles    

Verification and Application of Dairy Cow Mastitis Risk Assessment System in Pastures in North and South of China

BAI Yunfei1, ZHAO Tingting2, LI Wenlong1, XING Yue1, REN Xiaoli3, GUO Gang2, YU Ying1   

  1. 1. College of Animal Science and Technology, China Agricultural University, Beijing 100193, China;
    2. Beijing SunLon Livestock Development Co., Ltd., Beijing 100076, China;
    3. Henan Dairy Herd Improvement Center, Zhengzhou 450046, China
  • Received:2022-12-13 Published:2023-07-27

Abstract: 【Objective】 The aim of this study was to explore the accuracy and stability of dairy cow mastitis risk assessment system in pastures in North and South of China in recent years,and observe whether the system could be stably applied to predict mastitis of Chinese Holstein dairy.【Method】 Based on Logistic regression model of dairy cow mastitis (cow mastitis logistic regression model,CMLM) by the author’s research group had built,large scale horizontal and vertical data validation was conducted in 6 Holstein dairy farms in Beijing,Henan and Zhejiang.The dairy herd improvement (DHI) information of a total of 56 985 lactating dairy cows in 5 Holstein dairy farms in Northern region (Beijing and Henan) and 1 Holstein dairy farm in Southern region (Zhejiang) was used as the validation data.The prediction ability of risk assessment system for dairy cow mastitis was tested.More than 270 thousand DHI measurement records of pastures in Northern China from 2016 to 2021 were used to test the validation effect of CMLM on large scale vertical data in different time periods.【Result】 CMLM had a good accuracy in the validation data of pastures in Beijing,with an average accuracy of 71.85%.CMLM had a great actual application performance in the validation data of pastures in Henan which was new,with an average accuracy of 78.67%.CMLM had a stable actual application performance in the validation data of small-scale pasture in Southern region for 180 months,with an average accuracy of 77.63%.CMLM also had high predictive power at different time stages.The predictive value of the model was concentrated 0.70-0.90.It could be considered that CMLM was good.【Conclusion】 The CMLM system had excellent performance in the application of Holstein dairy herds in South and North of China,and had a high predictive effect in the verification of large scale horizontal and vertical data,which had a good promotion and application value.

Key words: Chinese Holstein dairy; mastitis; risk assessment; Logistic regression model; model validation

CLC Number: