Novel methodologies for design and analysis of dairy health surveillance programs

乳制品行业是加拿大经济中非常重要的一部分，2018年的净农业收入为66亿美元。传染病已经在我们的畜群中出现，对加拿大奶牛场的盈利能力和可持续性造成了重大损失。这些疾病还对动物福利、抗菌剂的使用和食品安全产生影响。此外，在畜群中引入新的疾病会对畜群的盈利能力产生破坏性影响，也会影响我们向国际市场出口产品、动物和基因的能力。由于这些原因，现在对加拿大乳制品生产商来说，实施基于强有力的科学证据的动物健康监测和控制计划是至关重要的。我正在进行我的研究项目，其长期总体目标是改善监测项目和生物安全实践，以控制各种动物疾病。具体的短期目标将是：1)量化不完善的抽样策略对动物卫生监测系统性能的影响，并制定抽样和统计建模策略以防止偏差；2)验证适用于奶牛生产中几种重要疾病的动物卫生监测诊断策略；3)开发新颖的工具，方便灵活地对包含主动和被动动物卫生监测信息的多个大型省和国家数据集进行数据可视化。为了实现这些短期目标，我们将在三个主题下进行一系列研究：1)抽样策略；2)统计建模策略；3)数据可视化。“抽样策略”主题下的研究将有助于比较国家奶业研究战略中确定为高度优先事项的五种奶牛疾病（传染性乳腺炎、白血病、牛病毒性腹泻、副肺结核、沙门氏菌病）的不同抽样方法（例如，抽样材料的选择、每个牛群抽样的动物数量、抽样频率、抽样的牛群数量、诊断测试的选择）。“统计建模策略”主题下的研究将评估现有的，并提出专门为动物卫生监测系统设计的新的尖端统计建模方法。最后，“数据可视化”主题下的研究活动将导致简单灵活的数据可视化界面的发展，这将彻底改变我们目前与主动和被动监测数据交互的方式。拟议的研究计划将对加拿大工业产生重大影响。它将支持建立最先进的动物健康监测和改进方案。此外，开发的新型数据可视化工具将使乳品从业者和其他最终用户能够实时获得监测数据，从而能够对传染病威胁作出迅速反应。

The dairy sector represents a very important part of the Canadian economy with net farm receipts of $ 6.6 billion in 2018. Infectious diseases already present in our herds take a significant toll on profitability and sustainability of Canadian dairies. These diseases also have an impact on animal welfare, use of antimicrobials, and food safety. Moreover, the introduction of new diseases in a herd can have devastating impacts on profitability of that herd, but also on our ability to export our products, our animals, and our genetics on international markets. For these reasons, it is now crucial for our Canadian dairy producers to implement animal health surveillance and control programs that are based on strong scientific evidences. I am pursuing my research program with the long-term overarching goal of improving surveillance programs and biosecurity practices for control of various animal diseases. Specific short-term objectives will be to: 1)quantify the impacts of imperfect sampling strategies on the performances of animal health surveillance systems and develop sampling AND statistical modelling strategies for preventing biases; 2)validate fit for purpose diagnostic strategies for animal health surveillance for a few diseases of great importance in dairy production; 3)develop novel tools for easy and flexible data visualisation of multiple, large provincial and national datasets containing active and passive animal health surveillance information. To achieve these short-term objectives we will conduct a series of studies under three themes: 1) sampling strategies; 2) statistical modelling strategies; and 3) data visualisation. Research under the `sampling strategies' theme will help comparing different sampling approaches (e.g., choice of sampling materials, number of animals sampled per herd, frequency of sampling, number of herds sampled, choice of diagnostic tests) for five dairy cattle diseases identified as high priorities in the National Dairy Research Strategy (contagious mastitis, leucosis, bovine viral diarrhea, paratuberculosis, salmonellosis). Research under the `statistical modelling strategies' theme will evaluate existing, and propose novel cutting-edge statistical modelling approaches designed specifically for animal health surveillance systems. Finally, research activities under the `data visualisation' theme will lead to the development of easy and flexible data visualisation interfaces that will completely transform the way we currently interact with active and passive surveillance data. The proposed research program will have an outstanding impact on the Canadian industry. It will support the establishment of state-of-the-art animal health surveillance and improvement programs. Moreover, the novel data visualisation tools developed will make the surveillance data available in real-time for dairy practitioners and other end users, thus allowing for prompt response to infectious disease threats.