Optimal Sample Size and Allocation in Disease Diagnostic Testing with Multi-Level Structures

• Activity Number: 533
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: International Chinese Statistical Association
• Author(s): Chong Wang* and Yinan Fang and Jeffrey Zimmerman
• Companies: Iowa State University and Iowa State University and Iowa State University
• Keywords: sample size ; sample allocation ; power ; disease freedom ; multi-level structures

Abstract

In animal disease testing at the population level, traditional sample size calculation methods are based on the assumption that individuals are independent and have the same chance of acquiring disease. In contemporary animal production systems, however, a hierarchical structure exists. For example, a farm may consist of multiple buildings and each building usually contains multiple pens of animals. Besides the accuracy concern of traditional sample size formulas with such hierarchical structures, there is also a lack of guidance about how to subsequently allocate sampling across strata. Despite the fact that strata are not necessarily homogeneous in terms of disease status, it is common for samples to be collected from selected strata, leaving other strata unsampled. For optimal disease detection at the farm level, a mathematically more intuitive way to allocate samples would be even distribution across the different strata. The objective of this project is to address the sample size and allocation questions with multi-level structures from a mathematical perspective.