Stochastic modelling as a tool for planning animal-health surveys and interpreting screening-test results.

Traditionally, the planning of surveys (in particular, sample-size calculations) has relied on assumptions including the assumption of perfect screening tests. This paper presents a novel approach that can be used for planning animal-health surveys and interpreting screening-test results in the context of these surveys. A stochastic simulation model developed to assess the properties of herd-level sampling schemes and surveys has been adapted for large surveys aimed at substantiating freedom from infection at a national or regional level. We use a Bayesian approach to derive the post-survey probability of freedom from infection from the pre-survey probability of freedom and the likelihood ratio that is associated with screening-test results. We applied the model to two consecutive surveys conducted in 1998 and 1999 in Switzerland to substantiate freedom from infectious bovine rhinotracheitis (IBR) in the cattle population of about 56000 herds (median herd size of 15 cattle > 2 yr of age in 1999). In 1998, serum samples were taken from five cattle > 2 yr in 4672 herds, and in 1999 from all cattle > 2 yr old in 648 herds; samples were analysed by ELISA. The survey of 1999 provided less evidence than that of 1998 to support a status of freedom from infection; also, the characteristics of both herd-level sampling schemes were similar. We argue that the rationale for survey planning depends on the pre-survey probability of freedom from infection (i.e. our level of confidence that the infection does not occur in the targeted animal population). In consequence, surveys should be tailored to individual populations in the respective countries or regions. The model has been developed in an Excel spreadsheet to allow flexibility of use, and adaptation to many other animal-health issues.