ABSTRACT
International trade regulations require disease freedom to be documented. This is generally achieved by national random surveys. Risk-based sample size calculations can reduce both the sample size and the costs for repeated surveys by one-third compared to traditional sample size calculations based on random sampling. In this approach, information from previous surveys can reduce the sample size of a new survey while maintaining its overall level of confidence. Risk-based sample size calculations have only formerly been described for one single survey and not when applied consecutively. Based on the existing deterministic origin of risk-based methods and sample size calculations, we devised a further development of sample size calculation to document freedom from non-highly contagious diseases. We show that for a repeated risk-based sample size calculation not only does the loss of confidence due to the risk of disease import and the risk of undetected disease within the country need to be calculated between the two surveys, but the time value of historical testing information also needs to be considered. As the order of surveys is important, a separate process, capable of calculating the time value of information from all conducted surveys, needs to be incorporated in the sample size calculation. This paper suggests how to use information from previous surveys in a risk-based approach for sample size calculations in consecutive surveys considering the adjusted time value of historical testing information, import risk assessments and the risk of disease spread within the country. Thus, the reduction of confidence between two surveys is adequately estimated. The results of sample size calculations for surveys over various years, under different conditions, show that the risk-based approach, including import risk assessments and adjusted time value of historical testing information, could reduce the sample size of annual surveys by 25% compared to the sample sizes calculated without a risk-based technique. The presented risk-based sample size calculation is appropriate for the requirements of consecutive surveys and outlines how the adjusted time value of historical testing information should be considered in this context. When planning to undertake repeated surveys to document freedom from disease, this method can easily be used to calculate sample sizes and consequently reduce them.
