Determination of sample size to serological surveillance plan for pullorum disease and fowl typhoid

Pak, Son-Il; Park, Choi-Kyu

Korean Journal of Veterinary Research 2008;48(4):457-462.

¹School of Veterinary Medicine and Institute of Veterinary Science, Kangwon National University
²National Veterinary Research and Quarantine Service

추백리-가금티푸스의 혈청학적 모니터링 계획수립을 위한 표본크기

박선일¹, 박최규²

¹강원대학교 수의학부대학 및 동물의학종합연구소
²국립수의과학검역원

Abstract

The objective of this study was to determine appropriate sample size that simulated different assumptions for diagnostic test characteristics and true prevalences when designing serological surveillance plan for pullorum disease and fowl typhoid in domestic poultry production. The number of flocks and total number of chickens to be sampled was obtained to provide 95% confidence of detecting at least one infected flock, taking imperfect diagnostic tests into account. Due to lack of reliable data, within infected flock prevalence (WFP) was assumed to follow minimum 1%, most likely 5% and maximum 9% and true flock prevalence of 0.1%, 0.5% and 1% in order. Sensitivity were modeled using the Pert distribution: minimum 75%, most likely 80% and maximum 90% for plate agglutination test and 80%, 85%, and 90% for ELISA test. Similarly, the specificity was modeled 85%, 90%, 95% for plate agglutination test and 90%, 95%, 99% for ELISA test. In accordance with the current regulation, flock-level test characteristics calculated assuming that 30 samples are taken from per flock. The model showed that the current 112,000 annual number of testing plan which is based on random selection of flocks is far beyond the sample size estimated in this study. The sample size was further reduced with increased sensitivity and specificity of the test and decreased WFP. The effect of increasing samples per flock on total sample size to be sampled and optimal combination of sensitivity and specificity of the test for the purpose of the surveillance is discussed regarding cost.

Key Words: fowl typhoid, pullorum disease, sample size, serosurveillance, simulation