Estimation of paratuberculosis prevalence in dairy cattle in a province of Korea using an enzyme-linked immunosorbent assay: application of Bayesian approach.

Abstract

To draw inferences about the sensitivity and specificity of the newly developed ELISA test for bovine paratuberculosis (PTB) diagnosis and posterior distribution on the prevalence of PTB in a province of Korea, we applied Bayesian approach with Gibbs sampler to the data extracted from the prevalence study in 1999. The data were from a single test results without a designated gold test. The prevalence estimates for PTB in study population ranged 3.2-5.3% for conservative and 6.7-7.1% for liberal, depending on the priors used. The simulated specificities of the ELISA close to one another, ranging 84.7-90.6%, whereas the sensitivity was somewhat spread out depending largely on the priors with a range of 46.4-88.2%. Our findings indicate that the ELISA method appeared useful as a screening tool at a minimum level in comparison to other diagnostic tests available for this disease in terms of sensitivity. However, this advantage comes at a cost of having low specificity of the test.