This study aimed to estimate the sensitivity (Se) and specificity (Sp) of loop-mediated isothermal amplification (LAMP) and single intradermal tuberculin (SIT) tests for the diagnosis of bovine tuberculosis (bTB) in dairy cattle in Thailand using a Bayesian approach. The SIT test was performed in 203 lactating dairy cattle from nine dairy farms located in Chiang Mai province, Thailand. Milk samples were collected for the LAMP test. Kappa analysis was performed to determine the agreement between the two tests. A one-population conditional independence Bayesian model was applied to estimate the Se and Sp of the two tests. Of 203 dairy cattle, 2 were positive for the SIT test using standard interpretation, whereas 38 were positive for the LAMP test. A poor agreement (kappa = 0) was observed between the two tests. The median Se and Sp of the SIT test using standard interpretation were 63.5% and 99.1%, respectively. The median Se and Sp of the LAMP test were 67.2% and 82.0%, respectively. The estimated true prevalence of bTB was 3.7%. The LAMP test with milk samples can potentially be used as a non-invasive screening test for the diagnosis of bTB in dairy cattle.
Keywords: Bayesian modeling; bovine tuberculosis; loop-mediated isothermal amplification; single intradermal tuberculin test; test performance.