BackgroundEnzyme‐linked immunosorbent assays (ELISAs) are the most widely used diagnostic tools in bovine paratuberculosis (bPTB) control. However, their diagnostic accuracy may be compromised by bovine tuberculosis (bTB) infection, as both diseases share diagnostic targets.MethodsThe bPTB and bTB infection status of 228 animals was determined using microbiological tissue culture as a reference test. The diagnostic performance (sensitivity, specificity, likelihood ratios and predictive values) of the bPTB‐ELISA on blood serum samples, taking into account the bPTB animal‐level prevalence of the area and the bTB status of the animals, was evaluated.ResultsA sensitivity of 40.7% (95% confidence interval [CI]: 27.5%–53.9%) and a specificity of 94.7% (95% CI: 91.4%–98.0%) were obtained for bPTB‐ELISA in all animals. A bPTB‐ELISA‐positive animal would have a post‐test probability of 70% or more of being infected in areas with a bPTB prevalence of 23% or more. A negative bPTB‐ELISA result, in areas with a bPTB prevalence of 41% or less, would rule out the disease with more than 70% certainty. In bTB‐positive animals, sensitivity increased (94.4% [95% CI: 81.4%–100%] vs. 25.1% [95% CI: 11.8%–38.4%]) and specificity decreased (82.6% [95% CI: 71.8%–93.4%] vs. 99.4% [95% CI: 98.0%–99.9%]). The bPTB‐ELISA is a good tool to rule out bPTB co‐infection in bTB‐positive animals, while in bTB‐negative animals, it allows confirmation of disease with more than 70% probability if disease prevalence is 6% or more.LimitationsThe observed differences could be enhanced by the effect of frequent application of the intradermal tuberculin test, which was unknown in the animals studied.ConclusionsThese results provide useful guidance for the application and interpretation of ELISA as a tool for bPTB disease control.