Bayesian latent class models to determine diagnostic sensitivities and specificities of two point of care rapid tests (Selma plus, Dipslide) for the detection of Streptococcus uberis associated with mastitis in dairy cows

Abstract

IntroductionDevelopment and validations of accurate mastitis diagnostics are crucial to make timely and evidence-based decisions on mastitis therapy in order to reduce its impact on productivity, animal welfare and practicing the prudent use of antimicrobials on dairy farms.MethodsThe objectives of this study were to assess the agreement between test results from reference laboratory and two point of care tests (Selma plus, Dipslide) and to estimate the test accuracies with Bayesian latent class models (BLCMs). In total of 509 single quarter milk samples from cows with mastitis were included in the study.ResultsAmong all analyzed mastitis pathogens, Streptococcus spp. was detected in up to one third of all analyzed samples and for Selma all Streptococcus samples were considered as Streptococcus uberis. The agreement (κ) when comparing two tests varied greatly depending on the bacteria, ranging from no agreement to good agreement (κ = negative to 0.86) depending on the prevalence of identified pathogens. Based on BLCMs to assess diagnostic test accuracies for the pathogen Streptococcus uberis, posterior sensitivities of 76, 71, and 64% for Selma plus, Dipslide and laboratory standard culture and specificities of 93%, 98% for Selma and Dipslide, respectively, were obtained.DiscussionThe two point of care rapid culture systems Dipslide and Selma plus plate can provide important preliminary pathogen identification for targeted mastitis therapy, especially when general information about growth and a rough classification of the bacteria into groups have an impact on treatment strategy. The two evaluated rapid culture systems, Dipslide and Selma plus plate, show good test accuracies for Streptococcus uberis at least at genus level. Therefore, using these tests may contribute to prudent use of antibiotics.