Supplementary data for: Sensitivity and specificity evaluation of a consensus-derived core outcome set for the diagnosis of respiratory disease in pre-weaned dairy calves: protocol for a diagnostic accuracy trial using a Bayesian latent-class modelling approach

The primary objective of this trial is to estimate (using a template from Center for Animal Disease Modeling and Surveillance at UC Davis Veterinary Medicine) the sensitivity (Se) and specificity (Sp) of a consensus derived clinical scoring system (CSS) and combination of CSS and LUS for the diagnosis of pneumonia in pre-weaned dairy calves, using a validated LUS methodology as the imperfect reference standard. A Bayesian latent-class modelling approach is planned.

Secondary objectives are to identify potential optimal diagnostic thresholds for continuously measured variables within the CSS and to estimate optimal weights for each component of the CSS alone and the CSS + LUS combined case definitions. The data set shows the effect of different parameter values (e.g., expected prevalence of 1cm of lung consolidation in replacement calves) on sample size calculation.