A Bayesian framework for estimating the incremental value of a diagnostic test in the absence of a gold standard

摘要: The absence of a gold standard, i.e., diagnostic reference standard having perfect sensitivity and specificity, is common problem in clinical practice research studies. There need for methods to estimate the incremental value new, imperfect test this context. We use Bayesian approach probability unknown disease status via latent class model extend two commonly-used measures based on predictive values [difference area under ROC curve (AUC) integrated discrimination improvement (IDI)] context where no exists. are illustrated using simulated data applied estimating novel interferon-gamma release assay (IGRA) over tuberculin skin (TST) tuberculosis (TB) screening. also show how IGRAs when decisions observed results rather than values. showed that greatest both specificity new better conditional dependence between tests reduces value. IGRA depends TST, as well prevalence TB, may thus vary different populations. Even statistics be estimated can aid about practical test.