Prevalence and sensitivities, specificities, and positive and negative predictive values (PPV's and NPV's) of diagnostic tests using the reference assay as gold standard and for final Bayesian latent class models.

aPrevalence and accuracy of each test was estimated by the observed proportion classified by considering that the reference assay was perfect (100% sensitivity and 100% specificity). 95% confidence intervals (CI) were obtained in Stata 11.1.bPrevalence and accuracy of each test was estimated by Bayesian latent class models by considering that the reference assay could be imperfect. Posterior estimates and 95% credible intervals (CrI) of each parameter were obtained in WinBUGs from 10,000 iterations of each of two chains starting from different initial values following a burn-in period of 5,000 iterations.cA combination considers that positivity of either test is positive for dengue infection.NA = Not applicable.