Els van de Vijver

55 Supplementary data 1: R scripts added as separate files JAGS script: ###Model correcting for Differential Verification Bias### model { ###Prior distributions### phi ~ dbeta(1,1) ST ~ dbeta(1,1) CT ~ dbeta(1,1) SR1 ~ dbeta(151.125,3.875)# Sensitivity range R1: 95-100% CR1 ~ dbeta(151.125,3.875) # Specificity range R1: 95-100% SR2 ~ dbeta(31.5,3.5) # Sensitivity range R2: 80-100% CR2 ~ dbeta(58.1,24.9) # Specificity range R2: 40-60% VT1R1 ~ dunif(0,1) VT0R1 ~ dunif(0,1) VT1R2 ~ dunif(0,1) VT0R2 ~ dunif(0,1) ###Probability data### p1 <- phi*ST+(1-phi)*(1-CT) p2 <- (SR1*phi*ST/(phi*ST+(1-phi)*(1-CT))+(1-CR1)*(1-phi)*(1-CT)/(phi*ST+(1-phi)*(1-CT))) p3 <- (SR1*phi*(1-ST)/(phi*(1-ST)+(1-phi)*CT)+(1-CR1)*(1-phi)*CT/(phi*(1-ST)+(1-phi)*CT)) p4 <- (SR2*phi*ST/(phi*ST+(1-phi)*(1-CT))+(1-CR2)*(1-phi)*(1-CT)/(phi*ST+(1-phi)*(1-CT))) p5 <- (SR2*phi*(1-ST)/(phi*(1-ST)+(1-phi)*CT)+(1-CR2)*(1-phi)*CT/(phi*(1-ST)+(1-phi)*CT)) ###Likelihood observed data### #Stage 1 T1 ~ dbin(p1,n) Head-to-head comparison Calgranulin C and Calprotectin 55