灵敏度sensitivity:真阳性率,Se=a/(a+c)
特异度specificity:真阴性率,Sp=d/(b+d)
误诊率(mistake diagnostic rate,α),假阳性率,α=b/(b+d)=1-Sp
漏诊率(omission diagnostic rate, β),假阴性率,β=c/(a+c)=1-Se
准确度accuracy,粗一致性crude agreement rate=(a+b)/(a+b+c+d)
阳性预告值(positive predictive value, +PV)=a/(a+b)
阴性预告值(negative predictive value, -PV)=d/(c+d)
似然比(likelihood ratio, LR)患者中出现的概率与非患者中出现概率之比
阴性似然比(-LR)=Sp/β,越小,患病可能因越小