Table 1.
Characteristics of study participants.
Fig 1.
STARD flow diagram showing results of dipstick LE and nitrite in the classical analyses.
Table 2.
Comparison of diagnostic test accuracy estimated by classical and Bayesian latent class analyses.
Fig 2.
Post-test probability of UTI across varying pre-test probability, for urine leukocyte esterase (Panel A), nitrite (Panel B), and urine culture (Panel C).
Solid squares connected by continuous line indicate post-test probability based on accuracy estimated using Bayesian latent class analysis; Solid circles connected by dashed line indicate post-test probability based on accuracy estimated using classical analysis; blue squares and circles indicate post-test probability when the test is positive, and pink squares and circles indicate post-test probability when the test is negative. Sensitivity and specificity of urine culture cannot be assessed in the classical analysis since it is used as the reference standard.