Skip to main content

Table 6 Bayesian Number Needed to Diagnose

From: A Bayesian decision support sequential model for severity of illness predictors and intensive care admissions in pneumonia

Pre-test Probability

Post- test (LR+)

ADG

B-NND (rounded)

CURB 65 (14.0%)

PCT (alone) 43.0%

29.0%

3.45 (3)

CURB 65 (14.0%)

PCT and Lactate 93.0%

79.0%

1.27 (1)

CURB 65 (14.0%)

Lactate (alone) 76.0%

62.0%

1.61 (2)

  1. 1LR+: Positive Likelihood Ratio
  2. 1ADG: Absolute Diagnosis Gain
  3. 1B-NND: Bayesian Number Needed to Diagnose