Table 3.

Performance assessment of multivariable models developed on development subset and validated on validation set (predicting outcomes based on measurements in validation set and comparing with known blood culture result)

Model name	Variables included	Percent correctly classified	Specificity (95 % CI)	NPV (95 % CI)	Sensitivity (95 % CI)	PPV (95 % CI)
Saturated multivariable model (based on univariate p < 0.15)	Malignancy, low systolic, low diastolic, high pulse, CRB 65 score, source	91 % (225/248)	0.98 (0.96, 1.00)	0.92 (0.88, 0.95)	0.05 (0.00, 0.25)	0.02 (0.01, 0.72)
Trimmed multivariable model (drop variables with multivariable p > 0.15)	Source, CRB 65 score	91 % (225/248)	Same as saturated	Same as saturated	Same as saturated	Same as saturated
Classification tree (only Source used)	All variables (continuous version if available)	90 % (225/250)	0.97 (0.94, 0.99)	0.92 (0.88, 0.95)	0.05 (0.00, 0.25)	0.14 (0.00, 0.45)