Table 2.

Bacteremia prediction capability indicated with AUCs^a of biomarkers (CRP/PCT) and models (random forest/logistic regression)

Methods/Group	CBC/DCb		CRP&CBC/DCc			PCT&CBC/DCd
	Cross-validation	Testing		Cross-validation	Testing		Cross-validation	Testing
Used biomarker	–	–	CRP	0.692 ± 0.017	0.699	PCT	0.748 ± 0.021	0.731
MLe models
Random forest	0.792 ± 0.010	0.802	CRP excludedf	0.797 ± 0.010	0.806	PCT excludedh	0.759 ± 0.022	0.767
			Includedg	0.806 ± 0.011	0.814	Included	0.777 ± 0.018	0.767
Logistic regression	0.763 ± 0.009	0.772	Excluded	0.769 ± 0.009	0.775	Excluded	0.735 ± 0.030	0.734
			Included	0.784 ± 0.011	0.790	Included	0.761 ± 0.024	0.745

^aAreas under the ROC curve

^bComplete blood count/differential leukocyte count

^cC-reactive protein and complete blood count/differential leukocyte count

^dProcalcitonin and complete blood count/differential leukocyte count

^eMachine learning

^fTrained and validated based on CBC/DC data of CRP&CBC/DC group (n = 253,009)

^gTrained and validated based on CBC/DC and CRP data of CRP&CBC/DC group

^hTrained and validated based on CBC/DC data of PCT&CBC/DC group (n = 17,033)