Development of a prediction model for bacteremia in hospitalized adults with cellulitis to aid in the efficient use of blood cultures: a retrospective cohort study

Table 5 Clinical scores and their associated probabilities of true bacteremia, sensitivity, and specificity

Clinical score	Probabilities of true bacteremia (%)^a	Percentage correctly having blood culture(s) (sensitivity)^b	Percentage correctly rejecting blood culture(s) (specificity)^c
7.0	N/A	0	100
6.0	N/A	0	100
5.5	66.7 (2/3)	6.1	99.7
5.0	38.6 (5/13)	15.2	97.5
4.5	42.1 (8/19)	24.2	96.5
4.0	41.2 (14/34)	42.4	93.7
3.5	32.1 (27/84)	81.8	82.1
3.0	31.0 (27/87)	81.8	81.1
2.5	21.6 (30/139)	90.9	65.7
2.0^d	20.7 (31/150)	93.9	62.6
1.5	17.4 (32/184)	97	52.2
1.0	11.9 (32/269)	97	25.5
0	9.4 (33/351)	100	0

^aIndicates the probabilities of true bacteremia in cellulitis cases using the clinical score in that row as a cut-off point to determine whether to order a blood culture. In parenthesis are the number of cellulitis cases with true bacteremia (numerator) and total number of cellulitis cases (denominator) using the clinical score in that row as a cut-off point
^bFraction of patients with true bacteremia who would correctly have blood culture(s) if the clinical score in that row was used as a cut-off point
^cFraction of patients without true bacteremia who would correctly reject blood culture(s) if the clinical score in that row was used as a cut-off point
^dClinical score of 2.0 was used to illustrate its use for decision-making