A new scoring system derived from base excess and platelet count at presentation predicts mortality in paediatric meningococcal sepsis

Table 4 Base excess and platelet count (BEP) performance for different cutoffs on the development, validation and replication datasets.

Dataset	BEP	N	Sensitivity	Specificity	Positive predictive value	Negative predictive value
	> 0.1	50	0.65	0.88	0.29	0.97
	> 0.2	21	0.44	0.96	0.45	0.96
Development	> 0.3	8	0.26	0.99	0.6	0.94
	> 0.4	4	0.15	1	0.75	0.94
	> 0.5	2	0.05	1	0.55	0.93
	> 0.1	110	0.67	0.87	0.29	0.97
	> 0.2	64	0.57	0.94	0.43	0.96
Validation	> 0.3	30	0.35	0.98	0.55	0.95
	> 0.4	21	0.29	0.99	0.68	0.95
	> 0.5	11	0.17	1	0.78	0.94
	> 0.1	25	0.85	0.9	0.4	0.99
	> 0.2	16	0.8	0.97	0.7	0.98
Replication	> 0.3	11	0.56	0.98	0.72	0.97
	> 0.4	9	0.48	0.99	0.79	0.96
	> 0.5	6	0.36	1	1	0.95

N denotes the number of individuals with BEP score greater than the selected cutoff (for example on the validation dataset there are 110 subjects with a BEP > 0.1).

ISSN: 1364-8535