TABLE 2.

Ability of the Model to Triage PICU Admissions

Predicted Probability of Death	Positive Likelihood Ratio (95% CI)	Negative Likelihood Ratio (95% CI)	Per 1000 Admissions (~63 of Which Would Die)				Percentage of Admissions Triaged As High-Acuity (Ratio of Incorrect to Correct High-Acuity Triages)
			True Positive (High-Risk Admission Triaged to High-Acuity Area)	False Positive (Low-Risk Admission Triaged to High-Acuity Area)	True Negative (Low-Risk Admission Triaged to Low-Acuity Area)	False Negative (High-Risk Admission Triaged to Low-Acuity Area)
2.5%	1.18 (1.14–1.21)	0.07 (0.01–0.46)	62	788	150	1	85.0% (13:1)
5%	2.31 (2.08–2.57)	0.23 (0.14–0.37)	54	347	590	9	40.1% (6:1)
7.5%	3.27 (2.73–3.91)	0.44 (0.33–0.57)	41	186	751	22	22.7% (5:1)
10%	5.75 (4.57–7.23)	0.47 (0.37–0.59)	36	94	844	26	13.0% (3:1)
15%	11.43 (8.22–15.88)	0.56 (0.46–0.67)	29	39	899	34	12.3% (1:1)

Performance of the model at five cutoffs (decision thresholds or threshold probabilities). A cutoff of 10% reflects a triage strategy whereby all admissions with a predicted probability of death ≥ 10% are directed to a high-acuity area and all other admissions managed on the main unit. A decrease in threshold probability (cutoff) is associated with an increase in the sensitivity of the triage strategy for identifying high-risk admissions, at the cost of a greater proportion of admissions being directed to the high-acuity area. Additional classification indices are provided in Appendix 18 (http://links.lww.com/PCC/C445).