Table 2

Measures of model calibration and discrimination in all admissions eligible for each model

Model

APACHE II

APACHE III

SAPS II

MPM II

ICNARC


Eligible admissions, n (percent)

8,344 (75.7)

9,021 (81.9)

8,220 (74.6)

7,267 (65.9)

10,285 (93.3)

AUC (95 percent CI)

0.73 (0.72, 0.75)

0.75 (0.74, 0.77)

0.81 (0.80, 0.82)

0.81 (0.79, 0.82)

0.82 (0.81, 0.83)

Hosmer-Lemeshow C*

χ2(10)a

316

4542

162

99.5

94.6

p value

<0.001

<0.001

<0.001

<0.001

<0.001

Cox's calibration regression

Intercept (95 percent CI)

-0.23 (-0.29, -0.18)

0.44 (0.36, 0.52)

0.09 (0.03, 0.15)

0.27 (0.20, 0.34)

0.18 (0.12, 0.24)

Slope (95 percent CI)

0.66 (0.62, 0.70)

0.58 (0.55, 0.62)

0.86 (0.82, 0.90)

1.00 (0.94, 1.05)

0.97 (0.93, 1.01)

χ2(2)

262

2,045

74.8

84.8

66.3

p value

<0.001

<0.001

<0.001

<0.001

<0.001

Brier's score

0.187

0.193

0.164

0.168

0.155


aNote that due to different sample sizes, the Hosmer-Lemeshow C*-statistic should not be directly compared between models. APACHE, Acute Physiology And Chronic Health Evaluation; AUC, area under the receiver operating characteristic curve; CI, confidence interval; ICNARC, Intensive Care National Audit & Research Centre; MPM, Mortality Probability Models; SAPS, Simplified Acute Physiology Score.

Hyam et al. Critical Care 2006 10(Suppl 2):S2   doi:10.1186/cc5066

Open Data