Skip to main content

Table 4 Discrimination and calibration of models in training dataset (n = 6798)

From: Comparison of measures of comorbidity for predicting disability 12-months post-injury

Model

Comorbidity measure

Area under curve (95% CI)

H-L statistic (p-value)

LR test (p-value)

Injury group

None

0.631 (0.618, 0.644)

0.67 (0.954)

-

Injury group and age group

None

0.704 (0.692, 0.716)

4.04 (0.854)

507.40 (<0.0001)

Injury group, age group and comorbidity

Number of ICD-10 chapters

0.728 (0.716, 0.740)

11.90 (0.156)

228.23 (<0.0001)*

Injury group, age group and comorbidity

ICD-10 chapters

0.729 (0.717, 0.741)

8.51 (0.386)

251.38 (<0.0001)*

Injury group, age group and comorbidity

Number of Haagsma conditions

0.724 (0.712, 0.736)

7.51 (0.482)

182.06 (<0.0001)*

Injury group, age group and comorbidity

Haagsma conditions

0.725 (0.713, 0.737)

7.05 (0.531)

194.03 (<0.0001)*

Injury group, age group and comorbidity

CCI weight category

0.720 (0.708, 0.732)

1.30 (0.996)

159.35 (<0.0001)*

Injury group, age group and comorbidity

FCI score

0.716 (0.704, 0.728)

7.10 (0.526)

102.39 (<0.0001)*

Injury group, age group and comorbidity

FCI conditions

0.716 (0.704, 0.728)

8.05 (0.429)

-

  1. *compared to model including age group and injury group.