Table 5

Multivariate logistic regression analysis for death within 1500 days after surgery for patients admitted in both the control and MET periods

Odds ratio

95% CI

P value

VIF


Age (per 10 years)

1.73

1.61–1.86

<0.001

1.09

MET perioda

0.74

0.60–0.92

0.005

1.02

Female sex

0.72

0.58–0.89

0.002

1.05

Nonscheduled surgery

1.51

1.21–1.89

<0.001

1.18

Admitting unit

Cardiac surgery

0.33

0.22–0.48

<0.001

1.28

Thoracic surgery

2.76

1.89–4.03

<0.001

1.36

Orthopaedic surgery

0.50

0.37–0.69

<0.001

1.43

Vascular surgery

0.66

0.46–0.94

0.023

1.20

Neurosurgery

1.60

1.03–2.48

0.036

1.63

Liver transplant unit

0.30

0.09–1.03

0.056

1.05

Oncology

10.48

1.2–91.65

0.034

1.01

General medicine

3.00

1.13–7.95

0.027

1.02

Procedure

Bronchoscopy

3.13

1.24–7.88

0.015

1.16

Breast surgery

0.20

0.04–0.95

0.043

1.04

Spinal fusion

0.34

0.17–0.68

0.002

1.21

Cerebral resection

3.30

1.6–6.77

0.001

1.33

Cystoscopy ± bladder tumour resection

2.90

1.04–8.08

0.042

1.02


The characteristics of the operation clusters are outlined in Table 1. For this model, the Hosmer-Lemeshow goodness-of-fit statistic was 13.1 (P = 0.11). The percentage of appropriately classified patients in the final model is 75.3%. aThe OR and 95% CI for introduction of the MET pertains to that derived from the multivariate model. CI, confidence interval; OR, odds ratio; VIF, variance inflation factor.

Jones et al. Critical Care 2007 11:R12   doi:10.1186/cc5673

Open Data