Skip to main content

Table 4 The predictors of thirty-day mortality

From: Clinical impact of multidisciplinary carbapenem stewardship interventions: a retrospective cohort study

Variables

Binary Logistic Regression

Multiple Logistics Regressiond

ORb (95% CI)

p value

ORc (95% CI)

p value

CCI > 3

2.49 (1.18, 5.25)

0.017a

2.84 (1.28, 6.29)

0.010

Febrile at day-7

4.02 (1.68, 9.64)

0.002a

4.58 (1.83, 11.5)

0.001

Increasing or elevated TWBC at day-7

2.46 (1.03, 5.85)

0.043a

  

Sepsis at baseline

1.94 (0.92, 4.09)

0.080a

  

Acceptance of intervention

0.87 (0.39, 1.92)

0.731

  

Bacteraemia at baseline

1.14 (0.43, 3.01)

0.787

  

Previous ICU admission

1.18 (0.33, 4.27)

0.797

  

Previous broad-spectrum antibiotics use

0.86 (0.31, 2.40)

0.769

  

Previous hospitalisation

0.97 (0.44, 2.17)

0.949

  

Previous MRO infection

1.02 (0.18, 5.80)

0.980

  
  1. No interactions and multi-collinearity detected. Correlation matrix = 0.183
  2. Model is fit; Hosmer–Lemeshow test (χ2 = 0.001, df = 2, p = 0.999)
  3. Classification table (overall correctly classified = 71.4%)
  4. Area under ROC curve = 0.690 (95% CI 0.560, 0.78; p < 0.001)
  5. aVariables included in the multiple logistic regression
  6. ORb: Crude odds-radio
  7. ORc: Adjusted odds-ratio
  8. CI: Confidence interval
  9. dFinal model after forward LR method applied (excluding sepsis and TWBC trend)