Skip to main content

Table 3 Independent determinants of mortality by means of logistic regression analysis

From: Health inequities in the diagnosis and outcome of sepsis in Argentina: a prospective cohort study

Variable

Multivariable analysis

OR [95%CI]

Model 1

Model 2

Charlson score

1.22 [1.13–1.33] p < 0.01

1.29 [1.16–1.42] p < 0.01

Previous health state (EQ-VAS)

 Previous duration of disease

 

1.005 [1.004–1.010] p = 0.047

 Lactate (mmol/L)

1.20 [1.10–1.31] p < 0.01

1.28 [1.15–1.41] p < 0.01

 SOFA 24 h

1.13 [1.07–1.20] p < 0.01

1.14 [1.06–1.22] p < 0.01

 Mechanical ventilation utilization

8.61 [5.21–14.23] p < 0.01

12.91 [6.84–24.35] p < 0.01

 Highly resistant microorganisms†

 

1.76 [1.05–2.95] p = 0.032

 Admission to a public hospital

1.47 [1.00–2.17] p = 0.048

1.24 [.78–1.96] p = 0.360

  1. Model calibration and discrimination: model 1, which includes public hospital as a variable, has an area under the receiver operating characteristic (ROC) curve of 0.83 [0.80–0.86], with a Hosmer-Lemeshow test of 0.99. For model 2, values are 0.86 [0.83–0.89], respectively. Bivariate analysis is presented in Additional file 1: Table S2
  2. OR odds ratio, CI confidence interval, EQ-VAS EuroQol visual analogue scale (from 100 points [best health state to 0 worst] self-evaluated health state, previously to the diagnosis of sepsis), SOFA Sequential Organ Failure Assessment
  3. †Highly resistant microorganisms include methicillin-resistant S. aureus, vancomycin-resistant Enterococcus, P. aeruginosa, A. baumannii, and β-lactamase-producing Klebsiellae