Skip to main content

Table 2 Odds ratio (95% CI) for hospital mortality across treatment groups derived from unadjusted, partial and fully adjusted logistic regression models

From: Prehospital intravenous access and fluid resuscitation in severe sepsis: an observational cohort study

Number (%) with outcome

Model

All patients (Number = 1350)

No catheter or fluid (Number = 948)

Catheter only (Number = 90)

Catheter and fluid (Number = 312)

Odds ratio (95% CI) for catheter alonea

Odds ratio (95% CI) for catheter and fluida

Hospital mortality number (%)

      

 Unadjusted

205 (15)

120 (13)

14 (16)

71 (23)

1.27 (0.71, 2.28)

2.05 (1.72, 2.46)

 Adjusted for select variablesb

    

0.98 (0.51, 1.86)

1.26 (0.98, 1.63)

 Adjusted for all covariatesc

    

0.31 (0.17, 0.57)

0.46 (0.23, 0.88)

Increasing organ failures during hospitalization

      

 Unadjusted

485 (36)

265 (28)

47 (52)

173 (55)

2.72 (1.90, 3.90)

3.05 (2.67, 3.49)

 Adjusted for select variablesb

    

1.34 (0.49, 3.67)

1.66 (0.87, 3.17)

 Adjusted for all covariatesc

    

0.43 (0.21, 0.90)

0.58 (0.34, 0.98)

ICU admission

      

 Unadjusted

712 (53)

405 (42)

64 (71)

243 (78)

3.35 (2.03, 5.58)

4.50 (3.63, 5.59)

 Adjusted for select variablesb

    

1.81 (1.03, 3.17)

4.50 (3.63, 5.59)

 Adjusted for all covariatesc

    

0.41 (0.24, 0.70)

0.64 (0.37, 1.10)

  1. aCompared to referent group: no intravenous catheter or fluid; bpartial adjustment variables include age, gender, and initial prehospital heart rate, respiratory rate, Glasgow Coma Scale score, pulse oximetry, systolic blood pleasure; cfull adjustment includes partial adjustment variables, transport mode from scene, total scene time, transport time to hospital, prehospital procedures (for example, intubation, EKG monitoring, supplemental oxygen, bag-valve mask ventilation), EMS disease category (for example, cardiac, neurologic, respiratory), EMS call urgency (for example, life threatening, urgent, non-urgent). All estimates used imputed data after Rubin’s rules, including generalized estimating equations to account for within hospital clustering. Models shown for primary and secondary outcomes. CI, confidence interval; EKG, electrocardiogram; EMS, emergency medical services.