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

Table 3 Quantitative bias analysis illustrating the odds of mortality under varying assumptions

Odds ratio of hypothetical confounder^a	Prevalence of unmeasured confounder among patients not receiving prehospital catheter or fluid
Odds ratio of hypothetical confounder^a	0.075	0.1	0.2	0.3
1.25	0.46 (0.23,0.88)	0.46 (0.23, 0.89)	0.47 (0.24, 0.92)	0.48 (0.24, 0.94)
1.50	0.46 (0.23, 0.89)	0.46 (0.23, 0.90)	0.48 (0.24, 0.95)	0.50 (0.26, 0.99)
1.75	0.46 (0.23, 0.90)	0.47 (0.64, 0.92)	0.50 (0.25, 0.98)	0.53 (0.27, 1.04)
2.00	0.46 (0.23, 0.90)	0.47 (0.24, 0.93)	0.51 (0.26, 1.01)	0.56 (0.28, 1.09)
2.25	0.46 (0.24, 0.91)	0.48 (0.24, 0.94)	0.53 (0.27, 1.04)	0.58 (0.30, 1.14)
2.50	0.47 (0.24, 0.91)	0.48 (0.24, 0.95)	0.54 (0.28, 1.07)	0.61 (0.31, 1.19)

^aAssumptions: 1) prevalence of unmeasured confounder among patients receiving prehospital fluid = 0.05; 2) no modification of the effect of prehospital fluid by the unmeasured confounder; 3) confounder uncorrelated with other variables in the model. The upper most left-hand corner is the observed odds ratio and bolded cells are conditions in which the observed odds ratio is no longer significant.

ISSN: 1364-8535