TABLE 2.
Model | Variable | AUC (95% CI), Derivation Cohort (n = 10,085) | AUC (95% CI), Validation Cohort (n = 5,037) | Logistic Regression, OR (95% CI)a |
---|---|---|---|---|
SOFAb | SOFA | 0.66 (0.65–0.67) | 0.66 (0.64–0.67) | 1.39 (1.35–1.42) |
SOFA categoriesc | ≥ 6 and < 9 | 0.55 (0.54–0.55) | 0.54 (0.54–0.55) | 3.42 (2.89–4.06) |
≥ 9 and < 12 | 4.73 (3–7.82) | |||
≥ 12 | 5.89 (1.96–25.32) | |||
Age | Age | 0.71 (0.7–0.72) | 0.71 (0.69–0.72) | 1.06 (1.06–1.06) |
Age + SOFA categories | Age | 0.73 (0.72–0.74) | 0.72 (0.71–0.73) | 1.06 (1.05–1.06) |
≥ 6 and < 9 | 3.18 (2.67–3.82) | |||
≥ 9 and < 12 | 4.92 (3.05–8.3) | |||
≥ 12 | 6.66 (2.13–29.31) | |||
Age + SOFA | Age | 0.75 (0.74–0.76) | 0.74 (0.73–0.76) | 1.06 (1.05–1.06) |
SOFA | 1.33 (1.3–1.36) | |||
SOFA + age + covariatesd | SOFA | 0.75 (0.74–0.76) | 0.74 (0.73–0.76) | 1.33 (1.29–1.36) |
Age | 1.06 (1.05–1.06) | |||
Gender (male vs female) | 1.15 (1.05–1.25) | |||
Obesity | 0.92 (0.79–1.08) | |||
Diabetes | 1.18 (1.05–1.32) | |||
Hypertension | 0.84 (0.75–0.93) | |||
SOFA + age + Elixhauser score | Age | 0.75 (0.74–0.76) | 0.74 (0.73–0.76) | 1.06 (1.05–1.06) |
Elixhauser score | 1 (0.97–1.03) | |||
SOFA | 1.33 (1.3–1.36) | |||
SOFA + Elixhauser score | SOFA | 0.66 (0.65–0.67) | 0.66 (0.65–0.68) | 1.38 (1.35–1.42) |
Elixhauser score | 1.04 (1.01–1.07) | |||
Categories SOFA + Elixhauser score | ≥ 6 and < 9 | 0.57 (0.56–0.58) | 0.56 (0.55–0.58) | 3.33 (2.82–3.96) |
≥ 9 and < 12 | 4.59 (2.91–7.59) | |||
≥ 12 | 5.75 (1.91–24.76) | |||
Elixhauser score | 1.06 (1.03–1.09) |
AUC = area under the receiver operating curve, OR = odds ratio, SOFA = Sequential Organ Failure Assessment.
Outcome variable has two levels: deceased and discharged, where discharged is served as the reference level.
SOFA and all components are the scores recorded within the 24 hr prior to the start of ventilation.
Variable SOFA category is created based on SOFA variable with the cutoffs of: < 6, ≥ 6 and < 9, ≥ 9 and < 12, and ≥ 12. In the logistic regression model, SOFA < 6 is served as the reference group.
Covariates include age, gender, obesity, hypertension, and diabetes.