Table 4.
Independent predictors of intensive care unit length of stay of 30 days or longer by logistic regression analysis
| Parameter | β | SE | P value | Exp (β) |
|---|---|---|---|---|
| Constant | -5.4 | 0.37 | <0.001 | 0.004 |
| Male gender | -0.53 | 0.22 | 0.02 | 0.59 |
| Injury severity score | 0.04 | 0.008 | <0.001 | 1.04 |
| Pneumonia | 1.58 | 0.20 | <0.001 | 4.87 |
| Arrest | 1.35 | 0.47 | 0.004 | 3.87 |
| Sepsis | 1.80 | 0.26 | <0.001 | 6.02 |
| Urinary tract infection | 1.56 | 0.22 | <0.001 | 4.75 |
| Deep vein thrombosis | 1.10 | 0.37 | 0.003 | 2.99 |
| Arrhythmias | 0.90 | 0.26 | 0.001 | 2.45 |
| Acute respiratory distress syndrome | 2.29 | 0.23 | <0.001 | 9.83 |
| Gastrointestinal bleed | 0.93 | 0.42 | 0.03 | 2.54 |
| Respiratory | 1.51 | 0.20 | <0.001 | 4.53 |
Cox-Snell R square = 0.16
Nagelkerke R square = 0.54.
β = logistic coefficient (parameter estimate); Exp (β) = odds ratio; SE = standard error of logistic coefficient.