Table 3.
Univariate Logistic Regression Analysis for Variables Associated With ICU Admission and Independent Predictors of ICU Admission Determined by Multivariable Logistic Regression Analysis (n = 1664)a
| Univariate | Multivariableb | |||||
|---|---|---|---|---|---|---|
| Variable | OR | 95% CI | P Value | OR | 95% CI | P Value |
| Age ≥80 y | 0.32 | 0.24–0.42 | <.001 | 0.31 | 0.23–0.41 | <.001 |
| Male sex | 1.33 | 1.04–1.71 | .024 | - | - | - |
| Previous inhaled corticosteroids | 0.98 | 0.74–1.29 | .884 | - | - | - |
| Previous systemic corticosteroids | 1.16 | 0.66–2.03 | .606 | - | - | - |
| Antibiotic use in the last week | 0.90 | 0.68–1.19 | .442 | - | - | - |
| Chronic respiratory disease | 0.96 | 0.76–1.21 | .708 | - | - | - |
| Chronic cardiovascular disease | 0.76 | 0.54–1.05 | .099 | - | - | - |
| Chronic renal disease | 0.63 | 0.40–0.99 | .046 | - | - | - |
| Chronic liver disease | 1.16 | 0.67–2.00 | .593 | - | - | - |
| Diabetes mellitus | 0.85 | 0.65–1.11 | .224 | - | - | - |
| Chronic neurologic disease | 0.63 | 0.46–0.87 | .005 | - | - | - |
| Previous pneumonia | 0.70 | 0.49–0.99 | .043 | - | - | - |
| Nursing home resident | 0.62 | 0.39–0.99 | .044 | - | - | - |
| Fever | 0.91 | 0.70–1.18 | .474 | - | - | - |
| Deterioration in sensorium | 0.90 | 0.68–1.19 | .464 | - | - | - |
| C-reactive protein ≥15 mg/dL | 1.55 | 1.21–1.98 | <.001 | 1.39 | 1.07–1.79 | .012 |
| Lymphopenia (<724 lymphocytes/mm3) | 1.32 | 1.04–1.67 | .022 | 1.37 | 1.07–1.76 | .013 |
| Pleural effusion | 1.85 | 1.35–2.53 | <.001 | 1.81 | 1.30–2.52 | <.001 |
| Acute respiratory distress syndrome | 5.74 | 3.63–9.09 | <.001 | 6.14 | 3.78–9.96 | <.001 |
| Acute renal failurec | 1.22 | 0.96–1.55 | .108 | - | - | - |
| Streptococcus pneumoniae | 1.36 | 1.03–1.79 | .033 | - | - | - |
| Respiratory virus | 1.23 | 0.85–1.77 | .267 | - | - | - |
Data are shown as estimated ORs (95% CIs) of the explanatory variables in the sepsis group. The OR represents the odds that the presence of ICU admission will occur given exposure of the explanatory variable, compared with the odds of the outcome occurring in the absence of that exposure. The P values are based on the null hypothesis that all ORs relating to an explanatory variable equal unity (no effect).
aExcluded 236 patients with septic shock, 34 patients with missing data regarding septic shock, 211 patients who had do-not-resuscitate orders, and 58 with missing data regarding a do-not-resuscitate order.
Abbreviations: ICU, intensive care unit; OR, odds ratio.
bHosmer-Lemeshow goodness-of-fit test, P = .597.
cVariable highly correlated with another independent variable and therefore not included in the multivariable model.