Table 4

Multiple regression analysis of factors predicting hospital death and discharge to a care facility


Wald chi-square
Odds ratio (95% confidence interval)
P

Hospital deatha



Preadmission comorbidities



     Age per decade
22
3.2 (1.9–5.6)
<0.0001
     Female
6
1.3 (1.0–1.5)
0.0176
     Malignancy
18
1.7 (1.3–2.2)
<0.0001
Hospital and intensive care



     Medical admission
112
3.2 (2.6–3.9)
<0.0001
     Mechanical ventilation
74
2.6 (2.1–3.2)
<0.0001
     Maximal daily Sequential Organ Failure Assessment score >8
136
3.8 (3.0–4.8)
<0.0001
     Corticosteroid exposure (hydrocortisone, dexamethasone, methyl prednisone or prednisone)
11
1.4 (1.1–1.7)
0.001
Hospital diagnosis category



     Acute renal failure
81
3.0 (2.4–3.9)
<0.0001
     Cardiovascular
62
2.2 (1.8–2.7)
<0.0001
     Neurological
69
2.6 (2.1–3.3)
<0.0001
     Respiratory
16
1.5 (1.2–1.8)
<0.0001
Discharge to care facilityb



Preadmission comorbidities



     Age per decade
137
7.9 (5.6–11.1)
<0.0001
     Female
19
1.3 (1.1–1.4)
<0.0001
     Cerebral vascular disease
24
1.6 (1.3–2.0)
<0.0001
     Degenerative brain disease
212
2.7 (2.3–3.0)
<0.0001
     Congestive heart failure
13
1.3 (1.1–1.5)
0.00040
     Chronic obstructive pulmonary disease
4
1.1 (0.9–1.3)
0.05880
     Cachexia
11
1.6 (1.2–2.1)
0.00040
     Diabetes mellitus
6
1.2 (1.0–1.4)
0.01
Hospital and intensive care



     Medical admission
118
1.9 (1.7–2.1)
<0.0001
     Corticosteroid exposure



          Total hydrocortisone equivalent >900 mg
6
1.2 (1.0–1.4)
0.01
          Maximal daily Sequential Organ Failure Assessment score ≥ 4
21
1.3 (1.2–1.5)
<0.0001
     Hospital-acquired infection
124
2.2 (1.9–2.6)
<0.0001
     Antifungal antibiotics
10
1.3 (1.1–1.6)
0.00160
     Tracheostomy
63
2.9 (2.3–3.8)
<0.0001
Hospital diagnosis category



     Neurological
57
1.7 (1.5–2.0)
<0.0001
     Respiratory
10
1.2 (1.1–1.4)
0.00200
     Postoperative complications
28
1.5 (1.3–1.7)
<0.0001
     Fluid and electrolyte abnormalities
13
1.2 (1.1–1.4)
0.00030

aThe calibration of the multiple logistic model (lack of fit P = 0.9) and discrimination characteristic (area under the receiver operating characteristic curve = 0.88); P < 0.01 for all predictive factors. bThe calibration of the multiple logistic model (lack of fit P = 0.9) and discrimination characteristic (area under the receiver operating characteristic curve = 0.76); P < 0.01 for all predictive factors.

Rady et al. Critical Care 2006 10:R101   doi:10.1186/cc4971