The Efficacy of an Oscillating Bed in the Prevention of Lower Respiratory Tract Infection: RESULTS part 2
Since including patients who died early in their course might confound interpretation of LOS (in SICU and hospital) and duration of intubation, we analyzed these variables in two ways (ie, for all patients and for survivors only) (Table 2); LOS in the SICU was shorter for patients in the RRKTT group. The difference between groups for this variable approached statistical significance (p = 0.0645 and p = 0.0535, for all patients and survivors, respectively). Hospital LOS and dura¬tion of intubation were significantly shorter for patients in the RRKTT group.
Shown in Table 3 are significant risk factors for LRTI and pneumonia by univariate analysis. The following conditions were not significant risk factors for either outcome: age >=60 years; APACHE-II >=15; male sex; pulmonary contusion; major intra-thoracic trauma; major intraabdominal trauma; femur fracture; intubation. Some of these conditions (particularly intubation) were probably not shown to be significant because the number of patients without the risk factor was very small. buy levitra 20 mg
Table 2—Effect of Postural Oscillation on Incidence of LRTI and LOS
|
|
Conventional Bed |
RRKTT Bed |
p Value |
|
All LRTI |
|
|
|
|
All patients* |
28/48 (58.3%) |
13/51 (25.5%) |
0.0011 |
|
Patients with |
17/22 (77.3%) |
6/14 (42.9%) |
0.0732 |
|
Patients without |
11/26(42.3%) |
7/37 (18.9%) |
0.0526 |
|
Pneumonia |
|
|
|
|
All patients* |
19/48 (39.6%) |
7/51 (13.7%) |
0.0056 |
|
Patients with |
11/22 (50.0%) |
1/14 (7.1%) |
0.0111 |
|
Patients without |
8/26(30.8) |
6/37 (16.2%) |
0.2229 |
|
LOS in SICU (days) |
|
|
|
|
All patientsf |
8 (2-74) |
5 (1-32) |
0.0645 |
|
Survivors onlyt |
8 (2-43) |
5 (2-32) |
0.0535 |
|
LOS in hospital (days) |
|
|
|
|
All patientsf |
37 (5-612) |
20 (2-201) |
0.0138 |
|
Survivors onlyt |
44.5 (9-246) |
20(6-80) |
0.0020 |
|
Days intubated |
|
|
|
|
All patientsf |
7 (0-74) |
4 (0-32) |
0.0501 |
|
Survivors onlyt |
7(0-41) |
4(0-32) |
0.0267 |
4 shows the results of a stepwise logistic regression analysis, using probability of LRTI as the dependent variable. Randomization to a conventional bed increased the risk of LRTI roughly twofold (p = 0.004). In this multivariate analysis, both major head trauma and severe, multiple anatomic injuries (defined by ISS >=35) were significant risk factors for LRTI. Using stepwise logistic regression to analyze the risk factors for pneumonia, only randomization to a conventional hospital bed remained a significant risk factor for the development of pneumonia. The ad-justed odds ratio for developing pneumonia for pa-tients randomized to a conventional bed was 2.02 (95 percent confidence interval, 1.56 to 2.61; p = 0.003, based on the improvement chi square statistic). Including ISS >=35 tended to improve the model (p = 0.058, based on the improvement chi square statistic); however, the improvement in predictive accuracy was not statistically significantly. Viagra Super Active
Table 3—Univariate Analysis of Bisk Factors Jor LRTI and Pneumonia
|
|
No. of Patients |
% |
|
% |
|
|
Risk Factor |
with Risk Factor |
LRTI |
p Value |
Pneumonia |
p Value |
|
ISS |
48 |
56.3 |
0.0033 |
35.4 |
0.0452 |
|
Major head trauma |
36 |
63.9 |
0.0005 |
33.3 |
NS |
|
Conventional bed |
48 |
58.3 |
0.0008 |
39.6 |
0.0032 |
