Characteristics of enrolled studies
A total 878 articles were obtained according to the search strategy. After screening, 31 articles were excluded due to duplicates. A total of 124 articles remained after screening based on their title and abstract. Among the 124 articles, 2 were case reports, 3 were letters, 48 were review articles, 17 were not randomized control trials, 16 had no relevant data, and 4 did not adopt low tidal volume ventilation strategy in the control groups. Finally 13 randomized-controlled trials involving 4410 patients were enrolled for the network meta-analysis. Patients with ARDS received one of the five mechanical ventilation strategies (recruitment maneuver combined with PEEP titration, recruitment maneuver combined with higher PEEP, recruitment maneuver combined with lower PEEP, higher PEEP, and lower PEEP). Literature screening and results are shown in Figure 1A. The main characteristics of all studies are reported in Table 1.
Risk of bias within studies
The risk of bias within studies was assessed using the Cochrane Collaboration Reviewer Manager 5.3 tool, and the results are shown in Figure 1B. An article (Lim 2003) did not describe random methods. For safety reasons, blinding treating clinicians to group assignment was not feasible. Although few studies reported blinding of participants, patients who needed mechanical ventilation received sedation, which did not affect the results. Three articles had high risk of other bias, and one of them (Brower 2004) modified the high-PEEP strategy by eliminating the steps with a PEEP of less than 12 cm of water and requiring a minimum PEEP of 14 cm of water for the first 48 h. An article reported by Meade et al. had a programming error occurring late in the study, thereby disrupting the specified randomization blocks. The bias in another research (Constantin 2019) was caused by different treatments between the focal ARDS and the non-focal ARDS.
Heterogeneity and inconsistency assessment
In the pairwise meta-analyses, moderate to high heterogeneity was detected; for death in 28 days, the I2 of RM+PEEP titration versus RM+lower PEEP is 64.5%. For ventilator-free days the I2 of RM+PEEP titration versus lower PEEP is 56.8%, and that for higher PEEP versus lower PEEP is 71.3%. For barotrauma, the I2 of RM+PEEP titration versus lower PEEP is 66.9%. The comparisons of other strategies show low heterogeneity. The results of the pairwise meta-analyses are shown in Table 2.
In the Bayesian framework, the fit of the consistency model for all outcomes was similar or better to that of inconsistency model (Table 3), suggesting low global inconsistency. The node-splitting analysis for primary and secondary outcomes in Bayesian framework showed no significant inconsistency between the direct effects and indirect effects (Table 4).
In the frequentist framework, the global inconsistency test showed low inconsistency for all outcomes (p>0.05) as shown in Figure 2. The node-splitting analysis in the frequentist framework showed that in terms of death in hospital, RM+higher PEEP versus higher PEEP and higher PEEP vs. lower PEEP had local inconsistency (p<0.05); otherwise, no significant inconsistency was detected between the direct effects and the indirect effects (Table 5).
Table 1. Characteristics of intensive care unit patients with acute respiratory distress syndrome included in randomized controlled trials
Study
[references]
|
Center
|
Sample size
(average age)
|
Ventilation
strategies
|
Female(%)
|
PaO2/FiO2
(cmH2O)
|
APACHEII score
|
Plateau pressure
(cmH2O)
|
pulmonary ARDS(%)
|
Lim 2003[18]
|
1
|
20(60)
|
RM+ higher PEEP
|
10%
|
|
|
|
75%
|
19(61)
|
RM+ lower PEEP
|
10.5%
|
|
|
|
78.9%
|
8(60)
|
Higher PEEP
|
12.5%
|
|
|
|
87.5%
|
Brower 2004[15]
|
23
|
276(49)
|
Higher PEEP
|
43%
|
151±67
|
|
27±6
|
58%
|
273(54)
|
Lower PEEP
|
47%
|
165±77
|
|
24±7
|
53%
|
Meade 2008[19]
|
30
|
475(54.5)
|
RM+ higher PEEP
|
40.6%
|
144.8±47.9
|
24.8±7.8
|
30.4±5.5
|
69%
|
508(56.9)
|
Lower PEEP
|
39.6%
|
144.6±49.2
|
25.9±7.7
|
29.3±6.0
|
75.9%
|
Mercat 2008[20]
|
37
|
385(60)
|
Higher PEEP
|
32%
|
144±58
|
|
23.7±4.9
|
70%
|
382(60)
|
Lower PEEP
|
33%
|
143±57
|
|
22.9±5.3
|
75%
|
Tamlor 2008[21]
|
1
|
30(54.5)
|
RM+PEEP titration
|
37%
|
147±56
|
26.3±6.4
|
29±7
|
23.3%
|
31(51.2)
|
RM+Lower PEEP
|
45%
|
145±57
|
26.8±6.5
|
29±5
|
16.1%
|
Huh 2009[22]
|
1
|
30(55)
|
RM+PEEP titration
|
37%
|
115.0±8.5
|
22.0±1.1
|
|
66.7%
|
27(62)
|
Lower PEEP
|
40%
|
110.8±6.3
|
20.0±1.4
|
|
66.7%
|
Xi 2010[23]
|
14
|
55(62.2)
|
RM+ lower PEEP
|
30.9%
|
104.8±60.2*
|
21.5±6.7
|
24.2±5.3
|
45.4%
|
55(65.5)
|
Lower PEEP
|
27.3%
|
115.9±38.3*
|
23.1±8.6
|
23.4±5.3
|
38.2%
|
Hodgson 2011[24]
|
1
|
10(60)
|
RM+PEEP titration
|
30%
|
155±8
|
20.1±3
|
28.9±1.2
|
50%
|
10(58)
|
Lower PEEP
|
40%
|
149±12
|
20.1±2
|
27.1±1.2
|
60%
|
Kacmarek 2016[25]
|
20
|
99(52.2)
|
RM+PEEP titration
|
42.4%
|
121±37
|
18±10
|
27±5
|
68%
|
101(53.4)
|
Lower PEEP
|
33.7%
|
114±33
|
17±6
|
27±5
|
66%
|
ART 2017[26]
|
120
|
501(51.3)
|
RM+PEEP titration
|
37.5%
|
119.5±43.5
|
|
25.8±4.7
|
62.5%
|
509(50.6)
|
Lower PEEP
|
37.5%
|
117.2±41.9
|
|
26.2±5.2
|
61.5%
|
Beitler 2019[27]
|
14
|
102(58)
|
RM+PEEP titration
|
37.3%
|
99.2±41.5*
|
27±8
|
28.0±5.9*
|
80.4%
|
98(57.5)
|
RM+lower PEEP
|
55.1%
|
94.2±40.0*
|
28±7
|
27.4±3.7*
|
89.8%
|
Kung 2019[28]
|
4
|
60(66.8)
|
RM+PEEP titration
|
45/15
|
133.4±47.0
|
20.4±5.8
|
25.3±3.9
|
86.7%
|
60(63.7)
|
Lower PEEP
|
44/16
|
129.7±42.0
|
21.5±6
|
25.7±4.6
|
76.7%
|
Constantin2019[29]
|
20
|
82(63)
|
RM+PEEP titration
|
21%
|
121±4
|
|
26±1
|
73%
|
204(61)
|
Lower PEEP
|
28%
|
115±4
|
|
24±1
|
73%
|
*means the data was transformed from median (interquartile range) to mean(SD)
|
Table 2. Results of head-to-head comparisons according to pairwise meta-analyses on different outcomes
Ventilation strategy
|
N
|
I2
|
Heterogeneity
|
RR/SMD(95%CI)
|
p-value
|
Death in hospital
|
Lower PEEP vs
|
RM+PEEP titration
|
5
|
0.0%
|
0.547
|
1.044(0.934,1.167)
|
0.445
|
Lower PEEP vs
|
Higher PEEP
|
2
|
0.0%
|
0.906
|
0.929(0.825,1.046)
|
0.222
|
Lower PEEP vs
|
RM+higher PEEP
|
1
|
-
|
-
|
0.917(0.801,1.050)
|
0.21
|
Lower PEEP vs
|
RM+lower PEEP
|
1
|
-
|
-
|
0.745(0.508,1.094)
|
0.133
|
Higher PEEP vs
|
RM+higher PEEP
|
1
|
-
|
-
|
0.647(0.417,1.004)
|
0.052
|
Higher PEEP vs
|
RM+lower PEEP
|
1
|
-
|
-
|
0.654(0.418,1.023)
|
0.063
|
RM+lower PEEP vs
|
RM+PEEP titration
|
1
|
-
|
-
|
1.04 (0.780,1.387)
|
0.79
|
RM+lower PEEP vs
|
RM+higher PEEP
|
1
|
-
|
-
|
0.95 (0.515,1.751)
|
0.95
|
Death in 28 day
|
RM+PEEP titration vs
|
Lower PEEP
|
5
|
0.0%
|
0.651
|
1.097(0.986,1.220)
|
0.09
|
RM+PEEP titration vs
|
RM+lower PEEP
|
2
|
64.5%
|
0.093
|
0.806(0.412,1.578)
|
0.53
|
Lower PEEP vs
|
RM+higher PEEP
|
1
|
-
|
-
|
0.908(0.785,1.051)
|
0.196
|
Lower PEEP vs
|
RM+lower PEEP
|
1
|
-
|
-
|
0.718(0.466,1.107)
|
0.134
|
Lower PEEP vs
|
higher PEEP
|
1
|
-
|
-
|
0.921(0.785,1.081)
|
0.316
|
Ventilator-free days
|
RM+PEEP titration vs
|
lower PEEP
|
3
|
56.8%
|
0.099
|
-0.001(-0.214,0.211)
|
0.989
|
RM+PEEP titration vs
|
RM+lower PEEP
|
2
|
0.0%
|
0.476
|
0.013(-0.229,0.256)
|
0.914
|
Higher PEEP vs
|
Lower PEEP
|
2
|
71.3%
|
0.062
|
0.043(-0.162,0.247)
|
0.682
|
RM+lower PEEP vs
|
Lower PEEP
|
1
|
-
|
-
|
0.338(-0.038,0.715)
|
0.078
|
Barotrauma
|
RM+PEEP titration vs
|
lower PEEP
|
5
|
55.5%
|
0.061
|
1.136(0.763,1.753)
|
0.013
|
higher PEEP vs
|
lower PEEP
|
2
|
0
|
0.879
|
1.072(0.854,1.345)
|
0.548
|
RM+PEEP titration vs
|
RM+lower PEEP
|
1
|
-
|
-
|
1.074(0.615,1.875)
|
0.802
|
RM+higher PEEP vs
|
lower PEEP
|
1
|
-
|
-
|
1.109(0.911,1.350)
|
0.303
|
Significant values (p < 0.05) or I2>50% are in bold and underlined, indicating a significant heterogeneity, and the random-effects model was chose.
|
Table 3 Comparisons of the fit of consistency and inconsistency models in Bayesian framework
model
|
Death in 28 day
Median(95%Crl)
|
Death in hospital
Median(95%Crl)
|
Ventilator-free days
Median(95%Crl)
|
Barotrauma
Median(95%Crl)
|
consistency
|
0.36 (0.04, 0.94)
|
0.19 (0.01, 0.69)
|
1.62 (0.22, 3.25)
|
0.60 (0.12, 1.21)
|
Inconsistency
|
0.32 (0.04, 1.06)
|
0.19 (0.02, 0.71)
|
1.57 (0.17, 3.21)
|
0.56 (0.04, 1.19)
|
The DIC is a Bayesian model evaluation criterion that measures model fit adjusted with complexity of the model; smaller DIC values correspond to more preferable models.
|
Table 4 Node-splitting analysis for Death in hospital, Death in 28 day, Ventilator-free days and Barotrauma in Bayesian framework
Ventilation strategy
|
Direct Effect
(95%CI)
|
Indirect Effect
((95%CI))
|
Overall
((95%CI))
|
p-Value
|
Death in 28 day
|
RM+PEEP titration vs RM+lower PEEP
|
0.27 (-0.46, 1.24)
|
-0.69 (-1.95, 0.65)
|
0.02 (-0.65, 0.80)
|
0.20
|
RM+PEEP titration vs lower PEEP
|
-0.04 (-0.52, 0.57)
|
0.92 (-0.46, 2.51)
|
0.08 (-0.39, 0.71)
|
0.18
|
RM+lower PEEP vs lower PEEP
|
0.68 (-0.53, 1.85)
|
-0.33 (-1.42, 0.66)
|
0.07 (-0.70, 0.86)
|
0.17
|
Death in hospital
|
RM+PEEP titration vs RM+lower PEEP
|
-0.08 (-0.89, 0.77)
|
-0.65 (-1.48, 0.26)
|
-0.29 (-0.93, 0.26)
|
0.33
|
RM+PEEP titration vs lower PEEP
|
-0.05 (-0.37, 0.45)
|
0.53 (-0.62, 1.66)
|
0.03 (-0.30, 0.47)
|
0.31
|
RM+higher PEEP vs RM+lower PEEP
|
-0.26 (-1.64, 1.14)
|
-0.05 (-0.89, 1.10)
|
-0.07 (-0.78, 0.67)
|
0.75
|
RM+higher PEEP vs higher PEEP
|
2.28 (0.14, 5.17)
|
0.07 (-0.60, 0.74)
|
0.16 (-0.39, 1.11)
|
0.05
|
RM+higher PEEP vs lower PEEP
|
0.18 (-0.55, 0.91)
|
0.93 (-0.41, 2.37)
|
0.24 (-0.23, 0.96)
|
0.34
|
RM+lower PEEP vs higher PEEP
|
2.25 (0.08, 6.04)
|
0.10 (-0.56, 0.88)
|
0.27 (-0.34, 1.14)
|
0.06
|
RM+lower PEEP vs lower PEEP
|
0.60 (-0.37, 1.58)
|
0.22 (-0.47, 1.08)
|
0.35 (-0.18, 0.99)
|
0.55
|
higher PEEP vs lower PEEP
|
0.15 (-0.27, 0.59)
|
-1.97 (-5.06, 0.26)
|
0.07 (-0.49, 0.47)
|
0.07
|
Ventilator-free days
|
RM+PEEP titration vs RM+lower PEEP
|
-0.51 (-5.14, 4.27)
|
3.80 (-2.41, 9.36)
|
1.28 (-2.67, 4.83)
|
0.26
|
RM+PEEP titration vs lower PEEP
|
0.41 (-2.73, 2.80)
|
-3.90 (-11.14, 2.86)
|
-0.14 (-3.17, 2.15)
|
0.25
|
RM+lower PEEP vs lower PEEP
|
-3.41 (-8.60, 1.71)
|
0.75 (-4.68, 6.09)
|
-1.42 (-5.37, 2.38)
|
0.29
|
Barotrauma
|
RM+PEEP titration vs RM+lower PEEP
|
-0.19 (-2.05, 1.67)
|
-9.01 (-35.46, 11.46)
|
-0.19 (-2.07, 1.56)
|
0.45
|
RM+PEEP titration vs lower PEEP
|
-0.32 (-1.18, 0.76)
|
8.50 (-9.60, 45.55)
|
-0.30 (-1.15, 0.75)
|
0.38
|
RM+lower PEEP vs lower PEEP
|
5.73 (-15.99, 38.53)
|
-0.13 (-2.13, 2.07)
|
-0.10 (-2.00, 2.11)
|
0.63
|
Significant values (p ≤ 0.05) are in bold and underlined, indicating a significant inconsistency between the direct effect and indirect effects
|
Table 5 Node-splitting analysis for Death in hospital, Death in 28 day, Ventilator-free days and Barotrauma in Frequentist framework
Ventilation strategy
|
Direct Effect
(Std.Err)
|
Indirect Effect
(Std.Err)
|
Difference
(Std.Err)
|
p-Value
|
Death in 28 day
|
RM+PEEP titration vs RM+lower PEEP
|
0.36 (0.75)
|
0.16 (1.16)
|
0.20 (1.38)
|
0.887
|
RM+PEEP titration vs lower PEEP
|
0.57 (0.52)
|
0.76 (1.28)
|
-0.19 (1.38)
|
0.888
|
RM+lower PEEP vs lower PEEP
|
0.41 (1.04)
|
0.21 (0.91)
|
0.20 (1.38)
|
0.887
|
Death in hospital
|
RM+PEEP titration vs RM+lower PEEP
|
-0.06 (0.22)
|
-0.39 (0.20)
|
0.34 (0.29)
|
0.246
|
RM+PEEP titration vs lower PEEP
|
-0.04 (0.11)
|
0.30 (0.27)
|
-0.34 (0.29)
|
0.246
|
RM+higher PEEP vs RM+lower PEEP
|
0.04 (0.33)
|
-0.12 (0.21)
|
0.16 (0.39)
|
0.685
|
RM+higher PEEP vs higher PEEP
|
0.59 (0.25)
|
0.005 (0.11)
|
0.59 (0.28)
|
0.036
|
RM+higher PEEP vs lower PEEP
|
0.10 (0.09)
|
0.54 (0.27)
|
-0.43 (0.29)
|
0.129
|
RM+lower PEEP vs higher PEEP
|
0.57 (0.25)
|
0.03 (0.16)
|
0.54 (0.29)
|
0.067
|
RM+lower PEEP vs lower PEEP
|
0.30 (0.23)
|
0.22 (0.19)
|
0.08 (0.30)
|
0.778
|
higher PEEP vs lower PEEP
|
0.10 (0.08)
|
-0.41 (0.22)
|
0.51 (0.23)
|
0.031
|
Ventilator-free days
|
RM+PEEP titration vs RM+lower PEEP
|
-0.02 (0.15)
|
0.038 (0.24)
|
-0.40 (0.28)
|
0.149
|
RM+PEEP titration vs lower PEEP
|
0.04 (0.10)
|
-0.36 (0.26)
|
0.40 (0.28)
|
0.149
|
RM+lower PEEP vs lower PEEP
|
-0.34 (0.22)
|
0.06 (0.17)
|
-0.40 (0.28)
|
0.149
|
Barotrauma
|
RM+PEEP titration vs RM+lower PEEP
|
-0.14 (0.79)
|
-0.28 (2.11)
|
0.14 (2.25)
|
0.952
|
RM+PEEP titration vs lower PEEP
|
-0.25 (0.46)
|
-0.14 (2.20)
|
-0.14 (2.25)
|
0.952
|
RM+lower PEEP vs lower PEEP
|
-5.9e-11 (2.06)
|
-0.14 (0.91)
|
0.14 (2.25)
|
0.952
|
Significant values (p≤ 0.05) are in bold and underlined, indicating a significant inconsistency between the direct effect and indirect effects
|
Network structure and geometry
The network plot for all outcomes is shown in Figure 3. The size of the node is proportional to the number of patients randomized to receive the treatment. The width of each line is proportional to the number of trials comparing the connected treatments. The most common comparison was RM+PEEP titration versus lower PEEP, and the most common subjects were RM+PEEP titration versus lower PEEP. The network plots for death in 28 days (Figure 3A) and barotrauma (Figure 3D) were similar. The difference is mainly the number of studies comparing the two strategies.
Network meta-analysis for outcomes
For the primary outcomes in terms of death in 28 days, five ventilation strategies were included. None of the ventilation strategies were significantly superior to others, and the 95% CI included 1 in the Bayesian (Fig. 4A) and the frequentist frameworks ( Fig. 4C)
For the secondary outcomes, the ventilation strategies were compared except the RM+higher PEEP in terms of ventilator-free day. The result of network meta-analysis indicated that the differences among the strategies were not significant in either Bayesian (Fig. 4B) or Frequentist framework (Fig.4D) with 95% CI of 0. In terms of the other two secondary outcomes, the results of the two frameworks suggested that no strategy was superior in terms of hospital deaths and barotrauma.
Rank probabilities
The relative ranking of the ventilation strategies was estimated using SUCRA. The Bayesian and the Frequentist frameworks were used to calculate the ranking probability and SURCA. The Bayesian framework indicates that the lower the SUCRA, the more superior the strategy in terms of the outcome of death in 28 days, death in hospital, and barotrauma. The outcomes of ventilator-free days indicate that the higher the SUCRA, the more superior the strategy. The frequentist framework indicates that for all outcomes, the higher the SUCRA, the more superior the strategy. The ranking results in both frameworks are shown in Table 6 and Table 7, respectively, and the SUCRA in both frameworks are shown in Figure 5.
In terms of death in 28 days, the lower PEEP was the worst strategy in the Bayesian and frequentist frameworks. However, the superior strategy was different in the two frameworks. Higher PEEP was the best strategy in the Bayesian framework, whereas RM+PEEP titration was the best one in the frequentist framework.
In terms of hospital deaths, the results were the same for the Bayesian framework and the frequentist framework, that is, RM+lower PEEP was the superior strategy and lower PEEP was the worst strategy.
In terms of ventilator-free days, the results were similar to the results of hospital death. RM+lower PEEP had the highest SUCRA value in the Bayesian framework and the frequentist framework, whereas lower PEEP strategy had the lowest SUCRA value in the two frameworks.
In terms of barotrauma, the relative ranking was consistent in the Bayesian framework and the frequentist framework. Lower PEEP was the superior mechanical ventilation strategy and RM+PEEP titration was the worst mechanical ventilation strategy.
Table 6. Bayesian ranking results of network meta-analysis
Strategy
|
Rank of possibility
|
SUCRA
|
1
|
2
|
3
|
4
|
5
|
Death in 28d (Rank 1 is worst, rank N is best)
|
|
RM+PEEP titration
|
0.13
|
0.2
|
0.23
|
0.27
|
0.16
|
0.571
|
RM+higher PEEP
|
0.29
|
0.18
|
0.17
|
0.18
|
0.18
|
0.644
|
RM+lower PEEP
|
0.24
|
0.16
|
0.16
|
0.21
|
0.23
|
0.588
|
higher PEEP
|
0.2
|
0.13
|
0.12
|
0.17
|
0.39
|
0.5
|
lower PEEP
|
0.14
|
0.34
|
0.32
|
0.17
|
0.04
|
0.705
|
Death in hospital (Rank 1 is worst, rank N is best)
|
|
RM+PEEP titration
|
0.3
|
0.27
|
0.25
|
0.14
|
0.04
|
0.75
|
RM+higher PEEP
|
0.07
|
0.08
|
0.15
|
0.34
|
0.36
|
0.406
|
RM+lower PEEP
|
0.05
|
0.05
|
0.14
|
0.25
|
0.52
|
0.339
|
higher PEEP
|
0.25
|
0.17
|
0.27
|
0.23
|
0.07
|
0.639
|
lower PEEP
|
0.33
|
0.43
|
0.19
|
0.04
|
0
|
0.844
|
Ventilator-free days (Rank 1 is best, rank N is worst)
|
|
RM+PEEP titration
|
0.12
|
0.29
|
0.25
|
0.34
|
|
0.543
|
RM+lower PEEP
|
0.56
|
0.19
|
0.12
|
0.13
|
|
0.8
|
higher PEEP
|
0.28
|
0.29
|
0.21
|
0.23
|
|
0.663
|
lower PEEP
|
0.04
|
0.22
|
0.43
|
0.3
|
|
0.49
|
Barotrauma (Rank 1 is worst, rank N is best)
|
|
RM+PEEP titration
|
0.22
|
0.32
|
0.22
|
0.17
|
0.07
|
0.71
|
RM+higher PEEP
|
0.27
|
0.21
|
0.2
|
0.17
|
0.15
|
0.661
|
RM+lower PEEP
|
0.28
|
0.14
|
0.11
|
0.12
|
0.36
|
0.56
|
higher PEEP
|
0.19
|
0.19
|
0.2
|
0.2
|
0.21
|
0.584
|
lower PEEP
|
0.02
|
0.13
|
0.28
|
0.35
|
0.21
|
0.468
|
The number in each cell represents the probability of each ventilation strategy. The value of SUCRA with biggest probability of ranking best is in bold and underlined.
|
Table 7 Frequentist ranking results of network meta-analysis
Strategy
|
Rank of possibility
|
SUCRA
|
1
|
2
|
3
|
4
|
5
|
Death in 28d (Rank 1 is best, rank N is worst)
|
|
RM+PEEP titration
|
0.36
|
0.38
|
0.18
|
0.07
|
0.01
|
0.749
|
RM+higher PEEP
|
0.23
|
0.15
|
0.14
|
0.17
|
0.31
|
0.453
|
RM+lower PEEP
|
0.18
|
0.25
|
0.23
|
0.18
|
0.16
|
0.529
|
higher PEEP
|
0.22
|
0.15
|
0.14
|
0.17
|
0.32
|
0.445
|
lower PEEP
|
0.01
|
0.08
|
0.3
|
0.42
|
0.19
|
0.323
|
Death in hospital (Rank 1 is best, rank N is worst)
|
|
RM+PEEP titration
|
0.01
|
0.11
|
0.3
|
0.21
|
0.37
|
0.296
|
RM+higher PEEP
|
0.31
|
0.45
|
0.15
|
0.06
|
0.04
|
0.736
|
RM+lower PEEP
|
0.64
|
0.26
|
0.07
|
0.02
|
0.01
|
0.871
|
higher PEEP
|
0.04
|
0.15
|
0.28
|
0.21
|
0.32
|
0.345
|
lower PEEP
|
0
|
0.03
|
0.21
|
0.49
|
0.27
|
0.252
|
Ventilator-free days (Rank 1 is best, rank N is worst)
|
|
RM+PEEP titration
|
0.12
|
0.3
|
0.27
|
0.31
|
|
0.409
|
RM+lower PEEP
|
0.52
|
0.21
|
0.13
|
0.14
|
|
0.704
|
higher PEEP
|
0.3
|
0.26
|
0.2
|
0.24
|
|
0.542
|
lower PEEP
|
0.06
|
0.23
|
0.4
|
0.31
|
|
0.345
|
Barotrauma (Rank 1 is best, rank N is worst)
|
|
RM+PEEP titration
|
0.07
|
0.16
|
0.22
|
0.31
|
0.24
|
0.382
|
RM+higher PEEP
|
0.16
|
0.15
|
0.19
|
0.22
|
0.28
|
0.423
|
RM+lower PEEP
|
0.33
|
0.12
|
0.11
|
0.16
|
0.29
|
0.512
|
higher PEEP
|
0.22
|
0.21
|
0.2
|
0.19
|
0.18
|
0.521
|
lower PEEP
|
0.22
|
0.37
|
0.28
|
0.12
|
0.02
|
0.663
|
The number in each cell represents the probability of each ventilation strategy. The value of SUCRA with biggest probability of ranking best is in bold and underlined.
|