Table 1 Impact of sedation practice on patient safety outcomes

Marshall and colleagues 2008 [18]

Introduction of protocol (before-after study)

156

Before

Hospital mortality: 18%

10% (self-extubation)

11.5%

After active pharmacist intervention

Hospital mortality: 15%

P = 0.83

4% (self-extubation) P = 0.2

10.3%

P = 0.8

Quenot and colleagues 2007 [20]

Introduction of protocol (before-after study)

423

Before

ICU mortality: 39%

Hospital mortality: 45%

15% (ventilator -- associated pneumonia)

7 (unscheduled self-extubation)

After introduction of nurse-led sedation protocol

ICU mortality: 32%

P = 0.19

Hospital mortality: 38% P = 0.22

6% (ventilator -- associated pneumonia)

P = 0.005

11 (unscheduled self-extubation) P = 0.09

Arabi and colleagues 2007 [10]

Introduction of protocol (before-after study)

207

Before education -- no protocol

ICU mortality: 20%

Hospital mortality: 24%

28%

Before education -- protocol

ICU mortality: 18%

Hospital mortality: 24%

29%

After education -- no protocol

ICU mortality: 23%

Hospital mortality: 36%

11%

After education - protocol

ICU mortality: 13% P = 0.64 for all comparisons

Hospital mortality: 23%

P = 0.35 for all comparisons

11%

P = 0.02

Chanques and colleagues 2006 [13]

Introduction of protocol (before-after study)

230

Before

ICU mortality: 12%

14%

4% (self-extubation)

After monitoring of agitation

ICU mortality: 15%

P = 0.76

9%

P = 0.31; P = 0.03 for events/1000 ventilation days

2% (self-extubation) P = 0.65

Burns and colleagues 2003 [3]

Introduction of protocol (before-after study)

1105

Before

38% (timepoint unclear)

10%

After introduction of outcomes management protocol

31%

P = 0.02 (timepoint unclear)

7%

Mascia and colleagues 2000 [19]

Introduction of protocol (before-after study)

156

Before

16.7% (timepoint unclear)

After introduction of protocol

17.6%

P = 0.89 (timepoint unclear)

Jakob and colleagues 2007 [16]

Introduction of protocol (before-after study)

300

After implementation of intervention 1 (change in ICU organisation)

ICU mortality: 19%

After implementation of intervention 2 (introduction of protocols for weaning)

ICU mortality: 8%

After implementation of intervention 1 (change in ICU organisation)

ICU mortality: 7%

P = 0.02

De Jonghe and colleagues 2005 [14]

Introduction of protocol (before-after study)

102

Before

8%

After introduction of protocol

20.4%

P = 0.1

Brattebo and colleagues 2002; 2004 [11, 28]

Introduction of protocol (before-after study)

285

Before

ICU mortality: 27%

0

After introduction of protocol

ICU mortality: 22% (not significant)

0

Brook and colleagues 1999 [12]

Introduction of protocol (before-after study)

321

Before

Hospital mortality: 36%

9%

After introduction of protocol

Hospital mortality: 30%

13% P = 0.213

Kress and colleagues 2000; 2001; Schweickert and colleagues 2004 [4, 22, 29]

RCT of daily interruption of sedation

128

Continuous sedation

Hospital mortality: 47%

8% (ventilator-associated pneumonia)

Sedation interrupted daily

Hospital mortality: 36%

P = 0.25

3% (ventilator-associated pneumonia)

Girard and colleagues 2008 [6]

RCT of spontaneous breathing trials with and without daily interruption of sedation

335

Continuous sedation + spontaneous breathing trial

28-day mortality: 35%

1-year mortality: 58%

10% self-extubations

3% required re-intubation

14%

Sedation interrupted daily + spontaneous breathing trial

28-day mortality: 28%

P = 0.21

1-year mortality: 44%

P = 0.01

4% self-extubation P = 0.03

2% required re-intubation

P = 0.47

13%

P = 0.73

Kollef and colleagues 1998 [5]

Continuous sedation vs intermittent sedation

242

Continuous sedation

Hospital mortality: 34%

15%

Sedation interrupted daily

Hospital mortality: 30%

P = 0.58

5% P = 0.005

Carson and colleagues 2006 [8]

RCT of intermittent lorazepam vs. continuous propofol

132

Lorazepam

Hospital mortality: 38%

2% self-extubations

16% reintubations

Propofol

Hospital mortality: 37%

5% self-extubations P = 0.62

12% reintubations P = 0.59

Pandharipande and colleagues 2007a, b; 2006 [24, 30, 31]

RCT of dexmedetomidine vs lorazepam (RASS individualised to each patient)

103

Dexmedetomidine

28-day mortality: 17%

8%

Lorazepam

28-day mortality: 27%

P = 0.18

4% P = 0.41