- Email: [email protected]
Effects of staff training on the care of mechanically ventilated patients: a prospective cohort study
British Journal of Anaesthesia 103 (2): 232–7 (2009)
doi:10.1093/bja/aep114
Advance Access publication May 20, 2009
CRITICAL CARE Effects of staff training on the care of mechanically ventilated patients: a prospective cohort study F. Bloos1*, S. Mu¨ller1, A. Harz1, M. Gugel1, D. Geil1, K. Egerland1, K. Reinhart1 and G. Marx1 2 1
Department of Anaesthesiology and Intensive Care Medicine, University Hospital Jena, Erlanger Allee 101, 07747 Jena, Germany. 2Department of Surgical Intensive Care Medicine, Medical Faculty, RWTH Aachen University, Pauwelsstr. 30, 52074 Aachen, Germany *Corresponding author. E-mail: [email protected]
Methods. This study was performed on a 50-bed intensive care unit of a tertiary care university hospital. Application of a ventilator bundle consisting of semirecumbent positioning, lung protective ventilation in patients with acute lung injury (ALI), ulcer prophylaxis, and deep vein thrombosis prophylaxis (DVTP) was assessed before and after staff training in post-surgical patients requiring mechanical ventilation for at least 24 h. Results. A total of 133 patients before and 141 patients after staff training were included. Overall bundle adherence increased from 15 to 33.8% (P,0.001). Semirecumbent position was achieved in 24.9% of patient days before and 46.9% of patient days after staff training (P,0.001). Administration of DVTP increased from 89.5 to 91.5% (P¼0.048). Ulcer prophylaxis of .90% was achieved in both groups. Median tidal volume in patients with ALI remained unaltered. Days on mechanical ventilation were reduced from 6 (interquartile range 2.0 – 15.0) to 4 (2.0 – 9.0) (P¼0.017). Rate of ventilator-associated pneumonia (VAP), ICU length of stay, and ICU mortality remained unaffected. In patients with VAP, the median ICU length of stay was reduced by 9 days (P¼0.04). Conclusions. Staff training by an ICU change team improved compliance to a pre-defined ventilator bundle. This led to a reduction in the days spent on mechanical ventilation, despite incomplete bundle implementation. Br J Anaesth 2009; 103: 232–7 Keywords: audit, trainings; complications, respiratory; intensive care, pulmonary; ventilation, mechanical Accepted for publication: April 14, 2009
Nosocomial infections are a major cause of morbidity and mortality in intensive care patients. Specifically, ventilatorassociated pneumonia (VAP) is known to increase the duration of mechanical ventilation, mortality, and length of stay and also costs of care.1 It has been estimated that 15% of ventilated patients develop pneumonia. The National Nosocomial Infections Surveillance System (NNIS) reports a median occurrence of VAPs of 4.6 –5.1 per 1000 ventilator days in medical-surgical intensive care units (ICUs).2 Besides the development of infectious complications, mechanically ventilated patients are also at risk
of developing several other complications. Mechanical ventilation with high tidal volumes or high plateau pressures may induce pulmonary injury by volutrauma, barotrauma or both.3 Other complications include deep vein thrombosis4 and the development of peptic ulceration.5 Therefore, low-tidal volume ventilation, stress-ulcer prophylaxis, and deep vein thrombosis prophylaxis are recommended in these patients. The concept of treatment bundles has been successfully used to introduce multiple therapies into clinical practice. A bundle is defined as a selected set of interventions or
# The Author [2009]. Published by Oxford University Press on behalf of The Board of Directors of the British Journal of Anaesthesia. All rights reserved. For Permissions, please email: [email protected]
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
Background. Adherence to guidelines to avoid complications associated with mechanical ventilation is often incomplete. The goal of this study was to assess whether staff training in predefined interventions (bundle) improves the quality of care in mechanically ventilated patients.
Staff training to improve care in the ICU
procedures distilled from evidence-based medicine.6 This concept is currently used in the care of patients with sepsis; however, there is still limited evidence for this approach in mechanically ventilated patients. The goal of this study was first to assess to what extent the pre-defined ventilator bundle applied to mechanically ventilated patients in a 50-bed interdisciplinary surgical ICU. Secondly, we wanted to investigate whether training of physicians and nurses in the application of this ventilator bundle improved the quality of care in these patients.
Methods
Definition of the ventilator bundle The ventilator bundle from the Institute of Health Care Improvement (IHI)8 9 was modified for application in this study. The daily sedation break was removed from the bundle because it has not been studied in surgical ICU patients.10 It was replaced by lung protective ventilation for patients with acute lung injury (ALI), as a German national survey showed that lung-protective ventilation is poorly applied.11 Thus, the bundle in this trial consisted of semirecumbent positioning, lung protective ventilation, stress ulcer prophylaxis, and deep vein thrombosis prophylaxis (DVTP). Correct positioning of the patient was defined as a semirecumbent positioning of at least 308. Lung-protective ventilation was defined as a maximum tidal volume of 6 ml kg21 of predicted bodyweight ( pbw) and a plateau pressure of ,30 cm H2O. DVTP or stress ulcer prophylaxis was defined as being achieved if the treating physician prescribed the appropriate medication in the medical record and the administration of the medication was signed by the nurse. Complete bundle adherence was defined as fulfilment of all four elements in patients with ALI and fulfilment of all
Study design The study was performed as a pre-/post-test design and consisted of three phases: (1) Description of the current state: between June and September 2005, a specially trained task force visited the enrolled patients every day randomly between 8:00 and 20:00 and recorded compliance to the treatment bundles (Audit I). (2) Training of staff in charge of mechanically ventilated patients: a change team consisting of the ICU manager, the ICU consultants and also interested ICU residents and nurses was formed. For a 2 month period, the results of Audit I, the scientific background and technique of the bundle, were taught to all nurses and residents of the ICU in daily seminars by the change team. Red marks were attached on the walls to indicate a correct semirecumbent position. After the 2 month teaching period, the change team monitored daily compliance to the treatment bundles and individually trained nurses and residents if the bundle was not correctly applied for another 2 months. (3) Control of performance: patient visits as in Audit I were repeated between March and June 2006 (Audit II).
Data acquisition The Acute Physiology and Chronic Health Evaluation II (APACHE II) score was calculated from data within 24 h of admission. During the daily audits, current tidal volume and plateau pressure were recorded from the ventilator. Maximum tidal volume and maximum plateau pressure were obtained from the Patient Data Management System (COPRAw, Sasbachwalden, Germany). Adherence to lungprotective ventilation was checked in patients with ALI (PaO2/FIO2 ,39.5 kPa, bilateral pulmonary infiltrates on the chest X-ray in the absence of left atrial hypertension). Angle for semirecumbent positioning was measured manually or, if applicable, by an indicator integrated in the ICU bed. Mean daily backrest elevation was calculated individually for each patient by the sum of measured angles divided by the number of observation. Mean daily plateau pressure and mean daily tidal volume were calculated accordingly. Medication for DVTP and for ulcer prophylaxis was taken from the medical record. The Clinical Pulmonary Infection Score (CPIS) was calculated daily.12 Onset of pneumonia was defined as the first CPIS of at least 6 combined with the presence of pulmonary infiltrates in the chest X-ray.13 VAP was assumed if criteria of pneumonia were fulfilled after 48 h of mechanical ventilation.
233
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
Patients were recruited between June 2005 and June 2006 in a 50-bed ICU of a tertiary care university hospital. Patients were included if they were mechanically ventilated for more than 24 h and were at least 10 years of age. A day on mechanical ventilation was defined as any 24 h period in which the patient required any mode of controlled or assisted ventilation, with the exception of intermittent application of continuous positive airway pressure for atelectasis prophylaxis.7 Standard measures of general critical care were applied to all patients, including hand washing before dealing with the patients, daily oral care, and sterile tracheal suction. The ventilation tubing systems remained unchanged. Standard operating procedures were in place for analgosedation, weaning from the ventilator, and antibiotic therapy. Assignment of patients to nursing staff and clinicians was not controlled by the study protocol. The local ethics committee of the University of Jena waived the need for informed consent.
but the lung-protective ventilation element in patients without ALI.
Bloos et al.
Table 1 Patient characteristics and sedation. Data are expressed as frequencies or as median (interquartile range). *Kolmogorov– Smirnow test
n/patient days Female sex Age (yr) Height (cm) Weight (kg) APACHE II score Mechanical ventilation on ICU admission Referring facility Operating theatre Other hospital Normal ward Direct admission Type of surgery Heart surgery General surgery Neurosurgery Traumatology Other
Audit I
Audit II
P-value
133/1389 37.6% 63 (47.0 –71.5) 172 (165 –180) 77.0 (70 –85) 24 (5.0 – 21.5) 85%
141/1002 30.7% 66 (52.0 –73.0) 172 (165 –179) 78.5 (67 –85) 25 (20.0 –31.0) 76.6%
0.31 0.08 0.97 0.98 0.39 0.09
51.1% 24.8% 14.3% 8.3%
36.2% 23.4% 26.2% 14.2%
0.07*
41.4% 19.5% 18.0% 12.8% 6.9%
45.4% 14.2% 30.5% 5.7% 4.2%
0.66*
n/patient days Midazolam (mg day21) Sufentanil (mg) Ketamine (mg) Propofol (mg)
Audit I
Audit II
P-value
133 / 1389 31.1 (11.9 – 75.3) n¼64 (48.1%) 460.1 (191.3 –992.4) n¼93 (69.9%) 363.5 (83.5 –1039.6) n¼22 (16.5%) 502.3 (130.2 –920.9) n¼99 (74.4%)
141/1002 27.1 (7.8 –46.3) n¼71 (51.1%) 365.8 (153.1 – 810.1) n¼89 (64.0%) 271.6 (8.0 –526.1) n¼28 (20.1%) 297.2 (102.7 – 629.2) n¼66 (47.5%)
0.22 0.63 0.16 0.18 0.27 0.53 0.09 ,0.01
Weaning was defined as the first day where the patient was either extubated without non-invasive ventilation, or, in case of patients with a tracheostomy, able to breath spontaneously without mechanical support. Intermittent mask CPAP for atelectasis prophylaxis did not count as mechanical ventilation.
Statistics The primary endpoint of the study was adherence to the ventilator bundle. Secondary endpoints were rate of pneumonia, days on mechanical ventilation, ICU survival, and length of stay on the ICU. Patients with VAP were analysed as subgroup. Inferential statistics were calculated by the Mann – Whitney U-test or in case of frequencies with the x2-test. Continuous data including mean daily backrest elevation are expressed as median and 25% and also 75% percentiles. A P-value of ,0.05 was considered as statistically significant. A conditional forward stepwise Cox regression analysis was performed to evaluate which parameters independently affect the duration of mechanical ventilation with a P-value of ,0.05. Data were analysed with SPSS 15.0 for Windows (SPSS Inc., Chicago, IL, USA).
Results One hundred and thirty-three patients (1389 patient-days) were included in Audit I and 141 patients (1002 patientdays) in Audit II. Patient characteristics data are given in Table 1 and were not statistically different between the two groups. Significantly less patients received propofol in Audit II when compared with Audit I (Table 2). Cumulative dosage of propofol per days on mechanical ventilation was also lower in Audit II. However, this difference did not reach statistical significance. There was no difference in usage of other sedatives.
Application of the ventilator bundle Application of the ventilator bundle is summarized in Figure 1. Complete bundle adherence in all patients increased from 15 to 33.8% after staff training (P,0.01). Bundle adherence differed between patients with and without ALI. In patients without ALI, the overall bundle adherence was 22.8% in Audit I and 47% in Audit II (P,0.01) when compared with 9.9 and 18.2% (P,0.01) in patients with ALI. Semirecumbent position with an angle of at least 308 was achieved on significantly more patient days in Audit
234
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
Table 2 Cumulative analgosedation per days on mechanical ventilation. Data are expressed as median (interquartile range). Mann –Whitney test was used for differences in analgosedation dosage. The n value gives the number of patients who actually received the sedative on at least 1 day (x2-test for differences)
Staff training to improve care in the ICU
II than in Audit I (49.6 vs 24.9%, P,0.01). This resulted in a median of the mean daily backrest elevations of 22.68 (25 and 75% percentile: 18.4; 26.2) in Audit I and 26.78 (25 and 75% percentile: 22.5; 29.5) in Audit II (P,0.01). Tidal volumes and plateau pressures are given in Table 3. Sixty-seven patients in Audit I and 61 patients in Audit II fulfilled criteria of ALI on at least 1 day. In the subgroup of patients with ALI, the median of the mean daily tidal volumes was unaffected by staff training and were above 6 ml kg21 pbw in both audits. Medians of the mean daily plateau pressures were ,30 cm H2O throughout the trial. DVTP was given significantly more in Audit II (91.9%) than in Audit I (89.5%, P¼0.048). However, contraindications for DVTP were present in 6.4% of patient days Audit I Audit II
Ulcer prophylaxis P=0.05
90
60 70
50
P<0.01 30 40
Semirecumb. positioning 20
P<0.01
0 10
Lung protective ventilation
80
Deep vein thromb. proph.
Frequency (%)
Fig 1 Frequency of successful implementation of treatment bundles before and after staff training. Data are given in percentage of patient days. Lung protective ventilation applies to patients with ALI only.
Outcome Median duration of mechanical ventilation was significantly shorter in Audit II when compared with Audit I (Table 4). This was associated with a faster liberation from mechanical ventilation (Fig. 2). ICU length of stay and overall mortality were not different between the audits. In Audit II, a higher tracheostomy rate was observed than in Audit I (30.9 vs 19.5%, P¼0.037). Duration until onset of VAP was longer after staff training without reaching statistical significance. Forty-four patients in Audit I and 45 patients in Audit II developed VAP. There was no significant difference in the frequency of VAP between Audit I (33.1%) and Audit II (32.4%, P¼0.68). Median APACHE II scores in patients with VAP were 24 (20; 29) in Audit I and 24 (20; 34) in Audit II (P¼0.49). A post hoc subgroup analysis on outcome was carried out in patients with VAP (Table 4). There was a significant reduction in the duration of mechanical ventilation by 8 days and of ICU length of stay by 9 days from Audit I to Audit II. The associated decrease in mortality rate from 43.2 to 35.6% did not reach statistical significance.
Table 3 Parameters of mechanical ventilation in all patients and in patients with acute lung injury. Data are expressed as median (interquartile range). Maximum tidal volume and maximum plateau pressure correspond to the daily maximum of this parameter. Pbw, predicted body weight
All patients n/patient days Mean daily tidal volume (ml kg21 pbw) Mean daily plateau pressure (cm H2O) Patients with acute lung injury n/patient days Mean daily tidal volume (ml kg21 pbw) Mean daily plateau pressure (cm H2O)
Audit I
Audit II
P-value
131 / 1198 7.8 (6.9 –9.2) 21.5 (18.3 – 24.4)
141 / 980 7.9 (7.1 – 8.9) 19.3 (17.3 –21.6)
0.65 ,0.01
67 / 497 7.2 (6.5 –9.8) 24 (21.9 –26.7)
61 / 315 7.5 (6.2 – 8.6) 21.2 (18.6 –25.3)
0.72 0.11
Table 4 Outcome in all patients, and in patients with ventilator-associated pneumonia. Data are expressed as median and interquartile range Audit I
Audit II
n¼133 6.0 (2.0 –15.0) 27.8% 12.0 (5.0 –21.5) 33.1%
n¼141 4.0 (2.0 –9.0) 25.5% 13.0 (6.0 – 21.0) 32.4%
n¼44 18 (13– 26) 43.2% 30.0 (17.3 –45) 5.3 (3.5 –8.9)
n¼45 10 (6 –17) 35.6% 21 (11.5 – 35) 6.4 (4.1 –11.0)
P-value
All patients Duration of mechanical ventilation (days) ICU mortality ICU length of stay (days) Frequency of VAP Patients with ventilator-associated pneumonia (subgroup analysis) Duration of mechanical ventilation (days) ICU mortality ICU length of stay (days) Duration until onset of VAP (days)
235
0.02 0.68 0.71 0.68
,0.01 0.52 0.04 0.18
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
P<0.01
Complete bundle
during Audit II when compared with 11.7% in Audit I (P,0.01). Ulcer prophylaxis was prescribed in 94.5% of patient days in Audit I and 94.9% patient days in Audit II (P¼0.71).
Bloos et al.
Fraction of successfully weaned patients
1.0 0.8
Audit I Audit II
0.6 0.4 Log-rank test: P=0.02
0.2 0.0 0
50 60 10 20 30 40 Duration of mechanical ventilation (days)
Fig 2 Kaplan– Meier curves for freedom from mechanical ventilation in patients from Audit I and Audit II.
Results from stepwise regression
Discussion The implementation of a ventilator bundle through staff training by an ICU change team led to a more than two-fold increase in overall bundle adherence. Staff training resulted in significantly better compliance in the semirecumbent positioning of mechanically ventilated patients. Additionally, application of DVTP was improved after staff training, although adherence to low tidal ventilation in patients with ALI was only moderate. However, improved compliance of the ventilator bundle was associated with a significant reduction in the duration of mechanical ventilation by 2 days. In a subgroup analysis, adherence to the bundle in patients with VAP was also associated with a significant reduction in days on mechanical ventilation and also on ICU length of stay. This confirms data from a smaller study where application of the IHI ventilator bundle reduced median duration of mechanical ventilation by 1 day.14 Backrest elevation was hardly implemented before staff training in our study population. An observational study showed that semirecumbent positioning is usually poorly applied in mechanically ventilated patients.15 In our study, staff training significantly improved adherence to backrest elevation but was far from complete. In a multicentre trial by Van Nieuwenhoven and colleagues,16 semirecumbent positioning as a single-study intervention could not be completely achieved, despite daily monitoring of the backrest position by research nurses.16
(1) The effects of semirecumbent positioning on VAP rates are known to be variable and may depend on the applied angle of elevation. A study, which only achieved a backrest elevation of 308 instead of 458, did not lead to a reduction in VAP rates.16 Another study has defined an angle of 308 as being the cut-off for effective reduction of VAP rates.15 The trial by Drakulovic and colleagues,20 which showed an effective reduction in VAP rates, used a backrest elevation of 458; therefore, the lack of effectiveness may have been caused by differences in positioning. However, 458 has been observed to be difficult to implement16 and the minimal backrest elevation to avoid microaspiration and thus VAP is unknown. (2) The diagnosis of VAP is difficult as there are no accurate diagnostic criteria reliably differentiating pneumonia from non-infectious causes of pulmonary infiltration. Thus, a degree of uncertainty in the VAP diagnosis may have negated the smaller effects of the ventilator bundle on outcome. The CPIS score is recommended by the International Sepsis Forum as a standardized approach for diagnosing pneumonia in clinical studies13 and has been applied in a similar setting.15 The high VAP rates observed in our study may be a result of the selection of the study population. APACHE II scores of 25 implies a high-risk group of patients while those low-risk patients ventilated for ,24 h were not included. (3) The application of the bundle may still increase pulmonary integrity by reducing microaspiration even if the bundle is implemented incompletely. Although we could not detect differences in VAP rates and are unable to prove this hypothesis, the observation that duration of mechanical ventilation and ICU length of stay decreased after staff training in the subgroup of VAP patients does lend some weight to this theory. (4) We cannot completely rule out that differences in the patient populations in this pre-/post-test design
236
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
In univariate statistics, tracheostomy rates, angle of semirecumbent position, number of patients receiving propofol, days with DVTP, and days where DVTP was contraindicated were different between Audit I and Audit II. These parameters were entered into a stepwise Cox-regression with duration of mechanical ventilation as a dependent variable. The algorithm included all parameters into the final model. The overall P-value of the final model was ,0.01. The angle of semirecumbent positioning was independently associated with days on mechanical ventilation (P¼0.02).
Staff training had little effect on the adherence to lungprotective mechanical ventilation for patients with ALI in the present study. This problem has been demonstrated in several other studies11 17 18 despite the trial by the ARDS-Network demonstrating a significant reduction in mortality when low tidal volume ventilation was applied.3 As we tried to implement a bundle with four separate treatments, the training relating to lung-protective ventilation may have remained relatively unnoticed in comparison with the other three treatments. A similar study, where low tidal volume ventilation was the only intervention, successfully reduced tidal volumes.19 In our study, decrease in the duration of mechanical ventilation and application of the ventilator bundle did not alter the frequency of VAP despite an increase in backrest elevation in Audit II. Several reasons may be responsible for this observation:
Staff training to improve care in the ICU
accounted for some differences in outcome. However, disease severity and patient characteristics were similar in the two audits.
References 1 Warren DK, Shukla SJ, Olsen MA, et al. Outcome and attributable cost of ventilator-associated pneumonia among intensive care unit patients in a suburban medical center. Crit Care Med 2003; 31: 1312 – 7 2 NNIS System. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 2004; 32: 470 – 85 3 Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000; 342: 1301 – 8 4 Samama MM, Cohen AT, Darmon JY, et al. A comparison of enoxaparin with placebo for the prevention of venous thromboembolism in acutely ill medical patients. Prophylaxis in Medical Patients with Enoxaparin Study Group. N Engl J Med 1999; 341: 793 – 800 5 Steinberg KP. Stress-related mucosal disease in the critically ill patient: risk factors and strategies to prevent stress-related bleeding in the intensive care unit. Crit Care Med 2002; 30: S362 – S364
237
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
There is some limitation in the interpretation of this study. Tracheostomy rates differed significantly between the two audits. ICU procedures to regulate the decision for tracheostomy in mechanically ventilated patients were not changed between the two audits nor were procedures for tracheostomy part of the teaching process. We therefore performed a multivariate analysis to clarify this issue which showed that both semirecumbent positioning and the presence of a tracheostomy independently affected the duration of mechanical ventilation. In conclusion, the ventilator bundle was not well applied before training. Implementation of a ventilator bundle through staff training by an ICU change team improved compliance to a pre-defined ventilator bundle and reduced days on mechanical ventilation even though the ventilator bundle was not implemented completely. Incompleteness of adherence indicates the fact that higher success rates may only be achieved by a continuous quality improvement process.
6 Levy MM, Pronovost PJ, Dellinger RP, et al. Sepsis change bundles: converting guidelines into meaningful change in behavior and clinical outcome. Crit Care Med 2004; 32: S595 – S597 7 Schoenfeld DA, Bernard GR. Statistical evaluation of ventilatorfree days as an efficacy measure in clinical trials of treatments for acute respiratory distress syndrome. Crit Care Med 2002; 30: 1772 – 7 8 Youngquist P, Carroll M, Farber M, et al. Implementing a ventilator bundle in a community hospital. Jt Comm J Qual Patient Saf 2007; 33: 219 – 25 9 Resar R, Pronovost P, Haraden C, et al. Using a bundle approach to improve ventilator care processes and reduce ventilatorassociated pneumonia. Jt Comm J Qual Patient Saf 2005; 31: 243 – 8 10 Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med 2000; 342: 1471 – 7 11 Brunkhorst FM, Engel C, Ragaller M, et al. Practice and perception—a nationwide survey of therapy habits in sepsis. Crit Care Med 2008; 36: 2719 – 25 12 Pugin J, Auckenthaler R, Mili N, et al. Diagnosis of ventilatorassociated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic ‘blind’ bronchoalveolar lavage fluid. Am Rev Respir Dis 1991; 143: 1121 – 9 13 Calandra T, Cohen J. The international sepsis forum consensus conference on definitions of infection in the intensive care unit. Crit Care Med 2005; 33: 1538 – 48 14 Crunden E, Boyce C, Woodman H, Bray B. An evaluation of the impact of the ventilator care bundle. Nurs Crit Care 2005; 10: 242 – 6 15 Grap MJ, Munro CL, Hummel RS, III, et al. Effect of backrest elevation on the development of ventilator-associated pneumonia. Am J Crit Care 2005; 14: 325 – 32 16 van Nieuwenhoven CA, Vandenbroucke-Grauls C, van Tiel FH, et al. Feasibility and effects of the semirecumbent position to prevent ventilator-associated pneumonia: a randomized study. Crit Care Med 2006; 34: 396– 402 17 Young MP, Manning HL, Wilson DL, et al. Ventilation of patients with acute lung injury and acute respiratory distress syndrome: has new evidence changed clinical practice? Crit Care Med 2004; 32: 1260 – 5 18 Weinert CR, Gross CR, Marinelli WA. Impact of randomized trial results on acute lung injury ventilator therapy in teaching hospitals. Am J Respir Crit Care Med 2003; 167: 1304 – 9 19 Wolthuis EK, Korevaar JC, Spronk P, et al. Feedback and education improve physician compliance in use of lung-protective mechanical ventilation. Intensive Care Med 2005; 31: 540 –6 20 Drakulovic MB, Torres A, Bauer TT, et al. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999; 354: 1851 – 8
doi:10.1093/bja/aep114
Advance Access publication May 20, 2009
CRITICAL CARE Effects of staff training on the care of mechanically ventilated patients: a prospective cohort study F. Bloos1*, S. Mu¨ller1, A. Harz1, M. Gugel1, D. Geil1, K. Egerland1, K. Reinhart1 and G. Marx1 2 1
Department of Anaesthesiology and Intensive Care Medicine, University Hospital Jena, Erlanger Allee 101, 07747 Jena, Germany. 2Department of Surgical Intensive Care Medicine, Medical Faculty, RWTH Aachen University, Pauwelsstr. 30, 52074 Aachen, Germany *Corresponding author. E-mail: [email protected]
Methods. This study was performed on a 50-bed intensive care unit of a tertiary care university hospital. Application of a ventilator bundle consisting of semirecumbent positioning, lung protective ventilation in patients with acute lung injury (ALI), ulcer prophylaxis, and deep vein thrombosis prophylaxis (DVTP) was assessed before and after staff training in post-surgical patients requiring mechanical ventilation for at least 24 h. Results. A total of 133 patients before and 141 patients after staff training were included. Overall bundle adherence increased from 15 to 33.8% (P,0.001). Semirecumbent position was achieved in 24.9% of patient days before and 46.9% of patient days after staff training (P,0.001). Administration of DVTP increased from 89.5 to 91.5% (P¼0.048). Ulcer prophylaxis of .90% was achieved in both groups. Median tidal volume in patients with ALI remained unaltered. Days on mechanical ventilation were reduced from 6 (interquartile range 2.0 – 15.0) to 4 (2.0 – 9.0) (P¼0.017). Rate of ventilator-associated pneumonia (VAP), ICU length of stay, and ICU mortality remained unaffected. In patients with VAP, the median ICU length of stay was reduced by 9 days (P¼0.04). Conclusions. Staff training by an ICU change team improved compliance to a pre-defined ventilator bundle. This led to a reduction in the days spent on mechanical ventilation, despite incomplete bundle implementation. Br J Anaesth 2009; 103: 232–7 Keywords: audit, trainings; complications, respiratory; intensive care, pulmonary; ventilation, mechanical Accepted for publication: April 14, 2009
Nosocomial infections are a major cause of morbidity and mortality in intensive care patients. Specifically, ventilatorassociated pneumonia (VAP) is known to increase the duration of mechanical ventilation, mortality, and length of stay and also costs of care.1 It has been estimated that 15% of ventilated patients develop pneumonia. The National Nosocomial Infections Surveillance System (NNIS) reports a median occurrence of VAPs of 4.6 –5.1 per 1000 ventilator days in medical-surgical intensive care units (ICUs).2 Besides the development of infectious complications, mechanically ventilated patients are also at risk
of developing several other complications. Mechanical ventilation with high tidal volumes or high plateau pressures may induce pulmonary injury by volutrauma, barotrauma or both.3 Other complications include deep vein thrombosis4 and the development of peptic ulceration.5 Therefore, low-tidal volume ventilation, stress-ulcer prophylaxis, and deep vein thrombosis prophylaxis are recommended in these patients. The concept of treatment bundles has been successfully used to introduce multiple therapies into clinical practice. A bundle is defined as a selected set of interventions or
# The Author [2009]. Published by Oxford University Press on behalf of The Board of Directors of the British Journal of Anaesthesia. All rights reserved. For Permissions, please email: [email protected]
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
Background. Adherence to guidelines to avoid complications associated with mechanical ventilation is often incomplete. The goal of this study was to assess whether staff training in predefined interventions (bundle) improves the quality of care in mechanically ventilated patients.
Staff training to improve care in the ICU
procedures distilled from evidence-based medicine.6 This concept is currently used in the care of patients with sepsis; however, there is still limited evidence for this approach in mechanically ventilated patients. The goal of this study was first to assess to what extent the pre-defined ventilator bundle applied to mechanically ventilated patients in a 50-bed interdisciplinary surgical ICU. Secondly, we wanted to investigate whether training of physicians and nurses in the application of this ventilator bundle improved the quality of care in these patients.
Methods
Definition of the ventilator bundle The ventilator bundle from the Institute of Health Care Improvement (IHI)8 9 was modified for application in this study. The daily sedation break was removed from the bundle because it has not been studied in surgical ICU patients.10 It was replaced by lung protective ventilation for patients with acute lung injury (ALI), as a German national survey showed that lung-protective ventilation is poorly applied.11 Thus, the bundle in this trial consisted of semirecumbent positioning, lung protective ventilation, stress ulcer prophylaxis, and deep vein thrombosis prophylaxis (DVTP). Correct positioning of the patient was defined as a semirecumbent positioning of at least 308. Lung-protective ventilation was defined as a maximum tidal volume of 6 ml kg21 of predicted bodyweight ( pbw) and a plateau pressure of ,30 cm H2O. DVTP or stress ulcer prophylaxis was defined as being achieved if the treating physician prescribed the appropriate medication in the medical record and the administration of the medication was signed by the nurse. Complete bundle adherence was defined as fulfilment of all four elements in patients with ALI and fulfilment of all
Study design The study was performed as a pre-/post-test design and consisted of three phases: (1) Description of the current state: between June and September 2005, a specially trained task force visited the enrolled patients every day randomly between 8:00 and 20:00 and recorded compliance to the treatment bundles (Audit I). (2) Training of staff in charge of mechanically ventilated patients: a change team consisting of the ICU manager, the ICU consultants and also interested ICU residents and nurses was formed. For a 2 month period, the results of Audit I, the scientific background and technique of the bundle, were taught to all nurses and residents of the ICU in daily seminars by the change team. Red marks were attached on the walls to indicate a correct semirecumbent position. After the 2 month teaching period, the change team monitored daily compliance to the treatment bundles and individually trained nurses and residents if the bundle was not correctly applied for another 2 months. (3) Control of performance: patient visits as in Audit I were repeated between March and June 2006 (Audit II).
Data acquisition The Acute Physiology and Chronic Health Evaluation II (APACHE II) score was calculated from data within 24 h of admission. During the daily audits, current tidal volume and plateau pressure were recorded from the ventilator. Maximum tidal volume and maximum plateau pressure were obtained from the Patient Data Management System (COPRAw, Sasbachwalden, Germany). Adherence to lungprotective ventilation was checked in patients with ALI (PaO2/FIO2 ,39.5 kPa, bilateral pulmonary infiltrates on the chest X-ray in the absence of left atrial hypertension). Angle for semirecumbent positioning was measured manually or, if applicable, by an indicator integrated in the ICU bed. Mean daily backrest elevation was calculated individually for each patient by the sum of measured angles divided by the number of observation. Mean daily plateau pressure and mean daily tidal volume were calculated accordingly. Medication for DVTP and for ulcer prophylaxis was taken from the medical record. The Clinical Pulmonary Infection Score (CPIS) was calculated daily.12 Onset of pneumonia was defined as the first CPIS of at least 6 combined with the presence of pulmonary infiltrates in the chest X-ray.13 VAP was assumed if criteria of pneumonia were fulfilled after 48 h of mechanical ventilation.
233
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
Patients were recruited between June 2005 and June 2006 in a 50-bed ICU of a tertiary care university hospital. Patients were included if they were mechanically ventilated for more than 24 h and were at least 10 years of age. A day on mechanical ventilation was defined as any 24 h period in which the patient required any mode of controlled or assisted ventilation, with the exception of intermittent application of continuous positive airway pressure for atelectasis prophylaxis.7 Standard measures of general critical care were applied to all patients, including hand washing before dealing with the patients, daily oral care, and sterile tracheal suction. The ventilation tubing systems remained unchanged. Standard operating procedures were in place for analgosedation, weaning from the ventilator, and antibiotic therapy. Assignment of patients to nursing staff and clinicians was not controlled by the study protocol. The local ethics committee of the University of Jena waived the need for informed consent.
but the lung-protective ventilation element in patients without ALI.
Bloos et al.
Table 1 Patient characteristics and sedation. Data are expressed as frequencies or as median (interquartile range). *Kolmogorov– Smirnow test
n/patient days Female sex Age (yr) Height (cm) Weight (kg) APACHE II score Mechanical ventilation on ICU admission Referring facility Operating theatre Other hospital Normal ward Direct admission Type of surgery Heart surgery General surgery Neurosurgery Traumatology Other
Audit I
Audit II
P-value
133/1389 37.6% 63 (47.0 –71.5) 172 (165 –180) 77.0 (70 –85) 24 (5.0 – 21.5) 85%
141/1002 30.7% 66 (52.0 –73.0) 172 (165 –179) 78.5 (67 –85) 25 (20.0 –31.0) 76.6%
0.31 0.08 0.97 0.98 0.39 0.09
51.1% 24.8% 14.3% 8.3%
36.2% 23.4% 26.2% 14.2%
0.07*
41.4% 19.5% 18.0% 12.8% 6.9%
45.4% 14.2% 30.5% 5.7% 4.2%
0.66*
n/patient days Midazolam (mg day21) Sufentanil (mg) Ketamine (mg) Propofol (mg)
Audit I
Audit II
P-value
133 / 1389 31.1 (11.9 – 75.3) n¼64 (48.1%) 460.1 (191.3 –992.4) n¼93 (69.9%) 363.5 (83.5 –1039.6) n¼22 (16.5%) 502.3 (130.2 –920.9) n¼99 (74.4%)
141/1002 27.1 (7.8 –46.3) n¼71 (51.1%) 365.8 (153.1 – 810.1) n¼89 (64.0%) 271.6 (8.0 –526.1) n¼28 (20.1%) 297.2 (102.7 – 629.2) n¼66 (47.5%)
0.22 0.63 0.16 0.18 0.27 0.53 0.09 ,0.01
Weaning was defined as the first day where the patient was either extubated without non-invasive ventilation, or, in case of patients with a tracheostomy, able to breath spontaneously without mechanical support. Intermittent mask CPAP for atelectasis prophylaxis did not count as mechanical ventilation.
Statistics The primary endpoint of the study was adherence to the ventilator bundle. Secondary endpoints were rate of pneumonia, days on mechanical ventilation, ICU survival, and length of stay on the ICU. Patients with VAP were analysed as subgroup. Inferential statistics were calculated by the Mann – Whitney U-test or in case of frequencies with the x2-test. Continuous data including mean daily backrest elevation are expressed as median and 25% and also 75% percentiles. A P-value of ,0.05 was considered as statistically significant. A conditional forward stepwise Cox regression analysis was performed to evaluate which parameters independently affect the duration of mechanical ventilation with a P-value of ,0.05. Data were analysed with SPSS 15.0 for Windows (SPSS Inc., Chicago, IL, USA).
Results One hundred and thirty-three patients (1389 patient-days) were included in Audit I and 141 patients (1002 patientdays) in Audit II. Patient characteristics data are given in Table 1 and were not statistically different between the two groups. Significantly less patients received propofol in Audit II when compared with Audit I (Table 2). Cumulative dosage of propofol per days on mechanical ventilation was also lower in Audit II. However, this difference did not reach statistical significance. There was no difference in usage of other sedatives.
Application of the ventilator bundle Application of the ventilator bundle is summarized in Figure 1. Complete bundle adherence in all patients increased from 15 to 33.8% after staff training (P,0.01). Bundle adherence differed between patients with and without ALI. In patients without ALI, the overall bundle adherence was 22.8% in Audit I and 47% in Audit II (P,0.01) when compared with 9.9 and 18.2% (P,0.01) in patients with ALI. Semirecumbent position with an angle of at least 308 was achieved on significantly more patient days in Audit
234
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
Table 2 Cumulative analgosedation per days on mechanical ventilation. Data are expressed as median (interquartile range). Mann –Whitney test was used for differences in analgosedation dosage. The n value gives the number of patients who actually received the sedative on at least 1 day (x2-test for differences)
Staff training to improve care in the ICU
II than in Audit I (49.6 vs 24.9%, P,0.01). This resulted in a median of the mean daily backrest elevations of 22.68 (25 and 75% percentile: 18.4; 26.2) in Audit I and 26.78 (25 and 75% percentile: 22.5; 29.5) in Audit II (P,0.01). Tidal volumes and plateau pressures are given in Table 3. Sixty-seven patients in Audit I and 61 patients in Audit II fulfilled criteria of ALI on at least 1 day. In the subgroup of patients with ALI, the median of the mean daily tidal volumes was unaffected by staff training and were above 6 ml kg21 pbw in both audits. Medians of the mean daily plateau pressures were ,30 cm H2O throughout the trial. DVTP was given significantly more in Audit II (91.9%) than in Audit I (89.5%, P¼0.048). However, contraindications for DVTP were present in 6.4% of patient days Audit I Audit II
Ulcer prophylaxis P=0.05
90
60 70
50
P<0.01 30 40
Semirecumb. positioning 20
P<0.01
0 10
Lung protective ventilation
80
Deep vein thromb. proph.
Frequency (%)
Fig 1 Frequency of successful implementation of treatment bundles before and after staff training. Data are given in percentage of patient days. Lung protective ventilation applies to patients with ALI only.
Outcome Median duration of mechanical ventilation was significantly shorter in Audit II when compared with Audit I (Table 4). This was associated with a faster liberation from mechanical ventilation (Fig. 2). ICU length of stay and overall mortality were not different between the audits. In Audit II, a higher tracheostomy rate was observed than in Audit I (30.9 vs 19.5%, P¼0.037). Duration until onset of VAP was longer after staff training without reaching statistical significance. Forty-four patients in Audit I and 45 patients in Audit II developed VAP. There was no significant difference in the frequency of VAP between Audit I (33.1%) and Audit II (32.4%, P¼0.68). Median APACHE II scores in patients with VAP were 24 (20; 29) in Audit I and 24 (20; 34) in Audit II (P¼0.49). A post hoc subgroup analysis on outcome was carried out in patients with VAP (Table 4). There was a significant reduction in the duration of mechanical ventilation by 8 days and of ICU length of stay by 9 days from Audit I to Audit II. The associated decrease in mortality rate from 43.2 to 35.6% did not reach statistical significance.
Table 3 Parameters of mechanical ventilation in all patients and in patients with acute lung injury. Data are expressed as median (interquartile range). Maximum tidal volume and maximum plateau pressure correspond to the daily maximum of this parameter. Pbw, predicted body weight
All patients n/patient days Mean daily tidal volume (ml kg21 pbw) Mean daily plateau pressure (cm H2O) Patients with acute lung injury n/patient days Mean daily tidal volume (ml kg21 pbw) Mean daily plateau pressure (cm H2O)
Audit I
Audit II
P-value
131 / 1198 7.8 (6.9 –9.2) 21.5 (18.3 – 24.4)
141 / 980 7.9 (7.1 – 8.9) 19.3 (17.3 –21.6)
0.65 ,0.01
67 / 497 7.2 (6.5 –9.8) 24 (21.9 –26.7)
61 / 315 7.5 (6.2 – 8.6) 21.2 (18.6 –25.3)
0.72 0.11
Table 4 Outcome in all patients, and in patients with ventilator-associated pneumonia. Data are expressed as median and interquartile range Audit I
Audit II
n¼133 6.0 (2.0 –15.0) 27.8% 12.0 (5.0 –21.5) 33.1%
n¼141 4.0 (2.0 –9.0) 25.5% 13.0 (6.0 – 21.0) 32.4%
n¼44 18 (13– 26) 43.2% 30.0 (17.3 –45) 5.3 (3.5 –8.9)
n¼45 10 (6 –17) 35.6% 21 (11.5 – 35) 6.4 (4.1 –11.0)
P-value
All patients Duration of mechanical ventilation (days) ICU mortality ICU length of stay (days) Frequency of VAP Patients with ventilator-associated pneumonia (subgroup analysis) Duration of mechanical ventilation (days) ICU mortality ICU length of stay (days) Duration until onset of VAP (days)
235
0.02 0.68 0.71 0.68
,0.01 0.52 0.04 0.18
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
P<0.01
Complete bundle
during Audit II when compared with 11.7% in Audit I (P,0.01). Ulcer prophylaxis was prescribed in 94.5% of patient days in Audit I and 94.9% patient days in Audit II (P¼0.71).
Bloos et al.
Fraction of successfully weaned patients
1.0 0.8
Audit I Audit II
0.6 0.4 Log-rank test: P=0.02
0.2 0.0 0
50 60 10 20 30 40 Duration of mechanical ventilation (days)
Fig 2 Kaplan– Meier curves for freedom from mechanical ventilation in patients from Audit I and Audit II.
Results from stepwise regression
Discussion The implementation of a ventilator bundle through staff training by an ICU change team led to a more than two-fold increase in overall bundle adherence. Staff training resulted in significantly better compliance in the semirecumbent positioning of mechanically ventilated patients. Additionally, application of DVTP was improved after staff training, although adherence to low tidal ventilation in patients with ALI was only moderate. However, improved compliance of the ventilator bundle was associated with a significant reduction in the duration of mechanical ventilation by 2 days. In a subgroup analysis, adherence to the bundle in patients with VAP was also associated with a significant reduction in days on mechanical ventilation and also on ICU length of stay. This confirms data from a smaller study where application of the IHI ventilator bundle reduced median duration of mechanical ventilation by 1 day.14 Backrest elevation was hardly implemented before staff training in our study population. An observational study showed that semirecumbent positioning is usually poorly applied in mechanically ventilated patients.15 In our study, staff training significantly improved adherence to backrest elevation but was far from complete. In a multicentre trial by Van Nieuwenhoven and colleagues,16 semirecumbent positioning as a single-study intervention could not be completely achieved, despite daily monitoring of the backrest position by research nurses.16
(1) The effects of semirecumbent positioning on VAP rates are known to be variable and may depend on the applied angle of elevation. A study, which only achieved a backrest elevation of 308 instead of 458, did not lead to a reduction in VAP rates.16 Another study has defined an angle of 308 as being the cut-off for effective reduction of VAP rates.15 The trial by Drakulovic and colleagues,20 which showed an effective reduction in VAP rates, used a backrest elevation of 458; therefore, the lack of effectiveness may have been caused by differences in positioning. However, 458 has been observed to be difficult to implement16 and the minimal backrest elevation to avoid microaspiration and thus VAP is unknown. (2) The diagnosis of VAP is difficult as there are no accurate diagnostic criteria reliably differentiating pneumonia from non-infectious causes of pulmonary infiltration. Thus, a degree of uncertainty in the VAP diagnosis may have negated the smaller effects of the ventilator bundle on outcome. The CPIS score is recommended by the International Sepsis Forum as a standardized approach for diagnosing pneumonia in clinical studies13 and has been applied in a similar setting.15 The high VAP rates observed in our study may be a result of the selection of the study population. APACHE II scores of 25 implies a high-risk group of patients while those low-risk patients ventilated for ,24 h were not included. (3) The application of the bundle may still increase pulmonary integrity by reducing microaspiration even if the bundle is implemented incompletely. Although we could not detect differences in VAP rates and are unable to prove this hypothesis, the observation that duration of mechanical ventilation and ICU length of stay decreased after staff training in the subgroup of VAP patients does lend some weight to this theory. (4) We cannot completely rule out that differences in the patient populations in this pre-/post-test design
236
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
In univariate statistics, tracheostomy rates, angle of semirecumbent position, number of patients receiving propofol, days with DVTP, and days where DVTP was contraindicated were different between Audit I and Audit II. These parameters were entered into a stepwise Cox-regression with duration of mechanical ventilation as a dependent variable. The algorithm included all parameters into the final model. The overall P-value of the final model was ,0.01. The angle of semirecumbent positioning was independently associated with days on mechanical ventilation (P¼0.02).
Staff training had little effect on the adherence to lungprotective mechanical ventilation for patients with ALI in the present study. This problem has been demonstrated in several other studies11 17 18 despite the trial by the ARDS-Network demonstrating a significant reduction in mortality when low tidal volume ventilation was applied.3 As we tried to implement a bundle with four separate treatments, the training relating to lung-protective ventilation may have remained relatively unnoticed in comparison with the other three treatments. A similar study, where low tidal volume ventilation was the only intervention, successfully reduced tidal volumes.19 In our study, decrease in the duration of mechanical ventilation and application of the ventilator bundle did not alter the frequency of VAP despite an increase in backrest elevation in Audit II. Several reasons may be responsible for this observation:
Staff training to improve care in the ICU
accounted for some differences in outcome. However, disease severity and patient characteristics were similar in the two audits.
References 1 Warren DK, Shukla SJ, Olsen MA, et al. Outcome and attributable cost of ventilator-associated pneumonia among intensive care unit patients in a suburban medical center. Crit Care Med 2003; 31: 1312 – 7 2 NNIS System. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 2004; 32: 470 – 85 3 Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000; 342: 1301 – 8 4 Samama MM, Cohen AT, Darmon JY, et al. A comparison of enoxaparin with placebo for the prevention of venous thromboembolism in acutely ill medical patients. Prophylaxis in Medical Patients with Enoxaparin Study Group. N Engl J Med 1999; 341: 793 – 800 5 Steinberg KP. Stress-related mucosal disease in the critically ill patient: risk factors and strategies to prevent stress-related bleeding in the intensive care unit. Crit Care Med 2002; 30: S362 – S364
237
Downloaded from http://bja.oxfordjournals.org/ at colostateuniv on July 10, 2015
There is some limitation in the interpretation of this study. Tracheostomy rates differed significantly between the two audits. ICU procedures to regulate the decision for tracheostomy in mechanically ventilated patients were not changed between the two audits nor were procedures for tracheostomy part of the teaching process. We therefore performed a multivariate analysis to clarify this issue which showed that both semirecumbent positioning and the presence of a tracheostomy independently affected the duration of mechanical ventilation. In conclusion, the ventilator bundle was not well applied before training. Implementation of a ventilator bundle through staff training by an ICU change team improved compliance to a pre-defined ventilator bundle and reduced days on mechanical ventilation even though the ventilator bundle was not implemented completely. Incompleteness of adherence indicates the fact that higher success rates may only be achieved by a continuous quality improvement process.
6 Levy MM, Pronovost PJ, Dellinger RP, et al. Sepsis change bundles: converting guidelines into meaningful change in behavior and clinical outcome. Crit Care Med 2004; 32: S595 – S597 7 Schoenfeld DA, Bernard GR. Statistical evaluation of ventilatorfree days as an efficacy measure in clinical trials of treatments for acute respiratory distress syndrome. Crit Care Med 2002; 30: 1772 – 7 8 Youngquist P, Carroll M, Farber M, et al. Implementing a ventilator bundle in a community hospital. Jt Comm J Qual Patient Saf 2007; 33: 219 – 25 9 Resar R, Pronovost P, Haraden C, et al. Using a bundle approach to improve ventilator care processes and reduce ventilatorassociated pneumonia. Jt Comm J Qual Patient Saf 2005; 31: 243 – 8 10 Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med 2000; 342: 1471 – 7 11 Brunkhorst FM, Engel C, Ragaller M, et al. Practice and perception—a nationwide survey of therapy habits in sepsis. Crit Care Med 2008; 36: 2719 – 25 12 Pugin J, Auckenthaler R, Mili N, et al. Diagnosis of ventilatorassociated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic ‘blind’ bronchoalveolar lavage fluid. Am Rev Respir Dis 1991; 143: 1121 – 9 13 Calandra T, Cohen J. The international sepsis forum consensus conference on definitions of infection in the intensive care unit. Crit Care Med 2005; 33: 1538 – 48 14 Crunden E, Boyce C, Woodman H, Bray B. An evaluation of the impact of the ventilator care bundle. Nurs Crit Care 2005; 10: 242 – 6 15 Grap MJ, Munro CL, Hummel RS, III, et al. Effect of backrest elevation on the development of ventilator-associated pneumonia. Am J Crit Care 2005; 14: 325 – 32 16 van Nieuwenhoven CA, Vandenbroucke-Grauls C, van Tiel FH, et al. Feasibility and effects of the semirecumbent position to prevent ventilator-associated pneumonia: a randomized study. Crit Care Med 2006; 34: 396– 402 17 Young MP, Manning HL, Wilson DL, et al. Ventilation of patients with acute lung injury and acute respiratory distress syndrome: has new evidence changed clinical practice? Crit Care Med 2004; 32: 1260 – 5 18 Weinert CR, Gross CR, Marinelli WA. Impact of randomized trial results on acute lung injury ventilator therapy in teaching hospitals. Am J Respir Crit Care Med 2003; 167: 1304 – 9 19 Wolthuis EK, Korevaar JC, Spronk P, et al. Feedback and education improve physician compliance in use of lung-protective mechanical ventilation. Intensive Care Med 2005; 31: 540 –6 20 Drakulovic MB, Torres A, Bauer TT, et al. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999; 354: 1851 – 8