International Journal of Nursing Studies 47 (2010) 1425–1431

The use of bed-dials to maintain recumbent positioning for critically illmechanically ventilated patients (The RECUMBENT study): Multicentrebefore and after observational study

Louise Rose a,*, Ian Baldwin b,c, Tom Crawford d

a Lawrence S. Bloomberg Faculty of Nursing, University of Toronto, 155 College St, Suite 276, Toronto, ON, Canada M5T IP8b Department of Intensive Care, Austin Hospital, Victoria, Australiac RMIT University, Melbourne, Australiad Intensive Care Unit, Geelong Hospital, Victoria, Australia

A R T I C L E I N F O

Article history:

Received 11 December 2009

Received in revised form 2 March 2010

Accepted 12 April 2010

Keywords:

Ventilator-associated pneumonia

Semirecumbent

Backrest elevation

Nursing

Positioning

Nosocomial infection

A B S T R A C T

Background: Observational studies continue to report poor compliance with positioning

recommendations for prevention of ventilator-associated pneumonia. Inability to

accurately measure backrest elevation may contribute to this poor compliance.

Objective: To determine if provision of an accurate, simple to use angle measurement

device with an accompanying education program improved compliance with semi-

recumbency at 458 over time.

Design: Using a prospective pre- and post-design we implemented angle measurement

devices and an associated education intervention in three Australian ICUs. Backrest

elevation, contraindications to semirecumbency at 458, mean arterial pressure (MAP),

inotrope use, enteral feeding and weaning status were recorded 3-times daily using a pre-

determined randomization schedule for 7 consecutive days prior to implementation and

again at 1, 3 and 6 months post-implementation. Illness severity and a clinical pulmonary

infection score were recorded for each day of ventilation.

Results: Backrest elevation measurements (n = 1154) were recorded for 141 mechanically

ventilated patients. Contraindications to semirecumbency at 458 were noted for 163/1154

(14.1%) measurements the proportion of measurements at 458 rose from baseline by 10.1%

(P = 0.03) 1-month following implementation, however this change was not sustained

over time. The proportion of measurements 308 increased by 43.8% at 1-month and

remained above 70% 6-months after implementation (P< 0.001).

For measurements recorded in the absence of a contraindication to semirecumbency,

and adjusted for covariates (MAP, inotropic support, sequential organ failure assessment

maximum score, clinical pulmonary infection score maximum, and indication for

ventilation), decreased backrest elevation was associated with higher severity of illness

(0.38 [95% CI 0.1–0.5] for every 1-point increase in APACHE II score). Increased mean

backrest elevation was noted for older patients (0.88 [95% CI 0.1–1.5] for each 10-year

increment) and measurements recorded during weaning (2.78 [95% CI 1.2–4.1]).

Conclusions: Bedside implementation of an angle measurement device and associated

educational intervention did not result in a sustained improvement to compliance with

458 semirecumbency, questioning the clinical feasibility of this nursing intervention. A

sustained increased in semirecumbency at 308 or greater was achieved.

� 2010 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

International Journal of Nursing Studies

journal homepage: www.elsevier.com/ijns

* Corresponding author. Tel.: +1 416 978 3492; fax: +1 416 946 0665.

E-mail address: louise.rose@utoronto.ca (L. Rose).

0020-7489/$ – see front matter � 2010 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ijnurstu.2010.04.002

L. Rose et al. / International Journal of Nursing Studies 47 (2010) 1425–14311426

What is already known about the topic?

� S

emirecumbent positioning at 458 prevents ventilator-associated pneumonia. � P oor compliance with 458 semirecumbency has been

documented by several studies.

� In the absence of a standardized measurement tool,

nurses overestimate backrest elevation.

What this paper adds

� T

his paper provides further support to the notion thatsemirecumbent positioning of critically ill mechanicallyventilated patients at 458 may not be clinically feasible. � W eaning status was the strongest predictor of increased

mean backrest elevation.

� B efore new quality initiatives are developed to overcome

poor compliance with 458 semirecumbency, furtherresearch is required to confirm the superiority of 458semirecumbency for VAP prevention over lower, moreclinically feasible degrees of backrest elevation.

1. Introduction

Multiple practice guidelines are now available thatrecommend semirecumbent positioning with bed backrestelevation equivalent to or greater than 308 for preventionof ventilator-associated pneumonia (VAP) (AmericanAssociation of Critical Care Nurses, 2005; AmericanThoracic Society, 2005; Dodek et al., 2004; Tablen et al.,2004). These guidelines are based on clinical studiesdemonstrating increased aspiration of gastric contents bypatients in the supine position (Orozco-Levi et al., 1995;Torres et al., 1992), and reports including a singlerandomized controlled trial (Drakulovic et al., 1999)linking prolonged supine positioning, with little to nobackrest elevation, to increased risk of VAP (Grap et al.,2005; Kollef, 1993).

Although there is strong evidence indicating failure toplace patients in a semirecumbent position may worsenoutcomes for mechanically ventilated patients, there isuncertainty as to the exact degree of backrest elevationrequired for prevention of VAP (Wip and Napolitano,2009). This is partly due to a lack of studies comparing theeffect of various degrees of backrest elevation and VAPincidence but also due to uncertain or poor compliancewith targeted positioning in existing studies. The afore-mentioned randomized, controlled trial (Drakulovic et al.,1999) only assessed correctness of position (>458) dailyand did not report compliance. A subsequent randomized,controlled trial implemented continuous monitoring ofbackrest elevation via a transducer placed on the bed-frame and a dedicated research nurse to restore correctpositioning three times a day for 7 days. In this study only15% of patients achieved backrest elevation to 458 andthere was no reduction in VAP incidence for the experi-mental group (van Nieuwenhoven et al., 2006).

To conduct further studies examining the effects ofvarious degrees of semirecumbency and VAP prevention aswell as the potential for pressure ulcer formation (Peterson

et al., 2008), compliance with study targets must beassured. Improved compliance with positioning recom-mendations has been demonstrated using standardizedorders and an associated education program in a singlecentre (Helman et al., 2003). Other reported educationaland quality assurance interventions target multiple VAPprevention strategies resulting in limited reporting ofpositioning practices that may inform further research orguide clinical application (Babcock et al., 2004; Blamounet al., 2009; Bloos et al., 2009; Zack et al., 2002).

Accurate measurement of semirecumbency is essential.Estimation studies report clinicians may overestimate theangle of semirecumbency resulting in inadequate eleva-tion (McMullin et al., 2002; Peterlini et al., 2006) thoughresults are not consistent (Dillon et al., 2002). Some bedsdesigned for critically ill patients have angle measurementdevices incorporated into their design (Hill-Rom, Bateville,IN), however availability of these beds is not standardizedacross intensive care units (ICUs). Nurses should not haveto rely on estimation methods and require accurate,reliable and easy to use tools to guide semirecumbentpositioning.

We hypothesized that implementation of an educationprogram and provision of an accurate, simple to use anglemeasurement device available at every bed would improvecompliance with semirecumbent positioning and compli-ance would be maintained over time. We conducted aprospective, multicentre study with repeated observationsto evaluate this hypothesis.

2. Methods

2.1. Sample

During four observation periods over 7-days eachduring 2008 (January, March, June and November) allpatients 16 years and over admitted to the 3 participatingICUs and receiving mechanical ventilation were eligible forthe study. Patients receiving non-invasive ventilation wereexcluded. Patients became ineligible for further measure-ment of backrest elevation on extubation. Participatingunits were all mixed population ICUs in university-affiliated hospitals. The Institutional Review Boards ofthe participating hospitals and the University of Torontoapproved the study, waiving the requirement for informedconsent.

2.2. Study procedures

We prospectively evaluated compliance with semire-cumbent positioning at baseline then 1, 3 and 6 monthsafter implementation of the angle measurement devicesand accompanying education program. Participating ICUs,selected as a convenience sample, were provided with abed elevation measurement device for each bed-spacewith additional devices available in case of loss or damage.The measurement device consisted of an angle indicatordial attached to the bed head-frame using magnets. Theangle indicator dial used gravity to indicate, with a plasticmovable arrow, thus measuring the degree of backrestelevation. Accuracy of angle detection was confirmed prior

L. Rose et al. / International Journal of Nursing Studies 47 (2010) 1425–1431 1427

to study implementation using a mounted protractor andmovable metal arm.

To promote awareness of the study and necessaryeducation for semirecumbent positioning, we providedlarge colour picture posters to each ICU detailing studyaims and requirements, use of the device, and contra-indications to semirecumbency. A 15-slide power point(Microsoft 2007) presentation with background informa-tion highlighting the importance of semirecumbency,detail on study goals, and how to use the measurementdevices was provided to each ICU. These resources wereused for group education sessions, email communicationsent to all staff, and were made easily accessible oncomputers located either at the bedside or nursing stationfor all clinical staff to view at their own convenience. Thenumber of education sessions varied according to sitebased on assessed need. Site coordinators also conductedindividual bedside education as required and questionedpositioning practices during ICU rounds. Nurses wereinstructed to elevate the head of bed to 458 unlesscontraindicated. Absolute contraindications to semire-cumbency were defined as suspected spinal injury,intracranial hypertension, hemodynamic instability,unstable pelvic injury, intra-aortic balloon pump in situ,and prone positioning. Regular in-service educationsessions that reinforced the need for semirecumbentpositioning and gave feedback on previous audits wereprovided by site champions throughout the 6-month studyperiod.

During each observation period the degree of backrestelevation was measured at 3 time points per day for sevenconsecutive days using a pre-determined, randomizedmeasurement schedule. Using a random number gen-erator, measurement times were selected from 3 8-h timeblocks (08:00–15:59, 16:00–23:59 and 00:00–07:59) foreach 24-h period to ensure measurements were distrib-uted across all nursing shifts and to prevent nursesanticipating the time of observation and altering position-ing accordingly. Site investigators, not directly responsiblefor patient positioning, performed measurements. Nursesremained unaware of the measurement schedule.

Prior to data collection all site investigators were sent astudy pack including a detailed description of datacollection requirements, including definitions of all dataitems, study procedures and pictorial representation ofaccurate use of the bed-dials. Site coordinators wereregistered nurses employed as research coordinators withsubstantial training and experience in the collection ofresearch data or senior nurse educators experienced indata collection for project management purposes.

To examine variables possibly associated with semi-recumbency we collected data on feeding and mechanicalventilator weaning status, age, mean arterial bloodpressure (MAP), use of inotropes, and illness severityorgan dysfunction (Sepsis-related Organ Failure Assess-ment [SOFA] score) (Vincent et al., 1996), APACHE II scorerecorded during the first 24 h of ICU admission, and thereason for ventilation. We also recorded the modifiedclinical pulmonary infection score (CPIS) (Luna et al., 2003;Pugin et al., 1991) for each day of ventilation. The SOFAmaxscore (Moreno et al., 1999) was determined by summing

the worst daily scores for each of the six components of theSOFA score over the data collection period. A CPISmaxscore was calculated by summing the worst daily scores ofthe five elements of the modified CPIS score. Presence ofVAP determined by the attending clinician on clinicalfindings or microbiological confirmation was assesseddaily. Contraindications for semirecumbent positioning at458 from the horizontal were also noted.

2.3. Statistical analysis

To assess the effectiveness of our intervention we usedrepeated measures regression analysis, controlling forcorrelation among measurements taken for the samepatient and those recorded in the same ICU, to examinechange in backrest elevation over the four observationperiods. We examined univariate and multivariate rela-tionships between the degree of backrest elevation andpatient variables considered potentially influential onpatient positioning. Prior to inclusion in the multivariatemodel, the variables of interest were assessed for thepresence of multicollinearity using a tolerance statistic.Continuous variables were reported as means andstandard deviations or medians and interquartile rangesdependent on distribution. Categorical variables werereported as proportions and compared using chi-squaretests. All analyses were conducted using SAS Version 9.1(SAS Institute, Cary, NC, USA) by a statistician.

3. Results

We recorded data on 141 patients during the 4observation periods. Demographic characteristics of the141 patients are shown in Table 1 according to eachmeasurement period. Postoperative respiratory failure wasthe most frequent indication for mechanical ventilation.

VAP was diagnosed in 6/141 (4.2%) patients; 3/6 (50.0%)VAP diagnoses were confirmed microbiologically and 3/6(50.0%) diagnoses were based only on clinical findings.

A total of 1154 backrest elevation measurements wererecorded during the four observation periods. Contra-indications to semirecumbency at �458 were noted for163/1154 (14.1%) measurements. The most frequentcontraindications were hemodynamic instability (76/1154, 6.6%) and a procedure underway (41/1154, 3.6%)(Table 2).

The mean degree of backrest elevation increased frombaseline and remained above 308 across the 4 observationperiods for measurements recorded in the absence of acontraindication to �458 semirecumbency (P< 0.01). Theproportion of measurements at�458 rose from baseline by10.1% (P = 0.03) 1-month following implementation, how-ever this change was not sustained over time. Theproportion of measurements �308 increased by 43.8% at1-month and remained above 70% 6-months after imple-mentation (P< 0.001) (Table 3). When examining posi-tioning data by site, all 3 ICUs demonstrated a significantincrease in the proportion of patients with a backrestelevation �308. Only 1 ICU increased the proportion ofpatients positioned at �458 but this was not maintained atthe 6-month evaluation (Table 4).

Table 1

Demographic characteristics of patients.

Time 0 (n = 36) Time 1 (n = 33) Time 2 (n = 40) Time 3 (n = 32)

Age (years)a 69 (56–76) 60 (48–74) 66 (49–73) 69 (54–76)

Male gender 21 (58) 17 (52) 22 (55) 20 (63)

APACHE II scorea 17 (15–20) 15 (13–20) 19 (15–26) 18 (13–26)

SOFAmax scorea 6 (3–8) 7 (4–11) 9 (5–11) 9 (6–12)

CPISmax scorea 5 (3–6) 5 (3–6) 5 (3–6) 4 (3–5)

Indication for ventilation

Postoperativeb 15 (41.7) 16 (48.5) 17 (42.5) 19 (59.4)

Acute respiratory failurec 6 (16.7) 1 (3.0) 8 (20.0) 2 (6.3)

Trauma 4 (11.1) 0 (0.0) 1 (2.5) 4 (12.5)

Cardiac arrest 4 (11.1) 3 (9.1) 4 (10.0) 1 (3.1)

Sepsis 3 (8.3) 3 (9.1) 3 (7.5) 3 (9.4)

Overdose 0 (0.0) 4 (12.1) 2 (5.0) 0 (0.0)

Primary neurological disorder 1 (2.8) 1 (3.0) 3 (7.5) 1 (3.1)

Heart failure 0 (0.0) 2 (6.1) 0 (0.0) 0 (0.0)

COPD 0 (0.0) 1 (3.0) 0 (0.0) 1 (3.1)

Otherd 2 (5.5) 2 (6.1) 2 (5.0) 1 (3.1)

Not reported 1 (2.8) 0 (0.0) 0 (0.0) 0 (0.0)

APACHE: acute physiology and chronic health evaluation, SOFAmax: sequential organ failure assessment maximum, CPISmax: clinical pulmonary infection

score maximum, COPD: chronic obstructive pulmonary disease.a Median and (interquartile range). All other values are n (%).b Postoperative: refers to those patients who require continuation of mechanical ventilation after surgery because of underlying medical problems,

advanced age, airway maintenance or the high risk of the operative procedure.c Acute respiratory failure: includes pneumonia, aspiration, acute respiratory distress syndrome, and asthma.d Other: anaphylaxis, liver failure, diabetic ketoacidosis, acute renal failure, altered mental status due to hyponatremia, empyema, and pleural effusion.

L. Rose et al. / International Journal of Nursing Studies 47 (2010) 1425–14311428

Of the 1154 measurement episodes, 349 (30.2%) wererecorded during administration of inotropes, 741 (64.2%)while the patient received enteral feeding and 518 (44.9%)during weaning. Univariate analysis adjusting for the effectof measurements recorded in the same ICU and the samepatient using generalized estimating equations found themean degree of backrest elevation was increased duringweaning by 3.478 (95% CI 1.87–5.08, P< 0.0001). No other

Table 2

Contraindications to semirecumbent positioning at �458.

Time 0

(n = 344)

No contraindication 302 (87.8)

Hemodynamic instability 10 (2.9)

Undergoing procedure 17 (4.9)

Intracranial hypertension 3 (0.9)

Suspected or existing spinal injury 0 (0)

Femoral catheter in situ 3 (0.9)

Intra-aortic balloon pumping 0 (0.0)

Unstable pelvic fracture 0 (0.0)

Abdominal wound 0 (0.0)

Not reported 9 (2.6)

Table 3

Outcomes measured by study period.

Outcome Time 0

(n = 344a)

(n = 302b)

Degree of semirecumbency (no contraindications)c 26.6� 13.8

Degree of semirecumbency (all measurements)c 26.3� 13.9

Angle �458 (no contraindications), n (%) 34 (11.2)

Angle �308 (no contraindications), n (%) 97 (32.1)a All recorded measurements.b Measurements recorded in the absence of a defined contraindication.c Mean and standard deviation.

patient variable was associated with a change in backrestelevation. Multivariate analysis found weaning status,patient age and measurement period were all indepen-dently associated with increased mean backrest elevationwhen controlling for the effects of SOFAmax score, MAP,indication for ventilation, inotropic support, and CPISmax.Increasing APACHE II score was independently associatedwith decreased backrest elevation (Table 5).

Time 1

(n = 192)

Time 2

(n = 340)

Time 3

(n = 278)

174 (90.6) 282 (82.9) 233 (83.8)

5 (2.6) 12 (3.5) 17 (6.1)

8 (4.2) 17 (5.0) 7 (2.5)

0 (0.0) 5 (1.5) 0 (0.0)

2 (1.0) 3 (0.9) 5 (1.8)

0 (0.0) 0 (0.0) 4 (1.4)

0 (0.0) 11 (3.2) 5 (1.8)

0 (0.0) 1 (0.3) 0 (0.0)

0 (0.0) 9 (2.6) 2 (0.7)

3 (1.6) 0 (0.0) 5 (1.8)

Time 1

(n = 192a)

(n = 174b)

Time 2

(n = 340a)

(n = 282b)

Time 3

(n = 278a)

(n = 233b)

35.6� 15.8 33.2� 9.6 31.6� 9.5

34.3� 16.3 31.3� 10.6 30.3� 9.7

37 (21.3) 35 (12.4) 26 (11.2)

132 (75.9) 206 (73.0) 164 (70.4)

Table 4

Outcomes by site for measurements in the absence of contraindications.

Baseline, n (%) 1 month, n (%) 3 months, n (%) 6 months, n (%) P-Value

Unit 1 �45 19/160 (11.9) 9/36 (25.0) 9/97 (9.3) 11/75 (14.7) 0.11

Unit 2 �45 0/22 (0) 9/28 (32.1) 1/30 (3.3) 1/36 (2.8) <0.001

Unit 3 �45 15/120 (12.5) 19/110 (17.3) 25/155 (16.1) 14/132 (10.6) 0.39

Unit 1 �30 52/160 (32.5) 22/36 (61.1) 61/97 (62.9) 47/75 (62.7) <0.001

Unit 2 �30 2/22 (9.1) 21/28 (75.0) 17/30 (56.7) 15/36 (41.7) <0.001

Unit 3 �30 43/120 (35.8) 89/110 (80.9) 128/155 (82.5) 102/132 (77.3) <0.001

Table 5

Variables associated with backrest elevation in the absence of contraindications.

Variable Degree of backrest elevation change

Univariate mean degrees (95% CI) Multivariate mean degrees (95% CI)

Age (10-year increment) +0.16 (�0.54 to +0.86) +0.79 (+0.12 to +1.45)

Enteral feeding +0.35 (�1.56 to +2.25) +1.60 (�0.20 to +3.39)

Weaning +3.47 (+1.87 to +5.08) +2.66 (+1.17 to +4.15)

Inotropic support �0.62 (�2.66 to +1.42) +0.58 (�1.36 to +2.53)

Mean arterial pressure (10 mm Hg increment) +0.20 (�0.35 to +0.75) +0.33 (�0.23 to +0.89)

Reason for mechanical ventilation

Non-intrapulmonarya,b 1 1

Intrapulmonaryc +2.57 (�0.74 to +5.87) +1.05 (�1.61 to +3.71)

Postoperative +1.54 (�1.47 to +4.54) �0.33 (�2.95 to +2.28)

SOFAmax score increment +0.09 (�0.23 to +0.41) �0.10 (�0.43 to +0.24)

APACHE II score increment �0.07 (�0.26 to +0.12) �0.30 (�0.48 to �0.11)

CPISmax score increment +0.31 (�0.30 to +0.93) 0.39 (�0.17 to +0.96)

Measurement period

6-monthb 1 1

3-month +2.44 (�0.93 to 5.81) +3.28 (+0.46 to +6.10)

1-month +2.93 (�0.81 to +6.68) +1.45 (�1.73 to +4.62)

Baseline �5.29 (�8.68 to �1.91) �6.68 (�9.61 to �3.76)a Non-intrapulmonary reasons of mechanical ventilation include trauma, cardiac arrest, sepsis, overdose, traumatic head injury, primary neurological

disorders, heart failure, and neuromuscular disease.b Reference categoryc Intrapulmonary reasons of mechanical ventilation include pneumonia, aspiration, COPD, and ARDS.

L. Rose et al. / International Journal of Nursing Studies 47 (2010) 1425–1431 1429

4. Discussion

Our prospective, multicentre study evaluating theeffect of a bedside angle measurement device andaccompanying education program found compliance withsemirecumbency at �458 was not achieved. A significantincrease did occur in the proportion of patients positionedwith the backrest elevation at �308, an increase that wassustained over time. Our findings support those of aprevious study suggesting semirecumbency at 458 may notbe feasible (van Nieuwenhoven et al., 2006). Our study wasnot designed to examine the effect of our intervention onVAP incidence but to determine if relatively simplemeasures could improve and sustain compliance withcurrent positioning recommendations.

Guidelines for the prevention of VAP recommendsemirecumbency at �308 (American Association of CriticalCare Nurses, 2005), 30–458 (American Thoracic Society,2005; Tablen et al., 2004), and�458 (Muscedere et al., 2008),yet high level evidence for VAP prevention using semi-recumbency exists only for positioning at�458 (Drakulovicet al., 1999; Orozco-Levi et al., 1995; Torres et al., 1992). Theoriginal rationale for evaluating 458 semirecumbency forprevention of aspiration of gastric contents and VAP asopposed to the potentially more feasible backrest elevation

of 308 in existing studies is uncertain. New studies arerequired to confirm the equivalence or lack of inferiority oflower degrees of backrest elevation to strict 458 semire-cumbency. We would recommend future studies considerprograms such as our own involving multiple strategiesincluding objective measurement devices, education pro-grams that incorporate ongoing audit and feedback as wellas standardized orders to maximize compliance withsemirecumbency targets.

We found independent predictors of increased meanbackrest elevation were weaning status and older age.Weaning status has previously been shown to influencebackrest elevation (Cook et al., 2002; Rose et al., in press),yet the rationales for increased elevation during weaningare unclear. Patients receive less sedation and may be moreawake during weaning which may influence positioningpractices if semirecumbency is perceived to result in poorpatient comfort during sleep (Helman et al., 2003).However, no identified study has examined the associationbetween positioning and sedation score or use of sedativeagents. Semirecumbency also improves oxygenation andmaximizes ventilatory efforts due to decreased abdominalcompression of lung bases and increased recruitment ofdependent lung areas (Klingstedt et al., 1990; Richard et al.,2006).

L. Rose et al. / International Journal of Nursing Studies 47 (2010) 1425–14311430

The APACHE II score recorded for the first 24 h of ICUadmission was associated with decreased backrest eleva-tion, a finding also observed in previous observationalstudies (Evans, 1994; Grap et al., 2005). Increased illnessseverity may result in cardiovascular and respiratoryinstability as well as an increase in the number ofprocedures that restrict the use of semirecumbentpositioning. As identified in previous observational studies(Grap et al., 1999, 2003; Rose et al., in press), enteralfeeding did not influence backrest elevation despiteemphasis in current clinical guidelines that these patientsmay be at increased risk of aspiration.

Subjective estimation of backrest elevation may con-tribute to poor compliance with positioning guidelines.Peterlini et al. (2006) found only 15% of angles presented toa cohort of registered nurses, nursing students and nurseassistants were estimated accurately, with overestimationoccurring most frequently. Interestingly accuracy declinedas the angle of elevation increased. Though estimationstudies do not demonstrate consistent results (Dillon et al.,2002), provision of objective measurement devices willavoid under or overestimation of backrest angle. Imple-mentation of an angle measurement device demonstratedincreased compliance with backrest elevation recommen-dations in a previous single centre study (Williams et al.,2008).

Our study is limited in that we did not collect reasonsfor non-compliance with semirecumbency directly fromnursing staff at each measurement episode. We did collectpreviously described contraindications to semirecum-bency at 458 (Cook et al., 2002) and excluded from ouranalysis of compliance with positioning recommendationsthose measurements recorded when a contraindicationwas present. A previous single centre study implementingan education program and standardized order set butwithout an angle measurement device also demonstratedpoor compliance with 458 backrest elevation (Helmanet al., 2003). In this study, when asked to report on thereasons why compliance was not achieved, nursesexpressed concerns about the risk of patients slidingdown the bed, impairment of lateral rotation, poor patientcomfort particularly during sleep, increased risk of skinbreakdown and the effect of semirecumbency on hemo-dynamic monitoring.

Another limitation of our study was the use of an anglemeasurement device that described backrest elevation ofthe bed as opposed to the elevation of the patient’s trunk.Increased backrest elevation can cause patients to slideinto a position that impedes semirecumbency and maycompromise chest expansion and potentially worsen gasexchange due to atelectasis. Due to the non-randomized,quasi-experimental before and after design of this studywe were unable to account for the influence on our resultsof confounders including changes to the staffing andpatient profile within participating units, interactiveeffects of changes to unit policy or clinical procedures,as well as other VAP prevention interventions or programsincorporating positioning recommendations that mayhave influenced our results. In the Australian State ofVictoria the Safer Systems Saving Lives initiative of theDepartment of Human Services advocates semirecumbent

positioning in their toolkit for prevention of ventilator-associated complications (Department of Human Services,2005).

5. Conclusions

In this multicentre prospective study we could notdemonstrate compliance with semirecumbency at 458suggesting that this degree of backrest elevation is notfeasible for critically ill patients requiring mechanicalventilation. However we did demonstrate an improvementin the degree of backrest elevation to greater than 308, andthis was maintained over time. Reasons for the inability toachieve semirecumbency at �458 remain unclear andwarrant further investigation. By using an objectivemeasurement device and excluding measurementsrecorded when predefined contraindications to semire-cumbency were present we attempted to remove twopossible explanations for non-compliance. Potentiallyother non-measured nurse perceived contraindicationssuch as patient discomfort or perceived risk of pressureulceration remained influential.

Before new quality initiatives are developed to over-come poor compliance with 458 semirecumbency there areseveral questions that need answers. Future studies needto confirm the superiority of 458 semirecumbency for VAPprevention over lower, more clinically feasible degrees ofbackrest elevation. The incidence of adverse eventsassociated with 45 and 308 positioning need to bequantified. In addition, nurse decision making with regardto semirecumbent positioning requires more in-depthexploration.

Acknowledgements

The authors would like to thank the study coordinatorsof the three Intensive Care Units for their assistanceprovided to this study. In addition, the authors thank DrAlex Kiss for his statistical guidance.

Conflicts of interest

None declared.

Funding

Funding from the Connaught Start-up Award of theUniversity of Toronto supported this study.

Ethical approval

Ethics approval was obtained from the University ofToronto Ref no. 21874 and the 3 participating sites.

References

American Association of Critical Care Nurses, 2005. AACN practice alert:ventilator associated pneumonia. AACN Clin. Issues 16, 105–109.

American Thoracic Society, 2005. Guidelines for the management ofadults with hospital-acquired, ventilator-associated, and health-care-associated pneumonia. Am. J. Respir. Crit. Care Med. 171,388–416.

Babcock, H., Zack, J., Garrison, T., Trovillion, E., Jones, M., Fraser, V., Kollef,M., 2004. An educational intervention to reduce ventilator-associatedpneumonia in an integrated health system. Chest 125, 2224–2231.

Blamoun, J., Alfakir, M., Rella, M., Wojcik, J., Solis, R., Khan, M., VeBari, V.,2009. Efficacy of an expanded ventilator bundle for the reduction of

L. Rose et al. / International Journal of Nursing Studies 47 (2010) 1425–1431 1431

ventilator-associated pneumonia in the medical intensive care unit.Am. J. Infect. Control 37, 172–175.

Bloos, F., Muller, S., Harz, A., Gugel, M., Geil, D., Egerland, K., Reinhart, K.,Marx, G., 2009. Effects of staff training on the care of mechanicallyventilated patients: a prospective cohort study. Br. J. Anaesth. 103,232–237.

Cook, D., Meade, M., Hand, L., McMullin, J., 2002. Toward understandingevidence uptake: semirecumbency for pneumonia prevention. Crit.Care Med. 30, 1472–1477.

Department of Human Services, 2005. Preventing Ventilator-AssociatedComplications Toolkit. Department of Human Services, Melbourne.Availability confirmed 5th December 2009 at http://www.health.-vic.gov.au/sssl/downloads/prev_ventilator.pdf.

Dillon, A., Munro, C., Grap, M., 2002. Nurses’ accuracy in estimatingbackrest elevation. Am. J. Crit. Care 11, 34–37.

Dodek, P., Keenan, S., Cook, D., Heyland, D., Jacka, M., Hand, L., Muscedere,J., Foster, D., Mehta, N., Hall, R., Brun-Buisson, C., 2004. Evidence-based clinical practice guideline for the prevention of ventilator-associated pneumonia. Ann. Int. Med. 141, 305–313.

Drakulovic, M., Torres, A., Bauer, T., Nicolas, J., Nogue, S., Ferrer, M., 1999.Supine body position as a risk factor for nosocomial pneumonia inmechanically ventilated patients: a randomised trial. Lancet 354,1851–1858.

Evans, D., 1994. The use of position during critical illness: current practiceand review of the literature. Aust. Crit. Care 7, 16–21.

Grap, M., Cantley, M., Munro, C., Corley, M., 1999. Use of backrest eleva-tion in critical care: a pilot study. Am. J. Crit. Care 8, 475–480.

Grap, M., Munro, C., Bryant, S., Ashanti, B., 2003. Predictors of backrestelevation in critical care. Intensive Crit. Care Nurs. 19, 68–74.

Grap, M., Munro, C., Hummell, R., Elswick, R., McKinney, J., Sessler, C.,2005. Effect of backrest elevation on the development of ventilator-associated pneumonia. Am. J. Crit. Care 14, 325–332.

Helman, D., Sherner, J., Fitzpatrick, T., Callendar, M., Shorr, A., 2003. Effect ofstandardized orders and provider education on head-of-bed position-ing in mechanically ventilated patients. Crit. Care Med. 31, 2285–2290.

Klingstedt, C., Hedenstierna, G., Lundquist, H., Strandberg, A., Tokics, L.,Brismar, B., 1990. The influence of body position and differentialventilation on lung dimensions and atelectasis formation in anaes-thetized man. Acta Anaesthesiol. Scand. 34, 315–322.

Kollef, M., 1993. Ventilator-associated pneumonia: a multivariate analy-sis. JAMA 270, 1965–1970.

Luna, C., Blanzaco, D., Niederman, M., Matarucco, W., Baredes, N., Desm-ery, P., Palizas, F., Menga, G., Rios, F., Apezteguia, C., 2003. Resolutionof ventilator-associated pneumonia: prospective evaluation of theclinical pulmonary infection score as an early clinical predictor ofoutcome. Crit. Care Med. 31, 676–682.

McMullin, J., Cook, D., Meade, M., Weaver, B., Letelier, L., Kahmamoui, K.,Higgins, D., Guyatt, G., 2002. Clinical estimation of trunk positionamong mechanically ventilated patients. Intensive Care Med. 28,304–309.

Moreno, R., Vincent, J., Matos, R., Mendonca, A., Cantraine, F., Thijs, L.,Takala, J., Sprung, C., Antonelli, M., Bruining, H., Willatts, S., 1999. Theuse of maximum SOFA score to quantify organ dysfunction/failure in

intensive care. Results of a prospective, multicentre study. IntensiveCare Med. 25, 686–696.

Muscedere, J., Dodek, P., Keenan, S., Fowler, R., Cook, D., Heyland, D., 2008.Comprehensive evidence-based clinical practice guidelines for ven-tilator-associated pneumonia: prevention. J. Crit. Care 23, 126–137.

Orozco-Levi, M., Torres, A., Ferrer, M., Piera, C., el-Ebiary, M., de laBellacasa, J., Rodriguez-Roisin, R., 1995. Semirecumbent positionprotects from pulmonary aspiration but not completely from gastro-esophageal reflux in mechanically ventilated patients. Am. J. Respir.Crit. Care 152, 1387–1390.

Peterlini, M., Rocha, P., Kusahara, D., Pedreira, M., 2006. Subjectiveassessment of backrest elevation: magnitude of error. Heart Lung35, 391–396.

Peterson, M., Scwab, W., McCutcheon, K., van Oostrom, J., 2008. Effects ofelevating the head of bed on interface pressure in volunteers. Crit.Care Med. 36, 3038–3042.

Pugin, J., Auckenthaler, R., Mili, N., Janssens, J., Lew, P., Suter, P., 1991.Diagnosis of ventilator-associated pneumonia by bacteriologic ana-lysis of bronchoscopic and non-bronchoscopic ‘‘blind’’ bronchoalveo-lar lavage. Am. Rev. Respir. Dis. 143, 1121–1129.

Richard, J.-C.M., Maggiore, S.M., Mancebo, J., Lemaire, F., Jonson, B.,Brochard, L., 2006. Effects of vertical positioning on gas exchangeand lung volumes in acute respiratory distress syndrome. IntensiveCare Med. 32, 1623–1626.

Rose, L., Baldwin, I., Crawford, T., Parke, R., in press. A multicenter,observational study of semirecumbent positioning in mechanicallyventilated patients. Am. J. Crit. Care.

Tablen, O., Anderson, L., Besser, R., Bridges, C., Hajjeh, R., 2004. Guidelinesfor prevention of health care-associated pneumonia, 2003: recom-mendations of the CDC and the Healthcare Infection Control PracticesAdvisory Committee. MMWR Recomm. Rep. 53, 1–36.

Torres, A., Serra-Battlles, J., Ros, E., Piera, C., Puig de la Bellacasa, J., Cobos,A., Lomena, F., Rodriguez-Roisin, R., 1992. Pulmonary aspiration ofgastric contents in patients receiving mechanical ventilation: theeffect of body position. Ann. Int. Med. 116, 540–543.

van Nieuwenhoven, C., Vandenbroucke-Grauls, C., van Tiel, F., Joore, H.,Strack van Schijndel, R., van der Tweel, I., Ramsay, G., Bonten, M.,2006. Feasibility and effects of the semirecumbent position to preventventilator-associated pneumonia: a randomized study. Crit. CareMed. 34, 396–402.

Vincent, J., Moreno, R., Takala, J., Willatts, S., De mendonca, A., Bruining, H.,Reinhart, C., Suter, P., Thijs, C., 1996. The SOFA (Sepsis-related OrganFailure Assessment) score to describe organ dysfunction/failure.Intensive Care Med. 22, 707–710.

Williams, Z., Chan, R., Kelly, E., 2008. A simple device to increase rates ofcompliance in maintaining 30-degree head-of-bed elevation in ven-tilated patients. Crit. Care Med. 36, 1155–1157.

Wip, C., Napolitano, L., 2009. Bundles to prevent ventilator-associatedpneumonia: how valuable are they? Curr. Opin. Infect. Dis. 22 .

Zack, J., Garrison, T., Trovillion, E., Clinkscale, D., Coopersmith, C., Fraser,V., Kollef, M., 2002. Effect of an education program aimed at reducingthe occurrence of ventilator-associated pneumonia. Crit. Care Med.30, 2407–2412.