infection control and hospital epidemiology

october 2008, vol. 29, no. 10

original article

Preventing Ventilator-Associated Pneumonia in the United States:

A Multicenter Mixed-Methods Study
Sarah L. Krein, PhD, RN; Christine P. Kowalski, MPH; Laura Damschroder, MS, MPH; Jane Forman, ScD, MHS;
Samuel R. Kaufman, MA; Sanjay Saint, MD, MPH

objective. To determine what practices are used by hospitals to prevent ventilator-associated pneumonia (VAP) and, through qualitative
methods, to understand more fully why hospitals use certain practices and not others.
design.

Mixed-methods, sequential explanatory study.

methods. We mailed a survey to the lead infection control professionals at 719 US hospitals (119 Department of Veterans Affairs [VA]
hospitals and 600 non-VA hospitals), to determine what practices are used to prevent VAP. We then selected 14 hospitals for an in-depth
qualitative investigation, to ascertain why certain infection control practices are used and others not, interviewing 86 staff members and
visiting 6 hospitals.
results. The survey response rate was 72%; 83% of hospitals reported using semirecumbent positioning, and only 21% reported using
subglottic secretion drainage. Multivariable analyses indicated collaborative initiatives were associated with the use of semirecumbent
positioning but provided little guidance regarding the use of subglottic secretion drainage. Qualitative analysis, however, revealed 3 themes:
(1) collaboratives strongly influence the use of semirecumbent positioning but have little effect on the use of subglottic secretion drainage;
(2) nurses play a major role in the use of semirecumbent positioning, but they are only minimally involved with the use of subglottic
secretion drainage; and (3) there is considerable debate about the evidence supporting subglottic secretion drainage, despite a meta-analysis
of 5 randomized trials of subglottic secretion drainage that generally supported this preventive practice, compared with only 2 published
randomized trials of semirecumbent positioning, one of which concluded that it was ineffective at preventing the development of VAP.
conclusion. Semirecumbent positioning is commonly used to prevent VAP, whereas subglottic secretion drainage is used far less often.
We need to understand better how evidence related to prevention practices is identified, interpreted, and used to ensure that research
findings are reliably translated into clinical practice.
Infect Control Hosp Epidemiol 2008; 29:933-940

Ventilator-associated pneumonia (VAP) is the most common

type of hospital-acquired infection seen in the intensive care
unit (ICU).1 VAP occurs in 10%20% of patients receiving
mechanical ventilation for more than 48 hours and is associated with substantial morbidity, mortality, and excess
healthcare costs.2 Various strategies to prevent VAP have been
evaluated, most focusing on preventing oropharyngeal or gastric colonization by pathogens and the aspiration of contaminated secretions. Published guidelines3-6 describe several specific preventive practices, including the use of semirecumbent
positioning of ventilated patients, aspiration of subglottic secretions by using specialized endotracheal tubes, selective digestive tract decontamination, use of antimicrobial mouth
rinse, and use of specialized kinetic beds. Recently, interest
has focused on practice bundles, sets of practices implemented together.7-9

Unfortunately, few data exist regarding the strategies that

US hospitals use to prevent VAP. Moreover, little is known
about what factors influence the use of these strategies. The
purpose of this mixed-methods, sequential explanatory study
was to determine, by use of a quantitative survey, what practices US hospitals use for the prevention of VAP and to understand, by use of qualitative analysis, why hospitals are
using some practices but not others.

methods
Study Design
This study is part of a larger project focusing on the prevention of hospital-acquired infections in US hospitals.10-12
We combined quantitative and qualitative data to allow for
a more detailed analysis of the complex phenomena that are

From the Department of Veterans Affairs Ann Arbor Health Services Research and Development Center of Excellence, Department of Veterans Affairs
Ann Arbor Healthcare System (S.L.K., C.P.K., L.D., J.F., S.S.), the Department of Internal Medicine, University of Michigan Medical School (S.L.K., S.R.K.,
S.S.), and the Ann Arbor Department of Veterans AffairsUniversity of Michigan Patient Safety Enhancement Program (S.R.K., S.S.), Ann Arbor, Michigan.
Received April 14, 2008; accepted June 22, 2008; electronically published August 20, 2008.
2008 by The Society for Healthcare Epidemiology of America. All rights reserved. 0899-823X/2008/2910-0005$15.00.DOI: 10.1086/591455

934

infection control and hospital epidemiology

october 2008, vol. 29, no. 10

inherent in the delivery of healthcare services.13,14 First, in

May 2005, we sent a survey to the lead infection control
professionals (ICPs) at 719 US hospitals. The survey sample
consisted of all Department of Veterans Affairs (VA) medical
centers (n p 119), which are part of the countrys largest
centralized healthcare delivery system, and a national, stratified, random sample of non-VA general medical and surgical
hospitals with 50 or more beds and an ICU (n p 600). Data
were also obtained from the 2005 American Hospital Association (AHA) database and the 2003 Area Resource File.15
The survey responses were then used to implement a stratified purposeful sampling strategy,13 resulting in the selection
of 14 hospitals for an in-depth qualitative investigation. We
first stratified by hospital size, then looked at our quantitative
results to determine the typical infection control practices
used by hospitals in each stratum, and selected hospitals with
typical uses as well as hospitals with atypical uses of specific
practices. The selection of hospitals also focused on achieving
the maximum variation in the other relevant variables, such
as whether the hospital was a VA or non-VA facility. Semistructured telephone interviews were conducted with key
stakeholders at all 14 hospitals, and 6 of the 14 hospitals were
subsequently selected for a visit.
Institutional review board approval was obtained from the
VA Ann Arbor Healthcare System as well as from each of the
hospitals that were visited.

Qualitative Data Collection

Survey Measures

Statistical Analysis

At each acute care facility, respondents were asked how frequently certain infection control practices were used for
adults who received mechanical ventilation. Frequency of
practice was measured on a scale from 1 to 5 (in which 1
indicated never and 5 indicated always), with regular
use defined as a rating of 4 or 5. Practices of interest included
the use of semirecumbent positioning, the use of antimicrobial mouth rinse, aspiration of subglottic secretions by using
specialized endotracheal tubes, and the use of specialized kinetic beds that oscillate. These practices were selected because
there is supporting evidence for each in the literature, and
they are all discussed in the Centers for Disease Control and
Prevention (CDC) guidelines for preventing healthcare-associated pneumonia.3
We also asked respondents whether the facility had a hospital epidemiologist, whether the lead ICP was certified in
infection control and epidemiology, and whether the facility
was participating in a collaborative effort to encourage use
of infection control practices. Academic affiliation was defined as having residency training approval by the Accreditation Council for Graduate Medical Education as specified
in the AHA database. The AHA database was also used to
derive a measure of nurse staffing (full-time equivalent nurses
per adjusted average daily census), which has been shown to
be related to patient outcomes and safety.16-18

Given our stratified sampling approach, survey data were

analyzed by using sample weights based on the probability
of selection in each stratum, which resulted in estimates that
represented the full population of VA hospitals and non-VA
acute care hospitals with 50 or more hospital beds and an
ICU. Results were reported either as weighted proportions or
as weighted mean values, with 95% confidence intervals (CIs).
We used weighted logistic regression to determine which hospital characteristics (as described above) were associated with
the use of each infection prevention practice, while simultaneously adjusting for other factors such as the facilitys
number of ICU beds. Logistic regression results were presented as odds ratios (ORs) with 95% CIs. All reported P
values are 2 tailed, and all analyses were conducted using
Stata, version 9.0 (StataCorp).
The qualitative analysis was conducted using rigorous qualitative procedures.13,19 First, we established a preliminary codebook that was based on our studys conceptual model. This
codebook was revised by having 4 members of the research
team independently code 2 transcripts and then meet to discuss the codes. Because of the complex and dense nature of
our data and our use of interpretive codes, we chose a consensus approach to coding the data.20,21 Two team members
coded each transcript independently and then discussed the
coded transcript and resolved any discrepancies. The interviews and codes were entered into a qualitative computer

A total of 38 semistructured telephone interviews were conducted at 14 hosptals. The interviews were audio recorded
and transcribed. Multiple interviews were conducted at each
hospital (24 interviews ranging from 29 to 92 minutes) to
get the perspectives of staff in different positions. The telephone interviews were conducted during the period from July
2005 to May 2006. The first point of contact was the lead
ICP. Thereafter, we used a snowball technique whereby interviewees were asked to recommend other potential interviewees, particularly individuals who were involved in decisions related to the use of at least 1 of the infection control
practices of interest. Interviewees were asked to identify which
practices their organization used to prevent VAP and to discuss the decision-making process that led their organization
to use those practices (see Appendix). Interviewees were also
asked about practices that were not used and why.
During the period from October 2006 to October 2007,
there were 48 additional interviews conducted during visits
to 6 hospitals. The hospital visits provided an opportunity to
gather more in-depth information and to confirm or explore
issues mentioned during the telephone interviews. Interviewees from a wide variety of staff positions were purposefully
selected because they provided different perspectives (eg, senior executives) or because they were directly involved in infection control (eg, hospital epidemiologists).

preventing ventilator-associated pneumonia

table 1. Characteristics of Hospitals Responding to a Survey Determining Practices Used to Prevent Ventilator-Associated
Pneumonia
Characteristic

Weighted value,
% or mean (CI)

Approved for residency training

Have hospitalists
Participate in a collaborative
Have a hospital epidemiologist
Have an ICP certified in infection control
RN staffinga
VA hospital
In a metropolitan area
County population 10,000
Regularly use semirecumbent positioningb
Regularly use antimicrobial mouth rinse
Regularly use subglottic secretion drainage
Regularly use oscillating or kinetic beds

26
56
41
39
58
1.2 (1.11.2)
4
72
69.2 (56.282.3)
83
40
21
18

note.

Percentages are derived using sample weights to reflect the total

population of hospitals represented by the respondent sample, which includes
119 Veterans Affairs (VA) hospitals and 600 non-VA general medical and
surgical hospitals with 50 or more beds and an intenstive care unit. CI,
confidence interval; ICP, infection control professional; RN, registered nurse.
a
Full-time equivalent RNs per adjusted average daily census.
b
Regular use was defined as receiving a rating of 4 or 5 on a scale from 1
to 5 (in which 1 indicated never and 5 indicated always).

analysis program (NVivo7; QSR International). Our qualitative analysis was guided by quantitative findings and sought
to explain and confirm those findings. We also identified
factors, in addition to those included in our quantitative analysis, that influenced the use of VAP prevention practices, and
we explicitly sought out and highlighted examples that were
contrary to expected findings, to ensure the rigor of our
analysis.

results
Quantitative Survey
The survey response rate was 72%. Table 1 shows the characteristics of the responding hospitals. In our sample, 83%

of the hospitals reported regularly using semirecumbent positioning, 40% regularly used antimicrobial mouth rinse, 21%
regularly used subglottic secretion drainage, and 18% regularly used kinetic beds.
Multivariable logistic regression analysis (Table 2) revealed
that hospitals participating in a collaborative were more than
twice as likely to use semirecumbent positioning (OR, 2.6
[95% CI, 1.444.80]). Higher nurse-staffing levels also appeared to be associated with the use of semirecumbent positioning (OR, 1.8 [95% CI, 0.913.58]), although this association was not statistically significant (P p .09). Finally,
if the ICP was certified in infection control, the facility was
significantly more likely to report regular use of subglottic
secretion drainage (OR, 2.4 [95% CI, 1.214.65]). None of
our measured characteristics were associated with use of kinetic beds or antimicrobial mouth rinse.
Qualitative Themes
The characteristics of the 14 hospitals where telephone interviews took place and of the 6 hospitals where in-person
interviews took place that were included in the qualitative
phase of the study are shown in Table 3, and the characteristics
of all interviewees are shown in Table 4. Of particular interest
with regard to our quantitative data was the contrast between
the use of semirecumbent positioning and the use of subglottic secretion drainage, because, according to the CDC
guidelines for preventing healthcare-associated pneumonia,
implementation of both practices is suggested and both are
supported by suggestive clinical or epidemiologic studies or
by strong theoretical rationale (ie, category II recommendations).3 Thus, we used qualitative data to identify underlying mechanisms that would help explain these results. The
following 3 themes emerged from the qualitative analysis: (1)
the use of collaboratives and practice bundles; (2) nursestaffing levels and the role played by nurses in the use of
infection controal practices; and (3) the perceived level of
evidence supporting a practice.
Collaboratives. Qualitative analysis confirmed the association between the participation in a collaborative and the

table 2. Results of Multivariable Logistic Regression Analysis to Determine Regular Use of Practices to Prevent
Ventilator-Associated Pneumonia
Variable

Semirecumbent
positioning

Antimicrobial
mouth rinse

Approved for residency training

Participate in a collaborative
Have a hospital epidemiologist
Have an ICP certified in infection control
RN staffinga
VA hospital

0.7
2.6
1.2
0.8
1.8
1.7

1.3
1.0
0.9
1.2
0.9
0.7

(0.351.48)
(1.444.80)
(0.612.21)
(0.431.50)
(0.913.58)
(0.674.55)

935

(0.762.22)
(0.641.57)
(0.561.47)
(0.711.97)
(0.651.23)
(0.351.23)

Subglottic
secretion drainage
0.7
1.2
1.4
2.4
0.9
1.2

(0.331.29)
(0.712.10)
(0.772.52)
(1.214.65)
(0.651.27)
(0.562.52)

Oscillating or
kinetic beds
1.1
1.3
1.0
0.7
1.0
1.7

(0.542.41)
(0.732.28)
(0.541.81)
(0.361.29)
(0.711.35)
(0.803.44)

note. Data are adjusted odds ratios (95% confidence intervals). Regular use was defined as receiving a rating of 4 or 5 on a scale
from 1 to 5 (in which 1 indicated never and 5 indicated always). Results are also adjusted for other variables, such as number of
beds in the intensive care unit, metropolitan location, and presence of hospitalists. ICP, infection control professional; RN, registered
nurse; VA, Veterans Affairs.
a
Full-time equivalent RNs per adjusted average daily census.

936

infection control and hospital epidemiology

october 2008, vol. 29, no. 10

table 3. Characteristics of the Hospitals Where Interviews Took

Place to Understand Regular Use of Practices to Prevent VentilatorAssociated Pneumonia
No. of hospitals,
by type of interview
Characteristics
Location
Northeast
Midwest
South
West
Type of hospital
VA
Non-VA
No. of beds
25250
251
Academic affiliation
Yes
No
Participated in a collaborative
Yes
No
note.

Telephone
(n p 14)

In person
(n p 6)

2
5
3
4

0
3
0
3

7
7

3
3

6
8

1
5

12
2

6
0

5
9

4
2

VA, Veterans Affairs.

use of semirecumbent positioning, but it revealed no association between collaborative participation and the use of
subglottic secretion drainage. In particular, participation in
a collaborative was identified as positively influencing the use
of semirecumbent positioning, by focusing attention on the

practice and by encouraging systematic monitoring, feedback,

and reinforcement. A physician administrator at one hospital
stated that for many years people have tried toadvocate
for semirecumbent positioning but its always been sort of a
flavor of the monththe [collaborative] initiative began to
focus on how to hardwire some of these not totally exotic
practices A nurse at another facility noted that it wasnt
until the [collaborative] that all the intensivists were fully
behind the whole projectonce the graphs started going up,
once people started seeing results, it was just part of their
normal daily practice.
Nurse staffing. Although staffing levels were not explicitly
discussed during our interviews, our qualitative analysis revealed, as suggested by our quantitative data, that nurses play
an important role in the use of semirecumbent positioning.
Likewise, there was essentially no mention in our interviews
of the active involvement of nurses in the use of subglottic
secretion drainage. When describing the use of semirecumbent positioning, the interviewees identified nurses as key
facilitators at several hospitals, although a few also mentioned
nurses as barriers to the practice (often due to nurses resistance or misperceptions).
While describing the role of nurses in successfully implementing the use of semirecumbent positioning, a chief of
staff noted that the nurses are really outstanding and
theyactually make it [semirecumbent positioning] happenit doesnt depend on the vagaries of which attending
is on call. Similarly, when asked who promoted the use of
semirecumbent positioning, a nurse at one facility stated that
it was just the nursing staff, the charge nurses, it was me
[a nurse] going around talking about it.

table 4. Characteristics of the 86 Hospital Staff Members Interviewed to Understand

Regular Use of Practices to Prevent Ventilator-Associated Pneumonia
No. of people,
by type of interview
Job title of interviewee
Infection control professional
Physician hospital epidemiologist or chief of infectious diseases
ICU nurse manager
ICU chief or other critical care physician
Chair or vice chair of medicine
Intravenous nurse clinician
Clinical nurse specialist or nurse practitioner
Quality manager or medical director
Chief of staff
Chief nurse or nurse executive
Chief of general internal medicine or other general internal
medicine physician
Hospital director, deputy director, or chief executive officer
Othera
note.

Telephone
(n p 38)

In person
(n p 48)

15
5
5
2
2
2
3
1
0
0

6
5
4
7
5
1
0
5
4
3

0
0
3

3
3
2

ICU, intenstive care unit.

Including risk manager, project coordinator, emergency department nurse, and respiratory therapy
supervisor.

preventing ventilator-associated pneumonia

Although nurses were identified as key facilitators, they

were also described by some as potential barriers to the regular use of semirecumbent positioning. The nurse manager
of a medical ICU explained how she encountered resistance
from some nurses: many of my nurses have been here
many yearsso when you have evidence-based practice and
you have to make changes, its just hard switching. In addition, nurses misperceptions were another common issue.
As described by an ICU nurse, probably the biggest challenge was that the nurses thought that they were already
putting the head of the bed up 30 but if you looked at the
angle of the bed it was only 20.
Nurses also facilitated the use of this practice (ie, semirecumbent positioning) by regular monitoring. One nurse
described how she and other staff nurses would ask why the
head of the bed was not up, thereby becoming, as she phrased
it, the head of the bed police. However, other professional
staff were also involved in monitoring this practice at some
hospitals. For example, a respiratory therapist described how
a pharmacist became involved with monitoring: he was
checking off all of these itemshe started out with just the
sedation and the weaning protocol and, he thought well, I
may as well add that [monitoring head of bed elevation].
Perceived level of evidence. None of our respondents described any debate about the evidence supporting semirecumbent positioning to prevent VAP. As the director of a
medical ICU observed, something like head of bed, youd
say, theres no downside. This general acceptance, with little
discussion of the evidence, is in strong contrast to how decisions to use or not use subglottic secretion drainage were
made at several hospitals. An ICP at one hospital that was
not using subglottic secretion drainage explained that there
certainly is interest by our intensiviststhey think it shows
promise[but] theyre not convinced theres definitive evidence that it prevents VAP. At another hospital, a critical
care physician explained the lack of use of subglottic secretion
drainage this way: Well because I think if you look at the
literature, its a little bit questionable and it really looks like
oral decontamination by frequent mouth washing is probably
sufficient.
At hospitals using subglottic secretion drainage, it was often
used only for certain patients or on a trial basis, and, even
at these hospitals, the perception of the evidence was mixed.
The hospital epidemiologist at one hospital that had used
subglottic secretion on a trial basis stated I think that the
recommendations in the CDC guidelinemight be a tad
more enthusiastic than the number of studies and the data
actually support. When asked about why they were using
subglottic secretion drainage, a physician explained I think
a series of articles over the years that suggested they significantly reduced VAP. It was the biologic plausibility that made
sense.
Other reasons for not using, or for selectively using, subglottic secretion drainage were cost and mechanical issues, such
as tube rigidity. However, as one physician noted, if you look

937

at them, theyre not that expensiveso I dont think honestly

that thats a very strong argument, although it is the argument
that everyone raises. Furthermore, in several cases, the concerns expressed were based on hearsay and not personal experience. For example, as described by a physician, Id heard
from everybody Id asked that these [subglottic secretion
drainage endotracheal tubes] were hard to keep patent, to
make actually work, that they took a lot of tinkeringthey
kept clogging and it was like a nightmare.

discussion
Several findings emerged from this multicenter, mixed-methods study. Specifically, the practice of semirecumbent positioning to prevent VAP is used by a large percentage of US
hospitals, whereas other practicesincluding the use of antimicrobial mouth rinse, kinetic beds, and subglottic secretion
drainageare used by far fewer hospitals. Particularly notable
is the contrast between the use of semirecumbent positioning
(used by more than 80% of hospitals) and the use of subglottic secretion drainage (used by approximately 20% of hospitals), given that recommendations in the CDC guidelines
indicate the same level of support for both practices.3
Quantitative and qualitative findings suggest that hospitals
participating in a collaborative greatly increased their routine
use of semirecumbent positioning but that this participation
had little effect on their use of subglottic secretion drainage.
Similarly, both sources of data (quantitative and qualitatiave)
support the key role played by nurses in instituting semirecumbent positioning of patients receiving mechanical ventilation and their minimal involvement in the use of subglottic
secretion drainage. Although our quantitative results provided few insights into the lack of use of subglottic secretion
drainage, our qualitative data were quite illuminating. Disagreement about the strength of the evidence supporting the
use of subglottic secretion drainage was a key factor limiting
its use, compared with semirecumbent positioning, which
generated no apparent dispute.
The study results highlight the important role that collaboratives play in addressing hospital-acquired infection but
also suggest that this effect may be specific to certain practices.
Collaborativeswhich bring together teams from different
organizations to work together in a structured fashionare
increasingly popular for facilitating the spread of evidencebased practices and improving quality of care.22-24 In the
United States, collaborative initiatives, such as the Institute
for Healthcare Improvement 100,000 Lives Campaign, have
focused on preventing VAP by use of a practice bundle, which
includes the use of semirecumbent positioning.7,8 As noted
in Krein et al.,11 collaborative membership is associated with
the use of several practices to prevent central venous catheter
related bloodstream infection. However, this association is
most evident when the practice is included in a practice bundle, which seems to be taken as evidence for its effectiveness.25
So, although practice bundles may be a powerful tool for

938

infection control and hospital epidemiology

october 2008, vol. 29, no. 10

change, this approach warrants careful investigation, to prevent rapid, but misguided, increases in the use of potentially
ineffective practices. Another possible risk is that a focus on
certain practices can divert attention from other novel technologies. In this analysis, however, we did not find explicit
evidence to suggest that participation in a collaborative was
related to decreased use of a newer practice that is not part
of a bundle, such as subglottic secretion drainage.
Studies have identified an association between higher
nurse-staffing levels and better patient outcomes, including
decreases in hospital-acquired infections.16-18 In our study, the
level of nurse staffing was positively associated with the use
of semirecumbent positioning. Although this relationship was
not statistically significant, the role nurses played in the use
of semirecumbent positioning was clearly evident in our qualitative data. Specifically, our results illustrate the importance
of having nurses directly engaged in promoting the use of
semirecumbent positioning and maintaining adherence. Our
results also suggest that nurses misperceptions about their
own adherence to this practice may be a barrier to its effective
use. Likewise, other research shows that self-reported adherence to semirecumbent positioning may be overestimated.
According to surveys of ICU nurses, reported adherence to
semirecumbent positioning (ie, head of bed raised 3045)
ranged from 76.4% to 86%.26,27 In contrast, a study in which
patient position was observed in 3 ICUs (using 502 measurements of 169 patients) found that 120 (71%) of the 169
patients were supine and that the mean backrest elevation
was 19 (16.8 for patients who received mechanical ventilation vs 23.1 for patients who did not).28 Therefore, monitoring may be an essential element to ensure the effective
use of semirecumbent positioning, and nurses as well as other
staff, including respiratory therapists, pharmacists, and ICPs,
can perform this important function.
Unlike the use of semirecumbent positioning, nurses had
minimal influence on the use of subglottic secretion drainage,
which is to be expected, given that nurses rarely perform
endotracheal intubation. However, a practice alert on VAP
issued by the American Association of Critical-Care Nurses
in February 2004 identified the use of subglottic secretion
drainage as one of 3 recommended practices, along with use
of semirecumbent positioning and not routinely changing the
ventilator circuit.29 Thus, nurses could conceivably be proponents of subglottic secretion drainage.
The extent to which evidence truly influences decisions
about the use of infection prevention practices is not straightforward, as others have also found,30,31 and merits further
investigation. Our qualitative results confirm that subglottic
secretion drainage is a controversial practice for preventing
VAP, and varying opinions and debates about the evidence
supporting this technology are common. In comparison, we
found no debate surrounding the use of semirecumbent positioning, even though the published CDC guidelines suggest
that the level of evidence is similar for both practices.3 Moreover, a meta-analysis of 5 randomized trials of subglottic

secretion drainage generally supported this preventive practice,32 whereas there are only 2 published randomized trials
of semirecumbent positioning,33,34 one of which concluded
that it was ineffective at preventing the development of VAP.33
Even though we employed national sampling and
achieved an excellent response rate, several important limitations of our study must be acknowledged. First, we relied
on self-reporting from the lead ICP at each hospital to determine what practices were being used to prevent VAP, and
we have no mechanism to validate these reports. Although
the lead ICP may have overstated or understated the use of
the various practices, we have no evidence of systematic
bias. Second, although our sampling strategy was designed
to be nationally representative, it is possible that participating hospitals were different from nonparticipating hospitals, thereby decreasing the external validity of our survey
results. Third, the qualitative portion of our study was conducted to understand why hospitals are doing what they are
doing; the results are not intended to be representative of
the more than 6,000 acute care hospitals in the United
States. However, although our findings are not generalizable
by statistical inference from the study sample to the population, they can be generalized to similar settings and can
point to areas for further exploration.35
Despite these limitations, our multicenter, mixed-methods
study provides valuable information for policy makers, clinicians, and patient safety personnel interested in reducing
VAP and other hospital-acquired infections. First, collaboratives and the promotion of practice bundles can be powerful
tools for increasing awareness and engaging hospitals in the
use of specific infection prevention practices. However, the
long-term sustainability and potential unintended consequences of these types of initiatives are not yet fully understood. Second, although nurses may not be involved with the
implementation of all prevention strategies, their involvement
could play an important role in the use of many infection
prevention practices. Finally, we need to understand better
how evidence related to prevention practices is identified,
interpreted, and used to ensure that research findings are
reliably translated into clinical practice.

acknowledgments
We thank Nasia Safdar, MD, MS, and Pat Rouen, MSN, RN, for their feedback
and comments during manuscript preparation, and all those who participated
as survey respondents or in interviews.
Financial support. This project was supported by the Department of Veterans Affairs Health Services Research and Development Service (grant SAF
04-031) and the Ann Arbor VA Medical Center/University of Michigan Patient Safety Enhancement Program. S.S. is supported by an Advanced Career
Development Award from the Health Services Research and Development
Service of the Department of Veterans Affairs.
Potential conflicts of interest. All authors report no conflicts of interest
relevant to this article.

preventing ventilator-associated pneumonia

appendix
The main questions asked during the telephone interviews
are shown below. The interviews were semistructured.
How did this hospital identify the need for the practice?
What kind of information did this hospital have on the
practice?
Who were the important people who helped this hospital
decide to use this practice?
What problems did this hospital run into getting this practice adopted?
Is this hospital experiencing any implementation problems?
What are the major barriers that prevented your hospital
from implementing practices for reducing healthcare-associated infections? What are the facilitators?
If I could fix one barrier here that makes it more difficult
for your organization to implement key practices for preventing healthcare-associated infections, what would you
have me fix?
If someone wants to adopt a new infection control practice,
what committees do they go through?
How does being part of a large healthcare system affect
the adoption and implementation of infection control practices? What are the benefits? The barriers?
Who are the main people who need to be on board for
change to occur?
Have collaboratives had a role in the adoption or implementation of infection control practices at this hospital?
How would you describe the prevailing management style
in this hospital?
What is your perception of how senior management is
involved in the adoption and implementation of infection
control practices?
Address reprint requests to Sarah L. Krein, PhD, RN, Veterans Affairs Ann
Arbor Health Services Research and Development Center of Excellence, Veterans Affairs Ann Arbor Healthcare System, PO Box 130170, Ann Arbor, MI
48113 (skrein@umich.edu).
The views expressed in this article are those of the authors and do not
necessarily reflect the position or policy of the Department of Veterans Affairs.
Presented in part: 18th Annual Meeting of the Society for Healthcare
Epidemiology of America; Orlando, Florida; April 58, 2008 (Abstract 378).

references
1. Richards MJ, Edwards JR, Culver DH, Gaynes RP. Nosocomial infections
in combined medical-surgical intensive care units in the United States.
Infect Control Hosp Epidemiol 2000; 21:510515.
2. Safdar N, Dezfulian C, Collard HR, Saint S. Clinical and economic consequences of ventilator-associated pneumonia: a systematic review. Crit
Care Med 2005; 33:21842193.
3. Tablan OC, Anderson LJ, Besser R, Bridges C, Hajjeh R. Guidelines for
preventing health-careassociated pneumonia, 2003: recommendations
of CDC and the Healthcare Infection Control Practices Advisory Committee. MMWR Recomm Rep 2004; 53(RR-3):136.
4. Dodek P, Keenan S, Cook D, et al. Evidence-based clinical practice guideline for the prevention of ventilator-associated pneumonia. Ann Intern
Med 2004; 141:305313.

939

5. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit
Care Med 2005; 171:388416.
6. Collard HR, Saint S, Matthay MA. Prevention of ventilator-associated
pneumonia: an evidence-based systematic review. Ann Intern Med 2003;
138:494501.
7. Berwick DM, Calkins DR, McCannon CJ, Hackbarth AD. The 100,000
Lives Campaign: setting a goal and a deadline for improving health care
quality. JAMA 2006; 295:324327.
8. Resar R, Pronovost P, Haraden C, Simmonds T, Rainey T, Nolan T. Using
a bundle approach to improve ventilator care processes and reduce ventilator-associated pneumonia. Jt Comm J Qual Patient Saf 2005; 31:243
248.
9. Youngquist P, Carroll M, Farber M, et al. Implementing a ventilator
bundle in a community hospital. Jt Comm J Qual Patient Saf 2007; 33:
219225.
10. Krein SL, Olmsted RN, Hofer TP, et al. Translating infection prevention
evidence into practice using quantitative and qualitative research. Am J
Infect Control 2006; 34:507512.
11. Krein S, Hofer T, Kowalski C, et al. Use of central venous catheter-related
bloodstream infection prevention practices by US hospitals. Mayo Clin
Proc 2007; 82:672678.
12. Saint S, Kowalski CP, Kaufman SR, et al. Preventing hospital-acquired
urinary tract infection in the United States: a national study. Clin Infect
Dis 2008; 46:243250.
13. Sandelowski M. Combining qualitative and quantitative sampling, data
collection, and analysis techniques in mixed-method studies. Res Nurs
Health 2000; 23:246255.
14. Ivankova NV, Creswell JW, Stick SL. Using mixed-methods sequential
explanatory design: from theory to practice. Field Methods 2006; 18:3
20.
15. National Center for Health Workforce Analysis Bureau of Health Professions, Health Resources and Services Administration, Department of
Health and Human Services. Area Resource File (ARF). Fairfax, VA:
Quality Resource Systems (QRS), Inc.; 2003.
16. Needleman J, Buerhaus P, Mattke S, Stewart M, Zelevinsky K. Nursestaffing levels and quality of care in hospitals. N Engl J Med 2002; 346:
17151722.
17. Putnam W, Twohig PL, Burge FI, Jackson LA, Cox JL. A qualitative study
of evidence in primary care: what the practitioners are saying. CMAJ
2002; 166:15251530.
18. Kane RL, Shamliyan T, Mueller C, Duval S, Wilt TJ. Nurse staffing and
quality of patient care. Evid Rep Technol Assess (Full Rep) 2007; (151):
1115.
19. Mason J. Qualitative researching. Thousand Oaks, CA: Sage Publications;
2002.
20. Sandelowski M, Barroso J. Writing the proposal for a qualitative research
methodology project. Qual Health Res 2003; 13:781820.
21. Forman J, Damschroder LJ. Qualitative content analysis. In: Jacoby L,
Siminoff L, eds. Empirical Research for Bioethics: A Primer. Oxford, UK:
Elsevier Publishing; 2008:3962.
22. Ovretveit J, Bate P, Cleary P, et al. Quality collaboratives: lessons from
research. Qual Saf Health Care 2002; 11:345351.
23. Wilson T, Berwick DM, Cleary PD. What do collaborative improvement
projects do? Experience from seven countries. Jt Comm J Qual Saf 2003;
29:8593.
24. Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease
catheter-related bloodstream infections in the ICU. N Engl J Med 2006;
355:27252732.
25. Wachter RM, Pronovost PJ. The 100,000 Lives Campaign: a scientific
and policy review. Jt Comm J Qual Patient Saf 2006; 32:621627.
26. Ricart M, Lorente C, Diaz E, Kollef MH, Rello J. Nursing adherence
with evidence-based guidelines for preventing ventilator-associated
pneumonia. Crit Care Med 2003; 31:26932696.
27. Cason CL, Tyner T, Saunders S, Broome L. Nurses implementation of

940

28.
29.

30.

31.

infection control and hospital epidemiology

october 2008, vol. 29, no. 10

guidelines for ventilator-associated pneumonia from the Centers for Disease Control and Prevention. Am J Crit Care 2007;16:2836.
Grap MJ, Munro CL, Bryant S, Ashtiani B. Predictors of backrest elevation in critical care. Intensive Crit Care Nurs 2003; 19:6874.
American Association of Critical-Care Nurses Practice Alert. Ventilator
Associated Pneumonia. Available at: http://www.aacn.org/AACN/
practiceAlert.nsf/vwdoc/pa2. Accessed January 3, 2008.
Rello J, Lorente C, Bodi M, Diaz E, Ricart M, Kollef MH. Why do
physicians not follow evidence-based guidelines for preventing ventilator-associated pneumonia? A survey based on the opinions of an international panel of intensivists. Chest 2002; 122:656661.
Kaynar AM, Mathew JJ, Hudlin MM, et al. Attitudes of respiratory
therapists and nurses about measures to prevent ventilator-associated
pneumonia: a multicenter, cross-sectional survey study. Respir Care 2007;
52:16871694.

32. Dezfulian C, Shojania K, Collard HR, Kim HM, Matthay MA, Saint S.
Subglottic secretion drainage for preventing ventilator-associated pneumonia: a meta-analysis. Am J Med 2005; 118:1118.
33. 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;
34:396402.
34. Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogue S, Ferrer N.
Supine body position as a risk factor for nosocomial pneumonia in
mechanically ventilated patients: a randomised trial. Lancet 1999; 354:
18511858.
35. Forman J, Creswell JW, Damschroder L, Kowalski CP, Krein, SL. Qualitative research methods: key features and insights gained by their use
in infection prevention research. Am J Infect Control 2008. In press.