Care bundles: the holy grail of infectious risk management
in hospital?
Charis Marwick and Peter Davey
Quality and Safety Improvement Research Group,
Division of Community and Population Sciences and
Education, College of Medicine, Dentistry and Nursing,
University of Dundee, Dundee, Scotland
Correspondence to Peter Davey, Division of
Community and Population Sciences and Education,
Mackenzie Building, Kirsty Semple Way, Dundee DD2
4BF, Scotland
Tel: +44 1382 420000;
e-mail: p.g.davey@cpse.dundee.ac.uk
Current Opinion in Infectious Diseases 2009,
22:364369
Introduction
The Institute for Healthcare Improvement (IHI) devel-
oped the concept of bundles to help healthcare provi-
ders to improve the reliability of delivery of essential
healthcare processes [1]. The impact of a bundle depends
both on the evidence that supports the recommended
care processes and on the implementation and spread of
its recommendations. Not all clinical situations and con-
ditions require a bundle, and a bundle will not improve
quality of care if its development and application are
misguided. In this review, we discuss the design, imple-
mentation and testing of care bundles designed to
improve the management of sepsis and the prevention
of healthcare-associated infections.
What is a bundle?
The concept of care bundles has emerged from the
acknowledgement that healthcare delivery is too depen-
dent on individual clinicians knowledge, motivation and
skills, with the result that only about 50% of patients
receive the recommended care [2]. One explanation is
that processes in healthcare are rarely designed to meet
0951-7375 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Purpose of review
A care bundle is set of four or five processes that each individually improve patient
outcome and that should be performed together for every patient every time. We
describe how bundles should be designed, implemented and evaluated with
measurement designed for quality improvement rather than research or judgement.
Recent findings
A systematic review concluded that the relative risk reduction associated with the
introduction of a sepsis bundle exceeded 25%, and absolute risk reduction exceeded
9% in all studies. The number needed to treat to save one life in each study population
ranged from three to 11. Bundles for the prevention of infections have focused on
ventilator-associated pneumonia and catheter-associated bloodstream infections in the
ICU. The most persuasive evidence of effectiveness comes from multicentre studies,
but results from single ICUs provide valuable insights into how bundle implementation
fits within a broader quality improvement strategy.
Summary
Care bundles can be a powerful driver for improving the reliability of delivery of evidence-
based care and patient outcomes. It remains to be seen whether the success that has
been achieved in acute admissions and ICUs can be reproduced in general wards.
Keywords
bloodstream infections, care bundle, healthcare-associated infection, quality
improvement, sepsis, ventilator-associated pneumonia
Curr Opin Infect Dis 22:364369
2009 Wolters Kluwer Health | Lippincott Williams & Wilkins
0951-7375
specific, articulated reliability goals [2,3]. A care bundle is
a small set of practices that have been individually proven
to improve patient outcomes and when implemented
together are expected to result in a better outcome than
when implemented individually [1,4]. Bundles should
consist of three to six elements, which are well estab-
lished practices that are not being delivered consistently.
To improve consistency, these practices are then pack-
aged together into a bundle that should be delivered by
one healthcare team at one point in time to every patient
meeting the bundle criteria [1]. The point in time varies
according to the bundle aim, and some bundles may be
repeated for the same patient. For example, a central line
insertion bundle requires all the elements to be com-
pleted once, whereas a central line care bundle should be
repeated daily [5].
Developing a bundle is only the first step in making a
process more reliable and must be supported by under-
standing of habits and processes through measurement of
reliability and tests of change [1]. This is clearly illus-
trated by a study [6] of a bundle designed to reduce
ventilator-associated pneumonia (VAP) in a trauma ICU.
Introduction of the bundle in June 2002 had no impact on
DOI:10.1097/QCO.0b013e32832e0736

VAP, which only started to decline 3 months later after
the introduction of a process measurement program that
measured daily compliance with the bundle [6]. To avoid
confusion, we recommend that the term bundle is not
used to describe implementation strategies that include
more than one component, these should be called multi-
faceted interventions [7] rather than bundled beha-
vioural interventions [8].
The following steps have been recommended for the
development of a bundle, to combine the best of medical
science and improvement science [1,4]:
(1) Identify a set of four to six evidence-based interven-
tions that apply to a cohort of patients with a common
disease or a common location.
(2) Develop the will in the providers to deliver the
interventions every time they are indicated.
(3) Measure compliance as all or nothing.
(4) Redesign the delivery system to make it easy to
deliver the bundle, ideally to make reliable delivery
an integral part of the system.
(5) Measure related outcomes to ascertain the effects of
the changes in the delivery system.
What are the requirements for the inclusion of
a bundle element?
Ideally, the science behind the elements of a bundle
should be so well established that its implementation
should be considered a generally accepted practice, with
all the changes recommended based on randomized
controlled trials, and therefore supported by Level 1
evidence [1,4]. However, this level of evidence is not
always available for the link between processes of care
and patient outcome. Accordingly, the National Quality
Measures Clearinghouse (NQMC) does not stipulate that
trial evidence is required for a quality measure to be valid.
Evidence may consist of a clinical practice guideline or
other peer-reviewed synthesis of the clinical evidence, as
well as one or more research studies published in a
National Library of Medicine (NLM) indexed, peer-
reviewed journal [9].
Even when trial evidence for the link between a process of
care and outcome is available, knowing that something
works in another institution does not mean it will work in
ours. It is important to remember that the key question in
quality improvement is whether an intervention improves
our quality of care now [10]. The potential impact of an
intervention can be considered in terms of five dimensions:
(1) Reach: What proportion of the target population
participated?
(2) Efficacy: What is the success rate if implemented as
in protocol?
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Care bundles Marwick and Davey 365
(3) Adoption: What proportion of target hospitals/prac-
tices/clinicians adopts it?
(4) Implementation: To what extent is the intervention
implemented as intended/per protocol?
(5) Maintenance: To what extent is the program sus-
tained over time?
The impact of an intervention (impact reach  efficacy)
is filtered through the organizational dimensions of adop-
tion, implementation and maintenance [10,11]. The
influence of each of these dimensions on the effective-
ness of a bundle will vary greatly between hospitals, so
local adaptation of bundle elements and/or implementa-
tion strategies will be required.
With regards to the level of evidence behind the chosen
bundle elements, it is very important that this is made
clear, particularly when spread to other institutions is
planned. Good examples of this are the Surviving
Sepsis Campaign (SSC) severe sepsis management
guidelines [4,12] and the Scottish National Audit Pro-
ject on community-acquired pneumonia (SNAP-CAP)
[13]. An unusual, but helpful, feature of the SNAP-CAP
evidence review is that it includes discussion about
why two care processes were not included in the
bundle [13].
Bundles should be expected to evolve over time
because of new evidence and focusing on the aim of
improvement (Table 1). These five care bundles for
ventilated patients only have one common element:
elevation of the head of the bed [6,14,15,16,17]. Three
of the reasons are clear. First, the original IHI venti-
lator bundle was designed to improve the daily care of
patients who were being ventilated, not just to prevent
VAP [5,17], whereas more recent VAP bundles have
omitted deep venous thrombosis (DVT) prophylaxis in
order to focus on prevention of VAP [6,14,15,16].
Second, the more recent VAP bundles [14,15] have
omitted stress ulcer prophylaxis but included drainage
of oropharyngeal secretions as new evidence has been
collected and published [18,19]. Third, two of the
bundles include hand hygiene as one of the care
processes for prevention of VAP [6,15], whereas two
of the other VAP bundles were part of broader
improvement work in the ICU, in which hand hygiene
was an expected component of care for all patients
and was measured separately [14,16]. However, it
would be helpful if documentation of care bundles
always included some comparison with published
bundles with explicit discussion of why some care
processes have not been included. For example, two
of the bundles apparently do not include sedation
vacation or weaning [6,15]. Is this true and if so
why was sedation and weaning omitted from these
VAP bundles?
 366 Nosocomial and healthcare related infections

Table 1 Data from five studies that describe the content of care bundles for ventilated patients developed between 2005 and 2008
Jain
Resar et al. [16] Cocanour Bigham Anonymous
et al. [17] Adult medical et al. [6] et al. [15] (SICSAG) [14]
Type of ICU All adult and surgical Trauma Paediatric All adult

Number of ICUs 61 1 1 1 19
Process of care
Sedation and weaning Yes Yes No No Yes
Head of bed elevation Yes Yes Yes Yes Yes
DVT prophylaxis Yes Yes No No No
Stress ulcer prophylaxis Yes Yes Yes No No
Mouthcare No Yes Yes Yes Yes
Getting the patient out of bed No No Yes No No
Glove and apron use No No Yes Yes No
Endotracheal tube suctioning No No Yes No No
Changing nasopharyngeal irrigation fluids daily No No Yes No No
Use of sleeved suction catheters No No Yes No No
Clorhexidine baths twice weekly No No Yes No No
Strict glucose control No No Yes No No
Drainage of orophyaryngeal secretions No No No Yes Yes
Ventilator circuit and in-line suction catheter change No No No Yes No
Hand hygiene No No Yes Yes No
Storage of oral suction device No No No Yes No
DVT, deep venous thrombosis.

of each bundle element and of the whole bundle must be

Implementation through measurement and
tests of change
Successful implementation, as with all quality improve-
ment, requires engagement with all the staff involved in
its delivery. Measures for improvement must be designed
to provide just enough data to assess the reliability of
practice and must be sustainable by frontline staff with-
out additional resource [20]. Although this study is over
10 years old, it remains important because of the clear
distinction between the information needed for quality
improvement in comparison with research or account-
ability. This distinction is vital to engaging with frontline
clinical staff as drivers for change (Table 2). Completion
measurable as yes or no or completed or not completed
[1,4,21]. Compliance with the bundle means completion
of every component, with no score given for partial
completion. The use of rapid, repeated tests of change
has been described in detail for the implementation of
the IHI ventilator bundle in one ICU (Table 3 [22]).
Standardization of the care processes into a bundle should
be expected to reach a very high level of reliability. A
realistic first step is to aim for 80 or 90% success rate, then
identify and mitigate defects from the first step in order to
achieve more than 95% reliability [2]. Experience with
care bundles shows that these levels of reliability can be

Table 2 The differences between data for improvement, judgement and research in terms of aim, study design, analysis and
dissemination
Improvement Judgement Clinical research

Aim Provide front-line staff and service Compare different organizations in Produce new knowledge about
users with information about the order to select or reward service the relationships between
reliability of the care that they provide providers or investigate processes and outcome
poor performers
Test Observable, sequential No test Blinded, one big test
Bias Accept, for example, start tests of change in Adjust for bias in the analysis Design data collection and
areas where they are most likely to analysis to eliminate or
be successful and then spread to minimize bias
more challenging areas
Sample size Just enough to assess reliability, 100% of relevant data Collect as much data as
usually 10 observations if the target possible, just in case
is more than 95% reliability
Analysis: is change Run chart with predefined target and No change, data are usually Statistical test to determine how
an improvement? annotations to explain drivers for change displayed as a normative likely it is that change
comparison (e.g. league table) happened by chance
Confidentiality and Data only for stakeholders; consent is Publicly accessible information Subjects protected through
dissemination not required from individual patients or written consent based on
of information staff for data collection but the data should information agreed with an
only be made public if the subjects agree independent Research Ethics
Committee
Adapted from Solberg et al. [20].

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Care bundles Marwick and Davey 367

Table 3 Implementation and documentation of the Institute for Healthcare Improvement ventilator bundle in all patients in one ICU
Cycle Date Test of change Action

1 5 February Test sedation vacation with one patient Change time of day of sedation vacation
2 6 February Test sedation vacation with three patients Document exclusions from sedation vacations
3 8 February Test sedation vacation with all patients Include documentation of exclusions as a
standard element in multidisciplinary rounds
4 10 February Test head of bed elevation with one patient No modification required
5 14 February Test head of bed elevation with three patient No modification required
6 17 February Test head of bed elevation with all patients No modification required
7 22 February Test complete ventilator bundle with one patient New documentation required for completion
of whole bundle
8 24 February Test complete ventilator bundle with three patients No modification required
9 28 February Test complete ventilator bundle with all patients Bundle only completed for 50% of patients,
continue but aim for more than 95% reliability
10 14 March14 June Document bundle completion on all patients every day More than 95% reliability sustained over 4 months
Tests of change carried out over 23 days in February 2005 and the reliability of bundle implementation was then documented with a run chart updated
every week. On the basis of information from Table III and Fig. 2 of Pulcini et al. [22].

achieved and can yield measurable improvements in hospital-acquired infection was just one of the changes
patient outcome. For example, improvement of that were implemented (Fig. 1). For interventions that
reliability of a central line care bundle from 80 to more aim to reduce healthcare-associated infection, it is
than 95% in one ICU was followed by near abolition of important to document all of the key infection control
line-related infections [10]. Similarly, data from 35 processes, including those that are not included in the
ICUs that implemented the IHI ventilator bundle intervention [24].
showed that the most consistent reduction in VAP
occurred in six ICUs that achieved more than 80%
improvement in adherence with the bundle [17]. Distinguishing between quality improvement,
research and judgement
Data about the impact of quality improvement inter- We are increasingly realizing not only how critical
ventions should be presented as run charts that clearly measurement is to the quality improvement we seek
document change over time [23]. Annotation of run but also how counterproductive it can be to mix measure-
charts can be used to document the effect of multiple ment for accountability or research with measurement for
drivers for change, as most interventions have multiple improvement [20]. This critical study was published
components [16] (Fig. 1). This run chart clearly shows over 10 years ago, but the differences between data for
the contribution of an organizational change (multi- improvement and research are still not widely under-
disciplinary rounds and the appointment of an ICU stood. There are two critical issues for debate: ethics and
Medical Director to lead quality improvement) to the academic status.
reduction of adverse events in one adult ICU over
2 years. The implementation of bundles to reduce The ethical issues were graphically illustrated by con-
troversy surrounding an improvement project based on
a bundle for preventing bloodstream infections associ-
Figure 1 Run chart of the number of adverse events per ICU day
over 2 years and the impact of a multifaceted strategy for quality ated with intravenous catheters [25]. An anonymous
improvement complaint to the US Office for Human Research Pro-
tections (OHRP) after publication of the results led to
the cessation of improvement work for several months,
Number 30
of events
until public outcry caused the OHRP to reverse their
25 Multidisciplinary rounds decision [26]. Ethical supervision for quality improve-
20 Hand hygiene protocol ment projects should be integrated into the system of
15
Vent bundles accountability for the conduct of clinical care and
ICU medical director
should be clearly distinguished from approval for
10 UTI bundles
Central line bundles
research [26]. Intense scrutiny and the requirement
5 for specific informed consent from participants apply
0 when an activity involves additional risk. If the
2

02

03

03

improvement project aims to increase the reliability

00

00

00

00

00
20

20
/2

/2

/2

/2

/2

/2
0/

5/
/1

29

30

29

28

26
/3

/2
10

of delivery of evidence-based care, then risk to patients

1/

3/

5/

7/

9/
11

11

and professionals will be reduced by participation in

UTI, urinary tract infection. Reproduced with permission from Jain et al. the improvement work compared with standard care
[16].
[26].

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 368 Nosocomial and healthcare related infections
The current low academic status of quality improvement
work is likely to result in clinicians who are engaged in
poorly supported activities in academic medical centres
leaving to become directors of quality improvement in
nonteaching centres, which are investing heavily in qual-
ity improvement [27]. A fresh approach is required to
measuring and documenting academic medical activity in
quality improvement, particularly with respect to recog-
nizing innovation in local quality improvement work
within a single organization [27].
The distinction between quality improvement and jud-
gement is also important. Care bundles define important
processes for delivery to all eligible patients, but they are
only one component in a multifaceted strategy for con-
tinuous quality improvement [16]. When the same care
processes are included in centrally driven targets with
financial incentives, there is serious risk of unintended
consequences due to narrow focus on achieving the target
at the expense of other aspects of care [28]. One clear,
relevant study is administration of antibiotics to all
patients with CAP within 4 h of arrival in hospital. This
was adopted as a national target in the USA and resulted
in excessive administration of antibiotics to patients who
did not have CAP [29].
Bundles for the management of severe sepsis
and septic shock
The survival benefit of early goal-directed therapy
(EGDT) for the management of patients with severe
sepsis and septic shock was first established in a random-
ized trial [30]. The first SSC guidelines for the manage-
ment of severe sepsis and septic shock were published in
2004 [31] and have recently been updated [12]. These
extensive guidelines have been condensed into two
bundles: the Sepsis Resuscitation Bundle for com-
pletion within 6 hours and the Sepsis Management
Bundle, which should also begin immediately but must
be done within 24 h [4].
A recent review included 12 peer-reviewed publications
and 28 abstracts from 10 different countries, which eval-
uated EGDT as part of a sepsis quality initiative [32].
Combining the data identified 3042 patients before and
2956 after implementation of sepsis bundles. In the peer-
reviewed publications, the relative risk (RR) reduction
associated with the introduction of a sepsis bundle
exceeded 25%, and absolute risk reduction exceeded
9%, in all studies. The number needed to treat to save
one life in each study population ranged from three to 11
[32].
The success that has been achieved with improving the
management of severe sepsis needs to be reproduced for
management of less severely ill patients, both to improve
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
the timeliness of effective antibiotic treatment [33] and
to promote early review of antibiotic therapy in order to
minimize unnecessary use [34].
Bundles for prevention of hospital-acquired
infection
Most of the current evidence focuses on prevention of
catheter-related bloodstream infections [25] and VAP
[6,15,16,17] in ICUs. The evidence is most persuasive
when bundle implementation is associated with improve-
ment in outcome across multiple hospitals [17,25]. How-
ever, experience from individual hospitals can provide
additional important information about how change is
achieved, provided that data are presented appropriately.
There is little value to reports that only provide aggregate
information about processes and outcomes before and
after the intervention. For example, evaluation of a
ventilator care bundle as an uncontrolled before and after
study [35] actually provides no evidence that change was
due to the intervention [36]. In contrast, presentation of
data as run charts clearly shows how change occurred over
time [6,15,16].
We have selected one of these single hospital studies
(Fig. 1) to show how difficult it is to quantify the direct
impact of a care bundle within an ICU that is committed
to change and continuous improvement. The outcome
measured was the number of adverse events per ICU day,
which was measured by chart review and included
indicators of infection alongside indicators of other com-
plications. This ICU implemented three infection con-
trol bundles, that is, to prevent VAP, central line infec-
tions and catheter-associated urinary tract infection
(UTI). However, these were implemented after the
establishment of multidisciplinary rounds and a hand
hygiene protocol. The authors attributed the overall
reduction in adverse events to the change in culture
achieved through multidisciplinary rounds in 2002 and
to the appointment of an ICU Medical Director to lead
change for improvement in 2003 rather than to any one of
the three bundles [16].
The successes seen with care bundles in reducing
infection in the ICU have led to the development of
a broader set of bundles aimed at reducing surgical site
infection and infections in general wards [37]. It
remains to be seen whether the same success will be
achieved in these contexts in which teams are less well
defined and the ratio of staff to patients is much lower
than in ICU.
Conclusion
The answer to the question has to be no, care bundles are
not the holy grail of infectious risk management. Bundles

can be a powerful stimulus to focusing the multidisci-
plinary team on working together to deliver reliable care,
but they are only one component in an overall improve-
ment strategy. Bundles will not improve outcomes unless
their components are evidence based and they are deliv-
ered reliably to every patient every time.
References and recommended reading
Papers of particular interest, published within the annual period of review, have
been highlighted as:
 of special interest
 of outstanding interest
Additional references related to this topic can also be found in the Current
World Literature section in this issue (p. 413).
1 Haraden C. What is a bundle? 2006. http://www.ihi.org/IHI/Topics/Critical-
Care/IntensiveCare/ImprovementStories/WhatIsaBundle.htm. [Accessed 29
January 2009].
2 Resar RK. Making noncatastrophic healthcare processes reliable: learning to
walk before running in creating high-reliability organizations. Health Serv Res
2006; 41(4 (Part II)):16771689.
3 Nolan T, Berwick DM. All-or-none measurement raises the bar on perfor-
mance. JAMA 2006; 295:11681170.
4 SSC. Surviving sepsis campaign. Severe sepsis bundles; 2007. http://
www.survivingsepsis.org/node/89. [Accessed 20 February 2009].
5 Pronovost PJ, Berenholtz SM, Ngo K, et al. Developing and pilot testing quality
indicators in the intensive care unit. J Crit Care 2003; 18:145155.
6 Cocanour CS, Peninger M, Domonoske BD, et al. Decreasing ventilator-
associated pneumonia in a trauma ICU. J Trauma 2006; 61:122129.
7 Grimshaw JM, Thomas RE, MacLennan G, et al. Effectiveness and efficiency
of guideline dissemination and implementation strategies. Health Technol
Assess 2004; 8:iii-72.
8 Aboelela SW, Stone PW, Larson EL. Effectiveness of bundled behavioural
interventions to control healthcare-associated infections: a systematic review
of the literature. J Hosp Infect 2007; 66:101108.
9 NQMCH. The national quality measures clearinghouse: measure validity;
2008. http://www.qualitymeasures.ahrq.gov/resources/measure_domain-
s.aspx. [Accessed 3 March 2008].
10 Guthrie B, Wyke S. Clinical trials or quality improvement for evaluating
 effectiveness of interventions? University of Dundee and the Alliance for Self
Care Research; 2008. http://www.nhstayside.scot.nhs.uk/about_nhstay/
commitees/08_iaq/27012009/docs_024499.pdf [Accessed 29 January
2009].
This paper clearly sets out the criteria for deciding when quality improvement
methods rather than clinical trials should be used to assess effectiveness.
11 Glasgow R, McKay H, Piette J, Reynolds K. The RE-AIM Framework for
evaluating interventions: what can it tell us about approaches to chronic
illness management? Patient Educ Couns 2001; 44:119127.
12 Dellinger RP, Levy MM, Carlet JM, et al. Surviving sepsis campaign: interna-
tional guidelines for management of severe sepsis and septic shock 2008.
Crit Care Med 2008; 36:296327.
13 SNAP-CAP. The Royal College of Physicians of Edinburgh and the Royal
College of Physicians and Surgeons of Glasgow. Scottish National Audit
Project for Community Acquired Pneumonia (SNAP-CAP); 2006. http://
www.snaproject.org.uk/cap/cap-index.php. [Accessed 8 February 2009].
14 Anonymous. VAP prevention bundle guidance for implementation. Edinburgh:
Scottish Intensive Care Society Audit Group 2008. http://www.sicsag.s-
cot.nhs.uk/SubGroup/VAP_Prevention_Bundle_Guidance_For_Implementa-
tion1.pdf. [Accessed 24 February 2009].
15 Bigham MT, Amato R, Bondurrant P, et al. Ventilator-associated pneumonia in
 the pediatric intensive care unit: characterizing the problem and implementing
a sustainable solution. J Pediatr 2009; 154:582587.
This study shows the impact of a VAP bundle on the VAP rate per 1000 patient-
days over 3 years in one paediatric ICU.
16 Jain M, Miller L, Belt D, et al. Decline in ICU adverse events, nosocomial
infections and cost through a quality improvement initiative focusing on
teamwork and culture change. Qual Saf Health care 2006; 15:235239.
17 Resar R, Pronovost P, Haraden C, et al. Using a bundle approach to improve
ventilator care processes and reduce ventilator-associated pneumonia. Jt
Comm J Qual Patient Saf 2005; 31:243248.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Care bundles Marwick and Davey 369
18 Gastmeier P, Geffers C. Prevention of ventilator-associated pneumonia:
analysis of studies published since 2004. J Hosp Infect 2007; 67:18.
19 Krein SL, Kowalski CP, Damschroder L, et al. Preventing Ventilator-Asso-
 ciated Pneumonia in the United States: A Multicenter Mixed-Methods Study.
Infect Control Hosp Epidemiol 2008; 29:933940.
This study provides valuable quantitative and qualitative data about the variation in
use and professional support for interventions to reduce risk of VAP in ICUs from
the USA.
20 Solberg LI, Mosser G, McDonald S. The three faces of performance mea-
surement: improvement, accountability, and research. Jt Comm J Qual Improv
1997; 23:135147.
21 Nguyen HB. Implementation of a bundle of quality indicators for the early
management of severe sepsis and septic shock is associated with decreased
mortality. Crit Care Med 2007; 35:11051112.
22 Pulcini C, Crofts S, Campbell D, Davey P. Design, measurement, and
evaluation of an education strategy in the hospital setting to combat anti-
microbial resistance: theoretical considerations and a practical example. Dis
Manage Health Outcomes 2007; 15:151163.
23 Lloyd RC. Quality healthcare: a guide to developing and using indicators.
Sudbury, Massachusetts: Jones and Barrett; 2004.
24 Stone SP, Cooper BS, Kibbler CC, et al. The ORION statement: guidelines
for transparent reporting of Outbreak Reports and Intervention studies Of
Nosocomial infection. J Antimicrob Chemother 2007; 59:833840.
25 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.
26 Baily MA. Quality improvement methods in healthcare. In: Crowley M, editor.
 From birth to death and bench to clinic: the Hastings Center bioethics briefing
book for journalists, policymakers, and campaigns garrison. New York: The
Hastings Center; 2008. pp. 147152.
This publicly accessible text book chapter argues strongly for a clear, common
sense ethical framework that encourages data-guided management of organiza-
tional change, health services research on processes and outcomes and colla-
boration across organizations to share best practices and improve the standard of
care.
27 Shojania KG, Levinson W. Clinicians in quality improvement: a new career
pathway in academic medicine. JAMA 2009; 301:766768.
28 Custers T, Hurley J, Klazinga NS, Brown AD. Selecting effective incentive
structures in healthcare: a decision framework to support healthcare purcha-
sers in finding the right incentives to drive performance. BMC Health Serv Res
2008; 8:66.
29 Metersky ML, Sweeney TA, Getzow MB, et al. Antibiotic timing and diagnostic
uncertainty in Medicare patients with pneumonia: is it reasonable to expect all
patients to receive antibiotics within 4 h? Chest 2006; 130:1621.
30 Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the
treatment of severe sepsis and septic shock. N Engl J Med 2001; 345:1368
1377.
31 Dellinger RP, Carlet JM, Masur H, et al. Surviving sepsis campaign guidelines
for management of severe sepsis and septic shock [see comment][erratum
appears in Crit Care Med 2004; 32:1448]. Crit Care Med 2004; 32:858
873.
32 Rivers EP, Coba V, Whitmill M. Early goal-directed therapy in severe sepsis
 and septic shock: a contemporary review of the literature. Curr Opin Anae-
thesiol 2008; 21:128140.
Eleven peer-reviewed publications and 28 abstracts with a total of 5598 patients
are reviewed. The mean absolute and RR reduction with early goal-directed
therapy was greater than reported in the original clinical trial and shows that
the results are generalizable across multiple healthcare settings.
33 Minton J, Clayton J, Sandoe J, et al. Improving early management of blood-
stream infection: a quality improvement project. BMJ 2008; 336:440443.
34 Pulcini C, Defres S, Aggarwal I, et al. Design of a day 3 bundle to improve the
 reassessment of inpatient empirical antibiotic prescriptions. J Antimicrob
Chemother 2008; 61:13841388.
This study describes the development and testing of a bundle to improve the
reassessment of inpatient empirical antibiotic prescriptions over 15 months in one
infectious diseases ward. This shows that bundles can be implemented success-
fully in wards with lower staff to patient ratios than ICUs.
35 Crunden E, Boyce C, Woodman H, Bray B. An evaluation of the impact of the
ventilator care bundle. Nurs Crit Care 2005; 10:242246.
36 Davey P, Brown E, Fenelon L, et al. Interventions to improve antibiotic
prescribing practices for hospital inpatients. Cochrane Database Syst Rev
2005:CD003543.
37 Anonymous. Saving lives: high impact interventions (HIIs); 2007. http://
www.clean-safe-care.nhs.uk/index.php?pid=4 [Accessed 5 March 2009].