Evaluation & the Health
The online version of this article can be found at:
2012 35: 3 originally published online 24 May 2011 Eval Health Prof
Pamela Snow and Joachim Kugler
Thomas Rotter, Leigh Kinsman, Erica James, Andreas Machotta, Jon Willis,
Review and Meta-Analysis
Systematic Outcomes, Length of Stay, and Hospital Costs: Cochrane
The Effects of Clinical Pathways on Professional Practice, Patient
can be found
Evaluation & the Health Professions
Additional services and information for
What is This?
- May 24, 2011 OnlineFirst Version of Record
- Feb 14, 2012Version of Record >>
by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013ehp.sagepub.comehp.sagepub.comehp.sagepub.com ehp.sagepub.com ehp.sagepub.com ehp.sagepub.comehp.sagepub.comehp.sagepub.comehp.sagepub.com ehp.sagepub.com ehp.sagepub.comehp.sagepub.comehp.sagepub.comehp.sagepub.com ehp.sagepub.com ehp.sagepub.com ehp.sagepub.comehp.sagepub.com ehp.sagepub.com ehp.sagepub.comehp.sagepub.com ehp.sagepub.comehp.sagepub.com ehp.sagepub.comehp.sagepub.comehp.sagepub.comDownloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from
The Effects of
of Stay, and Hospital
Thomas Rotter1, Leigh Kinsman2, Erica James3,
Andreas Machotta4, Jon Willis5, Pamela Snow6, and
1Department of International Health, School for Public Health and Primary Care (CAPHRI),
Faculty of Health, Medicine & Life Sciences, Maastricht University, Maastricht, The Netherlands
2School of Rural Health, Monash University, Bendigo, Australia
of Newcastle, Newcastle, Australia
4Department of Anaesthesiology, Sophia Children’s Hospital, Erasmus University Rotterdam,
Rotterdam, The Netherlands
5School of Public Health, Faculty of Health Sciences, La Trobe University, Melbourne, Australia
6School of Psychology and Psychiatry, Monash University, Bendigo, Australia
7Department of Public Health, Dresden Medical School, University of Dresden, Dresden,
Thomas Rotter, Postbus 616, Maastricht NL-6200 MD, The Netherlands
Evaluation & the Health Professions
ª The Author(s) 2012
Reprints and permission:
This paper is a summary version of the previously published Cochrane
review. It may increase the reach of the topic to health researchers and
practitioners and encourage further discussion. The systematic review aims
to summarize the evidence and assess the effect of clinical pathways on
professional practice, patient outcomes, length of hospital stay, and hospital
costs. The authors searched the Database of Abstracts of Reviews of
Effectiveness, the Effective Practice and Organisation of Care Register, the
Cochrane Central Register of Controlled Trials and bibliographic databases
including MEDLINE, EMBASE, CINAHL, NHS EED, and Global Health.
for analysis. The main results were a reduction in in-hospital complications
11.95: 95% CI [4.72, 30.30]) associated with clinical pathways. Considerable
length of stay (LOS) and hospital costs. The authors concluded that clinical
clinical pathways, cochrane systematic review, meta analysis, patient
For the purpose of this review, clinical pathways (CPW) are defined as
structured multidisciplinary care plans which detail essential steps in the
care of patients with a specific clinical problem. They support the transla-
tion of clinical guidelines into local protocols and clinical practice and have
been implemented internationally since the 1980s (Campbell, Hotchkiss,
Bradshaw, & Porteous, 1998). Clinical pathways have the potential to align
with the case-mix model of health care delivery and to support cost minimi-
zation (Kimberly, de Pouvourville, & d’Aunno, 2009). In 2003, it was
reported that CPW had been implemented in more than 80% of hospitals
in the United States. This reflects an enormous resource commitment in the
development and implementation of pathways in hospitals. In this era of
evidence-informed practice, it is therefore problematic that individual stud-
ies of the impact of CPW are varied and contradictory (Saint, Hofer, Rose,
Kaufman, & McMahon, 2003). The plethora of study designs, settings, and
4 Evaluation & the Health Professions 35(1)
proposed pathway interventions make the relevance of individual studies
very difficult to evaluate and apply to particular clinical settings. This is fur-
ther complicated by the varying perceptions of what constitutes a CPW. A
recent literature review identified 84 different terms that may mean a CPW.
These included (among others) care map, care pathway, critical pathway,
integrated care pathway, protocol, and guideline (De Bleser et al., 2006).
Thus, it was timely to objectively define a CPW and catalogue and analyze
the existing literature via a rigorous systematic review in order to provide a
framework for doctors, nurses, physiotherapists, and other health care
professionals considering the effectiveness of CPWs. We summarized the
evidence and assessed the effect of CPW on professional practice (e.g.,
quality of documentation), patient outcomes (e.g., mortality and, complica-
tions), length of hospital stay, and hospital costs.
The purpose of this article is to provide a summary version of the
previously published Cochrane review. This version may increase the reach
of the topic to health researchers and practitioners and encourage further
and Organisation of Care (EPOC) group as well as the Publisher (Wiley)
gave their permission to this subsequent publication. This article is solely
published on the behalf of the individual review authors and all authors
approved the final version.
We followed the Cochrane systematic review methodology for considering
and analyzing studies (Bero et al., 2010).
Review Eligibility Criteria
We included randomized controlled trials (individual or cluster level),
controlled clinical trials (CCTs), controlled before and after studies
(CBAs), and studies using interrupted time series analysis (ITS) designs
were included after meeting EPOC methodological design and quality
criteria. The study population included (a) hospitalized patients (inpatient
and outpatient settings) with conditions managed on a CPW, irrespective
psychiatrists, speech pathologists, and dentists involved in CPW utilization
in the hospital setting, and (c) hospitals evaluating the impact of CPWs.
Rotter et al.5
and met three of the following four criteria:
1. The intervention was used to translate guidelines or evidence into local
The intervention detailed the steps in the course of treatment or care in
a plan, pathway, algorithm, guideline, and protocol of other ‘‘inventory
The intervention had time frames or criteria-based progression; and
procedure, or episode of health care in a specific population.
After meeting the minimum criteria for defining a CPW, we included
studies comparing (a) patients managed according to CPWs compared to
usual care and (b) patients managed within a multifaceted intervention
including a CPW compared to usual care. Patient outcome measures
of mechanical ventilation. Professional practice outcomes included quality
measures appropriate to the specific aim of the CPW, staff satisfaction, and
adherence toevidence-basedpractice,for example,qualityofdocumentation.
Length of stay (LOS) was assessed by extracting the duration of hospital stay
measured inhours ordaysthat were reportedin the included studies. Hospital
costs included cost of hospitalization and any appropriate resource utilization
data. We searched the Database of Abstracts of Reviews of Effectiveness
(DARE), the EPOC register, the Cochrane Central Register of Controlled
Trials (CENTRAL) and bibliographic databases such as MEDLINE,
tified terms that may relate to use of a CPW in a hospital environment.
ing reference lists and contacting authors of included studies, searching of
the ISI Web of Science and contact with professional associations, for
example, the European Pathways Association were performed to identify
potentially relevant studies. No language limitations were imposed.
The MEDLINE search strategy is presented in detail in the full version of
the EPOC review, recently published in the Cochrane Library (Rotter et al.,
2010). Two review authors independently screened all titles and abstracts
(LK and EJ for professional practice and patient outcomes; TR and AM for
relevanceregarding LOS and hospitalcosts).Unresolveddisagreementswere
6 Evaluation & the Health Professions 35(1)
referred to a third review author. The same review authors independently
assessed full-text articles for inclusion against criteria for defining a CPW
and study quality using a validated and standardized data extraction form.
Details on the number of retrieved references, the number of obtained full
text papers and the number of included and excluded articles are presented
in detail in the full version of the EPOC review, recently published in the
Cochrane Library(Rotteretal.,2010).Whennecessary,we soughtadditional
information from the authors of the primary studies. The methodological
quality of all included studies was also assessed independently by two
review authors (LK and EJ for patient outcomes and professional practice;
TR and AM for LOS and hospital costs) using the EPOC risk of bias tool
(Bero et al., 2010) and categorized them as low, moderate, or high risk.
We referred unresolved disagreements on risk of bias classification to a
third review author. We excluded studies classified as high risk of bias.
A summary of the risk of bias assessment is given in Table 1.
Data Analysis and Statistical Pooling of Studies
Statistical pooling was conducted when studies were sufficiently clinically
test). A Cochrane web-based program, Review Manager (RevMan), was
used to calculate a pooled estimate of the combined intervention effect
(Review-Manager, 2008). Cost/charges data are presented in $US for the
Version 1.0, a web-based tool that can be used to adjust an estimate of cost
expressed in one currency and price year to a target currency and/or price
year (Shemilt; Shemilt, Thomas, & Morciano, 2010). Costs/charges were
adjusted for inflation by applying Gross Domestic Product deflators as
‘‘GDPD values’’ or using government recommended rates and providing
a sensitivity analysis with a common discount rate recommended in the
literature (Drummond & Jefferson, 1996). Additionally, we provide the
undiscounted cost data to allow readers to recalculate the results using any
review recently published in the Cochrane Library (Rotter et al., 2010).
A final inclusion sample of 27 studies (see Table 1) was derived from the
3,214 abstracts identified by the search strategy. A summary of the key
Rotter et al.7
Table 1. Key Characteristics of Included Primary Studies
Type of Ward
Country Risk of Bias
Comparison 1: single CPW intervention versus usual care
Brook, 1999 Mechanical ventilation
Menorrhagia and urinary
Femural neck fracture
CPW Laparotomy and
Medical units of the hospitals
Hip and knee arthroplasty
Coronary Care unit/chest pain
Emergency and Pediatric
Table 1 (continued)
Type of Ward
Country Risk of Bias
Medical & Surgical ICU
Marrie, 2000 Pneumonia
CPW Chest pain/possible MI
Tilden, 1987 Identification of battered
Comparison 2: Multifaceted intervention including a CPW versus usual care
Bauer et al.,
Mental health outpatient clinic
Table 1 (continued)
Type of Ward
Country Risk of Bias
Care of delirium in older
Diabetic patients admitted
Philbin, 2000 Patients with heart failure
Note. Acute ¼ General acute hospital; AUS ¼ Australia; CBA ¼ controlled before and after study; CCT ¼ controlled clinical trial; COPD ¼ chronic
obstructive pulmonary disease; C-RCT ¼ cluster randomized clinical trial; ED ¼ Emergency department; Extended care ¼ Rehabilitation or palliative
facilities; ICU ¼ Intensive care unit; ITS ¼ interrupted time series; MI ¼ myocardial infarction; Other ¼ Psychiatric or mental health clinic/hospital;
P-RCT ¼ patient randomized clinical trial; UK ¼ United Kingdom; USA ¼ United States of America; TURP ¼ transurethral resection of the prostate.
characteristics of included studies is provided in Table 1. Results from the
study by Sulch were reported in two separate publications (2000 and 2002).
Included Study Characteristics
Of the 27 included studies, 19 were RCTs (Aizawa et al., 2002; Bauer et al.,
2006; Brook et al., 1999; Chen et al., 2004; Cole et al., 2002; Delaney et al.,
2003; Dowsey, Kilgour, Santamaria, & Choong, 1999; Falconer, Roth,
& Mooers, 1996; Johnson, Blaisdell, Walker, & Eggleston, 2000; Kampan,
2006; Kim et al., 2002; Kiyama et al., 2003; Kollef et al., 1997; Marelich
et al., 2000; Marrie et al., 2000; Philbin et al., 2000; Roberts et al., 1997;
Sulch, Melbourn, Perez, & Kalra, 2002; Sulch, Perez, Melbourn, & Kalra,
2000), four were CBAs (Bookbinder et al., 2005; Chadha et al., 2000;
Doherty & Jones, 2006; Smith et al., 2004), two were CCTs (Choong,
2004) and two were ITS (Brattebo et al., 2002; Tilden & Shepherd, 1987).
Of the 19 RCTs, two were cluster-randomized studies (Marrie et al., 2000;
Philbin et al., 2000).
Included studies targeted a large range of conditions (Please see Table 1
for a brief description). Across the 27 studies, there were 19 different con-
ditions targeted. Please see also effects of pathway interventions, subgroup
analysis by condition. Thirteen of the studies were conducted in the United
States (Bauer et al., 2006; Bookbinder et al., 2005; Brook et al., 1999;
Delaney et al., 2003; Falconer et al., 1993; Gomez et al., 1996; Johnson
et al., 2000; Kim et al., 2002; Kollef et al., 1997; Marelich et al., 2000;
Philbin et al., 2000; Roberts et al., 1997; Tilden & Shepherd, 1987) four
in Australia (Choong et al., 2000; Doherty & Jones, 2006; Dowsey et al.,
1999; Smith et al., 2004), three in Japan (Aizawa et al., 2002;
Kiyama et al., 2003; Usui et al., 2004), two each in the United Kingdom
(Chadha et al., 2000; Sulch et al., 2002; Sulch et al., 2000), and Canada
(Cole et al., 2002; Marrie et al., 2000) and one each in Thailand (Kampan,
studies were conducted in general acute wards (e.g., medical, surgical,
pediatrics, and gynecology; Aizawa et al., 2002; Chadha et al., 2000; Chen
et al., 2004; Choong et al., 2000; Cole et al., 2002; Doherty & Jones, 2006;
Dowsey et al., 1999; Gomez et al., 1996; Johnson et al., 2000; Kampan,
2006; Kiyama et al., 2003; Marrie et al., 2000; Philbin et al., 2000; Smith
tion; Bookbinder et al., 2005; Delaney et al., 2003; Falconer et al., 1993;
Rotter et al.11
Sulch et al., 2000; Sulch et al., 2002), four in an ICU (Brattebo et al., 2002;
Brook et al., 1999; Kollef et al., 1997; Marelich et al., 2000), three in an
Emergency Department (ED; Kim et al., 2002; Roberts et al., 1997; Tilden &
Shepherd, 1987), and one (Bauer et al., 2006) in a mental health outpatient
clinic. In nine studies, the CPW was designed for an invasive procedure
(Aizawa et al., 2002; Brattebo et al., 2002; Brook et al., 1999; Choong et al.,
2000; Delaney et al., 2003; Dowsey et al., 1999; Kiyama et al., 2003; Kollef
et al., 1997; Marelich et al., 2000). Sixteen described CPWs for a noninvasive
2000; Philbin et al., 2000; Smith et al., 2004; Sulch et al., 2002; Sulch et al.,
2000; Tilden & Shepherd, 1987; Usui et al., 2004) and two described
CPWs for combined invasive/noninvasive procedures, for example, for sus-
pected MI with or without percutaneous transluminal coronary angioplasty
(PTCA) (Chadha et al., 2000; Gomez et al., 1996).
Intervention Characteristics and Comparison
Twenty studies compared a stand-alone CPW to usual care (Aizawa et al.,
2002; Brook et al., 1999; Chadha et al., 2000; Choong et al., 2000; Delaney
et al., 2003; Doherty & Jones, 2006; Dowsey et al., 1999; Falconer et al.,
1993; Gomez et al., 1996; Johnson et al., 2000; Kim et al., 2002; Kiyama
et al., 2003; Kollef et al., 1997; Marelich et al., 2000; Marrie et al., 2000;
Roberts et al., 1997; Smith et al., 2004; Sulch et al., 2002; Sulch et al., 2000;
Tilden & Shepherd, 1987; Usui et al., 2004) and seven compared a multifa-
der et al., 2005; Brattebo et al., 2002; Chen et al., 2004; Cole et al., 2002;
combined with case management elements (Bauer et al., 2006; Bookbinder
et al., 2005; Chen et al., 2004; Cole et al., 2002; Kampan, 2006) or with com-
plex quality improvement programs (Bookbinder et al., 2005; Philbin et al.,
2000). Five studies were assessed as at low risk of bias with the remaining
22 rated as moderate. Table 1 presents a summary of the quality assessment.
Effects of Pathway Interventions
Comparison 1: Single CPW Intervention Versus Usual Care
In-hospital complications were measured in five invasive pathway studies
and all reported improved outcomes for the CPW group (Aizawa et al.,
12 Evaluation & the Health Professions 35(1)
2002; Choong et al., 2000; Delaney et al., 2003; Kiyama et al., 2003;
Marelich et al., 2000). The invasive CPW conditions included were TURP
(Aizawa et al., 2002), femoral neck fracture (Choong et al., 2000), laparot-
omy and intestinal resection (Delaney et al., 2003), gastrectomy (Kiyama
et al., 2003), and mechanical ventilation. The combined odds ratio (OR) for
in-hospital complications was 0.58 (95% CI [0.36, 0.94]) in favor of inva-
sive CPWs and statistically significant. The effect of CPWs on in-hospital
complications in the individual studies are depicted together with the
pooled OR in Figure 1. Statistical heterogeneity was not present among
the studies, indicating reliable pooled (in-hospital complications) results.
There was clinical variance in the range of follow-up periods that were used
as well as varying definitions of the term (in-hospital) complications.
The in-hospital complications included in the primary studies were post-
operative infections, thrombosis, bleeding, and pneumonia.
Six measures were comparable in terms of hospital readmission reported
for all causes and characterized with follow-up periods up to 6 months
(Aizawa et al., 2002; Choong et al., 2000; Delaney et al., 2003; Dowsey
et al., 1999; Gomez et al., 1996; Roberts et al., 1997). None of these
reported readmission rates reached statistical significance as reported in the
primary investigations. Statistical heterogeneity was not present (I2¼ 0%)
among the studies. The pooled OR for readmission was 0.6 (95% CI [0.32,
1.13]) in favor of clinical pathways but was not statistically significant.
Within the subgroup of single pathway interventions, three studies were
clinically and statistically comparable and reported in-hospital mortality
rates. The pooled OR for in-hospital mortality was 0.84 (95% CI [0.61,
1.11]) in favor of CPWs but did not reach a statistically significant level.
Three studies measured the impact of CPWs on quality and quantity
of documentation in medical records and two reported positive findings
(Doherty & Jones, 2006; Sulch et al., 2002; Tilden & Shepherd, 1987).
Doherty (2006) reported on the documentation of severity of asthma and
Sulch et al. (2002) measured the documentation of team goals for stroke
patients. Tilden & Shepherd, (1987) measured documented identification
of female victims of domestic violence in the emergency department and
found no statistically significant change when time-series analysis was
utilized. The studies by Doherty (2006) and Sulch et al. (2002) were clini-
cally and statistically comparable, resulting in a substantial and significant
result (OR 11.95: 95% CI [4.72, 30.30], see Figure 2) favoring improved
documentation with CPWs.
LOS was the most commonly employed outcome measure and the
majority of studies reporting LOS data showed a positive impact. Of the
Rotter et al. 13
20 studies categorized as single pathway interventions, 14 (70%) primary
studies examined the effect of CPWs on LOS (Aizawa et al., 2002; Brook
et al., 1999; Choong et al., 2000; Delaney et al., 2003; Dowsey et al., 1999;
Falconer et al., 1993; Gomez et al., 1996; Johnson et al., 2000; Kim et al.,
2002; Kiyama et al., 2003; Marrie et al., 2000; Roberts et al., 1997; Smith
et al., 2004; Sulch et al., 2000; Usui et al., 2004), with 11 showing significant
reductions (Aizawa et al., 2002; Brook et al., 1999; Choong et al., 2000;
Dowsey et al., 1999; Gomez et al., 1996; Johnson et al., 2000; Kim et al.,
2002; Kiyama et al., 2003; Marrie et al., 2000; Roberts et al., 1997; Usui
et al., 2004). The LOS reported by Kiyama et al. (2003) was calculated from
Heterogeneity within this subgroup analysis of studies reporting on LOS was
Primary studies were ordered in forest plots by country to examine possible
different market effects. We observed greater reported LOS effects from
Figure 1. Stand-alone clinical pathway versus usual care, outcome: OR in-hospital
complications. Source: Review-Manager.
Figure 2. Stand-alone clinical pathway versus usual care, outcome: OR professional
documentation. Source: Review-Manager.
14 Evaluation & the Health Professions 35(1)
Japanese studies with a pooled reduction of approximately three days
weighted mean difference (WMD) 3.01, followed by studies carried out
in Australia (WMD 1.6) and in the United States (WMD 0.8). Studies
carried out in the United States provided the majority of studies included
in the present review but reported the smallest decreases in LOS. A slightly
similar pattern was observed in hospital cost and charge outcomes reported.
There were four conditions or interventions for which there was more
than one included study. There were two studies evaluating pathway
management for stroke rehabilitation (Falconer et al., 1993; Sulch et al.,
2002; Sulch et al., 2000), pneumonia (Marrie et al., 2000; Usui et al., 2004),
mechanical ventilation (Brook et al., 1999; Kollef et al., 1997). Further
urethral resection of the prostate (Aizawa et al., 2002), menorrhagia and
urinary incontinency (Chadha et al., 2000), femoral neck fracture (Choong
et al., 2000), laparotomy and intestinal resection (Delaney et al., 2003),
asthma care (Doherty & Jones, 2006), hip and knee arthroplasty (Dowsey
et al., 1999), asthma in children (Johnson et al., 2000), atrial fibrillation
(Kim et al., 2002), gastrectomy (Kiyama et al., 2003), chronic obstructive
pulmonary disease or COPD (Smith et al., 2004) and a pathway instrument
designed for the better identification of female victims of domestic
violence (Tilden & Shepherd, 1987). Significant clinical and statistical
heterogeneity prevented the estimation of an overall pooled effect where
studies were grouped according to condition. Therefore, we concentrated
on subgroup analysis per pathway condition without a total estimate,
please see Figure 3. Falconer (1993) and Sulch et al. (2000) both reported
nonsignificant increased LOS associated with CPWs in stroke rehabilita-
tion units (see Figure 3). The combined OR for these two studies was
3.9 (95% CI [?0.29, 8.27]). Sulch et al. (2000) also compared mortality
at 26 weeks (13% vs. 8%) and found no statistically significant difference.
no differences in patient satisfaction but significant improvements in
the documentation of several processes, including nutritional assessment
(p < .001), multidisciplinary team goals (p < .001), and death (p ¼ .024),
as well as general practitioner (GP) notification of death or discharge
(p < .001).
Marrie et al. (2000) and Usui et al. (2004) both reported significant
reductions in LOS and duration of intravenous antibiotic infusion when
CPWs were implemented for the inpatient management of pneumonia.
Marrie reported a LOS of 8.2 (SD ¼ 1.9) days in the CPWs group versus
Rotter et al. 15
while duration of intravenous antibiotic infusion was also significantly
less in the CPWs group, 4.6 days (SD ¼ 0.9) versus 6.3 days (SD ¼ 1.4);
(WMD ?1.70; 95% CI [?2.01, ?1.39]). Usui reported a LOS of 8.0
(SD ¼ 4.2) days in the CPWs group versus 10.8 (SD ¼ 4.2) days in the
control group (WMD ?2.74; 95% CI [?4.84, ?0.64]) while duration of
intravenous antibiotic infusion was also significantly less in the CPWs
group, 6.5 days (SD ¼ 3.5) versus 8.2 days (SD ¼ 3.5); (WMD [?1.75;
95% CI [?3.52, 0.02]). When Marrie et al. (2000) and Usui et al.’s
(2004) results were statistically combined, LOS was 1.67 days (95% CI
(95% CI [?2.01, ?1.40]) in the CPW group. Gomez et al. (1996) and
Roberts et al. (1997) both reported decreases in LOS for CPWs implemen-
ted in emergency departments for suspected myocardial infarction
(see Figure 3). The combined LOS for the Gomez et al. (1996) and Roberts
et al. (1997) studies was WMD ?0.90 days (95% CI [?1.98, 0.18]) but did
not reach statistical significance. No evidence of a statistically significant
difference in 30-day readmission was found in the Gomez et al. (1996)
study (6% vs. 6%) or for 8-week readmission in the Roberts et al. (1997)
study (6.1% vs. 4.8%).
Three studies (Brook et al., 1999; Kollef et al., 1997; Marelich et al.,
2000) reported similar reductions in the total time patients required
Figure 3. Stand-alone clinical pathway versus usual care, outcome: LOS grouped
by CPW condition. Source: Review-Manager.
16 Evaluation & the Health Professions 35(1)
(95% CI [?55.73, ?11.71; see Figure 4). Marelich et al. (2000) also found a
statistically significant reduction in ventilation hours for the intervention
group (p ¼ .0001) but reported medians and interquartile ranges from which
the primary data could not be obtained for calculation of means. The median
(interquartile range 33–164) versus 124 hours (interquartile range 54–334)
for the control group (n ¼ 169). The different reporting of Marelich’s data
prevented this study being combined in meta-analysis with other mechan-
ical ventilation studies (Marelich et al., 2000). However, the findings of
Marelich’s study were consistent with the findings from the other studies
measuring the impact of CPWs on mechanical ventilation.
eight reported on a highly varying set of cost/charge measures. Of the eight
studies considering cost outcomes or surrogates, six found significant lower
hospitalization costs/charges or insurance points for pathway groups
(Aizawa et al., 2002; Gomez et al., 1996; Johnson et al., 2000; Kiyama
et al., 2003; Roberts et al., 1997; Usui et al., 2004). The statistical inconsis-
(I2¼ 66%) was substantial and compromised the estimation of a pooled
effect. Even the high level of heterogeneity per subgroup may refer to the
varying CPW interventions included in the present analysis as well as to the
considerable methodological variation in the sort of hospital costs included
a detailed description of the costing methods used and cost measures
are presented in the full version of the EPOC review in tabular form as
to the year 2000 (Rotter et al., 2010).
Figure 4. Stand-alone clinical pathway versus usual care, outcome: duration of
mechanical ventilation. Source: Review-Manager.
Rotter et al.17
Table 2. Overview of Costing Methods Used and Costs Measures Included in the Primary Studies
Country Costing Methods
Comparison 1: Single CPW intervention versus usual care
Insurance data (points)
Total hospital charges:
Including variable and fixed costs
Dosage, injection, treatment, operation, and
anesthesia, examination, diagnostic, room, and medical care
Falconer, 1993 Hospital charges to
proxy direct costs of
Charges for hospital bed days, medical and
rehabilitation services (including professional
fees), equipment, drugs and procedures
(radiographs, laboratory tests, injections)
Room, nursing care, laboratory, therapeutic,
Room, medication, laboratory tests, and
Direct variable costs
Remains unclear, only ‘‘total direct costs’’ reported Professional
Direct variable costs
Total medical costs including medication and
examination (physician fees)
Total direct variable and
Insurance data (points)
Direct charges: including
Treatment (antibiotic infusion), laboratory,
and radiography tests
Comparison 2: Multifaceted intervention including a CPW versus usual care
Bauer et al.
Direct variable costs
Thailand Remains unclear, only
Comparison 2: Multifaceted Intervention Including a CPW
Versus Usual Care
Of the seven primary studies categorized as multifaceted interventions
including a CPW element, three studies (Chen et al., 2004; Kampan,
2006; Philbin et al., 2000) measured the impact of a CPW on rate of hospital
readmission and two studies used mortality as a study endpoint and reported
no statistically significant difference following implementation of the CPW
(Cole et al., 2002; Philbin et al., 2000). Three investigations reported
LOS measures for statistical comparison (Cole et al., 2002; Kampan,
2006; Philbin et al., 2000) and no significant between-group differences
were seen. Three studies reported on hospital costs/charges (Bauer et al.,
2006; Kampan, 2006; Philbin et al., 2000). None of these three studies
reported statistically significant differences in costs/charges outcomes.
In summary, multifaceted interventions did not differ from usual care for
mortality, length of hospital stay, or hospital costs/charges. One (n ¼ 65)
out of three randomized studies found that multifaceted interventions
reduced hospital readmissions more than usual care, but only a very small
sample size was available for analysis and the study was underpowered
(Kampan, 2006). The study by Kampan (2006) reported in particular a
significant reduction in 6-month readmissions for hypoglycemia in patients
with diabetes (6% vs. 34%; p ¼ .04).
Seven separate conditions were analyzed in this group and subgroup
analysis was not possible. The different pathway indications were bipolar
disorder (Bauer et al., 2006), palliative care (Bookbinder et al., 2005),
mechanical ventilation (Brattebo et al., 2002), asthma in children
diabetic patients admitted with hypoglycemia (Kampan, 2006) and heart
failure (Philbin et al., 2000).
We screened and analyzed over 3,000 published studies for this review of
the impact of CPWs in hospitals, and after applying inclusion criteria,
27 studies were included with a total of 11,398 participants. Included
studies arose from eight different countries for CPWs implemented in many
Rotter et al.19
The number of included studies, total number of participants, and breadth
of settings suggest that this review provides a solid profile of the impact of
CPWs. The results are relevant to a variety of settings worldwide.
heterogeneity that made meta-analysis inappropriate for many of the out-
comes extracted. Despite this limitation, some of our findings remain
meaningful for clinicians, managers, and researchers, and eliminate some
of the contradictory findings from individual studies.
A major finding was the significant reduction in in-hospital complica-
tions associated with the introduction of CPWs. All seven studies (Aizawa
et al., 2002; Bookbinder et al., 2005; Choong et al., 2000; Delaney et al.,
2003; Dowsey et al., 1999; Kiyama et al., 2003; Marelich et al., 2000) that
measured complications reported results that favored CPWs. Six of the
seven studies examined invasive conditions or interventions, for example,
surgery procedures or mechanical ventilation. This reflects the fact that
studies of CPWs for invasive conditions were more likely to use complica-
tion measures such as infection and bleeding as an objective outcome
measure rather than suggesting that CPWs only reduce complication rates
for invasive procedures. The pooled result of an absolute risk reduction
of 5.6% (n ¼ 5 trials) for patients recovering from surgery, who were
managed on a CPW corresponds to prevention of one complication for
every 18 patients treated (NNT ¼ 18). This strongly suggests that CPWs
have a substantial role to play in patient safety.
Documentation appears to improve with the implementation of a CPW.
Clinical and statistical homogeneity supported the pooling of the studies by
Doherty (2006) and Sulch et al. (2002) resulting in a substantial and
significant result (OR 11.95: 95% CI [4.72, 30.30]) favoring improved
documentation with CPWs. While improved documentation may not appear
to be an outcome that directly influences patient outcomes, any intervention
that enhances communication must have a favorable influence on patient
care (Jorm, White, & Kaneen, 2009).
Multiple studies measured the impact of CPWs on pneumonia (Marrie
et al., 2000; Usui et al., 2004), myocardial infarction (Gomez et al.,
1996; Roberts et al., 1997), and mechanical ventilation(Brattebo et al.,
2002; Brook et al., 1999; Kollef et al., 1997; Marelich et al., 2000). All
found that hospital resources were reduced while patient outcomes were
not adversely affected. This reinforces the notion that CPWs are associ-
ated with efficient use of resources and efficiency of care. There were
insufficient numbers of homogenous studies to draw other conclusions
at this stage.
20 Evaluation & the Health Professions 35(1)
The findings regarding LOS for the Falconer (1993) and Sulch et al.
(2000, 2002) studies were not statistically significant but did not support the
decreased LOS from CPWs reported in other studies (Falconer et al., 1993;
Sulch et al., 2000; Sulch et al., 2002). This may be explained by the reha-
bilitation settings in which these studies were conducted already delivering
optimal care without use of a clinical pathway. The Stroke Unit Trialists’-
Collaboration (2007) landmark Cochrane systematic review reported that
improved outcomes were associated with admission to a specialized
stroke unit and organized multidisciplinary care (Stroke-Unit-Trialists’-
Collaboration, 2007). The rehabilitation settings described in the Falconer
(1993) and Sulch et al. (2000, 2002) studies contained these elements
already and it is highly likely their type of care was optimizing stroke
management without the introduction of a CPW.
Despite being utilized in health care since the 1980s, no clear definition
for CPWs has been widely accepted. Subsequently, minimum content
criteria were developed for this review based on previous attempts to
empirically describe CPWs and pilot tested for reliability. This approach
maximized the identification and assessment of studies where the interven-
tion of interest could be considered a CPW, despite the wide variety of
terms used in the literature. However, the time and effort taken to identify
relevant studies for this review highlights the difficulty facing clinicians
andhealth care managerswhen tryingtoascertain andappraise the evidence
regarding CPWs. It is imperative that an internationally accepted definition
of a CPW is adopted in order for current literature to be easily and widely
accessed and compared.
The proportion of studies screened that were sufficiently well designed,
conducted, and reported to enable inclusion was very small. Of the 3,214
search hits, only 27 studies met inclusion criteria, once the inclusion and
EPOC design and quality criteria were applied. The majority of studies
excluded from the review after meeting CPW content criteria were simple
before and after studies, mostly comparing two or more yearly patient
cohorts. This simple study design can be useful for internal monitoring, but
it is very difficult and misleading to draw meaningful conclusions due to the
lack of control and inherent high level of bias. Poor reporting was also a
large obstacle in this review and better reporting of study methods could
have facilitated the inclusion of more studies for analysis. While experi-
mental methods such as randomized trials are recommended, they may be
considered beyond the capacity of many clinicians and researchers.
Another well-designed evaluation like time-series analysis that meets the
EPOC gold standard methodological criteria can produce meaningful,
Rotter et al. 21
rigorous results with the use of relatively few resources. Our screening and
appraising of studies indicate that a large amount of resources are being
used on studies that do not contribute to the evidence base regarding CPWs
because of their low study design quality.
Studies were ordered in forest plots by country and major differences
were observed. This refers to the country-specific market forces and the
problematic generalization of the conclusions drawn from this systematic
review. Replicating the results of this review in other settings could be
problematic. As an example, it could be highly difficult to replicate conclu-
sions drawn from Japanese settings into a U.S. American hospital setting
where LOS is historically lower. The market forces in form of the average
LOS in acute care by country (Organisation for economic co-operation and
development, 2008) indicate also a country-specific estimate for the poten-
tial LOS impact of CPW strategies and are evidenced by the observed LOS
patterns from the present review if grouped or sorted by country. LOS in
hospital as reported in 11 studies was significantly reduced when a CPW
was introduced. Seven other studies measured LOS and found no statisti-
cally significant differences. While statistical heterogeneity prevented
pooled analysis the extent of the reductionreported indicates that it is highly
likely that CPWs are associated with reduced LOS. This is important when
combined with the magnitude of the reduced costs associated with CPWs,
for which meta-analysis was also inappropriate. This means that the
improved patient outcomes (e.g., fewer complications) and process of care
measurements (e.g., improved documentation) do not occur in a setting of
increased use of hospital resources.
While the design and content of a CPW were recognized as important, it
was not possible to assess these aspects as the CPW document was rarely
provided in the journal article. Future evaluations should therefore follow
the SQUIRE guidelines to increase the standard of reporting on design,
content, and implementation criteria. Moreover, design and content attri-
butes of successful CPWs require further exploration for which document
review could be incorporated to elicit characteristics associated with CPW
Implications for Practice
Thisreviewhas establishedthat CPWsareassociatedwithreducedcomplica-
tions and improved documentation when implemented in hospitals without
with invasive interventions or surgical conditions such as fractured neck of
22 Evaluation & the Health Professions 35(1)
femur (Choong et al., 2000), intestinal resection (Delaney et al., 2003),
gastrectomy, mechanical ventilation (Marelich et al., 2000), transurethral
resection of the prostate (Aizawa et al., 2002), and hip or knee arthroplasty
(Dowsey et al., 1999).
Implications for Research
Studies measuring the impact of CPWs should incorporate EPOC standards
into design to maximize the quality of evidence underpinning this model
that is being utilized in a vast array of health care settings. The comparison
of LOS in days revealed the largest decrease in statistical heterogeneity
when grouped per pathway condition. This has implications for future
systematic reviews. Assuming a high number of primary pathway investi-
gations meeting the EPOC quality gold standard, future reviews should
focus on grouping and comparing within pathway conditions, for example,
CPWs for pneumonia. This strategy is highly supported by the low level of
heterogeneity observed by grouping per condition.
Theauthorsthank manypeoplewho provided valuedassistanceto thisreview,includ-
ing Boer Xu for Chinese language translation,Mikako Hayashi for Japanese language
the Thai Cochrane Network; EPOC personnel, particularly Emma Tavender, Russ
Gruen, Doug Salzwedel, Luke Vale, and Driss Ait Ouakram; Australasian Cochrane
Centre personnel, particularly Veronica Pitt, Miranda Cumpston, and Damien Jolley;
Robert Champion for advice pertaining to Interrupted Time-Series studies.
Declaration of Conflicting Interests
The authors declared no conflicts of interest with respect to the authorship and/or
publication of this article.
The authors disclosed receipt of the following financial support for the research and/
or authorship of this article: The study was funded by the Group of 8/DAAD Ger-
man Research Exchange program for facilitating essential face-to-face meetings of
the review team.
Aizawa, T., Kin, T., Kitsukawa, S. I., Mamiya, Y., Akiyama, A., Ohno, Y., ...
Tachibana, M. (2002). Impact of a clinical pathway in cases of transurethral
resection of the prostate. Japanese Journal of Urology, 93, 463-468.
Rotter et al.23
Cooperative Studies Program 430 Study Team. (2006). Collaborative care for
bipolar disorder: Part I (& II) Intervention and implementation in a randomized
effectiveness trial. Psychiatry Services, 57, 927-936.
Bero L, Deane K, Eccles M, Grimshaw J, Gruen RL, Mayhew A, ... Zwarenstein
About The Cochrane Collaboration (Cochrane Review Groups
(CRGs)). Cochrane Effective Practice and Organisation of Care Group, 2009.
Bookbinder, M., Blank, A. E., Arney, E., Wollner, D., Lesage, P., McHugh, M., ...
Portenoy, R. K. (2005). Improving end-of-life care: Development and pilot-test
of a clinical pathway. Journal of Pain and Symptom Management, 29, 529-543.
Brattebo, G., Hofoss, D., Flaatten, H., Muri, A. K., Gjerde, S., & Plsek, P. E. (2002).
Effect of a scoring system and protocol for sedation on duration of patients’ need
for ventilator support in a surgical intensive care unit. British Medical Journal,
Brook, A. D., Ahrens, T. S., Schaiff, R., Prentice, D., Sherman, G.,
Shannon, W., ... Kollef, M. H. (1999). Effect of a nursing-implemented
sedation protocol on the duration of mechanical ventilation. Critical Care
Medicine, 27, 2609-2615.
Campbell, H., Hotchkiss, R., Bradshaw, N., & Porteous, M. (1998). Integrated care
pathways. British Medical Journal, 316, 133-137.
Chadha, Y., Mollison, J., Howie, F., Grimshaw, J., Hall, M., & Russell, I. (2000).
BJOG: An International Journal of Obstetrics & Gynaecology, 107, 535-543.
Chen, S. H., Yeh, K. W., Chen, S. H., Yen, D. C., Yin, T. J., & Huang, J. L. (2004).
The development and establishment of a care map in children with asthma in
Taiwan. Journal of Asthma, 41, 855-861.
Choong, P. F., Langford, A. K., Dowsey, M. M., & Santamaria, N. M. (2000).
Clinical pathway for fractured neck of femur: A prospective, controlled study.
Medical Journal of Australia, 172, 423-426.
Cole, M. G., McCusker, J., Bellavance, F., Primeau, F. J., Bailey, R. F.,
Bonnycastle, M. J., ... Laplante, J. (2002). Systematic detection and multidis-
Medical Association Journal, 167, 753-759.
(2006). Defining pathways. Journal of Nursing Management, 14, 553-563.
rehabilitation with early ambulation and diet and traditional postoperative care
after laparotomy and intestinal resection. Disease of the Colon & Rectum, 46,
24Evaluation & the Health Professions 35(1)
Doherty, S. R., & Jones, P. D. (2006). Use of an ‘evidence-based implementation’
strategy to implement evidence-based care of asthma into rural district hospital
emergency departments. Rural and Remote Health, 6, 529.
Dowsey, M.M., Kilgour, M. L., Santamaria, N. M., &Choong, P. F.(1999). Clinical
pathways in hip and knee arthroplasty: A prospective randomised controlled
study. Medical Journal of Australia, 170, 59-62.
Drummond, M. F., & Jefferson, T. O. (1996). Guidelines for authors and peer
reviewers of economic submissions to the BMJ. The BMJ Economic Evaluation
Working Party. British Medical Journal, 313, 275-283.
Falconer, J. A., Roth, E. J., Sutin, J. A., Strasser, D. C., & Chang, R. W. (1993).
The critical path method in stroke rehabilitation: Lessons from an experiment
in cost containment and outcome improvement. QRBQuality Review Bulletin,
Gomez, M. A., Anderson, J. L., Karagounis, L. A., Muhlestein, J. B., &
Mooers, F. B. (1996). An emergency department-based protocol for rapidly
ruling out myocardial ischemia reduces hospital time and expense: Results of
a randomized study (ROMIO). Journal of the American College of Cardiology,
Johnson, K. B., Blaisdell, C. J., Walker, A., & Eggleston, P. (2000). Effectiveness
of a clinical pathway for inpatient asthma management. Pediatrics, 106,
Jorm, C., White, S., & Kaneen, T. (2009). Clinical handover: Critical communica-
tion. Medical Journal of Australia, 190, S108-S109.
Kampan, P. (2006). Effects of counseling and implementation of clinical pathway
on diabetic patients hospitalized with hypoglycemia. Journal of the Medical
Association of Thailand, 89, 619-625.
Kim, M. H., Morady, F., Conlon, B., Kronick, S., Lowell, M., Bruckman, D., ...
Eagle, K. A. (2002). A prospective, randomized, controlled trial of an emergency
department-based atrial fibrillation treatment strategy with low-molecular-
weight heparin. Annals of Emergency Medicine, 40, 187-192.
Kimberly, J., de Pouvourville, G., & d’Aunno, T. (2009). The globalization of
managerial innovation in healthcare. Cambridge, UK: Cambridge University
Kiyama, T., Tajiri, T., Yoshiyuki, T., Mitsuhashi, K., Ise, Y., Mizutani, T., ...
Hasegawa, S. (2003). Clinical significance of a standardized clinical pathway
in gastrectomy patients. Journal of Nippon Medical School, 70, 263-269.
Kollef, M. H., Shapiro, S. D., Silver, P., John, R. E., Prentice, D., Sauer, S., ...
Baker-Clinkscale, D. (1997). A randomized, controlled trial of protocol-
directedversus physician-directedweaningfrommechanical ventilation.Critical
Care Medicine, 25, 567-641.
Rotter et al. 25
Marelich, G. P., Murin, S., Battistella, F., Inciardi, J., Vierra, T., & Roby, M. (2000).
Protocol weaning of mechanical ventilation in medical and surgical patients by
respiratory care practitioners and nurses: Effect on weaning time and incidence
of ventilator-associated pneumonia. Chest, 118, 459-467.
Marrie, T. J., Lau, C. Y., Wheeler, S. L., Wong, C. J., Vandervoort, M. K., &
Feagan, B. G. (2000). A controlled trial of a critical pathway for treatment of
community-acquired pneumonia. CAPITAL Study Investigators. Community-
Acquired Pneumonia Intervention Trial Assessing Levofloxacin. Journal of the
American Medical Association, 283, 749-755.
Organisation for Economic Co-operation and Development. (2008). OECD-Health-
Data. Paris: Author.
Philbin, E. F., Rocco, T. A., Lindenmuth, N. W., Ulrich, K., McCall, M., &
Jenkins, P. L. (2000). The results of a randomized trial of a quality improve-
ment intervention in the care of patients with heart failure. The MISCHF
Study Investigators [see comments]. American Journal of Medicine, 109,
Review-Manager. (2008). RevMan (Version 5.0) [computer program]. Copenhagen:
The Nordic Cochrane Centre, The Cochrane Collaboration; John Wiley.
Roberts, R. R., Zalenski, R. J., Mensah, E. K., Rydman, R. J., Ciavarella, G.,
Gussow, L., ... Rao, R. (1997). Costs of an emergency department-based
accelerated diagnostic protocol vs hospitalization in patients with chest pain:
A randomized controlled trial [see comments]. Journal of the American Medical
Association, 278, 1670-1676.
Rotter, T., Kinsman, L., James, E., Machotta, A., Gothe, H., Willis, J., ... Kugler, J.
(2010). Clinical pathways: Effects on professional practice, patient outcomes,
length of stay and hospital costs. Cochrane Database of Systematic Reviews,
Saint,S., Hofer, T. P., Rose, J.S., Kaufman, S. R., &McMahon, L. F.Jr. (2003). Use
of critical pathways to improve efficiency: A cautionary tale. The American
Journal of Managed Care, 9, 758-765.
Shemilt, I (2009). CCEMG-EPPI-Centre Cost Converter; Version 1.0. The Campbell
and Cochrane Economics Methods Group (CCEMG) and the Evidence for Policy
and Practice Information and Coordinating Centre (EPPI-Centre). Retrieved
December 18, 2009, from http://eppi.ioe.ac.uk/costconversion/default.aspx
Shemilt, I., Thomas, J., & Morciano, M. (2010). A web-based tool for adjusting
costs to a specific target currency and price year. Evidence & Policy, 6, 51-59.
Smith, B. J., Cheok, F., Heard, A. R., Esterman, A. J., Southcott, A. M., Antic, R.,
... Ruffin, R. E. (2004). Impact on readmission rates and mortality of a
chronic obstructive pulmonary disease inpatient management guideline. Chronic
Respiratory Disease, 1, 17-28.
26Evaluation & the Health Professions 35(1)
Stroke-Unit-Trialists’-Collaboration. (2007). Organised inpatient (stroke unit) care Download full-text
for stroke. Cochrane Database of Systematic Reviews. Art. No.: CD000197.
Sulch, D., Melbourn, A., Perez, I., & Kalra, L. (2002). Integrated care pathways and
quality of life on a stroke rehabilitation unit. Stroke, 33, 1600-1604.
Sulch, D., Perez, I., Melbourn, A., & Kalra, L. (2000). Randomized controlled
trial of integrated (managed) care pathway for stroke rehabilitation. Stroke, 31,
Tilden, V. P., & Shepherd, P. (1987). Increasing the rate of identification of battered
women in an emergency department: Use of a nursing protocol. Research in
Nursing and Health, 10, 209-224.
Usui, K., Kage, H., Soda, M., Noda, H., & Ishihara, T. (2004). Electronic clinical
pathway for community acquired pneumonia (e-CP CAP). Nihon Kokyuki
Gakkai zasshi ¼ The Journal of the Japanese Respiratory Society, 42, 620-624.
Rotter et al. 27