DOI 10.
1515/dx-2014-0007      Diagnosis 2014; 1(2): 173–181
Original Article
Emilie S. Powell*, Lanty M. O’Connor, Anna P. Nannicelli, Lisa T. Barker, Rahul K. Khare,
Nicholas P. Seivert, Jane L. Holl and John A. Vozenilek
Failure mode effects and criticality analysis:
innovative risk assessment to identify critical
areas for improvement in emergency department
sepsis resuscitation
Abstract
Background: Sepsis is an increasing problem in the prac-
tice of emergency medicine as the prevalence is increasing
and optimal care to reduce mortality requires significant
resources and time. Evidence-based septic shock resuscita-
tion strategies exist, and rely on appropriate recognition and
diagnosis, but variation in adherence to the recommenda-
tions and therefore outcomes remains. Our objective was to
perform a multi-institutional prospective risk-assessment,
using failure mode effects and criticality analysis (FMECA),
to identify high-risk failures in ED sepsis resuscitation.
Methods: We conducted a FMECA, which prospectively
identifies critical areas for improvement in systems and
processes of care, across three diverse hospitals. A mul-
tidisciplinary group of participants described the process
of emergency department (ED) sepsis resuscitation to then
create a comprehensive map and table listing all process
steps and identified process failures. High-risk failures
in sepsis resuscitation from each of the institutions were
compiled to identify common high-risk failures.
Results: Common high-risk failures included limited avail-
ability of equipment to place the central venous catheter
and conduct invasive monitoring, and cognitive overload
leading to errors in decision-making. Additionally, we iden-
tified great variability in care processes across institutions.
Discussion: Several common high-risk failures in sepsis
care exist: a disparity in resources available across hospi-
tals, a lack of adherence to the invasive components of care,
and cognitive barriers that affect expert clinicians’ deci-
sion-making capabilities. Future work may concentrate on
dissemination of non-invasive alternatives and overcoming
cognitive barriers in diagnosis and knowledge translation.
Keywords: cognition; health services research; risk assess-
ment; sepsis.
©2014, Emilie S. Powell et al., published by De Gruyter.
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.
Open Access
*Corresponding author: Emilie S. Powell, MD, MS, MBA,
Northwestern University, Department of Emergency Medicine,
Feinberg School of Medicine, 211 E Ontario St, Suite 200, Chicago,
IL 60611, USA, Phone: +312-694-7000, Fax: +312-926-6274,
E-mail: emilie-powell@md.northwestern.edu
Emilie S. Powell and Rahul K. Khare: Northwestern University,
Department of Emergency Medicine, Feinberg School of Medicine,
Chicago, IL, USA; and Northwestern University, Center for
Healthcare Studies, Institute for Public Health in Medicine, Feinberg
School of Medicine, Chicago, IL, USA
Lanty M. O’Connor: Northwestern University, Center for Education
in Health Sciences, Institute for Public Health in Medicine, Feinberg
School of Medicine, Chicago, IL, USA
Anna P. Nannicelli, Nicholas P. Seivert and Jane L. Holl:
Northwestern University, Center for Healthcare Studies, Institute for
Public Health in Medicine, Feinberg School of Medicine, Chicago,
IL, USA
Lisa T. Barker and John A. Vozenilek: University of Illinois College
of Medicine at Peoria/OSF St. Francis Medical Center, Jump Trading
Simulation and Education Center, Peoria, IL, USA
Introduction
Sepsis, defined as an overwhelming body infection, is a
complex disease that includes a range of clinical condi-
tions. Sepsis begins with an infection and the body’s sys-
temic response to that infection (in the form of vital sign
abnormalities, the systemic inflammatory response syn-
drome, SIRS) and may progress quickly to severe sepsis,
sepsis with evidence of organ dysfunction, and septic
shock, severe sepsis with evidence of sepsis-induced hypo-
tension [1]. Sepsis affects more than 750,000 people in the
United States (US) each year [1, 2], carrying a 50%–60%
mortality rate when progressing to septic shock. Optimal
care for severe sepsis requires significant resources from
Unauthenticated
Download Date | 1/20/17 3:36 PM
174      Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation
the healthcare system as a whole, including prolonged
intensive care unit stays with complex treatments including
mechanical ventilation, hemodynamic management with
vasopressors and invasive monitoring, and possibly hemo-
dialysis [3]. Costs for severe sepsis have increased faster
than any other cause for hospitalization, with an annual
estimated cost of $54 billion [4]. The majority of patients
with sepsis are identified and initially resuscitated in the
emergency department (ED), making the ED the ideal place
for interventions to improve care and reduce patient mor-
tality [2, 5–7].
Sepsis is one of the most challenging, complex, and
time-sensitive disease states that clinicians will encoun-
ter. Similar to other critical diagnoses such as acute
myocardial infarction and stroke, the speed and appro-
priateness with which a clinical team is able to deliver
care can influence outcomes. But sepsis presents unique
challenges because it presents as a continuum of signs
and symptoms and, therefore, is a healthcare quality and
patient safety challenge to treating clinicians who must
recognize sepsis at its multiple stages, share their impres-
sions, and decide to act.
Evidence-based septic shock resuscitation strat-
egies exist, predominantly in the form of the sepsis
resuscitation bundle (Bundle) initiated in 2004 by the
Institute for Healthcare Improvement and the Surviv-
ing Sepsis Campaign, which has been shown to reduce
mortality by over 25% if applied within the first six
hours of presentation [3, 8–11]. The Bundle is based on
an “early-goal-directed care” strategy that was origi-
nally published in 2001, and was combined with other
evidence-based strategies, including adequate fluids
resuscitation and early antibiotic delivery. The inten-
tion is that physicians perform all Bundle tasks 100%
of the time in patients with severe sepsis or septic shock
[9, 10]. Adherence to the Bundle requires appropriate
recognition and diagnosis as the disease progresses
from sepsis to severe sepsis to septic shock. The origi-
nal Bundle was produced nearly 10 years ago. Although
the Bundle is evidence-based and widely disseminated,
variation in adherence to the recommendations still
remain, and this parallels variation in clinical outcomes
for severely septic patients [12, 13]. While some improve-
ment in compliance has been shown with traditional
education, alerts, the formation of collaboratives, and
sepsis response teams, compliance across the nation
remains low. Documented rates of Bundle compliance
remain as low as 50%, despite intervention [13–18].
To date, we are unaware of a published report docu-
menting the processes and systems of care that contrib-
ute to the persistently low Bundle compliance rates in
the ED. The objective of this investigation was to perform
a multi-institutional prospective risk-assessment, using
failure mode effects and criticality analysis (FMECA), to
identify high-risk critical barriers to 100% Bundle com-
pliance in ED severe sepsis and septic shock (sepsis)
resuscitation.
Materials and methods
Study design
We conducted a qualitative analysis, based on techniques of
grounded theory (i.e., there is no preconceived hypotheses, but
rather a continual comparative analysis of the data leads to the gen-
eration of theories and results) [19], of severe sepsis and septic shock
resuscitation in the ED setting. Specifically, we employed a special-
ized qualitative risk assessment approach: failure mode effects and
criticality analysis (FMECA), which prospectively identifies barriers
and critical areas for improvement in systems and processes [20,
21]. It is a tool initially described by the aviation and nuclear power
industries to determine the magnitude of adverse consequences and
events as well as the likelihood of the occurrence of these events
[14, 15]. Traditionally, FMECA has been used in the investigation of
a binary operational event (i.e., widget works vs. widget does not
work), but has been increasingly applied to complex healthcare
process such as intravenous drug administration, blood transfu-
sion, sterilization of surgical instruments, and acute ST-elevation
myocardial infarction “door-to-balloon” time in the ED [22–33]. This
approach provides benefits over traditional qualitative methodolo-
gies, such as those employing focus groups and surveys, in that we
are able to prospectively: (1) identify all action steps in a process of
care; (2) investigate potential failures at each process action step; (3)
evaluate the frequency and potential severity of harm of each fail-
ure to assign a criticality value; and (4) assess how each failure may
affect overall system performance [34, 35]. However, this analysis
uses a novel extension of the FMECA approach to evaluate a process
that focuses more on cognitive processes as opposed to institutional
processes. For example, we sought to examine how clinicians decide
that a patient is in septic shock and the resultant treatment decisions
as opposed to examining how a medication is administered, instru-
ments are sterilized, or the process of critical laboratory value notifi-
cation [23, 25, 32]. Institutional Review Board approval was obtained
at each of the study institutions’ affiliated IRBs.
Study setting and population
FMECAs were conducted at three hospitals with diverse institutional
characteristics. The first site is an urban academic teaching hospi-
tal with 876 inpatient beds. The ED consists of 54 beds and has an
annual census of over 85,000 ED visits with approximately 6000
intensive care unit admissions annually. Critically ill patients are
seen in the main ED, which is staffed by two to four emergency medi-
cine (EM) attending physicians, four to six EM and rotating resident
physicians per EM attending, and 14 to 27 nurses. The second site,
a tertiary care referral hospital, is a 616-bed academic tertiary care
Unauthenticated
Download Date | 1/20/17 3:36 PM
 Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation      175
referral center and serves a catchment area that covers 42 hospitals
within a 100-mile radius. The ED consists of 56 beds that accommo-
date approximately 85,000 patient visits per year with 1800 annual
intensive care unit (ICU) admissions through the ED. The ED is staffed
with up to six EM attending physicians, one to two EM and rotating
resident physicians per EM attending and eight to 26 nurses. The
third site is a rural hospital with an ED that has approximately 14,000
patient visits per year, is staffed by one attending physician, and the
hospital has limited on-site critical care capabilities. The majority of
patients presenting to the ED with severe sepsis or septic shock are
transferred to other facilities. Patients admitted to the on-site ICU
are managed during off-hours through a telemedicine service that is
staffed remotely by an intensivist.
Between November 2011 and May 2012, six FMECA sessions,
consisting of two meetings per hospital, were conducted across the
three study institutions. Eligible participants (i.e., providers involved
in ED sepsis resuscitation), were identified by site leadership and
then contacted by the research team to participate in the FMECA
being held at their respective institution. In total, we assembled a
convenience sample of 32 clinicians and staff (Urban Academic
Teaching Hospital: n = 13; Tertiary Care Referral Hospital: n = 10; Rural
Hospital: n = 9). Sessions were attended by a multidisciplinary group
of healthcare providers including EM physicians (i.e., attending
physicians and residents based on institutional staffing), ED nurses
and managers, patient safety officers, laboratory staff, and hospital
pharmacists, which helped ensure that a comprehensive description
of all process steps was elicited. Each session lasted approximately
90 min; all meetings were scheduled up to four weeks in advance
to accommodate department schedules. Participants received mod-
est remuneration in the form of a gift card for each session attended.
Verbal consent was obtained from all participants prior to the start of
the first FMECA session.
Study protocol
We adapted the FMECA process from that described previously by
Khare et  al. in the process of evaluating an institution’s door-to-
balloon process of care for patients presenting to the ED with ST-
elevation myocardial infarction [28]. For each participating site, the
boundaries of the FMECA were set to the sepsis resuscitation bundle
[9] and the discussion was limited to the care of patients with sepsis
who proceed to be diagnosed with severe sepsis and septic shock in
the ED from arrival in patient room to disposition from the ED. Triage
processes were excluded from the discussion. The FMECA sessions
were facilitated by study investigators and the process documented
by at least two members of the study team using sticky notes. To
ensure accurate representation of the process as described by the
participants, all sessions were audio-recorded and two members of
the research team took field notes at each session.
In the first session at each site, the facilitator began the dis-
cussion with a hypothetical patient presenting to the ED with vital
signs indicative of severe sepsis which progresses to septic shock
and concluded with completion of the Bundle or transfer out of the
ED. The surviving sepsis bundle was reviewed and provided to the
participants [9]. Participants were then asked to further describe the
process by which they care for patients with severe sepsis or septic
shock. The study team then translated the described processes into
formal Process Maps (Microsoft Visio 2010; Microsoft Corporation,
Redmond, WA, USA). Process Maps lay out, step-by-step all steps
and decision points in a process (see Supplemental Data, Appen-
dix 1, which accompanies the article at http://www.degruyter.com/
view/j/dx.2014.1.issue-2/issue-files/dx.2014.1.issue-2.xml). Prior to
the second session, FMECA participants were asked to review their
respective Process Map and offer any revisions. The final Process Map
for each site was then translated into a Risk Assessment Chart: an
Excel document listing all steps and decision points in the ED Sepsis
Resuscitation process (Microsoft Excel 2011; Microsoft Corporation,
Redmond, WA, USA).
During the second session, FMECA participants (1) systemati-
cally reviewed each process step and identified potential failures and
their causes; (2) ranked the frequency of each identified potential
failure (range: F1 [low] to F4 [high]); (3) outlined the consequences
of each potential failure [range: C0 (none) to C4 (certain)]; and (4)
discussed institutional safeguards already in place. Following the
two FMECA sessions, researchers met with individual participants
as needed to fill in any gaps in the Process Map or Risk Assessment
Chart. Results were used to develop a finalized Risk Assessment
Chart, which included all identified potential failure points (see
Supplemental Data, Appendices 1 and 2 for Process Maps and Risk
Assessment Chart).
Key outcome measures
The key outcome measures of this study were the identification of
common high-risk critical failures to 100% Bundle compliance across
the three institutions. Secondary outcomes included the identifica-
tion of the causes of each of these high-risk critical failures across the
three institutions.
Data analysis
To then identify the most critical systematic and process vulnerabili-
ties (high-risk critical failures), the research team used an established
method from the US Department of Energy, adapted for healthcare,
which categorizes each failure’s frequency and consequence scores
into different “Risk Bins” [36, 37]. Identified failure points were given
a criticality score of high, medium, or low based on the frequency
and consequence of the failure. Process steps were identified as high
risk if the step was either missed frequently or the consequence was
severe if missed. High-risk steps in Bundle compliance from each of
the three institutions were then compiled to identify common high-
risk steps and causes of compliance failures across the three institu-
tions.
Results
The results of this multi-institutional FMECA study of
ED sepsis resuscitation revealed several common high-
risk failure steps across all the three participating sites
(Figure 1). Six out of the eight identified high-risk failure
steps are central to clinician recognition, reassessment,
and the decision to proceed with the Bundle. Across all
participating sites, many of the identified failure causes
Unauthenticated
Download Date | 1/20/17 3:36 PM
176      Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation
were related to cognitive barriers or errors in decision-
making (Table 1). As shown in Table 1, clinicians’ cited
reasons for not pursuing treatment of severe sepsis and
septic shock were the same as for the decision not to
place a central venous catheter. In particular, physicians
discussed that they often do not choose to implement
the Bundle because they overanalyze the patient’s vital
signs, become convinced that they are not in septic shock
(e.g., “The patient does not have persistent hypotension
to indicate septic shock, their blood pressure always runs
low”) and therefore do not initiate aggressive care. Across
all sites, physicians cited a lack of consistent, immediate
availability of equipment necessary to place the central
venous catheter. The rural center reported that they did
not have access to invasive monitoring techniques, central
venous pressure (CVP) monitoring, or central venous
oxygen saturation (ScVO2) lab testing and therefore were
unable to complete the Bundle. Physicians also men-
tioned distractors including other critical patients and
high patient volumes as barriers to reassessing the septic
patient, leading to delays in their identification and inter-
vention. There were multiple other high-risk failure steps
at each individual institution, and many present at two
of the three institutions, that are not shown in Figure 1 or
Table 1. These steps include physician orders (laboratories
and IV fluids), additional reassessment steps early in the
sepsis process (i.e., during initial IV fluid administration),
the pharmacy process (including physician ordering and
pharmacy verification and delivery of antibiotics), and
nurse administration of vasopressors (data not shown).
Discussion
The results of this multidisciplinary, multi-institutional
prospective risk assessment identified several common
high-risk failures and causes of sub-optimal Bundle
Common high-risk steps across participating sites:
1. ED RN gives 2 L of IV fluid
2. ED team decision to pursue severe sepsis/septic shock
treatment
3. ED MD decision to place a central venous catheter
4. ED MD reassessment of patient and decision to give more
fluids or initiate vasopressors to treat shock
5. ED MD orders vasopressors
6. ED MD decision to order CVP monitoring
7. ED MD and RN reassessment of patient with CVP, BP, repeat
lactate
8. ED MD decision to order ScVO2 and treats the patient based
on the result
Figure 1 Common high-risk steps across participating sites.
compliance across the three institutions. Barriers to clini-
cal guideline adherence, such as the Bundle, have been
previously described through a framework of knowledge
(familiarity, awareness, cognitive biases/dilemmas), atti-
tudes (disagreements about outcomes, efficacy, or lack
of motivation/inertia of prior practice), and behaviors
(external barriers, environmental factors) [38]. Our find-
ings indicate barriers to successful Bundle compliance fall
into each of these categories: including limited availabil-
ity of equipment to place the central venous catheter and
conduct invasive monitoring, and cognitive overload and
biases leading to errors in decision-making by clinicians.
Additionally, we identified great variability in process of
care across the hospitals. It is important to note that the
goal of this work was not to quantify Bundle compliance
at various institutions, but rather to identify common
risks for sub-optimal Bundle compliance.
A number of previous studies have examined imple-
mentation of sepsis protocols and several have exam-
ined factors related to compliance with these protocols.
These studies all document low compliance rates with
the Bundle across institutions [13, 14, 16, 39–41]. Similar
to previous studies, we demonstrate failures associated
with placing the central venous catheter, and obtaining
and meeting goals of care through CVP and ScVO2 through
issues in the demands of nursing care and availability of
equipment, highlighting behavioral barriers [13, 39, 42].
Our results also highlight the decision to pursue severe
sepsis and septic shock treatment as a high-risk step,
highlighting knowledge and attitude barriers, as other
studies have. In addition, other studies report failure to
consider sepsis and initiate the Bundle at all, but this was
not considered in our study given that the process was
bound to patients with sepsis who proceed to have severe
sepsis or septic shock and clinicians were told to consider
sepsis at the start of the process [14, 39].
The FMECA results highlight institutional variations
in care processes and resource availability across the three
sites. For example, clinicians at the rural site endorse that
they only rarely place a central venous catheter in the ED,
a step in the Bundle that is necessary to deliver life-saving
vasopressor medication [3]. In this instance, a patient
with severe sepsis and septic shock is usually transferred
to an outside hospital for critical management and there is
a common belief that the patient will receive the necessary
aggressive management at the receiving institution (i.e.,
central venous catheter placement and vasopressor medi-
cations) and therefore does not immediately require the
placement of a central venous catheter at the rural site.
But, the transfer process often takes several hours and
places the patient out of the Bundle-prescribed six-hour
Unauthenticated
Download Date | 1/20/17 3:36 PM
 Table 1 Causes of common high-risk process step failures present at all three institutions.
High-risk process step failures  High-risk process step failure causes
ED RN does not give patient  
2L of IVF
ED team does not decide  
to pursue severe sepsis  
treatment
ED MDs do not decide to place
central venous catheter
ED MD does not order CVP  
monitoring
ED MDs do not decide to  
initiate vasopressors for
septic shock (and therefore
does not order them)
ED MD and RN do not reassess   – Difficulty interpreting CVP. False sense   – CVP Frequently Omitted.
and record patient CVP/BP, or
order repeat lactate (non-
invasive management)
ED MD does not order ScVO2  
Download Date | 1/20/17 3:36 PM
Unauthenticated
Urban Academic Hospital  
– RN not aware of Bundle.  
– RN has competing interests.
– Poor communication.
– Patient does not appear that ill.  
– The MD is hesitant to “pull the trigger”   – MD prefers more conservative treatment.   – MD has competing interests
and increase resource consumption.
– MD hesitant to go “overboard”.
– MDs are focused elsewhere in the ED
and wants patient to be managed in
ICU.
– MD does not believe in CVP monitoring   Process Step Usually Omitted: CVP
and it is not prioritized
– Lack of communication.
– CVP not charted or communicated.
– Lack of reassessment.  
– ED MD hesitant to order: patient’s
blood pressure “runs low.”
– Difficulty interpreting CVP.
– ED MD busy with other patient care.
of security for patient’s stability.
– Busy with other patients. Lack of
communication.
– Competing patient priorities.
– Patient away from ED for testing.
– ScVO2 not prioritized. Limited time  
(patient being transferred to ICU).
– Hesitancy to engage additional
resources.
Tertiary Care Referral Hospital  
– RN not aware of Bundle.  
– RN has competing interests.
– Poor communication.
– Patient does not appear that ill.  
– The MD is hesitant to “pull the trigger”
and increase resource consumption.
– MD hesitant to go “overboard.”
– MDs attention focused elsewhere, wants
patient to be managed in ICU.
 
Monitoring performed in ICU
– Lack of reassessment.  
– ED MD hesitant to order: patient’s blood
pressure “runs low.”
– Does not want to pursue aggressive
treatment.
– Believe that vasopressors will be
initiated in the ICU.
 
– False sense of security for patient’s
stability.
– Busy with other patients.
 Lack of communication.
– Competing patient priorities.
– Patient away from ED for testing.
Process Step Usually Omitted:  
Ordering ScVO2
  Cognitive barriers
Rural Hospital
– RN not aware of Bundle.  
– RN has competing interests.
– Poor communication.
– Patient does not appear that ill.   0
  0
given that it is a single coverage
department.
– MD is hesitant to place central
venous catheter.
– Most patients do not receive central
venous catheter in the ED.
– CVP monitoring equipment not  
available.
– Central venous catheter not placed.   0
– Lack of reassessment.
– ED MD hesitant to order: patient’s
blood pressure “runs low.”
– RN and MD distracted by other tasks.
– RN unclear on how to administer
vasopressors.
– CVP monitoring not available.   0
– False sense of security for patient’s
stability.
– RN busy with other tasks.
– Lack of communication.
– Lack of RN experience.
– Patient away from ED unmonitored.
– ScVO2 testing not available.  
Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation      177
178      Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation
window [3]. Additionally, if the patient is admitted to the
rural hospital ICU, there is often not an intensivist on-site
to place the central venous catheter to adhere to Bundle
compliance, making central venous catheter placement
in the ED even more important. In further exploring the
issue of invasive treatment and monitoring as a part of
the Bundle, the FMECA not only identified limited avail-
ability of invasive monitoring equipment at rural hospi-
tals but also minimal trust in CVP monitoring, hesitancy
to engage additional resources, and hesitancy to order the
ScVO2 across all institutions. These findings touch upon
knowledge, attitudinal, and behavioral barriers [38].
Given that all hospitals do not have the ability to meet
100% Bundle compliance, especially regarding the inva-
sive CVP and ScVO2 monitoring components, and that the
literature has questioned the value of CVP monitoring [43,
44], research in implementing non-invasive alternatives
for diagnosis in early goal directed therapy is certainly
warranted, and three multi-national trials are on-going
[45]. Non-invasive alternatives to measure CVP and ScVO2
were introduced in the latest 2012 Surviving Sepsis Cam-
paign Guidelines [3] (e.g., the use of ultrasound assess-
ment of inferior vena cava collapse correlates with CVP
[46] and clearance of serum lactate by 10% over two hours)
[47], however, the basic tenants of the Bundle remain the
same: early antibiotics and hemodynamic optimization
(still with an invasive central venous catheter if neces-
sary) [48]. Given these new advances in the management
of severe sepsis and septic shock resuscitation, it can
be argued that the same outcomes of the Bundle could
be achieved without achieving CVP and ScVO2 goals in
particular, if non-invasive goals are used and the other
Bundle elements are still met.
The majority of identified high-risk failure causes
were related to knowledge and attitudes or cognitive bar-
riers that clinicians experience in caring for patients with
sepsis, severe sepsis, and septic shock that impede them
from pursuing aggressive, invasive care. These cognitive
barriers have been described as failed heuristics, biases,
or cognitive dispositions to respond, that clinicians use to
form judgments and make decisions and can often lead to
medical errors [49]. The clinical environment, especially
the ED, is primed for cognitive impairment as there are
many barriers to effectively translating the knowledge of
the Bundle into practice: time pressure and constraints,
scattered information across multiple providers, shift
changes, a chaotic working environment, and clinician
bias (e.g., anchor bias, reliance on past experiential knowl-
edge) [50–53]. Furthermore, since humans do not always
accurately perceive and interpret information from their
environment, we must rely on biases and other cognitive
mechanisms to prioritize and filter information [51]. For
example, FMECA participants cited that despite vital signs
and laboratory values indicative of septic shock, they may
not diagnose the patient with septic shock if the patient
initially appeared well (Table 1). As a result, the sepsis
Bundle is not initiated. This is an example of the cognitive
error of “anchoring”: the tendency to lock onto features in
a patient’s initial presentation and not alter this percep-
tion despite evidence to the contrary [49]. This example
demonstrates issues with expert cognition: expert clini-
cians are required to work quickly and balance many dif-
ferent inputs and demands and therefore do not always
have the time to fully and logically consider all options.
They rely on experiential knowledge or the representative
heuristic (i.e., ‘the patient doesn’t look sick, therefore he/
she must not be in shock’) [49]. In the case of sepsis, oper-
ating under this experiential knowledge may result in the
clinician’s failure to initiate the Bundle, despite evidential
vital signs, because they do not appear ‘that ill’.
While this study focused on the diagnosis and care
of patients with sepsis who proceed to be diagnosed with
severe sepsis and septic shock in the ED from arrival in
patient room to disposition from the ED, and excluded
initial diagnostic steps made through the triage process,
we anticipate that similar high-risk failures would
be found, specifically focusing on cognitive barriers,
biases, and barriers to reassessment. Upon arrival to the
ED, triage nurses must act with limited knowledge and
only have access to chief complaint, initial vital signs,
and the physical appearance of the patient to make the
initial diagnosis of “too sick to wait” or “well enough
to wait”. The triage nurse is likely to rely more heavily
on past experience and be more vulnerable to cognitive
biases. Furthermore, as EDs become more crowded, and
patients are required to wait longer for transfer back to
an ED care space, reassessment for change in clinical
condition becomes more important, and more difficult,
leading clinicians in the waiting room to anchor on
initial presentation more.
Future interventions to improve sepsis care across
the country should be performed both on the individual
hospital level to improve equipment availability and on
the national level to increase research on non-invasive
alternatives to invasive Bundle monitoring and dissemi-
nate this knowledge. Future educational and operational
interventions might also incorporate an understanding
of expert clinician cognition – its capabilities and its
limitations – in an effort to guide clinicians’ decision
making. Cognitive debiasing strategies can be useful to
improve knowledge retention and sustain Bundle com-
pliance [49], but are often very complex to employ [54].
Unauthenticated
Download Date | 1/20/17 3:36 PM
 Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation      179
One alternative is to use simulation, perhaps counter
intuitively, to distort reality in order to better guide the
cognitive system through educational interventions.
Another alternative is to deploy robust decision support
systems to guide clinicians by indicating the critical
nature of multiple data points that exist with a patient’s
record [55].
Limitations
There are limitations inherent to the FMECA approach.
Although we selected institutions that may be repre-
sentative of the spectrum of hospitals across the US, the
identified high-risk barriers and causes may be institu-
tion-specific and others may experience alternate barriers
and causes. Further, the methodology relies on conveni-
ence sampling and cannot involve all clinicians involved
in each institution’s ED sepsis resuscitation process. It is
possible that if additional clinicians participated in the
FMECA sessions that other process steps, barriers, and
causes would have been revealed. It is also possible that
participants may have inaccurately gauged the frequency
and consequences of the identified potential failures.
Additionally, this study begins with the assumption that
clinicians recognize the possibility of sepsis in this case
and then have knowledge of and use the Bundle (as we
presented participants with the surviving sepsis bundle
prior to the FMECA). Clinicians were also told that the
patient would proceed to develop severe sepsis and then
septic shock. It is possible that additional failures exist
surrounding recognition of sepsis at all in clinical prac-
tice and in basic knowledge of the Bundle. To mitigate this
limitation, we did ask clinicians to consider all process
steps and possible failures in the diagnosis and treatment
of severe sepsis and then septic shock, including knowl-
edge of the Bundle components. An additional limitation
relates to the methodology’s development of individual
consequence scores and subsequent criticality rating (i.e.,
risk binning). Common issues include participants being
over-optimistic and tending to underestimate the potential
consequences since the failure is perceived as consider-
ably frequent and no severe consequences have actually
References
1. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D,
et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International
Sepsis Definitions Conference. Intens Care Med 2003;29:
530–8.
occurred; or being over-pessimistic, wherein although
the participants have never experienced the failure, they
cannot exclude that the failure will ever occur [20, 28,
56]. Furthermore, the frequency and consequence scores
could be perceived differently by various participants
and when one participant might rate a failure as low fre-
quency, another could rate it as common. We attempted
to mitigate this failure by coming to a group consensus
regarding the majority of discussed failures.
Conclusions
This multidisciplinary, multi-institutional prospective risk
assessment revealed several common high-risk failures
to successful implementation of the sepsis resuscitation
bundle: a disparity in resources available across hospitals, a
lack of adherence to the invasive components of the Bundle,
and cognitive barriers that affect expert clinicians’ decision-
making capabilities as they proceed through the steps of the
Bundle. Future work may concentrate on dissemination of
non-invasive alternatives to the invasive Bundle protocol
and overcoming cognitive barriers in diagnosis and knowl-
edge translation of the evidence-based Bundle.
Conflict of interest statement
Authors’ conflict of interest disclosure: The authors
stated that there are no conflicts of interest regarding the
publication of this article. Research funding played no
role in the study design; in the collection, analysis, and
interpretation of data; in the writing of the report; or in the
decision to submit the report for publication.
Research funding: Dr. Powell was supported by grant
F-32 HS 020766-01 from AHRQ and the Emergency Medi-
cine Foundation. Dr. Khare was supported by grant
K08HS019005–01 from AHRQ.
Employment or leadership: None declared.
Honorarium: None declared
Received February 7, 2014; accepted March 24, 2014; previously
published online April 12, 2014
2. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J,
Pinsky MR. Epidemiology of severe sepsis in the United States:
analysis of incidence, outcome, and associated costs of care. Crit
Care Med 2001;29:1303–10.
Unauthenticated
Download Date | 1/20/17 3:36 PM
180      Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation
3. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal
SM, et al. Surviving Sepsis Campaign: International Guidelines
for Management of Severe Sepsis and Septic Shock: 2012. Crit
Care Med 2013;41:580–637.
4. Friedman B, Henke RM, Wier LM. Most Expensive Hospitaliza-
tions, 2008. Agency for Healthcare Research and Quality:
Healthcare Cost and Utilization Project, 2010.
5. Niska R, Bhuiya F, Xu J. National Hospital Ambulatory Medical
Care Survey: 2007 emergency department summary. Nat Health
Stat Rep 2010:1.
6. Reed K, May R. HealthGrades Emergency Medicine in American
Hospitals Study. Golden, Colorado: Health Grades, Inc, 2010.
7. Wang HE, Shapiro NI, Angus DC, Yealy DM. National estimates
of severe sepsis in United States emergency departments. Crit
Care Med 2007;35:1928–36.
8. Levy MM, Dellinger RP, Townsend SR, Linde-Zwirble WT, ­Marshall
JC, Bion J, et al. The Surviving Sepsis Campaign: results of
an international guideline-based performance improvement
program targeting severe sepsis. Intens Care Med 2010;36:
222–31.
9. The Surviving Sepsis Campaign and Institute for Healthcare
Improvement. Sepsis Bundles. [December 30, 2013]. Available
from: http://www.ihi.org/IHI/topics/CriticalCare/Sepsis.
10. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R,
et al. Surviving Sepsis Campaign: International guidelines for
management of severe sepsis and septic shock: 2008. Intens
Care Med 2008;34:17–60.
11. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B,
et al. Early goal-directed therapy in the treatment of severe
sepsis and septic shock. N Engl J Med 2001;345:1368–77.
12. Powell ES, Khare RK, Courtney DM, Feinglass J. Volume of emer-
gency department admissions for sepsis is related to inpatient
mortality: results of a Nationwide Cross-Sectional Analysis. Crit
Care Med 2010;38:2161–8.
13. Mikkelsen ME, Gaieski DF, Goyal M, Miltiades AN, Munson JC,
Pines JM, et al. Factors associated with nonadherence to early
goal-directed therapy in the ED. Chest 2010;138:551–8.
14. Casserly B, Baram M, Walsh P, Sucov A, Ward NS, Levy MM.
Implementing a collaborative protocol in a sepsis intervention
program: lessons learned. Lung 2011;189:11–9.
15. Cannon CM, Holthaus CV, Zubrow MT, Posa P, Gunaga S, Kella V,
et al. The GENESIS project (GENeralized Early Sepsis Interven-
tion Strategies): a multicenter quality improvement collabora-
tive. J Intens Care Med 2013;28:355–68.
16. Ferrer R, Artigas A, Levy MM, Blanco J, Gonzalez-Diaz G,
Garnacho-Montero J, et al. Improvement in process of care and
outcome after a multicenter severe sepsis educational program
in Spain. J Am Med Assoc 2008;299:2294–303.
17. Larosa JA, Ahmad N, Feinberg M, Shah M, Dibrienza R, Studer S.
The use of an early alert system to improve compliance with
sepsis bundles and to assess impact on mortality. Crit Care Res
Pract 2012;2012:980369.
18. Nelson JL, Smith BL, Jared JD, Younger JG. Prospective trial of
real-time electronic surveillance to expedite early care of severe
sepsis. Ann Emerg Med 2011;57:500–4.
19. Glaser BG, Strauss AL. The discovery of grounded theory: strate-
gies for qualitative research. New Jersey: Transaction Publish-
ers, 1977.
20. DeRosier J, Stalhandske E, Bagian JP, Nudell T. Using health
care failure mode and effect analysis: the VA National Center
for Patient Safety’s prospective risk analysis system. Jt Comm J
Qual Improv 2002;28:248–67, 09.
21. Latino RJ, Flood A. Optimizing FMEA and RCA efforts in health
care. J Health Risk Manag 2004;24:21–8.
22. Esmail R, Cummings C, Dersch D, Duchscherer G, Glowa J, Lig-
gett G, et al. Using healthcare failure mode and effect analysis
tool to review the process of ordering and administrating
potassium chloride and potassium phosphate. Healthcare Quart
(Toronto, Ont) 2005;8 Spec No:73–80.
23. Kozakiewicz JM, Benis LJ, Fisher SM, Marseglia JB. Safe chemo-
therapy administration: using failure mode and effects analysis
in computerized prescriber order entry. Am J Health Syst Pharm
2005;62:1813–6.
24. Kunac DL, Reith DM. Identification of priorities for
medication safety in neonatal intensive care. Drug Saf
2005;28:251–61.
25. Saxena S, Kempf R, Wilcox S, Shulman IA, Wong L, Cunningham
G, et al. Critical laboratory value notification: a failure mode
effects and criticality analysis. Jt Comm J Qual Patient Saf
2005;31:495–506.
26. Steinberger DM, Douglas SV, Kirschbaum MS. Use of fail-
ure mode and effects analysis for proactive identification
of c­ ommunication and handoff failures from organ procure-
ment to transplantation. Prog Transplant (Aliso Viejo, Calif)
2009;19:208–14; quiz 15.
27. van Tilburg CM, Leistikow IP, Rademaker CM, Bierings MB, van
Dijk AT. Health care failure mode and effect analysis: a useful
proactive risk analysis in a pediatric oncology ward. Qual Saf
Health Care 2006;15:58–63.
28. Khare RK, Nannicelli AP, Powell ES, Seivert NP, Adams JG, Holl
JL. Use of risk assessment analysis by failure mode, effects,
and criticality to reduce door-to-balloon time. Ann Emerg Med
2013;62:388–98.e12.
29. Burgmeier J. Failure mode and effect analysis: an applica-
tion in reducing blood transfusion. Jt Comm J Qual Patient Saf
2002;28:331–9.
30. Apkon M, Leonard J, Probst L, DeLizio L, Vitale R. Design of a
safer approach to intravenous drug infusions: failure mode
effects analysis. Qual Saf Health Care 2004;13:265–71.
31. Coles G, Fuller B, Nordquist K, Kongslie A. Using failure mode
effects and criticality analysis for high-risk process at three
community hospitals. J Qual Patient Saf 2005;31:
132–40.
32. Linkin DR, Sausman C, Santos L, Lyons C, Fox C, Aumiller L,
et al. Applicability of healthcare failure mode and effects analy-
sis to healthcare epidemiology: evaluation of the sterilization
and use of surgical instruments. Clin Infect Dis 2005;41:1014–
9.
33. Robinson DL, Heigham M, Clark J. Using failure mode and
effects analysis for safe administration of chemotherapy to
hospitalized children with cancer. Jt Comm J Qual Patient Saf
2006;32:161–6.
34. Hollnagel E. Cognitive reliability and error analysis method:
CREAM, 1st ed. Oxford, New York: Elsevier, 1998.
35. Croskerry P. Patient safety in emergency medicine. Philadel-
phia, PA: Wolters Kluwer Health/Lippincott Williams & Wilkins,
2009.
36. DOE-STD-3009-94 U.S. Preparation Guide for U.S. Department
of Energy Nonreactor Nuclear Facility Documented Safety Analy-
ses. In: Energy Do, editor. Washington, DC, 2002.
Unauthenticated
Download Date | 1/20/17 3:36 PM
 Powell et al.: FMECA: identifying critical areas for improvement in ED sepsis resuscitation      181
37. HNF-8739. Hanford Safety Analysis and Risk Assessment
Handbook (SARAH) Fluor Hanford Inc Richland, Washington,
2003.
38. Cabana MD, Rand CS, Powe NR, Wu AW, Wilson MH, Abboud PA,
et al. Why don’t physicians follow clinical practice guidelines?
A framework for improvement. J Am Med Assoc 1999;282:1458–
65.
39. Carlbom DJ, Rubenfeld GD. Barriers to implementing protocol-
based sepsis resuscitation in the emergency department–
results of a national survey. Crit Care Med 2007;35:2525–32.
40. Castellanos-Ortega A, Suberviola B, Garcia-Astudillo LA,
Holanda MS, Ortiz F, Llorca J, et al. Impact of the Surviving
Sepsis Campaign protocols on hospital length of stay and
mortality in septic shock patients: results of a three-year
follow-up quasi-experimental study. Crit Care Med
2010;38:1036–43.
41. Gao F, Melody T, Daniels DF, Giles S, Fox S. The impact of
­compliance with 6-hour and 24-hour sepsis bundles on hospital
mortality in patients with severe sepsis: a prospective observa-
tional study. Crit Care 2005;9:R764–70.
42. Jones AE, Shapiro NI, Roshon M. Implementing early goal-
directed therapy in the emergency setting: the challenges and
experiences of translating research innovations into clinical
reality in academic and community settings. Acad Emerg Med
2007;14:1072–8.
43. Marik PE, Cavallazzi R. Does the central venous pressure predict
fluid responsiveness? An updated meta-analysis and a plea for
some common sense. Crit Care Med 2013;41:1774–81.
44. Marik PE, Baram M, Vahid B. Does central venous pressure pre-
dict fluid responsiveness? A systematic review of the literature
and the tale of seven mares. Chest 2008;134:172–8.
45. Huang DT, Angus DC, Barnato A, Gunn SR, Kellum JA, Staple-
ton DK, et al. Harmonizing international trials of early goal-
directed resuscitation for severe sepsis and septic shock:
methodology of ProCESS, ARISE, and ProMISe. Intens Care Med
2013;39:1760–75.
46. Haydar SA, Moore ET, Higgins GL, 3rd, Irish CB, Owens WB,
Strout TD. Effect of bedside ultrasonography on the certainty
of physician clinical decisionmaking for septic patients in the
emergency department. Ann Emerg Med 2012;60:346–58 e4.
47. Jones AE, Shapiro NI, Trzeciak S, Arnold RC, Claremont HA, Kline
JA. Lactate clearance vs central venous oxygen saturation as
goals of early sepsis therapy: a randomized clinical trial. J Am
Med Assoc 2010;303:739–46.
48. Non-invasive Sepsis Management 2011 [September 12, 2012].
Available from: emcrit.org/wp-content/uploads/2010/10/Non-
invasive-11-18-2011.pdf.
49. Croskerry P. The importance of cognitive errors in diagnosis and
strategies to minimize them. Acad Med 2003;78:775–80.
50. Croskerry P, Sinclair D. Emergency medicine: a practice prone to
error? CJEM 2001;3:271–6.
51. Dror I. A novel approach to minimize error in the medical
domain: cognitive neuroscientific insights into training. Med
Teach 2011;33:34–8.
52. Dror IE. The paradox of human expertise: why experts can get it
wrong. In: Kapur N, editor. The Paradoxical Brain. Cambridge,
UK: Cambridge University Press, 2011:177–88.
53. Chisholm CD, Collison EK, Nelson DR, Cordell WH. Emergency
department workplace interruptions: are emergency physi-
cians “interrupt-driven” and “multitasking”? Acad Emerg Med
2000;7:1239–43.
54. Croskerry P. From mindless to mindful practice–cognitive bias
and clinical decision making. N Engl J Med 2013;368:2445–8.
55. Arnott D. Cognitive biases and decision support systems
development: a design science approach. Inform Syst J
2006;16:55–78.
56. Pasquini A, Pozzi S, Save L. A critical view of severity classification
in risk assessment methods. Reliab Eng Syst Safe 2011;96:53–63.
Unauthenticated
Download Date | 1/20/17 3:36 PM