Analysis of the literature on emergency department throughput.
ABSTRACT The purpose of this paper was to review and analyze all the literature concerning ED patient throughput. The secondary goal was to determine if certain factors would significantly alter patients' ED throughput.
A MEDLINE search was performed from 1966 to 2007 using the terms "turnaround," "emergency departments," "emergency medicine," "efficiency," "throughput," "overcrowding" and "crowding." Studies were graded using a scale of one to four based on the ACEP paper quality criteria. Inclusion criteria were English language and at least a level four or better on the quality scale. An analysis of successful procedures and techniques was performed.
Literature search using the key terms found 29 articles on turnaround times, 129 on ED efficiency, 3 on throughput, 64 on overcrowding and 52 on crowding. Twenty-six articles were found to meet the inclusion criteria. There were three level I studies, thirteen level II studies, five level III studies and five level IV studies. The studies were categorized into five areas: determinants (7), laboratories processes (4), triage process (3), academic responsibilities (2), and techniques (10). Few papers used the same techniques or process to examine or reduce patient throughput precluding a meta-analysis.
An analysis of the literature was difficult because of varying study methodologies and less than ideal quality. EDs with combinations of low inpatient census, in-room registration, point of care testing and an urgent care area demonstrated increased patient throughput.
- [Show abstract] [Hide abstract]
ABSTRACT: Triage standing orders are used in emergency departments (EDs) to initiate evaluation when there is no bed available. This study evaluates the effect of diagnostic triage standing orders on ED treatment time of adult patients who presented with a chief complaint for which triage standing orders had been developed. We conducted a retrospective nested cohort study of patients treated in one academic ED between January 2007 and August 2009. In this ED, triage nurses can initiate full or partial triage standing orders for patients with chest pain, shortness of breath, abdominal pain, or genitourinary complaints. We matched patients who received triage standing orders to those who received room orders with respect to clinical and temporal factors, using a propensity score. We compared the median treatment time of patients with triage standing orders (partial or full) to those with room orders, using multivariate linear regression. Of the 15,188 eligible patients, 25% received full triage standing orders, 56% partial triage standing orders, and 19% room orders. The unadjusted median ED treatment time for patients who did not receive triage standing orders was 282 minutes versus 230 minutes for those who received a partial triage standing order or full triage standing orders (18% decrease). Controlling for other factors, triage standing orders were associated with a 16% reduction (95% confidence interval -18% to -13%) in the median treatment time, regardless of chief complaint. Diagnostic testing at triage was associated with a substantial reduction in ED treatment time for 4 common chief complaints. This intervention warrants further evaluation in other EDs and with different clinical conditions and tests.Annals of emergency medicine 02/2011; 57(2):89-99.e2. · 4.23 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Despite consensus regarding the conceptual foundation of crowding, and increasing research on factors and outcomes associated with crowding, there is no criterion standard measure of crowding. The objective was to conduct a systematic review of crowding measures and compare them in conceptual foundation and validity. This was a systematic, comprehensive review of four medical and health care citation databases to identify studies related to crowding in the emergency department (ED). Publications that "describe the theory, development, implementation, evaluation, or any other aspect of a 'crowding measurement/definition' instrument (qualitative or quantitative)" were included. A "measurement/definition" instrument is anything that assigns a value to the phenomenon of crowding in the ED. Data collected from papers meeting inclusion criteria were: study design, objective, crowding measure, and evidence of validity. All measures were categorized into five measure types (clinician opinion, input factors, throughput factors, output factors, and multidimensional scales). All measures were then indexed to six validation criteria (clinician opinion, ambulance diversion, left without being seen (LWBS), times to care, forecasting or predictions of future crowding, and other). There were 2,660 papers identified by databases; 46 of these papers met inclusion criteria, were original research studies, and were abstracted by reviewers. A total of 71 unique crowding measures were identified. The least commonly used type of crowding measure was clinician opinion, and the most commonly used were numerical counts (number or percentage) of patients and process times associated with patient care. Many measures had moderate to good correlation with validation criteria. Time intervals and patient counts are emerging as the most promising tools for measuring flow and nonflow (i.e., crowding), respectively. Standardized definitions of time intervals (flow) and numerical counts (nonflow) will assist with validation of these metrics across multiple sites and clarify which options emerge as the metrics of choice in this "crowded" field of measures.Academic Emergency Medicine 05/2011; 18(5):527-38. · 1.76 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: In 2006, the Institute of Medicine (IOM) advanced the concept of "coordinated, regionalized, and accountable emergency care systems" to address significant problems with the delivery of emergency medical care in the United States. Achieving this vision requires the thoughtful implementation of well-aligned, system-level structures and processes that enhance access to emergency care and improve patient outcomes at a sustainable cost. Currently, the delivery of emergency medical care is supported by numerous administrative systems, including economic; reimbursement; legal and regulatory structures; licensure, credentialing, and accreditation processes; medicolegal systems; and quality reporting mechanisms. In addition, many regionalized systems may not optimize patient outcomes because of current administrative barriers that make it difficult for providers to deliver the best care. However, certain administrative barriers may also threaten the sustainability of integration efforts or prevent them altogether. This article identifies significant administrative challenges to integrating networks of emergency care in four specific areas: reimbursement, medical-legal, quality reporting mechanisms, and regulatory aspects. The authors propose a research agenda for indentifying optimal approaches that support consistent access to quality emergency care with improved outcomes for patients, at a sustainable cost. Researching administrative challenges will involve careful examination of the numerous natural experiments in the recent past and will be crucial to understand the impact as we embark on a new era of health reform.Academic Emergency Medicine 12/2010; 17(12):1330-6. · 1.76 Impact Factor
Western Journal of Emergency Medicine 104 Volume X, no. 2 : May 2009
Leslie S. Zun, MD, MBA
Analysis of the Literature on Emergency
Rosalind Franklin University of Medicine and Science/Chicago Medical School,
Department of Emergency Medicine
Mount Sinai Hospital, Department of Emergency Medicine, Chicago, Illinois
Supervising Section Editor: Christopher A. Kahn, MD, MPH
Submission history: Submitted March 31, 2008; Revision Received September 15, 2008; Accepted November 24, 2008.
Reprints available through open access at www.westjem.org
Introduction: The purpose of this paper was to review and analyze all the literature concerning ED
patient throughput. The secondary goal was to determine if certain factors would significantly alter
patients’ ED throughput.
Methods: A MEDLINE search was performed from 1966 to 2007 using the terms “turnaround,”
“emergency departments,” “emergency medicine,” “efficiency,” “throughput,” “overcrowding” and
“crowding.” Studies were graded using a scale of one to four based on the ACEP paper quality
criteria. Inclusion criteria were English language and at least a level four or better on the quality scale.
An analysis of successful procedures and techniques was performed.
Results: Literature search using the key terms found 29 articles on turnaround times, 129 on ED
efficiency, 3 on throughput, 64 on overcrowding and 52 on crowding. Twenty-six articles were found
to meet the inclusion criteria. There were three level I studies, thirteen level II studies, five level III
studies and five level IV studies. The studies were categorized into five areas: determinants (7),
laboratories processes (4), triage process (3), academic responsibilities (2), and techniques (10). Few
papers used the same techniques or process to examine or reduce patient throughput precluding a
Conclusions: An analysis of the literature was difficult because of varying study methodologies and
less than ideal quality. EDs with combinations of low inpatient census, in-room registration, point of
care testing and an urgent care area demonstrated increased patient throughput.
Improving efficiency and throughput in the emergency
department (ED) has multiple benefits. Better efficiency
should increase patient satisfaction, enhance revenue and
reduce ambulance diversion. The need to focus on ED
efficiency has become more acute in recent years due to
increasing litigation, including a case where a patient in
Chicago died while waiting for care.1
EDs across the U.S. struggle to provide efficient care
in a timely fashion. Increasing patient volumes, a reduction
in the number of EDs, higher inpatient census and ED staff
reduction all exacerbate the struggle. The purpose of this
paper is to review the literature and summarize strategies used
nationwide to deal with this crisis. Proven techniques could be
used by hospital and ED managers.
We searched MEDLINE from 1966 to March 2007 for
English language articles using the keywords turnaround,
efficiency, throughput, overcrowding and crowding. No other
restrictions in the search fields were used. We also reviewed
references from these articles to ensure that we included all
We required one or more factors related to throughput to
Volume X, no. 2 : May 2009 105 Western Journal of Emergency Medicine
include an article in further analysis. We used a classification
system modified from the American College of Emergency
Physicians to assess the study’s methodology and quality.2
To be more inclusive in the review, a fourth parameter was
added to the classification system (Table 1). Each article
was graded one to four based on this classification scheme.
Those studies with confounding variables, problematic study
design, limited data or poor presentations were downgraded
to the next lower class. Due to the lack of uniformity and
consistency within the literature, studies of similar design and
technique could only be identified and grouped into five broad
categories: throughput determinates, academic responsibilities,
laboratories, triage, and techniques. A table of the findings was
produced to summarize the class, design, analysis, conclusion
and limitations of each study (Table 2).
The literature search using the keywords crossed with
“emergency departments” and “emergency medicine” (EM)
found 29 articles related to turnaround, 129 articles related
to efficiency, four articles related to throughput, 52 articles
related to crowding and 64 articles related to overcrowding.
Twenty-six articles were found to meet the inclusion
criteria. Studies that lacked data, had poor scientific design or
provided limited information were not rated. There were three
level I studies, 13 level II studies, five level III studies and
five level IV studies (Table 2). We then sorted them, using the
five broad categories throughput determinants (seven articles),
laboratories processes (four articles), academic responsibilities
(two articles), triage process (three articles), and throughput
reduction techniques (10 articles).
Several articles focused on the correlation between
throughput time and ED factors. The articles showed that ED
length of stay (LOS) increased substantially with increased
admissions, number of ambulance arrivals, number of
pediatric patients and ED census.3-8 Rathlev et al.8 found that
daily mean LOS was increased not only by number of ED
admission and hospital occupancy but also by elective surgical
admission. Interestingly, two of the studies did not find a
significant correlation between the throughput time and hours
of nursing coverage, day of the week or urgent care hours.3,4
Saunders et al.9 performed a computer simulation study of ED
operations and found that throughput times correlated directly
with laboratory service times and inversely with number of
physicians and nurses. This latter relationship had a ceiling
where a continued increase in providers demonstrated no
change in throughput time.9
Two studies examined the effect of teaching on ED
throughput. Chan et al.10 examined how medical students
affected ED throughput and found that fourth year medical
students’ precepting for four weeks in the ED did not change
the LOS for patients. A similar study11 looking at the effects
of adding EM residents found that the residents increased the
total throughput time an average of seven to 39 minutes.
In a study of 11 hospital EDs, Holland et al.12 found
that addressing the laboratory outliers rather than the mean
turnaround time can reduce the ED LOS. In a study of 690
hospital laboratories, Steindel and Howanitz13 found that
faster throughputs were related to lab control of the specimen
handling and rapid transport times. Murray et al.14 performed a
randomized controlled trial comparing point-of-care testing to
central laboratory testing and found that point-of-care testing
reduced the median stay by 54 minutes. Study supplies and
equipment but not grant funding were provided for this
potentially biased study. In a comparison study of the use
of a pneumatic tube delivery system versus human couriers,
Fernandes et al.15 found that a tube system reduced lab
reporting time by 8-10 minutes.
Partovi et al.16 compared the LOS in triage with and
without an emergency physician, and found an 18% reduction
in LOS when a physician began patient evaluation and
treatment in triage. This paper reported a significant cost
of physicians in triage, which may outweigh the benefits of
reduced LOS. Choi and Claudius17 studied the use of pulse
oximetry on bronchiolitis patients in triage and found that it
could reduce throughput by 50 minutes. The authors did not
Table 1. Literature classification schema*
1Randomized, controlled trial
Prospective cohort using standard metrics
Population prospective cohort
3Case series, case report, consensus Case series, case report, consensusCase series, case report, consensus
4Expert opinion, design flaws,
Expert opinion Expert opinion
* Adapted from: American College of Emergency Physicians/Physician Consortium: Emergency medicine physician performance measurement set.2
Zun Analysis of Literature on ED Throughput
Western Journal of Emergency Medicine 106 Volume X, no. 2 : May 2009
study the reason for decreased throughput time nor describe
how the evaluation and treatment may have been altered
with pulse oximetry measurement; however, they suggested
that identification of hypoxia changes management, and
proper patient placement to urgent care or main ED may have
accounted for this time reduction.
Multiple studies described techniques used to reduce ED
LOS. Spaite et al.18 examined one ED that employed a rapid
process redesign, and found that it led to a 76-minute reduction
in average patient LOS. The rapid redesign focused on staffing
and internal processes, triage and registration procedures and
diagnostic radiology, laboratory and bed availabilities. This
rapid improvement process occurred over three months and
cost the hospital over $1 million annually. This cost was offset
by increased revenue, providing a net annualized profit of
$300,000. Purnell et al.19 surveyed 185 hospitals and found that
an urgent care unit reduced patient wait times by 20%. This
limited survey study performed in 1989 found that the mean
wait time was 72 minutes in EDs with fast track and 90 minutes
for those without.
In a comparison of multiple interventions, Cardin et al.20
found that increased emergency physician (EP) coverage,
designation of a physician coordinator and changes in hospital
policies on laboratory, consultations and admission procedures
could reduce ED mean LOS from 13.8 hours to 5.9 hours. The
article focused on the effect of the interventions on return visits
and hospital readmissions and not on the interventions used
and associated costs. There were ten total interventions noted
in the appendix with transfer-to-ward within one hour of bed
assignment having the most impact.
Patel and Vinson21 used an ED team concept, which joined
an EP with two nurses and one technician. This novel change
lead to improved patient satisfaction with an increase of 3.1% in
reported “very good” or “excellent” ratings, a reduction in the
time required to see a physician, a 7.7% increase in number of
patients seen within one hour, and a 0.7% decrease in patients
who left without being seen.
Another published approach to reducing the ED LOS was
to use a 72-hour admission unit on an existing medical unit with
16 beds designated for ED overflow patients.22 This study, which
used a short-stay, 72-hour unit found that chest pain and asthma
patients had a significant reduction in ED throughput times.
Mean ED time was reduced from 7.3 to 5.5 hours per chest pain
patient, and 5.0 to 2.9 hours per patient with asthma; however,
patients with sickle-cell disease or seizures showed no decrease.
The article notes that no other changes in protocols, staffing
or processes occurred during the study period. Although this
study examined the effect of a short stay unit, in essence it was
evaluating the effect of increased inpatient capacity on LOS in
the ED. Gorelich, Yen and Yun23 found that in-room registration
reduced the length of ED stay by 15.0 minutes or 9.3%.
After a thorough review of the literature, we were unable
to find consensus on techniques to improve ED efficiency and
thereby decrease LOS. This is most likely due to environmental,
demographic, or institutional variations. One could conjecture
that there are significant differences between teaching and non-
teaching, small community versus large university, trauma versus
non-trauma centers, and large-volume versus small-volume
hospitals that prevent agreement on specific techniques. In other
studies, the conclusions were not intuitive or widely accepted.
For instance, two studies found that residents slowed patient
throughput but medical students did not. Unfortunately, there are
no comparisons of those institutions with both medical students
and residents, level of student or residents, or the effect of
residents from other services.
Despite a lack of consensus, this analysis demonstrates
that there are a number of scientifically-based procedures to
reduce ED patient LOS that could be useful. Certain strategies
appear to be universally accepted. These include pulse oximetry
determination in triage, bedside registration, point-of-care testing,
use of an urgent care area, and efficient lab, radiograph and
hospital admission processes. Furthermore, the use of physicians
in triage was found to be effective, although a cost versus benefit
analysis is needed. The ability to apply and implement many
of these procedures in other EDs is dependent on local factors,
politics and resources.
Many other articles were reviewed that were not included
in this study either because they did not meet the study
requirements or were not found in MEDLINE. The study
required that the article include some type of research rather
than a description of process improvement techniques. Valuable
information on throughput is frequently published in hospital or
management journals that discuss process improvement.
Based on a review of the literature on reducing patient LOS
in the ED, the best means for improvement is first to select the
appropriate determinants that drive patient throughput at the
local level, such as number of admissions, number of ambulance
arrivals and ED census, and then review and revise the processes
that drive the throughput determinants and monitor the data to
ensure that the changed processes accomplish the goal to improve
In the author’s experience, the critical success factors to
implement the necessary changes were to obtain accurate and
timely throughput data to review, obtain buy-in to the process from
senior management as well as staff who will have to implement
the changes, and determine what the cost/benefit ratio will be.
At the author’s hospital, LOS was reduced by 31% and the left-
without-treatment rate was reduced from 10% to 2% without any
additional costs in a three-month time period. The keys to success
were a rapid redesign process involving all hospital departments
and services, as well as senior management and line staff, using
accurate and correct ED data and having managers focus on a
self–initiated process improvement methodology.
Analysis of Literature on ED Throughput Zun
Volume X, no. 2 : May 2009 107 Western Journal of Emergency Medicine
Table 2. Analysis of literature
Compared days with
students to those
Medical students do
not alter the throughput
Average throughput was 145
minutes with and 151 without
students. Multiple biases
limited the study.
Average LOS for central lab
was four hours 22 minutes
and point-of-care testing
was three hours 28 minutes.
Reduced time was found only
in discharged patients.
lab to point-of-care
Compared impact of
faculty doing triage
to nurse only
Moderate reduction in
Average LOS with faculty was
363 minutes and 445 minutes
with nurses. Faculty was
added to complement nurses.
analysis of input/
output variables for
outcomes and ED inputs
Average main LOS 445
minutes, urgent 265 minutes,
entire 385 minutes. Significant
differences found between 20
and 80 percentile ED arrivals
2Before and after trial
and after change in
IT, staff revisions and
Value in rapid ED entry
Pre to post reduction of 31
minutes. Average LOS was
five hours. Process change
was not well described.
et al; 2006.
2 Cross-sectional study
of an institution with
and without tube
Compared two EDs,
one with pneumatic
tube system and one
with human couriers
Reduced lab turnaround
from 8-10 minutes
for Hgb and K with
pneumatic tube system.
Average turnaround time
varied from 33 to 72 minutes.
Courier was called to transport
specimens. Limited by two
EDs in Canada.
Average ED LOS was 7.96
hours and hospital LOS was
5.63 days. Austrailian study.
Liew et al.;
Compared ED LOS
ED LOS correlates
strongly with inpatient
2 Before and after
between presence of
PY-3s and LOS.
Average LOS before residents
was 123 minutes and 162
minutes at year three.
2 Before and after study
with and without triage
Reduced throughput by
Average turnaround for
pediatric bronchiolitis patients
pre-intervention 159 minutes
and post-intervention was 89
2Before and after
study of multiple
in increased MD
reduced the mean LOS
by 7.9 hours.
ED LOS reduced from 13.8
hours to 5.9 hours. Limited
by multiple interventions in
2 Before and after
effect on patient
reduced by 9.5 minutes.
Average throughput varied
from 239-257 minutes.
Limited by multiple personnel
changes during study periods.
Zun Analysis of Literature on ED Throughput
Western Journal of Emergency Medicine 108 Volume X, no. 2 : May 2009
Table 2. Analysis of literature
surgery - 21 minutes,
additional admission -
2.2 minutes, every 5%
increase in hospital
occupancy - 4.1
Mean LOS 241 minutes.
Limited number of variables
2 Before and after study
short stay unit
Reduced number of
patients waiting to go up
from 9.6 to 2.3 patients
Average LOS reduced from
6.5 hours to 5.6 hours. One
2Before and after study
LOS to in-room
Average LOS was reduced
by an average of 15 minutes.
Average length of ED LOS
dependent on inpatients,
daily census, pediatric
Average throughput 330
minutes for admitted and 123
for discharged. The correlation
coefficients ranged from .54-
3 Survey study of
physicians done by
for 690 hospital
times for labs correlated
specimen handling and
rapid transport time.
Average order to reporting
time mean was 50-60 minutes.
Limted by survey study.
Daily ED LOS increased
by 18 minutes with
a 10% increase in
Average throughput was 354
minutes for admitted patients.
Canadian sudy limited by lack
of correlation coefficients.
4Before and after
very varied and the
effect on throughput
decreased by 76
Average throughput time
was 175 minutes. Limited by
multiple factors and descriptive
Varied number of
treatment beds and
Increasing number of
nurses and physicians
Number of exam
rooms had no effect.
Laboratory time had a
ED LOS was correlated
with total lab outliers
rather than lab
A number of variables not
taken into account. Computer
4 Observational study
No throughput times provided
for the ED.
4 Survey study
throughput time with
and without fast track
Facilities with a Fast
Track had reduced
waiting time by 18
Average waiting time with Fast
Track was 72 minutes and
without Fast Track was 90
minutes. Unvalidated survey
sent to some east coast
ED, emergency department; LOS, length of stay; IT, information technology.
Analysis of Literature on ED Throughput Zun