Process-Improvement Cost Model

for the Emergency Department

Sheila R. Dyas, process improvement engineer, Department of Emergency Medicine,
Huntsville Hospital, Huntsville, Alabama; Eric Greenfield, DO, medical director,
Department of Emergency Medicine, Huntsville Hospital; Sherri Messimer, PhD,
associate professor, Department of Industrial and Systems Engineering and Engineering
Management, College of Engineering, The University of Alabama in Huntsville; Swati
Thotakura, MD, chairman, Department of Emergency Medicine, Huntsville Hospital;
Sampson Gholston, PhD, associate professor, Department of Industrial and Systems
Engineering and Engineering Management, College of Engineering, The University of
Alabama in Huntsville; Tracy Doughty, BSN, senior vice president of operations,
Huntsville Hospital; Mary Hays, DSN, RN, associate professor, College of Nursing,
The University of Alabama in Huntsville; Richard Ivey, director, Management
Engineering, Huntsville Hospital Health System; Joseph Spalding, systems engineer,
Command and Control Division, Northrop Grumman, Huntsville; and Robin Phillips,
BSN, RN, quality improvement analyst, Department of Emergency Medicine,
Huntsville Hospital

E X E C U T I V E S U M M A R Y
The objective of this report is to present a simplified, activity-based costing approach
for hospital emergency departments (EDs) to use with Lean Six Sigma costbenefit
analyses. The cost model complexity is reduced by removing diagnostic and condition-
specific costs, thereby revealing the underlying process activities cost inefficiencies.
Examples are provided for evaluating the cost savings from reducing discharge delays
and the cost impact of keeping patients in the ED (boarding) after the decision to
admit has been made. The process-improvement cost model provides a needed tool
in selecting, prioritizing, and validating Lean process-improvement projects in the
ED and other areas of patient care that involve multiple dissimilar diagnoses.

For more information about the concepts in this article, contact Ms. Dyas at
sheila.dyas@uah.edu.

442
 P rocess -I mprove ment C ost M odel
INTRODUCTION
Emergency departments (EDs) are
seeing an increasing number of patients
with below-cost to no-payment capabil-
ity such as self-pay and public insurance
(Medicare and Medicaid). The Congres-
sional Budget Office (2015, p. 123)
estimated that the number of people
enrolled in Medicaid would increase by
11 million in 2015. Medicaid beneficia-
ries use the ED two times more than do
patients with private insurance (Mann,
2014; Taubman, Allen, Wright, Baicker,
& Finkelstein, 2014). A study conducted
by the Colorado Hospital Association
(2014) showed that the 61% of ED visits
across 62 Colorado hospitals in the first
quarter of 2014 were made by publicly
insured and self-pay patients, which is a
4% increase from the previous year. This
increase in uncompensated care poses a
threat to ED survival.
Healthcare has turned to the
manufacturing industries and continu-
ous improvement approaches, such as
Lean Six Sigma, to provide a conduit
for reducing costs while maintaining
or improving quality of care (Dickson,
Singh, Cheung, Wyatt, & Nugent,
2009). The goal of Lean is to provide
more value to customers through the
elimination of waste and increases in
quality, leading to an increase in profit
(Womack & Jones, 2003). In manufac-
turing, cost reductions typically can be
directly correlated with processing time
reductions (Hansen & Mowen, 2010).
Therefore, if processing time is reduced
through the elimination of waste or
the optimization of the process, the
costs associated with performing the
process will be reduced. Emphasis is
placed on measuring the time savings
443
for the E mergency D epartment
rather than the cost savings (Rother &
Shook, 2003).
The application of Lean to health-
care is widely debated. Leaders have had
concerns such as the effects on the
patients perception and the ability to
standardize physician and nurse activi-
ties without sacrificing patient care
(Dean, 2013; Waring & Bishop, 2010).
Holden (2011) summarized the docu-
mented implementation of Lean prac-
tices in EDs during the period from
2006 to 2009 as promising but incon-
clusive because of missing information
regarding patient safety, employee
satisfaction, sustainment, and manage-
ment participation. Holdens results
were supported by a 2009 case study of
the effect of Lean in four EDs (Dickson,
Anguelov, Vetterick, Eller, & Singh,
2009). The fact that these efforts
resulted in time reductions has been an
impetus to many hospital systems to
implement the Lean philosophy (Dart,
2011; Graban, 2014; Kenney, 2010). For
example, ThedaCare, a health system in
Wisconsin that introduced Lean think-
ing in 2003, implemented 10,000
continuous daily improvements in
2011 and more than 20,000 in 2012
(Graban, 2014; Mannon, 2014). Theda-
Care has used Lean tools in the ED to
reduce the door-to-balloon time for
patients who have had heart attacks
from 90 to 37 minutes and to increase
the percentage of patients with ischemic
stroke who undergo computed tomogra-
phy within 25 minutes from 51% to
89% (Toussaint & Gerard, 2010).
In the ED, a correlation between
time and cost does not always exist.
Instead, the costs associated with patient
care are predominately condition
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015
driven. For example, a patient with an
emergency severity index (ESI) acuity of
1 and a patient with an ESI acuity of 4/5
may have the same length of stay (LOS)
in the ED but not the same cost of care.
This, coupled with the high ED patient
variability, makes using Lean tools such
as value stream mapping (i.e., diagram
of steps in process and associated time
management information) and DMAIC
(define, measure, analyze, improve, and
control) challenging (George, Maxey,
Rowlands, & Price, 2004; Locher, 2013).
A key Lean reference points out that
many enterprises jump right into elimi-
nating waste without understanding the
entire value stream (Rother & Shook,
2003). The improvements used in one
part of the stream can negatively affect
another with regard to time, cost, or
both. For the DMAIC process, process-
improvement team members perform a
costbenefit analysis as part of the
project charter for comparison with
hospital leadershipestablished guide-
lines (Stamatis, 2010). The costbenefit
analysis is repeated in the improve
phase to ensure that the proposed
process changes are justifiable (Stamatis,
2010). An understanding of the costs is
necessary when using Lean Six Sigma
methodology, particularly when the
relationship between time and cost can
be inconsistent.
We developed an ED process-
improvement cost model and present it
here. This model will enable healthcare
leaders to evaluate the anticipated cost
savings and/or revenue increases result-
ing from a process-improvement project
and will ensure that the anticipated time
savings results in an acceptable cost
savings for the ED. In addition, the cost
444
model provides a tool to help prioritize
the Lean projects competing for ED staff
time and capital funding.
C O S T- M O D E L I N G A P P R O A C H
Determining the cost or revenue
changes per process change is possible
for manufacturing applications because
of the use of activity-based cost account-
ing that provides detailed cost informa-
tion per part and assembly. Cost data
are continually updated to reflect the
actual expenditures associated with each
stage of the manufacturing process.
Managers compare per-product revenue
with costs to ensure that an adequate
profit margin is maintained.
People have attempted to apply cost
accounting methods used in manufac-
turing to healthcare. In the 1990s,
activity-based costing (ABC) became the
forefront approach for healthcare
(Lawson, 2005). In ABC, costs are
assigned directly to the activity associ-
ated with them, and overhead is no
longer a percentage mark-up but is
allocated on the basis of use of the
overhead contributor. For example, the
cost of a blood transfusion more than
doubles when the costs associated with
the activities required to perform a
transfusion are included compared with
the cost of blood acquisition alone
(Shander et al., 2010). These activities
include hospital blood bank manage-
ment, pretransfusion process, delivering
components from the blood bank to the
transfusion site, and overhead activities.
In 2005, Raef Lawson (2005) from
the Institute of Management Accoun-
tants published a study comparing the
use of ABC in healthcare in 1994 with
its use in 2004. He found that its use
 P rocess -I mprove ment C ost M odel
had declined and that fewer healthcare
organizations were considering its
implementation. One of the key reasons
for the lack of interest was the extensive
data and data management
requirements.
As information technology has
progressed, allowing greater quantities of
data to be accessed and quickly ana-
lyzed, data management processes have
become less of a concern. Therefore,
activity-based accounting, or something
similar, has resurfaced in recent years as
a means of understanding the cost of
healthcare. The time-driven activity-
based costing (TDABC) approach, as
applied to healthcare, has been used by
the Head and Neck Center at MD
Anderson Cancer Center in Houston,
Texas; the Cleft Lip and Palate Program
at Childrens Hospital in Boston, Massa-
chusetts; the knee replacement units at
Bringham & Womens Hospital in
Boston, and Schn Klinik in Germany;
and at Magusa Yasam Hospital in Cyprus
(Kaplan & Porter, 2011; Kaplan et al.,
2014; ker & zyapc, 2013). MD process regardless of the patients
Anderson Cancer Center automated the
TDABC approach for use in other clinical
care areas. Use of the automated model
in their Anesthesia Assessment Center
reduced costs and improved process
efficiency without compromising patient
care. Identified process optimizations
reduced the average patient processing
time by 33% and the associated person-
nel costs by 46% (Kaplan et al., 2014).
MD Anderson Cancer Center plans to
implement TDABC institute wide in the
near future (Institute for Cancer Care
Innovation, 2013/2014).
The TDABC approach simplifies
activity-based costing by using a cost per
445
for the E mergency D epartment
unit of time rate based on total
resources used and average time
expended for each activity instead of
allocating total costs expended to each
transaction of an activity based on the
percentage of total time expended for
that activity (Kaplan & Anderson, 2004).
This unit cost coupled with the time
required to perform the activity provides
an estimate of cost per time for each
activity. The unit cost of supplying
capacity requires an estimation of all the
expenses involved in performing that
activity and their respective utilization
rates. Totaling the costs for each activity
yields a total cost for the predefined
patient type or process, which can be
compared directly with the invoiced
cost. Although the approach is easily
implemented for a specific diagnosis,
the process becomes quite complex
when evaluating an ED in which the
number of diagnoses is substantial
(Hooze, 2010).
The cost modeling approach pre-
sented here looks at the underlying
condition. By considering only the costs
associated with the process under
investigation, the model is greatly
simplified. Diagnosis-specific costs that
are invoiced directly to the patient, such
as inventory used and services per-
formed outside the ED, are not
included. This exclusion minimizes the
variability in cost per patient as a
function of processing time. For
instance, if we compared the costs
associated with the imaging process for
a high-acuity trauma patient with those
for a low-acuity soon-to-be discharged
patient, diagnosis-specific costs per time
or patient would be significantly
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015
different. The cost difference associated
with computed tomography with
contrast for the high-acuity patient
versus an X-ray for the low-acuity
patient would overshadow the cost
reductions associated with reducing the
staff time required to perform the
process. By including only the process-
specific costs associated with imaging
for the two types of patients (such as
increased staff and additional time), the
potential savings resulting from imaging
optimization are discernible.
The cost model is used to determine
the current process costs and to estab-
lish a baseline for comparison with the
future process costs associated with
Lean-derived process changes. The
process-improvement team members
analyze the recommended changes to
determine the effect on current time
expenditures, resources required, and
patients who can be accommodated.
The cost inputs, time inputs, or both are
then adjusted to reflect the proposed
changes. The team members then
determine the potential gains and
compare them with the baseline and
other proposed process-improvement
options in the costbenefit analysis. This
allows them to verify that the time
savings will result in cost savings and
that the costs of implementing the
changes are justified.
Identifying and mapping the com-
plete set of activities across an organiza-
tion is one principle of Lean.
Participants must understand the
complete process and its effects on other
processes (Womack & Jones, 2003). The
cost model supports this principle and
is to be developed for the ED value
stream map or another all-inclusive
446
process map first. As Lean advances in
stages and requires a more complete
understanding of specific activities, the
model is adjusted as necessary. Costs
can be further broken down, changes to
specific process activities can be evalu-
ated, and other details can be integrated
into the cost model.
Once overall ED operations are
optimized , participants can evaluate
condition-specific processes for preva-
lent conditions or those of higher
priority in the ED. Costing approaches
similar to the TDABC approach, which
includes condition-specific costs, can be
used (Kaplan & Porter, 2011).
C O S T- M O D E L I N G P R O C E S S
Defining the process is the first step in
creating a cost model. The process needs
to be as inclusive as possible to ensure
that changes made in one step do not
negatively affect other steps. Value
stream mapping is a Lean tool com-
monly used to define the process
(George et al., 2004). Process flow
charts, spaghetti diagrams, and SIPOC
(suppliers, inputs, processes, outputs,
and customers) diagrams are other Lean
tools used for mapping (George et al.,
2004). Any process mapping technique
is acceptable as long as complete
information is presented regarding the
steps in the process and their interaction
with one another. We have devised a
simple process mapping technique that
includes the human inventory (i.e.,
patient wait) as an activity, thereby
enabling a cost to be assigned to the
patient wait time.
Figure 1 displays the process map for
a factitious hospital, A-1. We used a
process coding system to categorize the
 P rocess -I mprove ment C ost M odel
paths based on an EDs desires or needs.
The black and gray arrows represent the
ideal path and currently acceptable paths
from a patients perspective, respectively,
and vary among EDs depending on their
operations strategies. The process
depicted represents a direct-bedding
strategy in which triage adds value only
when a bed is not available. Process-
improvement team members should
challenge and justify the identification of
a path as black or gray. The dashed
arrows represent paths for which process
planning is required but minimal patient
travel is desired. These paths typically
represent waste. The dotted arrows
indicate paths associated with a loss in
revenue and need to be eliminated.
COST MODEL
Using the activities identified in the
process map, the process-improvement
team develops a cost model. The model
algorithm is based on distributing the
activity-based costsevenly or per an
allocation schemeto each hour spent
FIGURE 1
A-1 Hospital Emergency Department Process Map
447
for the E mergency D epartment
by patients in the ED. The cost per
patient-hour equals the total activity
costs divided by the multiple of the total
number of patients over the period in
which the costs were incurred and the
average time spent by a patient in the
activity. Although the math is simple,
the multitude of cost factors and the
allocation method requires an off-the-
shelf spreadsheet program to facilitate
the computation.
The following data are required to
create the model: (1) ED budget con-
verted in dollars per day, (2) facility and
equipment costs that should be assigned
to the ED, (3) lost revenue estimates
associated with patients leaving before
treatment is started or completed and
missed opportunities from patients
going to a competitors ED, (4) time
data associated with the process activity
to be optimized, (5) type of patient
data, if required, associated with the
time data, and (6) number of patients
leaving the ED before starting or com-
pleting treatment.
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015
The advancement and availability
of ED information systems and the
hospitals required cost accounting and
budgeting information make most of
these data easily accessible (Weinstock,
2014). Typically, only the staffing and
inventory costs are part of the EDs
budget. All other items are included in
the hospitals budget or are not consid-
ered, such as overhead and lost revenue
(Nowicki, 2006). A process is needed
to determine the allocated amount or
to estimate these costs. Possible meth-
ods include allocating facility deprecia-
tion, maintenance, utilities, insurance,
and property tax by the square footage
of the ED versus that of the entire
hospital and define equipment mainte-
nance as a percentage of the purchase
cost. The important point is that the
estimates are reasonable and substanti-
ated and that equivalent methods are
used for all projects.
Items not included in the cost
model are directly billable costs that
involve staff from other departments,
directly billable inventory, physician
incentive pay based on relative value
units, and care by consulting physicians
or specialists. These costs, although
applicable to total costs, are not under
the EDs control or are diagnosis spe-
cific. As discussed earlier, these costs are
highly variable, depending on the
patient and his or her condition, and
they dwarf the costs needed to evaluate
ED processes.
The next step is determining the cost
per patient-hours spent in the ED on the
basis of the current process and the
proposed changes to cost and time. If
one assumes no change in cost per
patient-hour or in the total number of
448
patients seen, savings are calculated by
multiplying the cost per patient-hour by
the time reduction attributed to the
Lean process. However, this amount
underestimates the actual savings.
The distribution of costs, given a
reduction in total patient-hours, should
be determined in the cost model in one
of two ways: (1) an increase in patient
capacity attributed to better service or
(2) a decrease in costs caused by a
reduction in resources required, revenue
lost, or both. Decreasing costs involves a
simple adjustment to the model and is
reflected in the cost per patient-hour.
Increasing patient capacity is more
complex and requires a trial and error
approach or simple linear programming
to allocate the realized time savings to
patient type (admitted, discharged, or
other), if segregated. The number of
patients according to type and the LOS
are used to allocate the time savings and
determine the total increase in patient
capacity.
To demonstrate use of the model,
we present an example of a process-
improvement project for reducing
discharge delays at A-1 Hospitals ED.
This process-improvement project can
be summarized as follows:
The A-1 Hospitals process-
improvement team decided to reduce
the wait to leave ED time for patients
being discharged. The team prepared the
process map and gathered time statistics.
The team discovered that 25% of all
patients discharged from the ED were
delayed by an average of 25 minutes to
complete registration. Meetings were
held, the registration process was
reviewed, and steps were taken to reduce
the registration process time. However,
 P rocess -I mprove ment C ost M odel
the greatest delay in the process was
ascertaining when the patient was
available to complete registration. Team
members discussed making capital
purchases such as a real-time location
system to provide information regarding
patient availability (Drazen & Rhoads,
2011). The team then questioned how
much money would be saved.
The first question to be asked is this:
How much does it currently cost for a
discharged patient to wait to leave while
occupying an ED bed? To answer this
question, we first gather controllable
costs incurred by the ED. These costs
and the methods used to estimate them
are listed in Table 1.
A-1 Hospitals ED is a 90-bed,
55,000-square-foot facility that receives
90,000 patients per year; the average
revenue per ED visit is $900. The costs
equate to $300 per patient ($27 million
per year). This can be further broken
down on the basis of specific areas in
the ED, such as fast-track care versus
regular care, type of bed (such as trauma
or regular), or patient type (such as
admitted versus discharged).
Any breakdown of costs is done on
the basis of the Lean project being
considered. For A-1 Hospitals ED, the
Lean project is evaluating the wait to
leave ED activity for discharged patients
with a registration delay. Figure 1 shows
the process, which is identified by the
wait to leave ED process box and the
gray path to home. This process requires
a breakdown of costs between admitted
and discharged patients. The relevant
time statistics are presented in Table 2.
According to the equations pre-
sented in Table 3, A-1 Hospitals ED
experiences 32,567 patient-hours per
449
for the E mergency D epartment
month, 65% of which are for discharged
patients. Although 70% of the patients
were discharged, they used only 65% of
the ED resources. The cost for a dis-
charged patient is $270 ($17 million
per year).
For the wait to leave ED activity, the
difference in cost is even greater. Dis-
charged patients used only 29% of ED
resources. Using the same model but
considering only the percentage of LOS
attributed to wait to leave ED, the
current cost for a discharged patient
would be $23 ($1.4 million per year).
These differences highlight the impor-
tance of appropriately defining the
model to represent the costs associated
with the process-improvement project.
Using $274 instead of $23 per patient
greatly affects the conclusion.
A-1 Hospitals process-improvement
team would like to eliminate any waits
to leave the ED caused by registration
delays. Discharged patients who are
delayed because of registration use ED
resources 31% of the time and account
for $450,000 of the costs associated
with discharged patients waiting to leave
the ED. If A-1 Hospitals ED eliminates
the registration delay, the wait to leave
ED time for discharged patients would
be reduced to 14 minutes, and $450,000
per year could be saved.
The $450,000 savings assumes that
there is no change to the patient count
because of the time savings, and that the
cost per patient-hour remains constant.
To properly evaluate the expected cost
savings, the process-improvement team
must decide how this improvement in
performance will affect the number of
patients seen and calculate any changes
to the costs presented in Table 1. The
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015

TA B L E 1
Applicable Emergency Department (ED) Costs

Description A-1 Hospital

Patient Care ED Staff Payroll Expenses per Day
Physician $25,500
Midlevel provider $3,500
Registered nurse $21,000
Technician $4,000
Registrar $3,250
Housekeeping $1,000
Supply stocker $750
Other $500
Inventory
Nondiagnosis-specific supplies including ordering 25% of total inventory + 3% order cost/
costs (e.g., blood pressure cuffs) week = $151/day
All inventory losses (expired, damaged, and stolen) 5% of total inventory = $30/day
All inventory carrying costs in ED 15% of total inventory = $90/day
Equipment
Depreciation Straight line per equipment type = $800/day
Maintenance % of equipment cost per equipment type =
$425/day
Facility
Depreciation $4.80/year/square feet = $725/day
Maintenance $4.23/year/square feet = $640/day
Utilities $3.23/year/square feet = $490/day
Insurance $2.00/year/square feet = $300/day
Property tax Nonprofit hospital
Laundry $3.22/patient = $800/day
Waste management $500/day
ED Overhead
Overhead staff $3,250/day
Inventory for overhead staff use Negligible
Facility costs (overhead staff) $14.26/year/square feet = $450/day
Equipment costs (overhead staff) $10/day
Lost Revenue
Left without being seen (LWBS) patients No. of LWBS patients per year times
percentage of ED revenue per patient
allocated to above costs = $2,500/day
Elopements (patients seen but left before care Same calculation as that for LWBS patients,
completed) adjusted for average percentage payment
received = $1,300/day
Missed opportunities 50% of LWBS cost = $1,250/day
Source: International Facility Management AssociationHealthcare, American Society for Healthcare Engineering (2010).

450
 P rocess -I mprove ment C ost M odel
TA B L E 2
A-1 Hospitals ED Monthly Patient Statistics
Variable
Total No. of Patients
Admitted
Discharged
Average Length of Stay, Minutes
Admitted
Discharged
Wait to Leave ED, No. of Patients
Admitted
Discharged
Average Wait to Leave ED, Minutes
Admitted
Discharged
Discharge Delay Caused by Registration Hold
Average, minutes
Discharged patients, %
team assumed that a 10% reduction in
revenue loss would occur and the
patient-hours saved would increase the
number of patients seen.
Using the new wait to leave ED time
for discharged patients, we calculate an
average LOS of 256 minutes. This
4-minute savings per patient would
enable 126 additional patients (37
admitted, 89 discharged) to be seen per
month. Using A-1 Hospitals net revenue
for the average patient (admitted and
discharged), plus the savings attributed
to the reduction in lost revenue, results in
a potential savings of $630,000 per year.
In summary, A-1 Hospitals process-
improvement project has the potential
of realizing a gain of $630,000 per year
if the registration discharge delay is
eliminated. The process-improvement
451
for the E mergency D epartment
Statistic
2,200
5,300
260
310
240
2,200
5,300
55
120
20
25
25
team decided to reduce the time
required to register patients by reducing
the total activities in the registration
process and the amount of information
obtained from patients. No capital
expenditures were made on equipment
to determine patient availability. The
team shifted work to other departments
and eliminated gathering of unnecessary
patient information. The process-
improvement team will reevaluate
registration-related discharge delays after
the changes are implemented. The
information provided by the cost model
was essential in deciding how much
time, effort, and assets to put forth for a
process-improvement project.
In addition, by examining the wait
to leave ED times, the process-improve-
ment team was able to determine how
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015

TA B L E 3
Formulas and Calculations

Formula Result Calculation

Discharged patient hours = discharged patient count 21,200 patient 5,300 240 60
discharged patient LOS hours
Admitted patient hours = admitted patient count 11,367 patient 2,200 310 60
admitted patient LOS hours
Total patient hours = discharged patient hour 32,567 patient 21,200 + 11,367
+ admitted patient hours hours per month
% discharge patient hours = discharged patient 65% 21,200 32,567
hours/total patient hours
% discharge patients = discharged patient 71% 5,300 7,500
count/total patient count
Total cost = sum of cost per day number of $2,196,330 per $73,211 30
days per time period month
$ per patient hour = total cost/total patient $67.44 per $2,196,330/32,567
hours patient hour
Cost per discharged patient = $ per patient hour $270 per patient $67.44 240 60
discharged patient LOS
Note: LOS = length of stay.

much time boarding added to LOS. C O S T B E N E F I T A N A LY S I S

Given the cost modeling approach
presented here, the team is able to
determine the cost associated with
boarding (see sidebar, How Much Is Bed
Boarding Costing Your ED?).
Lean and process improvement is a
continuous exercise (Pirasteh & Kan-
nappa, 2013). Whether it be in the
discovery of an area requiring improve-
ment or in monitoring the performance
of an area that has undergone a process
improvement, a cost model is crucial to
determine whether the changes to be
made are worth the disruption to
operations (Seidman & McCauley,
2012).
The ED process-improvement cost
model presented here offers a simplified
and, thus, cost-effective means for
understanding the financial resources
required to implement a Lean improve-
ment. By removing the activity costs
associated with diagnostic tests, proce-
dures performed, and medications
administered in the ED, the complexity
of the spreadsheet calculations and the
time required to create it are decreased
substantially.
The costs associated with A-1
Hospitals ED model can be broken
down into three areas: process mapping,
cost model development, and

452
 P rocess -I mprove ment C ost M odel for the E mergency D epartment

How Much Is Bed Boarding Costing Your ED?

At A-1 Hospital, 83% of admitted patients stay in an ED bed for 60 or more minutes,
with an average wait time of 140 minutes. The remaining 17% wait an average of 25
minutes. Using these numbers and Table 2 data, the authors find that boarders account
for $3 million per year, or 12% of the total ED costs examined in this report. Reducing
the average wait time from 140 to 60 minutes would enable the ED to see an additional
6,600 patients per year. The additional revenue would increase the gain realized to $4
million per year.

spreadsheet creation (using off-the-shelf developing the cost model, running

software). Process mapping is a prin-
ciple of Lean and is required regardless
of whether a cost model is used or not;
process mapping should not be consid-
ered in the costbenefit analysis of
developing a cost model.
On the basis of our experiences with
the cost model, less than 40 hours
would be required to gather data and
define the cost model algorithm for an
ED with an information system. How-
ever, the time needed to prepare the
spreadsheet depends on the hospital
and its processes. The greater the num-
ber of paths, beds, care units, and so
forth, the greater the time required to
create the spreadsheet. The time
required to evaluate a process change
proposal should be less than 8 hours.
Costs for both of these activities would
be a function of the hourly labor rate of
the people performing the tasks.
The process used to create A-1
Hospitals ED cost model made it
difficult to distinguish the costs of
model preparation from those of the
Lean project. The activities performed by
the team consisted of preparing a
process map, performing a DMAIC
process to eliminate registration delays,
interactions based on proposed changes,
and documenting the results. The
percentage of costs associated directly
with development of the cost model is
difficult to determine. However, once
the model is created, the costs associ-
ated with using it as a tool in other Lean
projects is minimal. In addition, the
knowledge gained by creating the cost
model will greatly reduce the time
required to make changes in the process
map in the current ED or to develop a
cost model for another hospitals ED.
The greatest benefit is the reduction
of resources required to conduct Lean
projects. Performing a costbenefit
analysis as part of the Lean project
charter ensures that the potential savings
will meet a hospital-defined minimum
before resources are expended (Stamatis,
2010). In addition, understanding the
costs up-front minimizes the time spent
on evaluating noncost-effective options
for process optimization. Finally,
understanding the interaction between
process map activities in the cost model
ensures that a positive cost effect in one
activity does not have a negative effect
in another. The cost model also serves as
a means to aid in prioritizing Lean

453
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015
projects, confirming that the Lean
project stays on track and achieves the
goal of the project, and providing an
assessment of how Lean methodology
works in the hospital setting.
S T U D Y L I M I TAT I O N S
Process changes made in an ED may
not be independent of other depart-
ments. A process improvement for the
ED may have a negative effect on costs
in other departments or the hospital as
a whole (Schultz, 2011). For A-1 Hospi-
tals process-improvement project,
registration activities were reduced by
shifting the work to another depart-
ment. This shifting of work to another
department was flagged as a change that
was not considered by the cost model
and that could have had a cost impact
to the hospital system. We assessed this
impact and determined it to be negli-
gible because the activities were already
being performed by the other depart-
ment. To ensure that overall costs are
reduced, managers must understand the
value stream or process map of the
entire enterprise or at least the connec-
tions between departments (Womack &
Jones, 2003).
In addition, improving a process
often involves factors that are difficult to
quantify. For instance, time reductions
could result in an increase in employee
turnover, staff reductions could lead to a
decrease in customer satisfaction ratings,
and inventory reductions could hamper
disaster preparedness (Beach, 2010;
Guilding, Lamminmaki, & McManus,
2014). The process-improvement team
must consider all factors that might be
affected by the changes and balance the
possible consequences against the cost
454
savings from the recommended process
changes.
Finally, process and patient statistics
are gathered over a defined period and
combined with actual or budgeted costs
to create the cost model. The cost model
represents the average patient popula-
tion and process performance. The
process-improvement team can perform
a sensitivity analysis to gain insight into
the cost differences associated with
patient and process variability. The
analysis can be as simple as changing a
few factors independently or as detailed
as a multifactorial design of experiments
(Hillier & Lieberman, 2010). The sensi-
tivities will identify the risk level associ-
ated with the process-improvement
decision with regard to daily versus
average operation.
A cost model, like any other Lean
tool, is a representation of current or
planned operations and is limited in its
comprehensiveness. The effects on
operations outside of the ED, factors not
included in the cost model, and atypical
operations need to be considered when
deciding to implement a process-
improvement recommendation.
CONCLUSION
The process-improvement cost model
provides activity-based costing of Lean
process improvements in the ED with
minimal complexity. Lean methodology
is increasingly being used in the ED to
improve patient care, increase patient
flow, and reduce costs. The process-
improvement team performs a cost
benefit analysis at various stages to
ensure that the cost savings resulting
from the projected time savings are
adequate to cover the time, effort, and
 P rocess -I mprove ment C ost M odel
capital expenditures required to opti-
mize the process. The process-
improvement team can reduce the
effort associated with performing a
costbenefit analysis by accepting the
premise that a reduction in time equates
to a reduction in cost. Although this is
typically the case in manufacturing, it is
not so in the ED environment.
Understanding the costs associated
with ED care is difficult when all treat-
ment options are included. By omitting
the costs outside the EDs control, such
as diagnosis-specific treatments, devel-
oping a cost model is simpler and the
underlying process easier to evaluate.
Although the total cost of patient care is
not known, the costs that the ED can
change are defined per activity per
patient-hour. This approach provides an
important metric to be added to the
evaluation, comparison, and verification
process associated with Lean process-
improvement projects in the ED.
Once the cost model is developed
for a hospitals ED, it can be easily
modified to reflect changes in the EDs
process map or used in another ED in
the hospital system. In addition, the cost
model can be used in other areas of
patient care that involve multiple
dissimilar diagnoses, such as intensive
care or medicalsurgery units. The initial
outlay to develop the cost model can be
readily justified by the benefits it may
provide to a hospital system.
ACKNOWLEDGMENT
The authors thank Huntsville Hospital in
Huntsville, Alabama, for providing representa-
tive ED data, thereby enabling A-1 Hospitals
ED to reflect actual hospitals. Huntsville
Hospital is a 941-bed, community-based,
not-for-profit institution. It is the second
for the E mergency D epartment
largest hospital in Alabama and part of the
third-largest publicly owned hospital system in
the nation. Huntsville Hospital serves as the
major referral hospital and trauma center for
patients in northern Alabama and southern
Tennessee and is a teaching facility for the
University of Alabama at Birminghams Family
Practice and Internal Medicine residency
programs (Huntsville Hospital, 2014).
REFERENCES
Beach, M. (2010). Disaster preparedness and
management. Philadelphia, PA: F.A. Davis
Co.
Colorado Hospital Association, Center for
Health Information and Data Analytics.
(2014, September). Impact of Medicaid
expansion on hospitals: Updated for second-
quarter 2014. Retrieved from http://www
.cha.com/Documents/CHA-Study/FINAL
-CHA-Medicaid-Expansion-Study-Q2
-Sept-2014.aspx
Congressional Budget Office. (2015, January).
The budget and economic outlook: 2015 to
2025. Publication 49892, Appendix B, 211.
Retrived from https://www.cbo.gov/sites
/default/files/114th-congress-2015-2016
/reports/49892-Outlook2015.pdf
Dart, R. C. (2011). Can Lean thinking trans-
form American health care? Annals of
emergency medicine, 57(3), 279281.
Dean, M. L. (2013). Lean healthcare deployment
and sustainability (pp. 2731). New York,
NY: McGraw Hill Education.
Dickson, E. W., Anguelov, Z., Vetterick, D.,
Eller, A., & Singh, S. (2009). Use of Lean
in the emergency department: A case series
of 4 hospitals. Annals of Emergency
Medicine, 54(4), 504510.
Dickson, E. W., Singh, S., Cheung, D. S., Wyatt,
C. C., & Nugent, A. S. (2009). Application
of Lean manufacturing techniques in the
emergency department. The Journal of
Emergency Medicine, 37(2), 177182.
Drazen, E., & Rhoads, J. (2011, April). Using
tracking tools to improve patient flow in
hospitals. California Healthcare Foundation
Issue Brief. Retrieved from http://www
.chcf.org/~/media/MEDIA%20
LIBRARY%20Files/PDF/PDF%20U
/PDF%20UsingPatientTrackingToolsIn
Hospitals.pdf
George, M. L., Maxey, J., Rowlands, D., & Price,
M. (2004). The Lean Six Sigma pocket
455
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015
toolbook: A quick reference guide to 100 tools
for improving quality and speed (pp. 3354).
New York, NY: McGraw-Hill Education.
Graban, M. (2014). Good change: Using kaizen
toward a culture of continuous improve-
ment humanizes the healthcare workforce
for better outcomes. Industrial Engineer,
46(2), 30.
Guilding, C., Lamminmaki, D., & McManus, L.
(2014). Staff turnover costs: In search of
accountability. International Journal of
Hospitality Management, 36(2014),
231243.
Hansen, D., & Mowen, M. (2010, January 1).
Cornerstones of cost accounting (pp. 123,
131186). Independence, KY: Cengage
Learning.
Hillier, F. S., & Lieberman, G. J. (2010).
Introduction to operations research (pp.
217258). New York, NY: McGraw-Hill
Higher Education.
Holden, R. J. (2011). Lean thinking in emer-
gency departments: A critical review.
Annals of Emergency Medicine, 57(3),
265278.
Hooze, S. (2010). On the determinants and
the role of costing accuracy of time-driven
activity-based costing (dissertation).
Ghent, Belgium: Ghent University.
Retrieved from http://hdl.handle.net
/1854/LU-970901
Huntsville Hospital. About. (2014). Retrieved
from http://www.huntsvillehospital.org
/about/
International Facility Management Association-
Healthcare, American Society for Health-
care Engineering. (2010). Operations and
maintenance benchmarks for health care
facilities report (pp. 2452). Chicago, IL:
Author.
Institute for Cancer Care Innovation.
(2013/2014, Winter). TDABC implementa-
tion proves fruitful. Institute for Cancer
Care Innovation Newsletter. Retrieved from
http://www.mdanderson.org/education
-and-research/research-at-md-anderson
/cancer-care-innovation/about-us
/icci-winter-13-14-newsletter-3.pdf
Kaplan, R. S., & Anderson, S. R. (2004).
Time-driven activity-based costing.
Harvard Business Review, 82(11), 131138.
Kaplan, R. S., & Porter, M. E. (2011). How to
solve the cost crisis in health care. Harvard
Business Review, 89(9), 4664.
456
Kaplan, R. S., Witkowski, M. L., Abbott, M.,
Guzman, A., Higgins, L., Meara, J., ...
Feeley, T. W. (2014). Using time-driven
activity-based costing to identify value-
improvement opportunities in health care.
Journal of Healthcare Management, 59(6),
399412.
Kenney, C. (2010). Transforming healthcare:
Virginia Mason Medical Centers pursuit of
the perfect patient experience (pp. 3348).
New York, NY: Taylor & Francis.
Lawson, R. A. (2005). The use of activity based
costing in the healthcare industry: 1994 vs.
2004. Research in Healthcare Financial
Management, 10(1), 7794.
Locher, D. (2013). Practicing Lean fundamen-
tals in an office environment. Retrieved
from http://www.lean.org/common
/display/?o=2215
Mann, C. (2014, January 16). Reducing
nonurgent use of emergency departments
and improving appropriate care in
appropriate settings. CMCS Informational
Bulletin. Retrieved from http://www
.medicaid.gov/Federal-Policy-Guidance
/Downloads/CIB-01-16-14.pdf
Mannon, M. (2014). Lean healthcare and
quality management: The experience of
ThedaCare. The Quality Management
Journal, 21(1), 710.
Nowicki, M. (2006, June 1). HFMAs introduc-
tion to hospital accounting (pp. 1823).
Chicago, IL: Health Administration Press.
ker, F., & zyapc, H. (2013). A new costing
model in hospital management: Time-
driven activity-based costing system. The
Health Care Manager, 32(1), 2336.
Pirasteh, R. M., & Kannappa., S. (2013). The
synergy of continuous process improve-
ment. Industrial Engineer, 45(6), 41.
Rother, M., & Shook, J. (2003). Learning to see:
Value stream mapping to add value and
eliminate muda (pp. 15, 4868).
Cambridge, MA: Lean Enterprise Institute.
Schultz, J. R. (2011). Making it all work: A pocket
guide to sustain improvement and anchor
change (pp. 18). New York, NY: Routledge.
Seidman, W., & McCauley, M. (2012). Restor-
ing executive confidence in performance
improvement. Performance Improvement,
51(4), 1420.
Shander, A., Hofmann, A., Ozawa, S.,
Theusinger, O. M., Gombotz, H., &
Spahn, D. R. (2010). Activity-based costs
 P rocess -I mprove ment C ost M odel for the E mergency D epartment

of blood transfusions in surgical patients
at four hospitals. Transfusion, 50(4),
753765.
Stamatis, D. H. (2010). Essentials for the
improvement of healthcare using Lean & Six
Sigma (pp. 179194). New York, NY:
Productivity Press.
Taubman, S. L., Allen, H. L., Wright, B. J.,
Baicker, K., & Finkelstein, A. N. (2014).
Medicaid increases emergency-department
use: Evidence from Oregons health
insurance experiment. Science, 343(6168),
263268.
Toussaint, J., & Gerard, R. A. (2010). On the
mend: Revolutionizing healthcare to save lives
and transform the industry (pp. 3260).
Cambridge, MA: Lean Enterprises Institute.
Waring, J. J., & Bishop, S. (2010). Lean health-
care: Rhetoric, ritual and resistance. Social
Science & Medicine, 71(7), 13321340.
Weinstock, M. (2014). 2014 most wired.
H&HN, 88(7), 3043.
Womack, J. P., & Jones, D. T. (2003). Lean
thinking: Banish waste and create wealth in
your corporation (pp. 15101, 275298).
New York, NY: Productivity Press.

PRACTITIONER A P P L I C AT I O N

Sherri Welch, BSN, FACHE, chief nursing officer, College Station Medical Center,
College Station, Texas

T he financial viability of hospital emergency departments (EDs) should be a focus

of all hospital administrators. As Dyas et al. indicate, the expansion of public
insurance has resulted in an increase in uncompensated care and a shift in payer mix
for hospitals. As more people use the ED as their sole source of medical care, it
becomes even more important for hospitals to focus on the elimination of waste and
reduction of costs.
The ED process improvement cost model, as presented by Dyas et al., is based on
a Lean Six Sigma costbenefit analysis and illustrates process improvements that
reduce costs by improving efficiencies. This study suggests that understanding the
costs associated with ED patient care helps staff, administrators, and physicians
identify and improve the processes controlled by the ED.
A good understanding of costs is necessary when using this methodology because
the relationship between time and cost can be inconsistent. Patients who visit the ED
have numerous diagnoses and complex conditions. As mentioned in the article,
many EDs use the emergency severity index, a five-level triage algorithm providing
clinically relevant stratification of patients, to assign patient acuity and for coding
purposes. This tool allows hospitals to bill for the higher cost of treating higher-
acuity patients. Dyas et al. did not use acuity levels or charges to quantify the cost of
care; instead, they considered the underlying processes for any ED patient, regardless
of the patients condition or acuity.
In the example of the A-1 Hospital ED, cost savings were realized by measuring
time savings. As a hospital administrator, I have participated in similar process-
improvement activities. For example, we used these tools to reduce door-to-doc

457
J o u r n al of H ealt hcare M anage ment 60:6 N ove mber /D ecember 2015

times in our ED. We set a goal of 30 minutes or less for a patient to see a provider. As
expected, staff and providers expressed much resistance. Their common request was
to hire more staff to accomplish the feat. After months of working with the ED staff,
ED physicians, and many other interdisciplinary players, we achieved an average wait
time of 13 to 15 minutes, a reduction from the 30-minute wait time 2 years ago. We
were able to accomplish this goal without increasing the number of staff members.
Similar to the time savings in the A-1 Hospital example, by focusing on door-to-
doc time, our time savings resulted in increased patient flow throughout the hospi-
tal, contributing to the cost reduction.
Any effects on operations outside the ED, factors not included in the cost model,
and atypical operations (e.g., variations in care by provider) need to be considered
when deciding to implement a process improvement recommendation. Hospitals are
complex and departments depend on other departments to continue the care of
patients. Any time a process improvement is begun, all of the departments that will
be affected should participate in the process.
The ED is often called the gateway or front door to the hospital, as it is many
patients first encounter with the hospital. Providing the most cost-effective care in
our EDs will assist in ensuring that they remain viable.

458