1-s2.0-s0377221715007912-main
DESCRIPTION
ffvccvbTRANSCRIPT
![Page 1: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/1.jpg)
European Journal of Operational Research 249 (2016) 327–339
Contents lists available at ScienceDirect
European Journal of Operational Research
journal homepage: www.elsevier.com/locate/ejor
Innovative Applications of O.R.
Use of a discrete-event simulation in a Kaizen event: A case study in
healthcare
Chantal Baril a,∗, Viviane Gascon b, Jonathan Miller c, Nadine Côté d
a Department of Industrial Engineering, Université du Québec à Trois-Rivières, 3351 Boulevard des Forges, Trois-Rivières, Québec G9A 5H7, Canadab Department of Management Science, Université du Québec à Trois-Rivières, 3351 Boulevard des Forges, Trois-Rivières, Québec G9A 5H7, Canadac Department of Industrial Engineering, Université du Québec à Trois-Rivières, 3351 Boulevard des Forges, Trois-Rivières, Québec G9A 5H7, Canadad Performance Management Office in a Health Facility, Canada
a r t i c l e i n f o
Article history:
Received 10 March 2014
Accepted 23 August 2015
Available online 31 August 2015
Keywords:
Discrete-event simulation
Business game
Lean approach
Kaizen event
Outpatient clinic
a b s t r a c t
To improve service delivery, healthcare facilities look toward operations research techniques, discrete event
simulation and continuous improvement approaches such as Lean manufacturing. Lean management often
includes a Kaizen event to facilitate the acceptance of the project by the employees. Business game is also
used as a tool to increase understanding of Lean management concepts. In this paper, we study how a business
game can be used jointly with discrete event simulation to test scenarios defined by team members during a
Kaizen event. The aim is to allow a rapid and successful implementation of the solutions developed during the
Kaizen. Our approach has been used to improve patients’ trajectory in an outpatient hematology–oncology
clinic. Patient delays before receiving their treatment were reduced by 74 percent after 19 weeks.
© 2015 Elsevier B.V. and Association of European Operational Research Societies (EURO) within the
International Federation of Operational Research Societies (IFORS). All rights reserved.
1
h
i
t
i
i
a
i
M
c
s
h
S
c
a
J
2
c
B
t
(
a
p
I
c
c
s
v
o
M
s
l
o
b
A
(
b
(
t
u
d
r
n
h
0
A
. Introduction
Considering the increasing needs for services in healthcare,
ospital services must review their practices to improve them and
ncrease their performance. Healthcare facilities look toward con-
inuous improvement approaches such as Lean manufacturing to
mprove service delivery (Houchens & Kim, 2014). Lean manufactur-
ng is a management approach aiming to improve the performance of
n organization by reducing waste, delays, etc. while involving staff
n decision-making. During the last few years, the DMAIC (Define,
easure, Analyze, Improve and Control) problem-solving approach
ombined to six sigma was used jointly with Lean to become Lean
ix sigma.
Operations research techniques and discrete event simulation
ave also been used by healthcare managers (Fone et al., 2003).
ince healthcare services are mostly dynamic and stochastic pro-
esses, discrete event simulation has been more often used to model
nd analyze flows in healthcare processes (Fone et al., 2003; Jun,
acobson, & Swisher, 1999; Mielczarek & Uzialko-Mydlikowska,
012). More recently researchers included a Lean approach to dis-
rete event simulation in a facilitated mode (Robinson, Worthington,
urgess, & Radnor, 2014). Robinson, Radnor, Burgess, and Wor-
hington (2012) describe the role of simulation in a Lean approach
before, during and after a Kaizen event). A Kaizen event is a group
∗ Corresponding author. Tel.: +1 819 376 5011; fax: +1 819 376 5152.
E-mail address: [email protected] (C. Baril).
w
p
O
a
ttp://dx.doi.org/10.1016/j.ejor.2015.08.036
377-2217/© 2015 Elsevier B.V. and Association of European Operational Research Societies (
ll rights reserved.
ctivity, commonly lasting 5 days, in which a team identifies and im-
lements a significant improvement in a process (Lean Enterprise
nstitute, 2014). It is a participative activity and it facilitates the ac-
eptance of the project by the employees. Tako and Kotiadis (2015)
ombine discrete-event simulation, a hard OR approach, with soft
ystems methodology (SSM) in order to incorporate stakeholder in-
olvement in the simulation study lifecycle.
Business games are also used as a tool to increase understanding
f Lean management concepts (Ashenbaum, 2010; Billington, 2004;
artin, 2007; Swanson, 2008). van der Zee and Slomp (2009) as-
ert that they could help workers find solutions for specific prob-
ems, or to familiarize themselves with and ease their acceptance
f new work methods or systems. Originally, business games have
een used to help find solutions in different business environment.
business game has been defined by Greco, Baldissin, and Nonino
2013) as “a game with a business environment that can lead to one or
oth of the following results: the training of players in business skills
hard and/or soft) or the evaluation of players’ performance (quan-
itatively and/or qualitatively)”. The business game allows a better
nderstanding of complex problems. The pedagogical principle un-
erlying the business game is involving participants in a virtual envi-
onment. Business games were originally developed to educate busi-
ess managers. They reproduce a process in a virtual environment
hile being inspired by reality. It can also be used to let employees
erform a task or a given operation for real (Ellis, Goldsby, Bailey, &
h, 2014). However business games can be helpful to educate man-
gers, employees and change agents in healthcare or education. The
EURO) within the International Federation of Operational Research Societies (IFORS).
![Page 2: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/2.jpg)
328 C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339
Ini�atesimula�on study
Define system Workshop 1
Specifyconceptual model
Workshop 2Model coding
Experimenta�onWorkshop 3
Implementa�onWorkshop 4
Modelling teamStakeholdersModelling team
StakeholdersModelling team Modeller
StakeholdersModelling team
StakeholdersModelling team
Fig. 1. PartiSim (Tako & Kotiadis, 2015).
R
e
c
Y
i
P
S
e
J
G
a
i
f
d
c
i
s
M
O
S
i
c
p
m
m
t
i
T
T
c
e
t
v
o
a
o
s
h
T
g
t
i
l
i
(
o
c
s
fl
o
t
u
h
play themes are not only related to enterprise strategy but address
other topics such as quality, work organization, planning, safety at
work or project management.
In this paper, we ask how it is possible to facilitate rapid imple-
mentation of solutions found in a kaizen event and reduce barriers
when implementing Lean in healthcare. We believe that discrete-
event simulation, business game and the involvement of the project
team can play a crucial role in achieving those goals. In doing so,
we present an approach to implement a Lean project according to
the DMAIC problem-solving procedure (de Mast & Lokkerbol, 2012).
The DMAIC approach has been used to analyze patients’ trajectory in
an outpatient hematology–oncology clinic in order to propose and
implement improvements aiming to reduce patients’ waiting time
when receiving a chemotherapy treatment. Data were gathered only
3 months after the Kaizen event to evaluate the impact of the modi-
fications implemented. A Kaizen event was organized at the Improve
step during which a business game was used to help find solutions.
During the Kaizen event, discrete event simulation was used to test
how the solutions could modify patient waiting times. We describe
the role of each stakeholder in this approach and how it is facilitative
and participative. We also explain the advantages of using a business
game and discrete event simulation during a Kaizen event.
The paper is organized as follows. Section 2 presents relevant liter-
ature while Section 3 presents the methodology. The implementation
of our approach and results are described in Section 4. Finally the re-
sults are discussed in Section 5 and Section 6 presents the conclusion
of our research.
2. Relevant literature
The heart of Lean consists in preserving value with less work by
the identification and elimination of “waste” and in developing stan-
dardized, reliable processes. This is performed in a context of con-
nectedness, respect, and growth of all employees who are trained to
identify waste and errors, and suggest possibilities for improvements
that will be tested using scientific methods. Lean seems to be an ef-
fective way of improving healthcare organizations and the growing
number of implementations and reports found in the literature rein-
force this view (Brando de Souza, 2009)
Lean implementation in healthcare requires adaptation and de-
velopment to fit the specific context and allow healthcare staff to
take ownership of the approach (Poksinska, 2010). Literature re-
view shows that there have been some significant tangible out-
comes in healthcare organizations that adopted Lean principles such
as increased patient throughput (Dickson, Singh, Cheung, Wyatt, &
Nugent, 2008; Van Lent, Goedbloed, & Van Harten, 2009), reduced
waiting times (Al-Araidah, Momani, Khasawneh, & Momani, 2010;
Lodge & Bamford, 2008) and improvements in work environment
(Kaplan & Patterson, 2008; Nelson-Peterson & Leppa, 2007). How-
ever, many papers identified barriers when implementing Lean man-
agement in healthcare organizations such as lack of ownership of
proposed processes, skepticism and resistance to change (Brandão de
Souza & Pidd, 2011; Proudlove, Moxham, & Boaden, 2008; Radnor,
Walley, Stephens, & Bucci, 2006).
In the last few years, discrete-event simulation has been con-
sidered as an interesting tool to help improving healthcare services
(Brailsford, Harper, Patel, & Pitt, 2009; Fone et al., 2003; Mielczarek &
Uzialko-Mydlikowska, 2012) as in outpatient clinics (Jun et al., 1999;
ohleder, Lewkonia, Bischak, Duffy, & Hendijani, 2011). Discrete-
vent simulation has been applied to solve a wide variety of health-
are problems such as patient appointment systems (Klassen and
oogalingam, 2009; Ogulata, Cetik, & Koyuncu, 2009), patient wait-
ng time (Paul, Reddy, & De Flitch, 2010; Santibanez, Chow, French,
utterman, & Tyldesley, 2009), patient flow, (Rohleder et al., 2011;
epulveda, Thompson, Baesler, Alvarez, & Cahoon, 1999; White
t al., 2011), operational performance (Berg et al., 2009; Griffiths,
ones, Read, & Williams, 2010) and others problems (Hagtvedt,
riffin, Keskinocak, & Roberts, 2009; Katsaliaki & Brailsford, 2007).
Too often, discrete event simulation models have been developed
nd used by experts to find solutions without involving stakeholders
n the development process. Recently, more work has been done on
acilitated modeling to involve stakeholders in the development of
iscrete event simulation models.
Facilitated modeling consists in developing models jointly with a
lient group: from defining the nature of the problem, to support-
ng the evaluation of priorities and development of plans for sub-
equent implementation (Franco & Montibeller, 2010). Franco and
ontibeller (2010) discuss in detail facilitated modeling as an
R intervention tool in organizations. Jahangirian, Taylor, Eatock,
tergioulas, and Taylor (2015) examine the stakeholder engagement
n the context of healthcare simulation. They find that “communi-
ation gap between simulation and stakeholder groups” is the top
rimary factor contributing the most to the poor stakeholder engage-
ent in healthcare simulation projects, followed by “poor manage-
ent support”, “clinician’s high workload” and “failure in producing
angible and quick results”. Recently managers began to be included
n problem definition and process modeling (Kotiadis et al., 2013;
ako, Kotiadis, & Vasilakis, 2010a; Tako, Kotiadis, & Vasilakis, 2010b).
his participation is especially important for studies in healthcare
haracterized by the presence of many stakeholders with tacit knowl-
dge of their part of the system and often multiple views and objec-
ives. Tako and Kotiadis (2015) combined the steps required to de-
elop a discrete event simulation model with the participative steps
f Soft Systems Methodology. Their whole procedure, called PartiSim,
llows stakeholders to be involved at every stage of the model devel-
pment and experiments (except for programming which requires a
pecific expertise). Fig. 1 presents the PartiSim steps in which stake-
olders are involved through workshops. Our approach differs from
ako and Kotiadis (2015) because it includes the Lean project steps to-
ether with the development of the discrete event simulation model.
There are few studies on the use of discrete event simulation
hrough a Lean approach, or other continuous improvement process,
n healthcare systems. Young et al. (2004) propose the use of simu-
ation to evaluate the benefices of a continuous improvement project
n healthcare before the implementation. Khurma, Bacioiu, and Pasek
2008) present a discrete event simulation model to study the impact
f a Lean project in an emergency unit. Even if they do not specifically
onsider Lean management, Proudlove, Black, and Fletcher (2007)
how how a simple simulation model can improve efficiently patient
ows. Robinson et al. (2012) explore potential complementary roles
f discrete event simulation and a Lean approach in healthcare sys-
ems. Their model, SimLean, defines three roles for discrete event sim-
lation used with Lean: education, facilitation and evaluation (Fig. 2).
According to Robinson et al. (2012), discrete event simulation can
ave an educational function in teaching Lean principles (Educate
![Page 3: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/3.jpg)
C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339 329
SimLean EducateRole: educate Lean event
SimLean FacilitateRole: engage/facilitate
SimLean EvaluateRole: experiment/evaluate
Before During A�er
Fig. 2. SimLean model: the roles of discrete-event simulation and Lean in healthcare (Robinson et al., 2012).
Overview of the current processIntroduc�on to simula�on (SimLean Educate)GembaMapping the processEs�mate �me for each ac�vity
Facilitated modelling using the SimLean approach (Robinson et al., 2014)
Model codingSimple model developed
Recap of the first dayModel briefly explainedFace valida�onProblem scopingImprovement
Day1kaizen
Day2 Day3Kaizen
SimLean Facilitate
Kaizen (2 days)
Fig. 3. Facilitated modeling using the SimLean approach (Robinson et al., 2014).
s
p
m
e
u
n
d
c
c
t
t
t
i
s
t
n
c
c
i
fi
(
K
c
s
w
e
s
t
t
a
a
T
o
i
a
(
t
s
m
t
v
s
m
f
m
o
3
o
i
p
c
K
p
i
p
a
c
fi
t
tep, Fig. 2). It is a mean to understand the dynamics of a healthcare
rocess. One or many predefined standard discrete event simulation
odels can be used to teach Lean principles before or during a Kaizen
vent. During Lean events, processes must be analyzed through the
se of a process mapping. Simulation can be used to create a dy-
amic version of the process mapping (Facilitate step, Fig. 2). A simple
iscrete-event simulation model developed during the Kaizen event
an be used to better understand the process dynamics and to en-
ourage participants to propose improvement solutions. With a de-
ailed discrete event simulation model, different scenarios can be
ested. Developing a detailed discrete event simulation model may
ake a long time since it requires obtaining data, modeling and val-
dating the process, validating the model and generating improving
olutions. These steps are usually performed after the Kaizen event
o test solutions found by the participants and to eventually propose
ew ones (Evaluate step, Fig. 2). The discrete event simulation model
an also be used during the implementation phase and plays a role in
ontinuous improvement. In fact, the model in Robinson et al. (2012)
mplies defining three discrete event simulation models: (1) prede-
ned models (Educate step), (2) simple model (Facilitate step) and
3) detailed model (Evaluate step).
Unlike Robinson et al. (2012) who use a simple model during the
aizen event to better understand the process dynamics and to en-
ourage participants to propose improvement solutions, a detailed
imulation model was considered to measure the impact on patients
aiting time of solutions found by the participants during the Kaizen
vent. Since Robinson et al.’s model implies developing the detailed
imulation model (Evaluate step) after the Kaizen event, it increases
he delay before implementing solutions, while it is crucial to start
he implementation on a short delay after the Kaizen event to guar-
nty its success (Martin & Orsterling, 2007).
The paper of Robinson et al. (2014) focuses on SimLean facilitate as
n example of facilitated modeling using discrete-event simulation.
he simulation model was developed and used within a 3 days period
f an improvement workshop. However, they had to build the model
n the “back office”, meaning that a fully facilitated model was not
chieved. During the Kaizen event the simple but not validated model
since it was built with estimated times) helped in suggesting solu-
ions. Consequently Kaizen participants were able to immediately get
ome feedback on how their ideas could improve the system perfor-
ance even though it could not be measure precisely. Fig. 3 presents
heir facilitated approach.
Unlike Robinson et al. (2014) our Kaizen activity is entirely de-
oted in finding solutions that could improve the whole process and
atisfy stakeholders (doctors, nurses, etc.). Data collection (process
apping, time study) and the simulation model are determined be-
ore the Kaizen event. As in Robinson et al. (2014), our simulation
odel was not defined during the Kaizen event but rather in “back
ffice”.
. Overview of our approach
This paper presents an approach to conduct a Lean project in an
ncology clinic during which a detailed simulation model is defined
n order to validate improving ideas proposed by the Kaizen event
articipants. The DMAIC solving problem method was used in a fa-
ilitate mode. Moreover a business game was performed during the
aizen event to facilitate employees’ involvement and be able to im-
lement improvement ideas more quickly. The approach is presented
n Fig. 4.
The DEFINE step allows specifying the project and determining
erformance indicators. In order to achieve this, a project charter
nd interviews with the personal clinic were realized. The project
harter defines the team vision: problem statement and objective,
nancial impact, project scope, schedule and team members. Once
he project charter finished, the project itself can begin. A meeting
![Page 4: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/4.jpg)
330 C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339
Define Measure Analyze Improve Control
Project charter
Day 1 Day 2 Day 3
Project launch and interviews
Project team
Process mapping
Time study
The clinic
Development of Simula�on model
Development of businessl game
Kaizen event
Lean and kaizen presenta�onsMeasure step presenta�onProject objec�veProcess irritants and wasteWorkshops iden�fica�on
Recap of the day beforeIdeal processWorkshopsBusiness gameSimula�on model
Recap of the day beforeWorkshopsIndicatorsAc�on plan
Our approach
SimLean and facilitated modelling
Fig. 4. Our approach.
o
(
t
y
t
t
c
4
t
4
4
4
f
t
s
r
with the clinic’s employees is then planned to explain the project and
present the team.
Next, each clinic employee (nurses, doctors, pharmacists, clerks,
assistant technicians) is met individually in order to
• Know how they perceive the problem and how it affects their
work• Identify other problems than the one identified in the project
charter• Better understand working relations and teamwork• Determine what they expect from the project
The aim of these interviews is to assure that the employees share
the same objective and work together to reach it rather than trying to
reach individual goals.
The MEASURE step consists in measuring the current process per-
formance.
Data collection is conducted during a representative working
week and included appointment scheduling, work schedules, treat-
ment capacity for a better understanding of the clinic. Data is next
used to describe the process and develop the discrete-event simula-
tion model and business game.
In the ANALYZE step are identified the causes of the problems on
which the team will work during the IMPROVE step. A detailed dis-
crete event simulation model is built to help find solutions. Because
human judgment is not taken into account in discrete event simula-
tion and has a great influence on process efficiency (Bok, 2007), we
developed a business game.
To IMPROVE the process efficiency, a Kaizen event was planned so
that each team member could participate in finding solutions. During
this Kaizen event, the business game and the detailed discrete event
simulation model were used to evaluate the different solutions pro-
posed by the team members. The main output of the Kaizen event is
an action plan to implement the selected scenario immediately after
the event.
The objective of the CONTROL step consists in making sure that
the new process will remain efficient. The performance indicators
must be measured over time to verify the process stability and take
actions if necessary.
4. Implementation of our approach in a hematology–oncology
clinic
Our approach was implemented in a hematology–oncology clinic
when carrying out a Lean project. The clinic under study offers
ncology treatments (chemotherapy) and hematology treatments
hemoglobin transfusions, blood transfusions, phlebotomy, coagula-
ion factors). On average, 8500 treatments are administered every
ear. The clinic is opened from 8h00 AM to 8h00 PM. Seven doc-
ors, six nurses, one clinical nurse, two to three pharmacists and two
o three assistant pharmacists (ATP) depending on the day, and two
lerks work in the clinic. Nurses work on 8-hour shifts (8h00 AM to
h00 PM, 9h00 AM to 5h00 PM, 10h00 AM to 6h00 PM and 12h00 PM
o 8h00 PM).
.1. Define
.1.1. Project team
The project team consists of:
• University team: has the expertise to collect data and build the
simulation model.• Lean facilitator: a member of the clinic who acts as a Lean ex-
pert. This person leads the project, makes the interviews and leads
meetings and the Kaizen event.• Intervention team: all employees (nurses, doctors, pharmacists,
clerks, assistant pharmacists).• Kaizen team: two nurses, one clerk, two doctors, one pharmacist,
one assistant pharmacist, one employee from the informatics de-
partment, the chief laboratory, the Lean facilitator and the univer-
sity team.• Clinic manager: to ease the implementation of the solutions and
make the connection between the hospital managers and the in-
tervention team.• Hospital managers: control decisions.
.1.2. Project charter
The project charter was developed by hospital managers, the Lean
acilitator, the clinic manager and the university team (Table 1). At
his step it was decided to create a detailed simulation model to
tudy the patient trajectory instead of a simple one for the following
easons:
• Encourage participation of many persons with varied expertise.• Make it easier for the intervention team to accept improvement
ideas since they will have been tested on a more realistic model.• Measure precisely the impact of the solutions proposed to reach
the objective knowing that with a detailed model the error margin
is less than with a simple one regarding the reduction of patients’
waiting time (Bowers, Ghattas, & Mould, 2012).
![Page 5: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/5.jpg)
C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339 331
Arrival
Registra�on Blood sample Mee�ng withthe doctor
Making an appointment
Taking charge treatment
Exit
Registra�on
Registra�on
Registra�on
Blood sampleTaking charge
treatment
Taking charge treatment
Mee�ng withthe doctor
Making an appointment
Day 1
1
2
3
4
Fig. 5. Mapping of patient trajectories in the outpatient hematology–oncology clinic.
Table 1
Project charter.
Problem The outpatient hematology–oncology clinic has experienced
an increase of patients waiting time since fall 2010. Some
patients wait up to 4 hours before receiving their
treatment, after their arrival at the clinic. Many patients
expressed their dissatisfaction by lodging a complaint to
the local complaint commissioner. A significant increase in
the personnel workload has also been noticed, resulting in
an increase in overtime the last 18 months
Objective Reduce patients waiting time
Scope Review the patient trajectory from the arrival up to the
departure from the clinic
Impact Managers and department chief wish to increase patients
and employees’ satisfaction and clinic’s performance
Schedule 9 months (January to September)
Team members Hospital managers
Lean facilitator
Clinic manager
Intervention team (nurses, doctors, clerks, pharmacists,
assistant pharmacists)
University team
4
w
t
w
a
t
4
4
s
(
(
(
(
(
o
m
p
c
c
a
(
a
w
q
t
b
4
i
p
p
e
m
(
t
o
s
s
r
t
m
t
t
n
.1.3. Project launch and interviews
The project launch meeting was led by the Lean facilitator during
hich the project charter was also presented. Employees showed en-
husiasm toward the project and a willingness to collaborate. After-
ards individual interviews were held with all employees involved
nd they led to following findings:
• Personnel agree on the main problem being too long patients
waiting time. Doctors must necessarily agree with the objective
of reducing patients waiting time and not have individual objec-
tives.• Doctors must be involved in each step of the project (DMAIC).
Good relations among personnel members should contribute to
the project’s success.
These interviews indicated that the intervention team was ready
o put the required effort to turn this project into a success.
.2. Measure
.2.1. Process mapping
The complete trajectory followed by a patient comprises five
teps:
1) The patient arriving at the clinic registers with the clerk.
2) The patient waits for a blood sample which is sent to the labora-
tory to be analyzed.
3) Once the blood tests results are available, the patient may need to
meet the doctor.
4) After meeting with the doctor, the patient makes another appoint-
ment.
5) The patient receives his treatment after being taken care of by the
nurse, if his health status allows for it, otherwise the treatments
are given another day.
Not all patients need to follow the five steps. Some of them come
nly for a blood sample and to meet the doctor, others only for treat-
ents or, only to meet the doctor if they previously had a blood sam-
le. As the process mapping shows (Fig. 5), four different trajectories
an be followed by patients:
1. Follow-up and taking charge (5 steps): registration, blood sam-
ple, meeting with the doctor, making an appointment and taking
charge; this trajectory is followed by 20 percent of patients.
2. Blood sample and taking charge (3 steps): registration, blood
sample and taking charge; this trajectory is followed by
14 percent of patients.
3. Treatment (2 steps): registration and taking charge; this pathway
is followed by 19 percent of patients.
4. Meeting with the doctor (3 steps): registration, meeting with the
doctor and making an appointment; this trajectory is followed by
47 percent of patients.
Treatments can last from 15 minutes to up to 8 hours. Patients
oming to the clinic to receive treatments represent 53 percent of
ll patients. The taking charge step is followed by three sub-steps:
1) meeting with the pharmacist, (2) hydration and premedication
nd (3) treatment.
The process mapping was first realized by the Lean facilitator. It
as next posted in the clinic for the intervention team to get ac-
uainted with it. Then during a meeting led by the Lean facilitator
hey validated the process mapping to make sure that no steps had
een forgotten.
.2.2. The clinic
Appointment scheduling provides information on patients com-
ng to the clinic. Three appointment lists are considered: (1) list of
atients needing a blood sample (only in the afternoon), (2) list of
atients needing to meet a doctor (fixed periods and last 20 minutes
ach from 8h00 AM to 1h30 PM) and (3) list of patients needing treat-
ents.
The first treatments of the day are scheduled every 15 minutes
regardless of the treatment duration). Later in the day, the time be-
ween scheduled appointments for treatments may vary depending
n the end of the previous one. Ad hoc rules are used by clerks to
chedule the appointments. Clerks must consider doctors’ working
chedules, patients ‘preferences and opening hours of the treatment
ooms. Appointment scheduling showed that the number of planned
reatments is on average equal to 37, the average daily number of ad-
inistrated treatments is 32 and the average percentage of canceled
reatments is equal to 14 percent. Results in Table 2 show that when
reatments are scheduled, the capacity (number of hours available) is
ot always considered.
![Page 6: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/6.jpg)
332 C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339
Table 2
Number of hours for treatments (planned and available) and number of hours available.
Monday Tuesday Wednesday Thursday Friday
Number of hours of
scheduled treatments
89.00 91.00 135.25 100.50 64.50
Number of hours
available (capacity)
128 128 128 128 128
Table 3
Nurses’ schedules and number of patients waiting.
Time Working schedules Number of Treatment Number of
Nurse 1 Nurse 2 Nurse 3 patients ready room capacity patients waiting
8h00 1 1 0
8h15 3 1 2
8h30 3 1 4
8h45 2 1 5
9h00 1 1 5
9h15 0 1 4
9h30 2 1 5
9h45 2 1 6
10h00 1 1 6
10h15 5 1 10
10h30 2 1 11
10h45 1 1 11
11h00 0 1 11
Table 4
Summary of the time study for the four trajectories (in minutes).
Trajectories Total lead time from Total lead time from
registration to start of registration to taking
meeting the doctor charge in treatment room
1. Follow up/treatment 74.76 174.60
2. Blood sample/treatment n/a 114.65
3. Treatment n/a 60.73
4. Meeting the doctor 50.12 n/a
p
T
s
4
c
c
b
w
w
l
t
w
j
c
f
fi
n
p
r
u
e
n
c
t
On Wednesday, for instance, treatments were scheduled for the
equivalent of 135.25 hours while there were only 128 hours available
(16 chairs × 8 hours/chair). However, on Thursday, more treatments
could have been scheduled.
The treatment room capacity depends on nurses’ schedules. Each
nurse is responsible for four patients (or 4 chairs). Sixteen chairs are
available for treatments. Treatment chairs are available when nurses
are working (from 8h00 AM to 4h00 PM, for instance). Between 6h00
and 8h00 PM there is only one nurse at work. However, at least two
nurses should be at work at the same time in case a patient has health
problems during his treatment. Table 3 shows an example of the link
between nurses’ schedules and patients waiting.
We can see that the number of patients ready for treatment often
exceeds the treatment room capacity in the morning. Patients ready
for treatment follow trajectories 1, 2 or 3. On trajectory 1, seven doc-
tors can receive patients. This implies that seven patients could be
ready for treatment at the same time while the treatment room has
the capacity to receive only one every 15 minutes from 8h00 to 11h00
AM. The number of patients waiting increases during the same time
leading in a long waiting time. It is thus important to coordinate ap-
pointments with doctors with treatment appointments and to take
into account the treatment room capacity.
Considering that treatments can be long (up to 8 hours), that there
must be at least two nurses to take care of one patient and that
meeting with doctors are from 8h00 AM to 1h30 PM, it is impor-
tant to schedule treatments as early as possible. However actually,
nurse schedules do not allow beginning treatments early enough in
the morning.
Patient trajectories were also analyzed. Let us consider trajec-
tory 1: registration, blood sample, meeting with the doctor, making
an appointment and treatment. A patient following this trajectory
must make two appointments: one to meet the doctor and one for
the treatment. We were interested in verifying if the treatment ap-
pointment was coherent with the appointment with the doctor. The
computerized appointment system does not allow scheduling more
than one treatment at the same time (even if there are four nurses
working at the same time). The clerk is forced by the system to en-
ter a fake appointment time in order to provide a list of patients for
nurses working in the treatment room. Nurse providing treatments
cannot rely on the appointment schedules to determine the next
atient to see. They rather see patients on a first come, first call basis.
he actual computerized appointment system is not consistent when
cheduling appointments with doctors and treatments.
.2.3. Time study
Our time study consisted in determining patient lead times ac-
ording to their trajectory (Fig. 5). Data collection was done at the
linic over 1 week. Each patient coming to the clinic received a num-
ered chip. At every step of his trajectory he had to identify himself
ith his number and the observer would take note of the time he
ent through the step. Table 4 provides a summary of the trajectory
ead times.
The total lead time from registration until meeting with the doc-
or could be computed for only two trajectories. For trajectory 1, total
aiting time is 69.31 minutes, 93 percent of total lead time. For tra-
ectory 4, it is 48.50 minutes, 97 percent of total lead time. In both
ases waiting time is considered too high. Reducing total lead time
rom registration to beginning preparing for treatment was identi-
ed as the main objective during the Kaizen event. Waiting times are
oticed at every step of the process. The time study provided other
erformance indicators (Table 5).
The average time required to prepare the patient in the treatment
oom is 20 minutes while appointments are scheduled every 15 min-
tes. This shows inconsistency in scheduling appointments. The av-
rage treatment chair utilization rate is 68 percent, showing that the
umber of treatment chairs is adequate and that many more patients
ould receive treatments.
Finally, patients are asked to arrive 30 minutes before meeting
he doctor, to leave enough time for blood sampling. Our time study
![Page 7: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/7.jpg)
C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339 333
Table 5
Other performance indicators.
Indicators Mean value
Average time to prepare patient in the treatment room 20 minutes
Average chair utilization rate in the treatment room 68 percent
Mean time to analyze a blood sample 31 minutes
Percentage of blood sample made the previous day 22 percent
Number of complaints in 2011–2012 9
s
3
m
e
d
a
w
w
f
4
4
A
(
(
e
s
T
s
s
o
i
t
i
t
r
p
a
t
t
t
f
L
w
i
a
o
s
t
t
t
H
t
t
p
4
s
K
m
r
u
l
a
b
w
u
t
o
e
d
4
4
t
K
fi
p
a
t
d
e
hows that the average time required to analyze a blood sample is
1 minutes. If we add the waiting time before the blood sample (21
inutes) to these 31 minutes, it is obvious that patients should arrive
arlier for the blood sample results to be ready before meeting the
octor. Patients having their blood sample taken the day before their
ppointment with the doctor or before their treatment don’t have to
ait for the blood sampling results. It also reduces the amount of
ork in the clinic. With only 22 percent of the blood samples per-
ormed the previous day, there is place for improvement.
.3. Analyze
Data analysis indicates that:
• Appointment schedules do not take into account the treatment
room capacity and treatment durations.• Taking charge of patients for treatments is done between 8h00
AM and 12h00 PM depending on nurses’ schedules.• Patients needing a blood sample before meeting the doctor or
before their treatment usually wait an additional 30 minutes for
blood analysis to come back from the laboratory.• A patient is taken in charge for treatment every 20 minutes but
appointments are planned every 15 minutes.• Each step generates delays.
.3.1. Development of a detailed discrete event simulation model
The discrete event simulation model was developed using the
rena software (Kelton, Sadowski, & Sturrock, 2007) in order to
1) evaluate improving scenarios before the Kaizen event and
2) measure the impact of the ideas expressed during the Kaizen
vent on patient lead times. The statistical distributions used in the
imulation model are based on data collected from the clinic process.
hese statistical distributions provided by Input analyzer of the Arena
oftware are the ones fitting best the data considering the mean-
quare error. In our simulation model, entities (patients) follow one
f the different trajectories, according to the percentages presented
n Section 4.1. The model was developed to be as close as possible
o the real process. Comparisons between simulated and real wait-
ng and lead times were performed for all trajectories. A ±10 percent
hreshold was considered to take into account the margin of error
elated to the statistic curves and the error induced by a simplified
rocess. Table 6 shows that the gap between our simulation model
nd the real process is at most ±10 percent confirming the validity of
he model to be used to test scenarios.
Table 6
Validation of the discrete event simulation model.
Steps Waiting time
Real Simulated Gap (p
Registration 3.68 3.63 −1.4
Blood sample 21.85 22.06 +1.0
Meeting the doctor 44.30 44.68 +0.8
Making an appointment 4.18 4.19 +0.2
Preparing for treatment 79.70 74.60 −6.4
The simulation model was built by the university team. However
he clinic personnel were involved in many aspects to understand all
he specifics of the process. The final model was validated by the Lean
acilitator, the clinic manager, one doctor and a nurse. Unlike the Sim-
ean approach (Fig. 1), the detailed discrete-event simulation model
as developed before the Kaizen event. It was used to make prelim-
nary tests to avoid unnecessary discussions during the event. The
nimator of the activity was able to evaluate objectively the impact
f adding a nurse to blood sampling and increasing the size of blood
amples shipments, on patients waiting times. These two proposi-
ions were tested with the simulation model. Results showed that
here was no significant improvement. Therefore discussion among
he Kaizen event participants could be moved toward other subjects.
owever different scheduling appointment rules, taking into account
he treatment room capacity, had to be evaluated with the simula-
ion model. They had to be determined with the help of the clinic
ersonnel.
.3.2. Development of a business game
A business game was developed to allow participants to really
chedule appointments according to the rules defined during the
aizen event and to measure with the discrete event simulation
odel how it impacted on patient waiting times. A set of cards, each
epresenting a patient with his characteristic, was prepared to sched-
le the appointments of a typical day. Each card contained the fol-
owing information:
1. Fictitious patient’s name
2. Need of a blood sample or not
3. Name of doctor to meet
4. Treatment type (chemotherapy or else) and duration
5. Need to make another appointment with the doctor or not
The game will work as follows: participant will choose randomly
card and schedule the appointment according to the patient’s need:
lood sample, meeting the doctor or treatment. Patients arrival rates
ill be modified in the simulation model depending on the sched-
les generated. A member of the university team will run the simula-
ion model to measure the impact of the new appointment schedules
n patients waiting time and present the results. For the game to be
fficient during the Kaizen event, it should be able to integrate the
ifferent arrival rates rapidly.
.4. Improve
.4.1. Kaizen event
To improve the process efficiency, a Kaizen event was planned so
hat each team member could participate in finding solutions. The
aizen team includes 10 persons: two nurses, one administration of-
cer, two doctors, one pharmacist, one assistant pharmacist, one em-
loyee from the department of informatics, one laboratory manager,
nd one clinic manager. The Kaizen event was led by the Lean facili-
ator together with a member of the university team.
The objective of the Kaizen event after data analysis was to re-
uce patient waiting times for treatments by 45 percent while light-
ning the whole process. Even though this objective may seem too
Lead time (waiting and service)
ercent) Real Simulated Gap (percent)
5.30 4.90 −7.5
31.00 29.90 −3.5
92.32 87.00 −5.8
98.90 97.60 −1.3
107.10 110.10 +2.8
![Page 8: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/8.jpg)
334 C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339
Arrival ExitRegistra�onTaking charge
treatment
Day 1
Arrival
Registra�onBlood sample
hospitalMee�ng with
the doctorMaking an
appointment
Exit
Registra�on
Registra�on
Blood samplehospital
Mee�ng withthe doctor
Making an appointment
Day 0
1
2
4
Blood sampleregion
3
Fig. 6. New mapping of patients trajectories.
a
n
w
l
a
w
u
w
T
l
a
t
D
r
d
m
F
t
t
d
p
s
w
p
a
c
d
i
t
a
m
t
p
t
w
t
f
j
ambitious, it must be given all resources and effort involved for the
project. This objective is related to trajectories 1 (follow-up and treat-
ment), 2 (blood sample and treatment) and 3 (treatment) from Fig. 3.
The targets to be achieved were evaluated using results from Table 3
and they are 96 minutes for trajectory 1, 63 minutes for trajectory 2
and 34 minutes for trajectory 3. The Kaizen event was held on three
consecutive days.
Day 1
The activity begins with the presentation of the objective by a
member of the management team: reduce patients waiting time by
45 percent. The Lean facilitator presents the Lean approach and the
results obtained after the MEASURE step (process mapping, time
study, etc.). These presentations assure that all Kaizen members re-
ceive the same information. Participants are invited to discuss and
modify the process mapping if necessary. Process irritants and waste
are next identified by teams of two to three persons. They are written
on post-it. Each team share their findings and explain each irritant
when a member places the paper on the process mapping poster.
Then participants agree on 4–6 main topics to group the irritants.
These topics are written on an Ishikawa diagram and each irritant
is put next to the appropriate topic.
Participants identified the following six topics to discuss in
workshops:
1. Nurses work organization
2. Patients trajectories
3. Equipments and work sites
4. Tools and information flow
5. Taking charge of patients by nurses
6. Scheduling appointments
Workshops whose results had more impact on patients waiting
time are 1, 2 and 6.
At the end of the day, the Kaizen team finds a name to the project.
It is a way for the team to take ownership of the project. Finally, the
hospital manager is invited to attend the meeting. One of the partici-
pants summarizes the work done during the day. Since the direction
members do not attend the Kaizen event it is a way to keep them in-
formed and show the Kaizen team how the project is important to
managers.
Day 2
Day 2 begins with the sum up of the previous day. Then the Lean
facilitator asks the Kaizen team to think of an ideal process and iden-
tify the constraints to reach it. For instance, the ideal process may
need a real time follow-up of the patient trajectory which can be
costly. However it might be possible to have a real time follow-up for
one of the critical steps of the trajectory. Identifying the ideal process
llows the Kaizen team to clearly determine actions and decisions
eeded to get solutions. Next three teams of three participants each
ork on three topics identified from the Ishikawa diagram to find so-
utions to the irritants. They present their solutions to the other teams
nd discussion is undergone to find new ideas. Here follows results of
orkshops 1, 2 and 6.
Workshop 1: Nurses work organization
During this workshop, participants proposed new nurses sched-
les. The treatment chairs’ availability depend on nurses’ schedules
hich was modified to be from 8h00 AM to 4h00 PM for all nurses.
his modification allows for the beginning of more treatments ear-
ier in the day. Real data showed that it was taking 20 minutes on
verage to prepare a patient (Table 5). This was due to the fact that
reatments were not ready when the patient was ready to receive it.
uring the workshop, pharmacists assured that treatments could be
eady by 8h00 AM for the first incoming patients. Consequently it was
ecided to continue scheduling appointments every 15 minutes.
Workshop 2: Patients trajectories
Participants proposed to divide trajectories 1 (follow-up and treat-
ent) and 2 (blood sample and treatment) over 2 days. As shown in
ig. 6, blood sample and meeting with the doctor (day 0) are planned
he day before treatments (day 1).
These new trajectories allow a reduction in treatment cancella-
ion rate on the same day (day 1) due to bad blood results (day 0). A
eal was made with healthcare providers in specific regions to allow
atients to have their blood sample close to their home (day 0) in-
tead of coming to the hospital, therefore eliminating the 30 minute
aiting time before meeting the doctor.
Workshop 6: Scheduling appointments
The business game was mainly used to test ideas to improve ap-
ointment scheduling. Participants to the game were a clerk, a doctor,
n assistant pharmacist technician and the clinic manager. Let us re-
all that the current appointment planning worksheet (treatment) is
ivided into 15 minute time slots from 8h00 AM to 8h00 PM and that
t is impossible to schedule two appointments at the same time. The
reatment room capacity is not taken into account.
The game is quite simple. First, participants were asked to cre-
te their own appointment planning worksheets for blood sample,
eeting doctors and treatment, according to different criteria. Pa-
ients used to arrive for blood sampling 30 minutes before their ap-
ointment with the doctor while it was taking around 50 minutes
o perform and analyze the blood samples. Therefore appointments
ere also added for blood samples (on morning) to reduce waiting
imes for blood sampling. The idea was to coordinate appointments
or blood samples and with doctors according to the different tra-
ectories. The appointment planning worksheets for treatments were
![Page 9: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/9.jpg)
C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339 335
Table 7
Results.
Patient waiting times before treatment
Mean observed Standard deviation Mean observed Standard deviation Simulated
(before Kaizen) (before Kaizen) (after Kaizen) (after Kaizen) (during Kaizen)
Trajectory 3: treatment (day 1) 61 minutes 52 minutes 16 minutes (−74 percent) 7 minutes 6 minutes (−90 percent)
d
n
i
s
a
i
t
o
d
a
p
n
t
s
i
d
l
t
p
s
T
t
a
t
b
t
C
t
p
t
p
r
t
i
w
i
t
o
p
h
t
i
4
p
T
t
b
m
r
r
t
1
k
e
b
i
n
e
u
i
o
(
t
4
b
i
b
5
5
p
S
K
v
t
v
n
m
s
K
i
o
d
s
t
a
s
s
p
s
f
c
fl
f
i
w
m
o
esigned to consider the number of treatment chairs available and
urses’ schedules (workshop 1). These modifications allowed tak-
ng into account the treatment room capacity when building new
chedules by planning as many patients at the same time as there
re nurses and considering treatment times. The computer special-
sts confirmed that these new worksheets would be included into to
he current appointment planning software. Moreover they could test
ther priority rules such as scheduling patients with longer treatment
urations at the beginning of the day.
Second, the cards were randomly sorted to simulate demands for
n appointment. Finally, clerks had to pick a card and assign an ap-
ointment time to the fictitious patient using the appointment plan-
ing worksheet. Results during business game with the cards show
hat 8 of 94 had to be scheduled on another day. It was impossible to
chedule all patients on the same day since capacity was now taken
nto account.
The appointment schedules for blood samples, meeting with the
octor and treatments were used as inputs in the discrete event simu-
ation model to measure their impact on patient waiting times. Since
rajectories 1 and 2 were divided over 2 days (workshop 2), it was im-
ossible to compare results before and after testing improving propo-
itions. Only trajectory 3 (treatment) could be analyzed more closely.
he discrete event simulation model was replicated 100 times. Pa-
ient waiting times before treatment were reduced by 90 percent on
verage (from 61 minutes to 6 minutes). Participants could see that
he new schedules had a significant impact on patient waiting times
efore treatments.
One of the participants summarizes the work done at the end of
he day to the direction members.
Day 3
Day 3 begins with the sum up of the work done the previous day.
onsidering the solutions defined during workshops participants try
o find a new process. They build the new process mapping and pre-
are two action plans: a short term action plan (20 days) and a long
erm action plan. Participants identify the actions required to put in
lace the new process (who, what, when and how) together with the
ole of each one. Eighty percent of all actions should be in the short
erm action plan to facilitate a quick implementation of the proposed
mprovement ideas and keep participants involved in the project.
During day 3, participants identify performance indicators that
ill be used to verify if the objective of reducing patients waiting time
s reached and determine how the required data will be collected. At
he end of the day, participants fill out an evaluative questionnaire
f the Kaizen event. They also present the new process to the hos-
ital manager who may evaluate the implementation cost since she
as received information every day of the Kaizen event. The hospi-
al manager can thus give her approbation to move forward with the
mplementation of the 20 day action plan the very next day.
.4.2. Results
At the end of the kaizen event, participants have built an action
lan to quickly implement those changes, ideally in 20 working days.
he 20 day action plan is the main output of the Kaizen event. Since
he improving scenarios were tested during the Kaizen event, it will
e easier to implement them rapidly. Nineteen weeks after imple-
enting the modifications, new data was collected. Since trajecto-
ies 1 and 2 were divided over 2 days, it was impossible to compare
esults before and after testing improving propositions. Only trajec-
ory 3 (treatment) could be analyzed more closely (Table 7).
Patient lead times for trajectory 3 were reduced from 61 to
6 minutes, a 74 percent reduction. Simulated results (during the
aizen) promised a 90 percent reduction. The gap between what was
xpected and the real value after implementation can be explained
y limits of the discrete event simulation model which did not take
nto account patients’ lateness, treatments beginning late, treatments
ot ready on time, patients not feeling well and other human related
vents that cannot be modeled. Consequently the discrete event sim-
lation model overestimated lightly the expected improvement but
t is still an appropriate manner to evaluate new ideas to reach the
bjective. Results show that the 34 minutes target for trajectory 3
55 percent × 61 minutes) has been reached. The challenge is now
o ensure that those results are maintained.
.5. Control
Control is assured by weekly meetings called weekly huddle, led
y the clinic manager with different members of the clinic. They ver-
fy patients waiting time and propose new actions if necessary to still
e able to reach the target.
. Discussion
.1. From a Lean and simulation perspective
Robinson et al. (2012) showed that Lean and simulation are com-
lementary methods even though they are often used independently.
imLean Educate involves the use of existing models before the
aizen event in teaching key Lean principles, SimLean Facilitate in-
olves rapid modeling during the Kaizen event to better understand
he dynamic of the process and SimLean Evaluate involves the de-
elopment of a detailed model after the Kaizen event to evaluate sce-
arios. Our approach proposes to develop a single detailed simulation
odel (Fig. 7).
The time required to develop a detailed simulation model is not
horter than Robinson et al. (2012). We used the model before the
aizen event (Educate step), to generate different improving scenar-
os and eliminate the least interesting ones (those having less impact
ver the reduction of patients waiting time). This leaves more time
uring the Kaizen event to provide guidance to the team to find fea-
ible solutions. The Kaizen event is used to find solutions to improve
he process performance and reach an identified target. Kaizen events
re costly considering the salary of 10 participants during 3 days. Con-
equently some steps (process mapping, time study and data analy-
is) are done before the Kaizen event leaving more time to partici-
ants to find solutions. During the Kaizen event, the discrete event
imulation model contributes to engage participants to discuss dif-
erent points of views and to provide evidence in order to achieve
onsensus. As Robinson et al.’s model, ours helps in managing con-
icts of interest between team members. Simulation model is thus a
undamental tool to evaluate the proposed solutions. Since the model
s valid participants are confident that the results will be closed to
hat is expected in real life. This help reducing barriers when imple-
enting Lean in the clinic such as skepticism and lack of ownership
f solutions (Brandao de Souza & Pidd, 2011).
![Page 10: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/10.jpg)
336 C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339
Detailed simula�on model Role: educate/facilitate Kaizen event
Detailed simula�on modelRole: experiment/evaluate
Implementa�on
Before During A�er
Fig. 7. Our approach integrating detailed discrete event simulation in a Kaizen event.
s
w
l
i
m
o
p
t
c
c
a
c
K
i
o
b
s
p
w
t
t
o
l
c
e
m
p
R
i
i
e
R
s
s
i
w
e
l
K
t
i
m
s
d
r
t
e
Our approach allows a rapid implementation of the solutions gen-
erated during the Kaizen event compared to Robinson et al. (2014). In-
deed the first steps of the DMAIC approach (DMA) have been realized
in 6 months. The last two steps (IC) needed 3 months, 1 month for im-
plementation (20 day action plan) and 2 months for follow-up. Our
approach allows reducing the implementation delays of Robinson’s
model (2007) caused by the development of a detailed simulation
model during the Evaluate step (Fig. 1). This is a major contribution
since a Kaizen event holds out an expectation from participants. If so-
lutions resulting from the Kaizen event are not implemented quickly,
team members may believe that their efforts have been worthless
and that the approach is not working. Indeed, failure in producing
tangible and quick results is a factor contributing to the poor stake-
holder engagement in healthcare simulation project (Jahangirian
et al., 2015)
5.2. From a facilitation and participation in simulation and Lean
perspective
In a facilitate mode, an intervention team with members of the
client’s organization are actively involved in determining the scope
of the project, analyzing and solving the problem. This team is sup-
ported by an operations research consultant who acts as a facilitator
(Franco & Montibeller, 2010). To better understand how our approach
is facilitating, the composition of the intervention team must be ana-
lyzed.
Our intervention team is similar to Kotiadis, Tako, & Vasilakis
(2014). The university team has the expertise to collect data and build
the simulation model (modeling team in Kotiadis et al., 2014). The
Lean facilitator serves as a Lean expert leading the project, meet-
ings and the Kaizen event. The intervention team consists of nurses,
pharmacists, doctors and clerks (stakeholder team). The clinic man-
ager is responsible for implementing solutions and maintaining re-
lations between the management team and the intervention team
(project champion). Finally the manager team can make all decision
(key stakeholders). Fig. 8 shows the steps and activities realized dur-
ing the project with the detailed schedule and involvement of the
team members.
Our approach has other similarities with Kotiadis et al. (2014)
since it includes interviews and workshops with stakeholders to in-
volve them in the Lean steps and the development of the simulation
model.
Because Lean healthcare project often involves many stakehold-
ers with plurality of opinions and objectives, we wanted to make
sure that all team members agree with the target and collaborate
to reach it and not individual goals. All personnel members were
met individually (DEFINE step): 5 nurses, 2 clerks, 8 doctors, 1 phar-
macist and 2 assistant pharmacists. These meetings are considered
beneficial for the success of the project while reducing resistance to
change (Brandao de Souza & Pidd, 2011).
Four group meetings were organized. The first meeting (DEFINE
tep) is to present the team project, their role and the steps of the
hole project including discrete event simulation to employees. Un-
ike Kotiadis et al. (2014) this meeting is informative instead of partic-
pative since the project objective has been previously defined by the
anagement team, the facilitator and the clinic manager. The sec-
nd meeting (MEASURE step) is used to validate the process map-
ing made by the Lean facilitator that was posted in the clinic so
hat everyone could see it. They all discuss about possible modifi-
ations. The presence of the university team to that meeting is cru-
ial to help them understand the process and model it adequately
ccording to the process mapping. This is a participative meeting
omparable to workshop 2 (stage 3: specify conceptual model) of
otiadis et al. (2014). In Robinson et al. (2014) the process mapping
s also realized before the Kaizen event but it is only finished day 1
f the Kaizen. In our case the whole process mapping is completed
efore the Kaizen event. During the third meeting (ANALYZE step) re-
ults from the MEASURE steps are presented: patients waiting time,
ersonnel tasks, treatment room capacity compared to the demand,
ork schedules, appointments schedules, etc. It is more an informa-
ive meeting led by the Lean facilitator. It allows the university team
o capture the details needed to build the simulation model. The goal
f the fourth meeting (ANALYZE step) consists in validating the simu-
ation model and presenting simulation to the team members who
an propose modifications that will be studied before the Kaizen
vent. It is led by the university team and the Lean facilitator. This
eeting is participative and comparable to workshop 3 (stage 5: ex-
erimentation) of Kotiadis et al. (2014). Like Kotiadis et al. (2014) and
obinson et al. (2014), model coding was done apart from the meet-
ngs and the Kaizen event. In our approach, model coding was real-
zed at the ANALYZE step right before the Kaizen event. In Kotiadis
t al. (2014), model coding is done at step 4 (model coding) and in
obinson et al. (2014), it is done between 2 days of the Kaizen event.
The Kaizen members received training on Lean principles and
imulation (ANALYZE step) from the Lean facilitator and the univer-
ity team. Our approach is different from Robinson et al. (2014) since
t is devoted to finding solutions, identifying the best scenario and
riting the action plan. Kotiadis et al. (2014) did not have a Kaizen
vent.
The organization of our Kaizen event differs from what is found in
iterature in two ways. It uses a detailed simulation model during the
aizen event to evaluate improving scenarios proposed by the Kaizen
eam in a unique way. Indeed Kotiadis et al. (2014) evaluate scenar-
os outside the workshops. Robinson et al. (2014) uses the simulation
odel to evaluate scenarios during the Kaizen event. However the re-
ults are not precise since the model was developed with estimated
ata. Therefore participants are not able to know if the goal has been
eached. Let us recall that the target consists in a 45 percent reduc-
ion of patients waiting time. Solutions at the end of the 3 day Kaizen
vent must allow reaching this goal. The simulation model is thus an
![Page 11: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/11.jpg)
C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339 337
Hos
pita
l man
ager
Lean
faci
litat
or
Clin
ic m
anag
er
Inte
rven
�on
team
Kaiz
en te
am
univ
ersi
ty T
eam
(ex
pert
)
Janu
ary
Febu
ary
Mar
ch
Apr
il
May
June
July
Aou
t
Sept
embe
r
DEFINEProject charter X X XProject launch X X X X XInterviews X X X X
MEASUREMapping process XMapping process va l ida�on X X X XTime study X XThe cl inic X X
ANALYZEData analys is and compi la�on X XModel coding XSimula�on model va l ida�on X X XPrel iminary tests with the s imual�on model X X XPresenta�on of the results from the measure s tep X X X XDevelopment of bus iness game X XTra ining on lean principles and s imula�on X X X X
IMPROVEKaizen event(3 days) X X X X X XAc�on plan implementa�on X X X
CONTROLWeekly huddle X X XMonitoring indicators X
Project team Timeline
Steps and ac�vi�es
Fig. 8. Schedule and roles of team members during the Lean project.
i
c
u
q
h
a
t
p
t
l
c
e
a
o
t
a
f
a
l
a
fi
n
n
s
d
i
5
p
l
s
q
i
i
i
t
m
s
u
q
a
H
u
e
p
a
t
p
t
e
c
t
o
c
r
a
s
5
m
t
K
i
nteresting tool. In Robinson et al. (2014) the target is not defined pre-
isely. Second using a business game to generate appointment sched-
les during the Kaizen event is an innovative contribution.
Scheduling appointments in a hematology–oncology clinic re-
uires human judgment even if there are guidelines. Business game
elped clerks to test different guidelines, better understand them
nd take note of the results. By scheduling appointments themselves
hrough business game, they realized that they were able to detect
otential problems for the future. Business game allows participants
o test in real-time their ideas and to detect rapidly potential prob-
ems. More realistic scenarios can then be proposed. Arrivals in dis-
rete event simulation are modeled as appointment schedule. The
valuation being performed during the Kaizen event, the best ones
re chosen immediately accelerating their implementation. The use
f business game during the Kaizen event encourages participants
o be more creative. Participants in the Kaizen event could evalu-
te the impact of the proposed changes on their work and anticipate
uture problems. They could then propose solutions immediately. It
lso allows taking into account human aspects when determining so-
utions and make the Kaizen event more dynamic. Like van der Zee
nd Slomp (2009), we conclude that the game could help workers
nd solutions for specific problems and facilitate their acceptance of
ew work methods or systems. Our results demonstrate that busi-
ess game combined with a discrete event simulation can be used to
upport participants during the Kaizen event.
Finally since the 20 day action plan is defined during the third
ay of the Kaizen event this is comparable to workshop 4 (stage 6:
mplementation) of Kotiadis et al. (2014).
.3. When to use this approach?
Our approach helps maintaining interactions between partici-
ants during a Lean project involving the development of a simu-
ation model during a Kaizen event to evaluate different improving
cenarios. It could be used not only in healthcare applications. It re-
uires a team eager to be involved in the project and opened to the
dea of using simulation models. Team members must accept to give
ndividual interviews and participate in group meetings during work-
ng hours. This implies coordination to avoid disturbing the clinic ac-
ivities and to obtain a high level of participation. The Lean facilitator
ust put a lot of effort to plan, coordinate meetings and follow the
teps rigorously.
Robinson et al. (2014) propose the development of a simple sim-
lation model in 2 days which is less time than what our model re-
uires. Our approach is closer to Kotiadis et al. (2014) since it implies
detailed simulation model requiring more time to be developed.
owever the model reproduces more adequately reality and can be
sed during the Kaizen event to help participants. Considering all the
ffort required in developing the simulation model and from partici-
ants, our approach is more convenient to solve complex problems.
Finding solutions during the Kaizen event is crucial in our
pproach. The scope of the problem must be well identified so
hat participants can work toward the same target. It requires the
articipation of at least one representative of each profession over
hree consecutive days for the Kaizen event. Otherwise the Kaizen
vent cannot be organized. Even though it can be perceived as a
onstraint it is a success factor. Participants become ambassadors of
he solution to their colleagues facilitating implementation. The use
f Kaizen and the business game encourages participants to be more
onfident in results and to rapidly implement the 20 day action plan.
The university and project teams are co-partners with roles and
esponsibilities well defined (Fig. 8). The DMAIC is well structured
nd rigorous for the project to progress adequately and meet the
chedules.
.4. Lessons
Four success factors have been identified. It needs the involve-
ent of doctors, pharmacists, managers and employees. It confirms
he importance to have doctors and pharmacists present during the
aizen event. Second it requires a culture of continuous improvement
n the organization and among the team members. Indeed before the
![Page 12: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/12.jpg)
338 C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339
G
H
H
J
J
K
K
K
K
K
K
K
L
L
M
M
M
N
O
P
P
P
P
R
R
R
R
S
S
S
T
T
Kaizen event the team had already begun to implement modifica-
tions to improve the process. Third the use of a structured approach
(DMAIC) and the presence of a Lean facilitator were comforting to the
team members. Finally being able to generate appointments sched-
ules and to measure their impact during the Kaizen even with the
simulation model was well received by participants. However the
20 day action plan to implement the selected scenario requires a sus-
tained rhythm and a lot of work that was underestimated.
6. Conclusion
As Robinson et al. (2014) and Tako and Kotiadis (2015) we showed
that participants involvement is crucial for the success of an ambi-
tious project. Even though our approach is similar to a facilitated
mode, a fully facilitated mode was not achieved since the client was
not involved in the model coding step. Our approach can be consid-
ered as participative because of the numerous meetings (individual
and group). Our study showed that the use of simulation and a busi-
ness game in a Kaizen even favors participation of all members. It also
has for principal advantage to help finding an adequate solution and
to measure its impact before the implementation. Given the nature
of the model, the result could be taken as an accurate result (contrary
to Robinson et al., 2014). The 20 day action plan can be implemented
immediately after the Kaizen. There is no delay after the Kaizen to
develop a detailed simulation model as for Robinson et al. (2014). Fi-
nally data have been collected 3 months after implementing the so-
lution. It shows that our approach provides an important advantage
by allowing a rapid implementation. Future work could be devoted
to implement our approach in another healthcare department or in
other activity sectors such as manufacturing.
References
Al-Araidah, O., Momani, A., Khasawneh, M., & Momani, M. (2010). Lead-time reduc-
tion utilizing lean tools applied to healthcare: The inpatient pharmacy at a localhospital. Journal for Healthcare Quality, 32, 59–66.
Ashenbaum, B. (2010). The twenty-minute just-in-time exercise. Decision Sciences Jour-
nal of Innovative Education, 8(1), 269–274.Berg, B., Denton, B., Nelson, H., Balasubramanian, H., Rahman, A., Bailey, A., et al.
(2009). A discrete event simulation model to evaluate operational performanceof a colonoscopy suite. Medical Decision Making, 30(3), 1–8.
Billington, P. J. (2004). A classroom exercise to illustrate lean manufacturing pull con-cepts. Decision Sciences Journal of Innovative Education, 2, 71–76.
Bok, B. M. (2007). Experiential foresight: Participative simulation enables social reflex-
ivity in a complex world. Journal of Futures Studies, 12(2), 111–120.Bowers, J., Ghattas, M., & Mould, G. (2012). Exploring alternative routes to realising the
benefits of simulation in healthcare. Journal of the Operational Research Society, 63,1457–1466.
Brailsford, S. C., Harper, P. R., Patel, B., & Pitt, M. (2009). An analysis of the academicliterature on simulation and modelling in health care. Journal of Simulation, 3(3),
130–140.
Brandão de Souza, L., & Pidd, M. (2011). Exploring the barriers to lean health care im-plementation. Public Money & Management, 31(1), 59–66.
Brandao de Souza, L. (2009). Trends and approaches in lean healthcare. Leadership inHealthcare, 22(2), 121–139.
de Mast, J., & Lokkerbol, J. (2012). An analysis of the Six Sigma DMAIC method from theperspective of problem solving. International Journal of Production Economics, 139,
604–614.
Dickson, E. W., Singh, S., Cheung, D. S., Wyatt, C. C., & Nugent, A. S. (2008). Application oflean manufacturing techniques in the emergency department. Journal of Emergency
Medicine, 37(2), 177–182.Ellis, S. C., Goldsby, T. J., Bailey, A. M., & Oh, J.-Y. (2014). Teaching lean six sigma within a
supply chain context: The airplane supply chain simulation. Decision Sciences Jour-nal of Innovative Education, 12(4), 287–319.
Fone, D., Hollinghurst, S., Temple, M., Round, A., Lester, N., Weightman, A., et al. (2003).Systematic review of the use and value of computer simulation modelling in pop-
ulation health and health care delivery. Journal of Public Health Medicine, 25(4),
325–335.Franco, L. A., & Montibeller, G. (2010). Facilitated modelling in operational research.
European Journal of Operational Research, 205, 489–500.Greco, M., Baldissin, N., & Nonino, F. (2013). An exploratory taxonomy of business
games. Simulation & Gaming, 44(5), 645–682.
riffiths, J. D., Jones, M., Read, M. S., & Williams, J. E. (2010). A simulation model of bedoccupancy in a critical care unit. Journal of Simulation, 4(1), 52–59.
agtvedt, R., Griffin, P., Keskinocak, P., & Roberts, R. (2009). A simulation model to com-pare strategies for the reduction of health-care-associated infections. Interfaces,
39(3), 256–270.ouchens, N., & Kim, S. C. (2014). The application of lean in the healthcare sector:
Theory and practical examples. In Wickramasinghe, et al. (Eds.), Lean thinking forhealthcare (pp. 43–53). New York: Springer.
ahangirian, M., Taylor, S. J. E., Eatock, J., Stergioulas, L. K., & Taylor, P. M. (2015). Casual
study of low stakeholder engagement in healthcare simulation projects. Journal ofthe Operational Research Society, 66(3), 369–379.
un, J. B., Jacobson, S. H., & Swisher, J. R. (1999). Application of discrete-event simulationin health care clinics: A survey. Journal of the Operational Research Society, 50(2),
109–123.aplan, G., & Patterson, S. (2008). Seeking perfection in healthcare: A case study in
adopting Toyota Production System methods. Healthcare Executive, 23, 16–21.
atsaliaki, K., & Brailsford, S. C. (2007). Using simulation to improve the blood supplychain. Journal of the Operational Research Society, 58(2), 219–227.
elton, W. D., Sadowski, R., & Sturrock, D. (2007). Simulation with arena (4th ed., p. 636).New York: McGraw-Hill.
hurma, N., Bacioiu, G. M., & Pasek, Z. J. (2008). Simulation-based verification of leanimprovement for emergency room process. In Proceedings of the 2008 winter simu-
lation conference (pp. 1490–1499).
lassen, K. J., & Yoogalingam, R. (2009). Improving performance in outpatient appoint-ment services with a simulation optimization approach. Production and Operations
Management, 18(4), 447–458.otiadis, K., Tako, A., & Vasilakis, C. (2014). A participative and facilitative conceptual
modelling framework for discrete event simulation studies in healthcare. Journalof the Operational Research Society, 65(2), 197–213.
otiadis, K., Tako, A., Rouwette, E., Vasilakis, C., Gandhi, P., & Wegstapel, H. (2013). Using
a model of the performance measures in Soft Systems Methodology (SSM) to takeaction: A case study in health care. Journal of the Operational Research Society, 64(1),
125–137.ean Enterprise Institute (2014). Lean lexicon: A graphical glossary for lean thinkers (5th
ed., p. 131). Cambridge, MA: Lean Enterprise Institute.odge, A., & Bamford, D. (2008). New development: Using lean techniques to reduce
radiology waiting times. Public Money & Management, 28(1), 49–52.
artin, C. H. (2007). A simulation based on Goldratt’s matchstick/die game. DecisionSciences Journal of Innovative Education, 5(2), 423–429.
artin, K., & Osterling, M. (2007). The Kaizen event planner: Achieving rapid improvementin office, service, and technical environments. New York: Productivity Press, pp. 240.
ielczarek, B., & Uzialko-Mydlikowska, J. (2012). Application of computer simulationmodeling in the health care sector: A survey. Simulation, 88(2), 197–216.
elson-Peterson, D., & Leppa, C. (2007). Creating an environment for caring using lean
principles of the Virginia Mason Production System. Journal of Nursing Administra-tion, 37(6), 287–293.
gulata, N. S., Cetik, M. O., & Koyuncu, E. (2009). A simulation approach for schedulingpatients in the department of radiation oncology. Journal of Medical Systems, 33,
233–239.aul, S. A., Reddy, M. C., & De Flitch, C. J. (2010). A systematic review of simulation
studies investigating emergency department overcrowding. Simulation, 86(8–9),559–571.
oksinska, B. (2010). The current state of lean implementation in health care: Literature
review. Quality Management in Health Care, 19(4), 319–329.roudlove, N., Moxham, C., & Boaden, R. (2008). Lessons for lean in health care from
using six sigma in the NHS. Public Money & Management, 28(1), 27–34.roudlove, N. C., Black, S., & Fletcher, A. (2007). OR and the challenge to improve the
NHS: Modelling for insight and improvement in in-patient flows. Journal of theOperational Research Society, 58, 145–158.
adnor, Z., Walley, P., Stephens, A., & Bucci, G. (2006). Evaluation of the lean approach
to business management and its use in the public sector. Edinburgh: The ScottishGovernment.
obinson, S., Radnor, Z. J., Burgess, N., & Worthington, C. (2012). SimLean: Utilising sim-ulation in the implementation of lean in healthcare. European Journal of Operational
Research, 219, 188–197.obinson, S., Worthington, C., Burgess, N., & Radnor, Z. J. (2014). Facilitated modelling
with discrete-event simulation: Reality or myth? European Journal of Operational
Research, 234, 231–240.ohleder, T. R., Lewkonia, P., Bischak, D. P., Duffy, P., & Hendijani, R. (2011). Using simu-
lation modeling to improve patient flow at an outpatient orthopaedic clinic. HealthCare Management Sciences, 14, 135–145.
antibanez, P., Chow, V. S., French, J., Putterman, M. L., & Tyldesley, S. (2009). Reducingpatient wait times and improving resource utilization at British Columbia Cancer
Agency’s ambulatory care unit through simulation. Health Care Management Sci-
ence, 12, 392–407.epulveda, J. A., Thompson, W. J., Baesler, F. F., Alvarez, M. I., & Cahoon, L. E. (1999).
The use of simulation for process improvement in a cancer treatment center. InProceedings of the 1999 winter simulation conference (pp. 1541–1548).
wanson, L. (2008). The lean lunch. Decision Sciences Journal of Innovative Education,6(1), 153–157.
ako, A., Kotiadis, K., & Vasilakis, C. (2010). A participative modelling framework for
developing conceptual models in healthcare simulation studies. In Proceedings ofthe 2010 winter simulation conference (pp. 500–512).
ako, A., Kotiadis, K., & Vasilakis, C. (2010). Developing a conceptual modelling frame-work for stakeholder participation in simulation studies. In Proceedings of the 5th
operational research society simulation workshop (SW10) (pp. 76–85).
![Page 13: 1-s2.0-S0377221715007912-main](https://reader033.vdocument.in/reader033/viewer/2022042608/5695cfd11a28ab9b028fae32/html5/thumbnails/13.jpg)
C. Baril et al. / European Journal of Operational Research 249 (2016) 327–339 339
T
v
V
W
Y
ako, A. A., & Kotiadis, K. (2015). PartiSim: A multi-methodology framework to supportfacilitated simulation modelling in healthcare. European Journal of Operational Re-
search, 244, 555–564.an der Zee, D. J., & Slomp, J. (2009). Simulation as a tool for gaming and training in
operations management: A case study. Journal of Simulation, 3, 17–28.an Lent, W. A. M., Goedbloed, N., & Van Harten, W. H. (2009). Improving the efficiency
of a chemotherapy day unit: Applying a business approach to oncology. EuropeanJournal of Cancer, 45(5), 800–806.
hite, D. L., Froehle, C. M., & Klassen, K. J. (2011). The effect of integrated schedulingand capacity policies on clinical efficiency. Production and Operations Management
Society, 20(3), 442–455.oung, T., Brailsford, S., Connell, C., Davies, R., Harper, P., & Klein, J. H. (2004). Using
industrial processes to improve patient care. British Medical Journal, 328, 162–164.