European Journal of Operational Research 249 (2016) 327339

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 Ct d
a

Department of Industrial Engineering, Universit du Qubec Trois-Rivires, 3351 Boulevard des Forges, Trois-Rivires, Qubec G9A 5H7, Canada
Department of Management Science, Universit du Qubec Trois-Rivires, 3351 Boulevard des Forges, Trois-Rivires, Qubec G9A 5H7, Canada
c
Department of Industrial Engineering, Universit du Qubec Trois-Rivires, 3351 Boulevard des Forges, Trois-Rivires, Qubec G9A 5H7, Canada
d
Performance Management Oce in a Health Facility, Canada
b

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 dened 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 hematologyoncology
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. Introduction
Considering the increasing needs for services in healthcare,
hospital services must review their practices to improve them and
increase their performance. Healthcare facilities look toward continuous improvement approaches such as Lean manufacturing to
improve service delivery (Houchens & Kim, 2014). Lean manufacturing is a management approach aiming to improve the performance of
an organization by reducing waste, delays, etc. while involving staff
in decision-making. During the last few years, the DMAIC (Dene,
Measure, Analyze, Improve and Control) problem-solving approach
combined to six sigma was used jointly with Lean to become Lean
six sigma.
Operations research techniques and discrete event simulation
have also been used by healthcare managers (Fone et al., 2003).
Since healthcare services are mostly dynamic and stochastic processes, discrete event simulation has been more often used to model
and analyze ows in healthcare processes (Fone et al., 2003; Jun,
Jacobson, & Swisher, 1999; Mielczarek & Uzialko-Mydlikowska,
2012). More recently researchers included a Lean approach to discrete event simulation in a facilitated mode (Robinson, Worthington,
Burgess, & Radnor, 2014). Robinson, Radnor, Burgess, and Worthington (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: chantal.baril@uqtr.ca (C. Baril).

activity, commonly lasting 5 days, in which a team identies and implements a signicant improvement in a process (Lean Enterprise
Institute, 2014). It is a participative activity and it facilitates the acceptance of the project by the employees. Tako and Kotiadis (2015)
combine discrete-event simulation, a hard OR approach, with soft
systems methodology (SSM) in order to incorporate stakeholder involvement in the simulation study lifecycle.
Business games are also used as a tool to increase understanding
of Lean management concepts (Ashenbaum, 2010; Billington, 2004;
Martin, 2007; Swanson, 2008). van der Zee and Slomp (2009) assert that they could help workers nd solutions for specic problems, or to familiarize themselves with and ease their acceptance
of new work methods or systems. Originally, business games have
been used to help nd solutions in different business environment.
A business game has been dened by Greco, Baldissin, and Nonino
(2013) as a game with a business environment that can lead to one or
both of the following results: the training of players in business skills
(hard and/or soft) or the evaluation of players performance (quantitatively and/or qualitatively). The business game allows a better
understanding of complex problems. The pedagogical principle underlying the business game is involving participants in a virtual environment. Business games were originally developed to educate business managers. They reproduce a process in a virtual environment
while being inspired by reality. It can also be used to let employees
perform a task or a given operation for real (Ellis, Goldsby, Bailey, &
Oh, 2014). However business games can be helpful to educate managers, employees and change agents in healthcare or education. The

http://dx.doi.org/10.1016/j.ejor.2015.08.036
0377-2217/ 2015 Elsevier B.V. and Association of European Operational Research Societies (EURO) within the International Federation of Operational Research Societies (IFORS).
All rights reserved.

328

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Modelling team

Iniate
simulaon study

Stakeholders
Modelling team

Dene system
Workshop 1

Stakeholders
Modelling team

Specify
conceptual model
Workshop 2

Modeller

Model coding

Stakeholders
Modelling team

Experimentaon
Workshop 3

Stakeholders
Modelling team

Implementaon
Workshop 4

Fig. 1. PartiSim (Tako & Kotiadis, 2015).

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 implementation of solutions found in a kaizen event and reduce barriers
when implementing Lean in healthcare. We believe that discreteevent 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 hematologyoncology 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 modications implemented. A Kaizen event was organized at the Improve
step during which a business game was used to help nd 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 literature while Section 3 presents the methodology. The implementation
of our approach and results are described in Section 4. Finally the results 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 identication and elimination of waste and in developing standardized, reliable processes. This is performed in a context of connectedness, respect, and growth of all employees who are trained to
identify waste and errors, and suggest possibilities for improvements
that will be tested using scientic methods. Lean seems to be an effective way of improving healthcare organizations and the growing
number of implementations and reports found in the literature reinforce this view (Brando de Souza, 2009)
Lean implementation in healthcare requires adaptation and development to t the specic context and allow healthcare staff to
take ownership of the approach (Poksinska, 2010). Literature review shows that there have been some signicant tangible outcomes 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). However, many papers identied barriers when implementing Lean management in healthcare organizations such as lack of ownership of
proposed processes, skepticism and resistance to change (Brando de
Souza & Pidd, 2011; Proudlove, Moxham, & Boaden, 2008; Radnor,
Walley, Stephens, & Bucci, 2006).
In the last few years, discrete-event simulation has been considered 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;

Rohleder, Lewkonia, Bischak, Duffy, & Hendijani, 2011). Discreteevent simulation has been applied to solve a wide variety of healthcare problems such as patient appointment systems (Klassen and
Yoogalingam, 2009; Ogulata, Cetik, & Koyuncu, 2009), patient waiting time (Paul, Reddy, & De Flitch, 2010; Santibanez, Chow, French,
Putterman, & Tyldesley, 2009), patient ow, (Rohleder et al., 2011;
Sepulveda, Thompson, Baesler, Alvarez, & Cahoon, 1999; White
et al., 2011), operational performance (Berg et al., 2009; Griths,
Jones, Read, & Williams, 2010) and others problems (Hagtvedt,
Grin, Keskinocak, & Roberts, 2009; Katsaliaki & Brailsford, 2007).
Too often, discrete event simulation models have been developed
and used by experts to nd solutions without involving stakeholders
in the development process. Recently, more work has been done on
facilitated modeling to involve stakeholders in the development of
discrete event simulation models.
Facilitated modeling consists in developing models jointly with a
client group: from dening the nature of the problem, to supporting the evaluation of priorities and development of plans for subsequent implementation (Franco & Montibeller, 2010). Franco and
Montibeller (2010) discuss in detail facilitated modeling as an
OR intervention tool in organizations. Jahangirian, Taylor, Eatock,
Stergioulas, and Taylor (2015) examine the stakeholder engagement
in the context of healthcare simulation. They nd that communication gap between simulation and stakeholder groups is the top
primary factor contributing the most to the poor stakeholder engagement in healthcare simulation projects, followed by poor management support, clinicians high workload and failure in producing
tangible and quick results. Recently managers began to be included
in problem denition and process modeling (Kotiadis et al., 2013;
Tako, Kotiadis, & Vasilakis, 2010a; Tako, Kotiadis, & Vasilakis, 2010b).
This participation is especially important for studies in healthcare
characterized by the presence of many stakeholders with tacit knowledge of their part of the system and often multiple views and objectives. Tako and Kotiadis (2015) combined the steps required to develop a discrete event simulation model with the participative steps
of Soft Systems Methodology. Their whole procedure, called PartiSim,
allows stakeholders to be involved at every stage of the model development and experiments (except for programming which requires a
specic expertise). Fig. 1 presents the PartiSim steps in which stakeholders are involved through workshops. Our approach differs from
Tako and Kotiadis (2015) because it includes the Lean project steps together with the development of the discrete event simulation model.
There are few studies on the use of discrete event simulation
through a Lean approach, or other continuous improvement process,
in healthcare systems. Young et al. (2004) propose the use of simulation to evaluate the beneces of a continuous improvement project
in healthcare before the implementation. Khurma, Bacioiu, and Pasek
(2008) present a discrete event simulation model to study the impact
of a Lean project in an emergency unit. Even if they do not specically
consider Lean management, Proudlove, Black, and Fletcher (2007)
show how a simple simulation model can improve eciently patient
ows. Robinson et al. (2012) explore potential complementary roles
of discrete event simulation and a Lean approach in healthcare systems. Their model, SimLean, denes three roles for discrete event simulation used with Lean: education, facilitation and evaluation (Fig. 2).
According to Robinson et al. (2012), discrete event simulation can
have an educational function in teaching Lean principles (Educate

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

SimLean Educate
Role: educate

Lean event

329

SimLean Evaluate
Role: experiment/evaluate

SimLean Facilitate
Role: engage/facilitate

During

Before

Aer

Fig. 2. SimLean model: the roles of discrete-event simulation and Lean in healthcare (Robinson et al., 2012).

Facilitated modelling using the SimLean approach (Robinson et al., 2014)

SimLean Facilitate

Kaizen (2 days)

Day1
kaizen

Day3
Kaizen

Day2

Overview of the current process

Introducon to simulaon (SimLean Educate)
Gemba
Mapping the process
Esmate me for each acvity

Model coding
Simple model developed

Recap of the rst day

Model briey explained
Face validaon
Problem scoping
Improvement

Fig. 3. Facilitated modeling using the SimLean approach (Robinson et al., 2014).

step, Fig. 2). It is a mean to understand the dynamics of a healthcare

process. One or many predened standard discrete event simulation
models can be used to teach Lean principles before or during a Kaizen
event. During Lean events, processes must be analyzed through the
use of a process mapping. Simulation can be used to create a dynamic version of the process mapping (Facilitate step, Fig. 2). A simple
discrete-event simulation model developed during the Kaizen event
can be used to better understand the process dynamics and to encourage participants to propose improvement solutions. With a detailed discrete event simulation model, different scenarios can be
tested. Developing a detailed discrete event simulation model may
take a long time since it requires obtaining data, modeling and validating the process, validating the model and generating improving
solutions. These steps are usually performed after the Kaizen event
to test solutions found by the participants and to eventually propose
new ones (Evaluate step, Fig. 2). The discrete event simulation model
can also be used during the implementation phase and plays a role in
continuous improvement. In fact, the model in Robinson et al. (2012)
implies dening three discrete event simulation models: (1) predened 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
Kaizen event to better understand the process dynamics and to encourage participants to propose improvement solutions, a detailed
simulation model was considered to measure the impact on patients
waiting time of solutions found by the participants during the Kaizen
event. Since Robinson et al.s model implies developing the detailed
simulation model (Evaluate step) after the Kaizen event, it increases
the delay before implementing solutions, while it is crucial to start
the implementation on a short delay after the Kaizen event to guaranty its success (Martin & Orsterling, 2007).
The paper of Robinson et al. (2014) focuses on SimLean facilitate as
an example of facilitated modeling using discrete-event simulation.

The simulation model was developed and used within a 3 days period
of an improvement workshop. However, they had to build the model
in the back oce, meaning that a fully facilitated model was not
achieved. During the Kaizen event the simple but not validated model
(since it was built with estimated times) helped in suggesting solutions. Consequently Kaizen participants were able to immediately get
some feedback on how their ideas could improve the system performance even though it could not be measure precisely. Fig. 3 presents
their facilitated approach.
Unlike Robinson et al. (2014) our Kaizen activity is entirely devoted in nding solutions that could improve the whole process and
satisfy stakeholders (doctors, nurses, etc.). Data collection (process
mapping, time study) and the simulation model are determined before the Kaizen event. As in Robinson et al. (2014), our simulation
model was not dened during the Kaizen event but rather in back
oce.
3. Overview of our approach
This paper presents an approach to conduct a Lean project in an
oncology clinic during which a detailed simulation model is dened
in order to validate improving ideas proposed by the Kaizen event
participants. The DMAIC solving problem method was used in a facilitate mode. Moreover a business game was performed during the
Kaizen event to facilitate employees involvement and be able to implement improvement ideas more quickly. The approach is presented
in Fig. 4.
The DEFINE step allows specifying the project and determining
performance indicators. In order to achieve this, a project charter
and interviews with the personal clinic were realized. The project
charter denes the team vision: problem statement and objective,
nancial impact, project scope, schedule and team members. Once
the project charter nished, the project itself can begin. A meeting

330

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Our approach

Improve

Dene

Measure

Analyze

Project charter

Process mapping

Development of
businessl game

Project launch and

interviews

Time study

Project team

The clinic

Kaizen event

Development of
Simulaon model

Day 2

Day 1
Lean and kaizen presentaons
Measure step presentaon
Project objecve
Process irritants and waste
Workshops idencaon

Recap of the day before

Ideal process
Workshops
Business game
Simulaon model

Control

SimLean and
facilitated modelling

Day 3
Recap of the day before
Workshops
Indicators
Acon plan

Fig. 4. Our approach.

with the clinics 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 identied 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 performance.
Data collection is conducted during a representative working
week and included appointment scheduling, work schedules, treatment capacity for a better understanding of the clinic. Data is next
used to describe the process and develop the discrete-event simulation model and business game.
In the ANALYZE step are identied the causes of the problems on
which the team will work during the IMPROVE step. A detailed discrete event simulation model is built to help nd solutions. Because
human judgment is not taken into account in discrete event simulation and has a great inuence on process eciency (Bok, 2007), we
developed a business game.
To IMPROVE the process eciency, a Kaizen event was planned so
that each team member could participate in nding solutions. During
this Kaizen event, the business game and the detailed discrete event
simulation model were used to evaluate the different solutions proposed 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 ecient. The performance indicators
must be measured over time to verify the process stability and take
actions if necessary.

oncology treatments (chemotherapy) and hematology treatments

(hemoglobin transfusions, blood transfusions, phlebotomy, coagulation factors). On average, 8500 treatments are administered every
year. The clinic is opened from 8h00 AM to 8h00 PM. Seven doctors, six nurses, one clinical nurse, two to three pharmacists and two
to three assistant pharmacists (ATP) depending on the day, and two
clerks work in the clinic. Nurses work on 8-hour shifts (8h00 AM to
4h00 PM, 9h00 AM to 5h00 PM, 10h00 AM to 6h00 PM and 12h00 PM
to 8h00 PM).
4.1. Dene
4.1.1. Project team
The project team consists of:

4.1.2. Project charter

The project charter was developed by hospital managers, the Lean
facilitator, the clinic manager and the university team (Table 1). At
this step it was decided to create a detailed simulation model to
study the patient trajectory instead of a simple one for the following
reasons:

4. Implementation of our approach in a hematologyoncology

clinic
Our approach was implemented in a hematologyoncology clinic
when carrying out a Lean project. The clinic under study offers

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 expert. 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 department, the chief laboratory, the Lean facilitator and the university team.
Clinic manager: to ease the implementation of the solutions and
make the connection between the hospital managers and the intervention team.
Hospital managers: control decisions.

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).

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Registraon

Blood sample

Registraon

Blood sample

Meeng with
the doctor

Making an
appointment

331

Taking charge
treatment

Day 1
2

Taking charge
treatment

Arrival
Exit

Taking charge
treatment

Registraon

Registraon

Meeng with
the doctor

Making an
appointment

Fig. 5. Mapping of patient trajectories in the outpatient hematologyoncology clinic.

Table 1
Project charter.
Problem

Objective
Scope
Impact
Schedule
Team members

The outpatient hematologyoncology 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 signicant increase in
the personnel workload has also been noticed, resulting in
an increase in overtime the last 18 months
Reduce patients waiting time
Review the patient trajectory from the arrival up to the
departure from the clinic
Managers and department chief wish to increase patients
and employees satisfaction and clinics performance
9 months (January to September)
Hospital managers
Lean facilitator
Clinic manager
Intervention team (nurses, doctors, clerks, pharmacists,
assistant pharmacists)
University team

4.1.3. Project launch and interviews

The project launch meeting was led by the Lean facilitator during
which the project charter was also presented. Employees showed enthusiasm toward the project and a willingness to collaborate. Afterwards individual interviews were held with all employees involved
and they led to following ndings:

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 objectives.
Doctors must be involved in each step of the project (DMAIC).
Good relations among personnel members should contribute to
the projects success.

These interviews indicated that the intervention team was ready

to put the required effort to turn this project into a success.
4.2. Measure
4.2.1. Process mapping
The complete trajectory followed by a patient comprises ve
steps:
(1) The patient arriving at the clinic registers with the clerk.
(2) The patient waits for a blood sample which is sent to the laboratory 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 appointment.

(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 ve steps. Some of them come
only for a blood sample and to meet the doctor, others only for treatments or, only to meet the doctor if they previously had a blood sample. As the process mapping shows (Fig. 5), four different trajectories
can be followed by patients:
1. Follow-up and taking charge (5 steps): registration, blood sample, 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
coming to the clinic to receive treatments represent 53 percent of
all patients. The taking charge step is followed by three sub-steps:
(1) meeting with the pharmacist, (2) hydration and premedication
and (3) treatment.
The process mapping was rst realized by the Lean facilitator. It
was next posted in the clinic for the intervention team to get acquainted with it. Then during a meeting led by the Lean facilitator
they validated the process mapping to make sure that no steps had
been forgotten.
4.2.2. The clinic
Appointment scheduling provides information on patients coming to the clinic. Three appointment lists are considered: (1) list of
patients needing a blood sample (only in the afternoon), (2) list of
patients needing to meet a doctor (xed periods and last 20 minutes
each from 8h00 AM to 1h30 PM) and (3) list of patients needing treatments.
The rst treatments of the day are scheduled every 15 minutes
(regardless of the treatment duration). Later in the day, the time between scheduled appointments for treatments may vary depending
on the end of the previous one. Ad hoc rules are used by clerks to
schedule the appointments. Clerks must consider doctors working
schedules, patients preferences and opening hours of the treatment
rooms. Appointment scheduling showed that the number of planned
treatments is on average equal to 37, the average daily number of administrated treatments is 32 and the average percentage of canceled
treatments is equal to 14 percent. Results in Table 2 show that when
treatments are scheduled, the capacity (number of hours available) is
not always considered.

332

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Table 2
Number of hours for treatments (planned and available) and number of hours available.

Number of hours of
scheduled treatments
Number of hours
available (capacity)

Monday

Tuesday

89.00

91.00

128

128

Wednesday

Thursday

135.25

100.50

128

128

Friday
64.50
128

Table 3
Nurses schedules and number of patients waiting.
Time

Working schedules
Nurse 1

8h00
8h15
8h30
8h45
9h00
9h15
9h30
9h45
10h00
10h15
10h30
10h45
11h00

Nurse 2

Nurse 3

Number of

Treatment

Number of

patients ready

room capacity

patients waiting

1
3
3
2
1
0
2
2
1
5
2
1
0

1
1
1
1
1
1
1
1
1
1
1
1
1

0
2
4
5
5
4
5
6
6
10
11
11
11

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 doctors 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 appointments 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 important 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 trajectory 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 appointment 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 enter 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

Table 4
Summary of the time study for the four trajectories (in minutes).
Trajectories

Total lead time from

registration to start of
meeting the doctor

Total lead time from

registration to taking
charge in treatment room

1. Follow up/treatment
2. Blood sample/treatment
3. Treatment
4. Meeting the doctor

74.76
n/a
n/a
50.12

174.60
114.65
60.73
n/a

patient to see. They rather see patients on a rst come, rst call basis.
The actual computerized appointment system is not consistent when
scheduling appointments with doctors and treatments.
4.2.3. Time study
Our time study consisted in determining patient lead times according to their trajectory (Fig. 5). Data collection was done at the
clinic over 1 week. Each patient coming to the clinic received a numbered chip. At every step of his trajectory he had to identify himself
with his number and the observer would take note of the time he
went through the step. Table 4 provides a summary of the trajectory
lead times.
The total lead time from registration until meeting with the doctor could be computed for only two trajectories. For trajectory 1, total
waiting time is 69.31 minutes, 93 percent of total lead time. For trajectory 4, it is 48.50 minutes, 97 percent of total lead time. In both
cases waiting time is considered too high. Reducing total lead time
from registration to beginning preparing for treatment was identied as the main objective during the Kaizen event. Waiting times are
noticed at every step of the process. The time study provided other
performance indicators (Table 5).
The average time required to prepare the patient in the treatment
room is 20 minutes while appointments are scheduled every 15 minutes. This shows inconsistency in scheduling appointments. The average treatment chair utilization rate is 68 percent, showing that the
number of treatment chairs is adequate and that many more patients
could receive treatments.
Finally, patients are asked to arrive 30 minutes before meeting
the doctor, to leave enough time for blood sampling. Our time study

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Table 5
Other performance indicators.
Indicators

Mean value

Average time to prepare patient in the treatment room

Average chair utilization rate in the treatment room
Mean time to analyze a blood sample
Percentage of blood sample made the previous day
Number of complaints in 20112012

20 minutes
68 percent
31 minutes
22 percent
9

The simulation model was built by the university team. However

the clinic personnel were involved in many aspects to understand all
the specics of the process. The nal model was validated by the Lean
facilitator, the clinic manager, one doctor and a nurse. Unlike the SimLean approach (Fig. 1), the detailed discrete-event simulation model
was developed before the Kaizen event. It was used to make preliminary tests to avoid unnecessary discussions during the event. The
animator of the activity was able to evaluate objectively the impact
of adding a nurse to blood sampling and increasing the size of blood
samples shipments, on patients waiting times. These two propositions were tested with the simulation model. Results showed that
there was no signicant improvement. Therefore discussion among
the Kaizen event participants could be moved toward other subjects.
However different scheduling appointment rules, taking into account
the treatment room capacity, had to be evaluated with the simulation model. They had to be determined with the help of the clinic
personnel.

shows that the average time required to analyze a blood sample is

31 minutes. If we add the waiting time before the blood sample (21
minutes) to these 31 minutes, it is obvious that patients should arrive
earlier for the blood sample results to be ready before meeting the
doctor. Patients having their blood sample taken the day before their
appointment with the doctor or before their treatment dont have to
wait for the blood sampling results. It also reduces the amount of
work in the clinic. With only 22 percent of the blood samples performed the previous day, there is place for improvement.

4.3.2. Development of a business game

A business game was developed to allow participants to really
schedule appointments according to the rules dened during the
Kaizen event and to measure with the discrete event simulation
model how it impacted on patient waiting times. A set of cards, each
representing a patient with his characteristic, was prepared to schedule the appointments of a typical day. Each card contained the following information:

4.3. Analyze
Data analysis indicates that:

333

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.

1.
2.
3.
4.
5.

Fictitious patients name

Need of a blood sample or not
Name of doctor to meet
Treatment type (chemotherapy or else) and duration
Need to make another appointment with the doctor or not

The game will work as follows: participant will choose randomly

a card and schedule the appointment according to the patients need:
blood sample, meeting the doctor or treatment. Patients arrival rates
will be modied in the simulation model depending on the schedules generated. A member of the university team will run the simulation model to measure the impact of the new appointment schedules
on patients waiting time and present the results. For the game to be
ecient during the Kaizen event, it should be able to integrate the
different arrival rates rapidly.

4.3.1. Development of a detailed discrete event simulation model

The discrete event simulation model was developed using the
Arena 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
event on patient lead times. The statistical distributions used in the
simulation model are based on data collected from the clinic process.
These statistical distributions provided by Input analyzer of the Arena
software are the ones tting best the data considering the meansquare error. In our simulation model, entities (patients) follow one
of the different trajectories, according to the percentages presented
in Section 4.1. The model was developed to be as close as possible
to the real process. Comparisons between simulated and real waiting and lead times were performed for all trajectories. A 10 percent
threshold was considered to take into account the margin of error
related to the statistic curves and the error induced by a simplied
process. Table 6 shows that the gap between our simulation model
and the real process is at most 10 percent conrming the validity of
the model to be used to test scenarios.

4.4. Improve
4.4.1. Kaizen event
To improve the process eciency, a Kaizen event was planned so
that each team member could participate in nding solutions. The
Kaizen team includes 10 persons: two nurses, one administration ofcer, two doctors, one pharmacist, one assistant pharmacist, one employee from the department of informatics, one laboratory manager,
and one clinic manager. The Kaizen event was led by the Lean facilitator together with a member of the university team.
The objective of the Kaizen event after data analysis was to reduce patient waiting times for treatments by 45 percent while lightening the whole process. Even though this objective may seem too

Table 6
Validation of the discrete event simulation model.
Steps

Registration
Blood sample
Meeting the doctor
Making an appointment
Preparing for treatment

Waiting time

Lead time (waiting and service)

Real

Simulated

Gap (percent)

Real

Simulated

Gap (percent)

3.68
21.85
44.30
4.18
79.70

3.63
22.06
44.68
4.19
74.60

1.4
+1.0
+0.8
+0.2
6.4

5.30
31.00
92.32
98.90
107.10

4.90
29.90
87.00
97.60
110.10

7.5
3.5
5.8
1.3
+2.8

334

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Day 0

Registraon

Blood sample
hospital

Registraon

Blood sample
hospital

Meeng with
the doctor

Making an
appointment

Arrival
Exit

3
4

Blood sample
region

Registraon

Meeng with
the doctor

Making an
appointment

Day 1
Arrival

Taking charge
treatment

Registraon

Exit

Fig. 6. New mapping of patients trajectories.

ambitious, it must be given all resources and effort involved for the
project. This objective is related to trajectories 1 (follow-up and treatment), 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 receive the same information. Participants are invited to discuss and
modify the process mapping if necessary. Process irritants and waste
are next identied by teams of two to three persons. They are written
on post-it. Each team share their ndings and explain each irritant
when a member places the paper on the process mapping poster.
Then participants agree on 46 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 identied the following six topics to discuss in
workshops:
1.
2.
3.
4.
5.
6.

Nurses work organization

Patients trajectories
Equipments and work sites
Tools and information ow
Taking charge of patients by nurses
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 nds 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 participants summarizes the work done during the day. Since the direction
members do not attend the Kaizen event it is a way to keep them informed 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 identify 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

allows the Kaizen team to clearly determine actions and decisions

needed to get solutions. Next three teams of three participants each
work on three topics identied from the Ishikawa diagram to nd solutions to the irritants. They present their solutions to the other teams
and discussion is undergone to nd new ideas. Here follows results of
workshops 1, 2 and 6.
Workshop 1: Nurses work organization
During this workshop, participants proposed new nurses schedules. The treatment chairs availability depend on nurses schedules
which was modied to be from 8h00 AM to 4h00 PM for all nurses.
This modication allows for the beginning of more treatments earlier in the day. Real data showed that it was taking 20 minutes on
average to prepare a patient (Table 5). This was due to the fact that
treatments were not ready when the patient was ready to receive it.
During the workshop, pharmacists assured that treatments could be
ready by 8h00 AM for the rst incoming patients. Consequently it was
decided to continue scheduling appointments every 15 minutes.
Workshop 2: Patients trajectories
Participants proposed to divide trajectories 1 (follow-up and treatment) and 2 (blood sample and treatment) over 2 days. As shown in
Fig. 6, blood sample and meeting with the doctor (day 0) are planned
the day before treatments (day 1).
These new trajectories allow a reduction in treatment cancellation rate on the same day (day 1) due to bad blood results (day 0). A
deal was made with healthcare providers in specic regions to allow
patients to have their blood sample close to their home (day 0) instead of coming to the hospital, therefore eliminating the 30 minute
waiting time before meeting the doctor.
Workshop 6: Scheduling appointments
The business game was mainly used to test ideas to improve appointment scheduling. Participants to the game were a clerk, a doctor,
an assistant pharmacist technician and the clinic manager. Let us recall that the current appointment planning worksheet (treatment) is
divided into 15 minute time slots from 8h00 AM to 8h00 PM and that
it is impossible to schedule two appointments at the same time. The
treatment room capacity is not taken into account.
The game is quite simple. First, participants were asked to create their own appointment planning worksheets for blood sample,
meeting doctors and treatment, according to different criteria. Patients used to arrive for blood sampling 30 minutes before their appointment with the doctor while it was taking around 50 minutes
to perform and analyze the blood samples. Therefore appointments
were also added for blood samples (on morning) to reduce waiting
times for blood sampling. The idea was to coordinate appointments
for blood samples and with doctors according to the different trajectories. The appointment planning worksheets for treatments were

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

335

Table 7
Results.
Patient waiting times before treatment

Trajectory 3: treatment (day 1)

Mean observed
(before Kaizen)

Standard deviation
(before Kaizen)

Mean observed
(after Kaizen)

Standard deviation
(after Kaizen)

Simulated
(during Kaizen)

61 minutes

52 minutes

16 minutes (74 percent)

7 minutes

6 minutes (90 percent)

designed to consider the number of treatment chairs available and

nurses schedules (workshop 1). These modications allowed taking into account the treatment room capacity when building new
schedules by planning as many patients at the same time as there
are nurses and considering treatment times. The computer specialists conrmed that these new worksheets would be included into to
the current appointment planning software. Moreover they could test
other priority rules such as scheduling patients with longer treatment
durations at the beginning of the day.
Second, the cards were randomly sorted to simulate demands for
an appointment. Finally, clerks had to pick a card and assign an appointment time to the ctitious patient using the appointment planning worksheet. Results during business game with the cards show
that 8 of 94 had to be scheduled on another day. It was impossible to
schedule all patients on the same day since capacity was now taken
into account.
The appointment schedules for blood samples, meeting with the
doctor and treatments were used as inputs in the discrete event simulation model to measure their impact on patient waiting times. Since
trajectories 1 and 2 were divided over 2 days (workshop 2), it was impossible to compare results before and after testing improving propositions. Only trajectory 3 (treatment) could be analyzed more closely.
The discrete event simulation model was replicated 100 times. Patient waiting times before treatment were reduced by 90 percent on
average (from 61 minutes to 6 minutes). Participants could see that
the new schedules had a signicant impact on patient waiting times
before treatments.
One of the participants summarizes the work done at the end of
the day to the direction members.
Day 3
Day 3 begins with the sum up of the work done the previous day.
Considering the solutions dened during workshops participants try
to nd a new process. They build the new process mapping and prepare two action plans: a short term action plan (20 days) and a long
term action plan. Participants identify the actions required to put in
place the new process (who, what, when and how) together with the
role of each one. Eighty percent of all actions should be in the short
term action plan to facilitate a quick implementation of the proposed
improvement ideas and keep participants involved in the project.
During day 3, participants identify performance indicators that
will be used to verify if the objective of reducing patients waiting time
is reached and determine how the required data will be collected. At
the end of the day, participants ll out an evaluative questionnaire
of the Kaizen event. They also present the new process to the hospital manager who may evaluate the implementation cost since she
has received information every day of the Kaizen event. The hospital manager can thus give her approbation to move forward with the
implementation of the 20 day action plan the very next day.
4.4.2. Results
At the end of the kaizen event, participants have built an action
plan to quickly implement those changes, ideally in 20 working days.
The 20 day action plan is the main output of the Kaizen event. Since
the improving scenarios were tested during the Kaizen event, it will
be easier to implement them rapidly. Nineteen weeks after implementing the modications, new data was collected. Since trajectories 1 and 2 were divided over 2 days, it was impossible to compare

results before and after testing improving propositions. Only trajectory 3 (treatment) could be analyzed more closely (Table 7).
Patient lead times for trajectory 3 were reduced from 61 to
16 minutes, a 74 percent reduction. Simulated results (during the
kaizen) promised a 90 percent reduction. The gap between what was
expected and the real value after implementation can be explained
by limits of the discrete event simulation model which did not take
into account patients lateness, treatments beginning late, treatments
not ready on time, patients not feeling well and other human related
events that cannot be modeled. Consequently the discrete event simulation model overestimated lightly the expected improvement but
it is still an appropriate manner to evaluate new ideas to reach the
objective. Results show that the 34 minutes target for trajectory 3
(55 percent 61 minutes) has been reached. The challenge is now
to ensure that those results are maintained.
4.5. Control
Control is assured by weekly meetings called weekly huddle, led
by the clinic manager with different members of the clinic. They verify patients waiting time and propose new actions if necessary to still
be able to reach the target.
5. Discussion
5.1. From a Lean and simulation perspective
Robinson et al. (2012) showed that Lean and simulation are complementary methods even though they are often used independently.
SimLean Educate involves the use of existing models before the
Kaizen event in teaching key Lean principles, SimLean Facilitate involves rapid modeling during the Kaizen event to better understand
the dynamic of the process and SimLean Evaluate involves the development of a detailed model after the Kaizen event to evaluate scenarios. Our approach proposes to develop a single detailed simulation
model (Fig. 7).
The time required to develop a detailed simulation model is not
shorter than Robinson et al. (2012). We used the model before the
Kaizen event (Educate step), to generate different improving scenarios and eliminate the least interesting ones (those having less impact
over the reduction of patients waiting time). This leaves more time
during the Kaizen event to provide guidance to the team to nd feasible solutions. The Kaizen event is used to nd solutions to improve
the process performance and reach an identied target. Kaizen events
are costly considering the salary of 10 participants during 3 days. Consequently some steps (process mapping, time study and data analysis) are done before the Kaizen event leaving more time to participants to nd solutions. During the Kaizen event, the discrete event
simulation model contributes to engage participants to discuss different points of views and to provide evidence in order to achieve
consensus. As Robinson et al.s model, ours helps in managing conicts of interest between team members. Simulation model is thus a
fundamental tool to evaluate the proposed solutions. Since the model
is valid participants are condent that the results will be closed to
what is expected in real life. This help reducing barriers when implementing Lean in the clinic such as skepticism and lack of ownership
of solutions (Brandao de Souza & Pidd, 2011).

336

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Detailed simulaon model

Role: educate/facilitate

Kaizen event

Implementaon

Detailed simulaon model

Role: experiment/evaluate

Before

During

Aer

Fig. 7. Our approach integrating detailed discrete event simulation in a Kaizen event.

Our approach allows a rapid implementation of the solutions generated during the Kaizen event compared to Robinson et al. (2014). Indeed the rst steps of the DMAIC approach (DMA) have been realized
in 6 months. The last two steps (IC) needed 3 months, 1 month for implementation (20 day action plan) and 2 months for follow-up. Our
approach allows reducing the implementation delays of Robinsons
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 solutions 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 stakeholder 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
clients organization are actively involved in determining the scope
of the project, analyzing and solving the problem. This team is supported 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 analyzed.
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, meetings and the Kaizen event. The intervention team consists of nurses,
pharmacists, doctors and clerks (stakeholder team). The clinic manager is responsible for implementing solutions and maintaining relations 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 during 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 involve them in the Lean steps and the development of the simulation
model.
Because Lean healthcare project often involves many stakeholders 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 pharmacist and 2 assistant pharmacists. These meetings are considered
benecial for the success of the project while reducing resistance to
change (Brandao de Souza & Pidd, 2011).

Four group meetings were organized. The rst meeting (DEFINE

step) is to present the team project, their role and the steps of the
whole project including discrete event simulation to employees. Unlike Kotiadis et al. (2014) this meeting is informative instead of participative since the project objective has been previously dened by the
management team, the facilitator and the clinic manager. The second meeting (MEASURE step) is used to validate the process mapping made by the Lean facilitator that was posted in the clinic so
that everyone could see it. They all discuss about possible modications. The presence of the university team to that meeting is crucial to help them understand the process and model it adequately
according to the process mapping. This is a participative meeting
comparable to workshop 2 (stage 3: specify conceptual model) of
Kotiadis et al. (2014). In Robinson et al. (2014) the process mapping
is also realized before the Kaizen event but it is only nished day 1
of the Kaizen. In our case the whole process mapping is completed
before the Kaizen event. During the third meeting (ANALYZE step) results from the MEASURE steps are presented: patients waiting time,
personnel tasks, treatment room capacity compared to the demand,
work schedules, appointments schedules, etc. It is more an informative meeting led by the Lean facilitator. It allows the university team
to capture the details needed to build the simulation model. The goal
of the fourth meeting (ANALYZE step) consists in validating the simulation model and presenting simulation to the team members who
can propose modications that will be studied before the Kaizen
event. It is led by the university team and the Lean facilitator. This
meeting is participative and comparable to workshop 3 (stage 5: experimentation) of Kotiadis et al. (2014). Like Kotiadis et al. (2014) and
Robinson et al. (2014), model coding was done apart from the meetings and the Kaizen event. In our approach, model coding was realized at the ANALYZE step right before the Kaizen event. In Kotiadis
et al. (2014), model coding is done at step 4 (model coding) and in
Robinson et al. (2014), it is done between 2 days of the Kaizen event.
The Kaizen members received training on Lean principles and
simulation (ANALYZE step) from the Lean facilitator and the university team. Our approach is different from Robinson et al. (2014) since
it is devoted to nding solutions, identifying the best scenario and
writing the action plan. Kotiadis et al. (2014) did not have a Kaizen
event.
The organization of our Kaizen event differs from what is found in
literature in two ways. It uses a detailed simulation model during the
Kaizen event to evaluate improving scenarios proposed by the Kaizen
team in a unique way. Indeed Kotiadis et al. (2014) evaluate scenarios outside the workshops. Robinson et al. (2014) uses the simulation
model to evaluate scenarios during the Kaizen event. However the results are not precise since the model was developed with estimated
data. Therefore participants are not able to know if the goal has been
reached. Let us recall that the target consists in a 45 percent reduction of patients waiting time. Solutions at the end of the 3 day Kaizen
event must allow reaching this goal. The simulation model is thus an

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Aout

September

July

June

May

April

March

Febuary

X
X

January

Intervenon team

X
X

Kaizen team

Clinic manager

X
X
X

Timeline

university Team (expert)

Lean facilitator

Steps and acvies

Hospital manager

Project team

337

DEFINE
Project cha rter
Project l a unch
Intervi ews

X
X

X
X
X

MEASURE
Ma ppi ng proces s
Ma ppi ng proces s va l i da on
Ti me s tudy
The cl i ni c

X
X
X
X

X
X
X

ANALYZE
Da ta a na l ys i s a nd compi l a  on
Model codi ng
Si mul a on model va l i da on
Prel i mi na ry tes ts wi th the s i mua l on model
Pres entaon of the res ul ts from the mea s ure s tep
Devel opment of bus i nes s ga me
Tra i ni ng on l ea n pri nci pl es a nd s i mul a on

X
X
X
X
X
X

X
X
X

X
X

X
X

X
X

X
X

X
X
X
X
X
X
X

IMPROVE
Ka i zen event(3 da ys )
Acon pl a n i mpl ementaon

CONTROL
Weekl y huddl e
Moni tori ng i ndi ca tors

Fig. 8. Schedule and roles of team members during the Lean project.

interesting tool. In Robinson et al. (2014) the target is not dened precisely. Second using a business game to generate appointment schedules during the Kaizen event is an innovative contribution.
Scheduling appointments in a hematologyoncology clinic requires human judgment even if there are guidelines. Business game
helped clerks to test different guidelines, better understand them
and take note of the results. By scheduling appointments themselves
through business game, they realized that they were able to detect
potential problems for the future. Business game allows participants
to test in real-time their ideas and to detect rapidly potential problems. More realistic scenarios can then be proposed. Arrivals in discrete event simulation are modeled as appointment schedule. The
evaluation being performed during the Kaizen event, the best ones
are chosen immediately accelerating their implementation. The use
of business game during the Kaizen event encourages participants
to be more creative. Participants in the Kaizen event could evaluate the impact of the proposed changes on their work and anticipate
future problems. They could then propose solutions immediately. It
also allows taking into account human aspects when determining solutions and make the Kaizen event more dynamic. Like van der Zee
and Slomp (2009), we conclude that the game could help workers
nd solutions for specic problems and facilitate their acceptance of
new work methods or systems. Our results demonstrate that business game combined with a discrete event simulation can be used to
support participants during the Kaizen event.
Finally since the 20 day action plan is dened during the third
day of the Kaizen event this is comparable to workshop 4 (stage 6:
implementation) of Kotiadis et al. (2014).

idea of using simulation models. Team members must accept to give

individual interviews and participate in group meetings during working hours. This implies coordination to avoid disturbing the clinic activities and to obtain a high level of participation. The Lean facilitator
must put a lot of effort to plan, coordinate meetings and follow the
steps rigorously.
Robinson et al. (2014) propose the development of a simple simulation model in 2 days which is less time than what our model requires. Our approach is closer to Kotiadis et al. (2014) since it implies
a detailed simulation model requiring more time to be developed.
However the model reproduces more adequately reality and can be
used during the Kaizen event to help participants. Considering all the
effort required in developing the simulation model and from participants, our approach is more convenient to solve complex problems.
Finding solutions during the Kaizen event is crucial in our
approach. The scope of the problem must be well identied so
that participants can work toward the same target. It requires the
participation of at least one representative of each profession over
three consecutive days for the Kaizen event. Otherwise the Kaizen
event cannot be organized. Even though it can be perceived as a
constraint it is a success factor. Participants become ambassadors of
the solution to their colleagues facilitating implementation. The use
of Kaizen and the business game encourages participants to be more
condent in results and to rapidly implement the 20 day action plan.
The university and project teams are co-partners with roles and
responsibilities well dened (Fig. 8). The DMAIC is well structured
and rigorous for the project to progress adequately and meet the
schedules.

5.3. When to use this approach?

5.4. Lessons

Our approach helps maintaining interactions between participants during a Lean project involving the development of a simulation model during a Kaizen event to evaluate different improving
scenarios. It could be used not only in healthcare applications. It requires a team eager to be involved in the project and opened to the

Four success factors have been identied. It needs the involvement of doctors, pharmacists, managers and employees. It conrms
the importance to have doctors and pharmacists present during the
Kaizen event. Second it requires a culture of continuous improvement
in the organization and among the team members. Indeed before the

338

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Kaizen event the team had already begun to implement modications 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 schedules 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 sustained 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 ambitious 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 considered as participative because of the numerous meetings (individual
and group). Our study showed that the use of simulation and a business game in a Kaizen even favors participation of all members. It also
has for principal advantage to help nding 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). Finally data have been collected 3 months after implementing the solution. 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 reduction utilizing lean tools applied to healthcare: The inpatient pharmacy at a local
hospital. Journal for Healthcare Quality, 32, 5966.
Ashenbaum, B. (2010). The twenty-minute just-in-time exercise. Decision Sciences Journal of Innovative Education, 8(1), 269274.
Berg, B., Denton, B., Nelson, H., Balasubramanian, H., Rahman, A., Bailey, A., et al.
(2009). A discrete event simulation model to evaluate operational performance
of a colonoscopy suite. Medical Decision Making, 30(3), 18.
Billington, P. J. (2004). A classroom exercise to illustrate lean manufacturing pull concepts. Decision Sciences Journal of Innovative Education, 2, 7176.
Bok, B. M. (2007). Experiential foresight: Participative simulation enables social reexivity in a complex world. Journal of Futures Studies, 12(2), 111120.
Bowers, J., Ghattas, M., & Mould, G. (2012). Exploring alternative routes to realising the
benets of simulation in healthcare. Journal of the Operational Research Society, 63,
14571466.
Brailsford, S. C., Harper, P. R., Patel, B., & Pitt, M. (2009). An analysis of the academic
literature on simulation and modelling in health care. Journal of Simulation, 3(3),
130140.
Brando de Souza, L., & Pidd, M. (2011). Exploring the barriers to lean health care implementation. Public Money & Management, 31(1), 5966.
Brandao de Souza, L. (2009). Trends and approaches in lean healthcare. Leadership in
Healthcare, 22(2), 121139.
de Mast, J., & Lokkerbol, J. (2012). An analysis of the Six Sigma DMAIC method from the
perspective of problem solving. International Journal of Production Economics, 139,
604614.
Dickson, E. W., Singh, S., Cheung, D. S., Wyatt, C. C., & Nugent, A. S. (2008). Application of
lean manufacturing techniques in the emergency department. Journal of Emergency
Medicine, 37(2), 177182.
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 Journal of Innovative Education, 12(4), 287319.
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 population health and health care delivery. Journal of Public Health Medicine, 25(4),
325335.
Franco, L. A., & Montibeller, G. (2010). Facilitated modelling in operational research.
European Journal of Operational Research, 205, 489500.
Greco, M., Baldissin, N., & Nonino, F. (2013). An exploratory taxonomy of business
games. Simulation & Gaming, 44(5), 645682.

Griths, J. D., Jones, M., Read, M. S., & Williams, J. E. (2010). A simulation model of bed
occupancy in a critical care unit. Journal of Simulation, 4(1), 5259.
Hagtvedt, R., Grin, P., Keskinocak, P., & Roberts, R. (2009). A simulation model to compare strategies for the reduction of health-care-associated infections. Interfaces,
39(3), 256270.
Houchens, N., & Kim, S. C. (2014). The application of lean in the healthcare sector:
Theory and practical examples. In Wickramasinghe, et al. (Eds.), Lean thinking for
healthcare (pp. 4353). New York: Springer.
Jahangirian, 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 of
the Operational Research Society, 66(3), 369379.
Jun, J. B., Jacobson, S. H., & Swisher, J. R. (1999). Application of discrete-event simulation
in health care clinics: A survey. Journal of the Operational Research Society, 50(2),
109123.
Kaplan, G., & Patterson, S. (2008). Seeking perfection in healthcare: A case study in
adopting Toyota Production System methods. Healthcare Executive, 23, 1621.
Katsaliaki, K., & Brailsford, S. C. (2007). Using simulation to improve the blood supply
chain. Journal of the Operational Research Society, 58(2), 219227.
Kelton, W. D., Sadowski, R., & Sturrock, D. (2007). Simulation with arena (4th ed., p. 636).
New York: McGraw-Hill.
Khurma, N., Bacioiu, G. M., & Pasek, Z. J. (2008). Simulation-based verication of lean
improvement for emergency room process. In Proceedings of the 2008 winter simulation conference (pp. 14901499).
Klassen, K. J., & Yoogalingam, R. (2009). Improving performance in outpatient appointment services with a simulation optimization approach. Production and Operations
Management, 18(4), 447458.
Kotiadis, K., Tako, A., & Vasilakis, C. (2014). A participative and facilitative conceptual
modelling framework for discrete event simulation studies in healthcare. Journal
of the Operational Research Society, 65(2), 197213.
Kotiadis, 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 take
action: A case study in health care. Journal of the Operational Research Society, 64(1),
125137.
Lean Enterprise Institute (2014). Lean lexicon: A graphical glossary for lean thinkers (5th
ed., p. 131). Cambridge, MA: Lean Enterprise Institute.
Lodge, A., & Bamford, D. (2008). New development: Using lean techniques to reduce
radiology waiting times. Public Money & Management, 28(1), 4952.
Martin, C. H. (2007). A simulation based on Goldratts matchstick/die game. Decision
Sciences Journal of Innovative Education, 5(2), 423429.
Martin, K., & Osterling, M. (2007). The Kaizen event planner: Achieving rapid improvement
in oce, service, and technical environments. New York: Productivity Press, pp. 240.
Mielczarek, B., & Uzialko-Mydlikowska, J. (2012). Application of computer simulation
modeling in the health care sector: A survey. Simulation, 88(2), 197216.
Nelson-Peterson, D., & Leppa, C. (2007). Creating an environment for caring using lean
principles of the Virginia Mason Production System. Journal of Nursing Administration, 37(6), 287293.
Ogulata, N. S., Cetik, M. O., & Koyuncu, E. (2009). A simulation approach for scheduling
patients in the department of radiation oncology. Journal of Medical Systems, 33,
233239.
Paul, S. A., Reddy, M. C., & De Flitch, C. J. (2010). A systematic review of simulation
studies investigating emergency department overcrowding. Simulation, 86(89),
559571.
Poksinska, B. (2010). The current state of lean implementation in health care: Literature
review. Quality Management in Health Care, 19(4), 319329.
Proudlove, N., Moxham, C., & Boaden, R. (2008). Lessons for lean in health care from
using six sigma in the NHS. Public Money & Management, 28(1), 2734.
Proudlove, N. C., Black, S., & Fletcher, A. (2007). OR and the challenge to improve the
NHS: Modelling for insight and improvement in in-patient ows. Journal of the
Operational Research Society, 58, 145158.
Radnor, 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 Scottish
Government.
Robinson, S., Radnor, Z. J., Burgess, N., & Worthington, C. (2012). SimLean: Utilising simulation in the implementation of lean in healthcare. European Journal of Operational
Research, 219, 188197.
Robinson, S., Worthington, C., Burgess, N., & Radnor, Z. J. (2014). Facilitated modelling
with discrete-event simulation: Reality or myth? European Journal of Operational
Research, 234, 231240.
Rohleder, T. R., Lewkonia, P., Bischak, D. P., Duffy, P., & Hendijani, R. (2011). Using simulation modeling to improve patient ow at an outpatient orthopaedic clinic. Health
Care Management Sciences, 14, 135145.
Santibanez, P., Chow, V. S., French, J., Putterman, M. L., & Tyldesley, S. (2009). Reducing
patient wait times and improving resource utilization at British Columbia Cancer
Agencys ambulatory care unit through simulation. Health Care Management Science, 12, 392407.
Sepulveda, 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. In
Proceedings of the 1999 winter simulation conference (pp. 15411548).
Swanson, L. (2008). The lean lunch. Decision Sciences Journal of Innovative Education,
6(1), 153157.
Tako, A., Kotiadis, K., & Vasilakis, C. (2010). A participative modelling framework for
developing conceptual models in healthcare simulation studies. In Proceedings of
the 2010 winter simulation conference (pp. 500512).
Tako, A., Kotiadis, K., & Vasilakis, C. (2010). Developing a conceptual modelling framework for stakeholder participation in simulation studies. In Proceedings of the 5th
operational research society simulation workshop (SW10) (pp. 7685).

C. Baril et al. / European Journal of Operational Research 249 (2016) 327339

Tako, A. A., & Kotiadis, K. (2015). PartiSim: A multi-methodology framework to support
facilitated simulation modelling in healthcare. European Journal of Operational Research, 244, 555564.
van 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, 1728.
Van Lent, W. A. M., Goedbloed, N., & Van Harten, W. H. (2009). Improving the eciency
of a chemotherapy day unit: Applying a business approach to oncology. European
Journal of Cancer, 45(5), 800806.

339

White, D. L., Froehle, C. M., & Klassen, K. J. (2011). The effect of integrated scheduling
and capacity policies on clinical eciency. Production and Operations Management
Society, 20(3), 442455.
Young, T., Brailsford, S., Connell, C., Davies, R., Harper, P., & Klein, J. H. (2004). Using
industrial processes to improve patient care. British Medical Journal, 328, 162164.