LSSGB SVES January 2018 Final PDF

1/1/2018
Excellence Through Innovative Research
WARNING
All rights reserved. No part of the course materials

Lean Six Sigma Green Belt
used in the instruction of this course may be
reproduced in any form or by any
Training
electronic or mechanical means, including the use
of information storage and retrieval systems,
without written approval from the copyright owner.
© Binghamton University
State University of New York
Binghamton University 1
Overall Agenda Meet The Instructor

Course Introduction Dr. Mohammad T. Khasawneh, Ph.D.
Meet The Instructor Professor & Chair, Systems Science and Industrial Engineering
Associate Director, Watson Institute for Systems Excellence
Course Objectives Director, Healthcare Systems Engineering Center
Course Outline Thomas J. Watson School of Engineering and Applied Science
State University of New York at Binghamton
Continuous Process Improvement
Binghamton, New York 13902, USA
Fundamental Concepts
Lean Contact Information:
[email protected]
Six Sigma
(607) 777-4408 (phone)
Lean Six Sigma (607) 777-4094 (fax)
DMAIC Framework
Case Studies
* SUNY Chancellor’s Award for Excellence in Teaching
Binghamton University 3 Binghamton University 4
1
1/1/2018
Instructor at a Glance Objectives

Joined Binghamton University, SUNY, in Fall of 2003
Research areas: Recognize key attributes of a successful lean six sigma
• Modeling and optimization of healthcare delivery systems
program
Leading/co-leading wide spectrum of projects with
multiple U.S. hospital systems
Understand the roles and responsibilities of a certified
35+ journal papers and 100+ conference papers
$7.00M+ Research funding Lean Six Sigma professional
$39M+ In-kind software grants
Research group Understand fundamentals of the DMAIC problem solving
• 20 PhD students
• 10 MS students methodology
Objectives (cont’d) Course Outline
Use basic lean six sigma tools for project definition and Continuous process improvement
Introduction to lean
process baseline
Introduction to six sigma
Understand the need for advanced problem solving and
DMAIC (Define, Measure, Analyze, Improve, and
improvement methodologies used by project teams Control)
• Define Phase: Project charter; process mapping; voice of
Integrate lean six sigma knowledge and tools to
the customer; 7 "new" management and planning tools; 7
successfully implement and deploy process "old" quality control tools; CTQ trees; SIPOC diagram; etc.
improvements
2
1/1/2018
Course Outline (cont’d) Course Outline (cont’d)
• Measure Phase: Basic probability and statistics for six • Control Phase: Statistical process control; process
sigma; Measurement System Analysis (MSA); Gauge R&R capability analysis; cost savings and Return On Investment
studies; process capability analysis; benchmarking; etc. (ROI) calculations; mistake-proofing; control plan; etc.
• Analyze Phase: Correlation analysis; regression analysis;

confidence intervals; hypothesis testing; etc. Spreadsheet-based statistics for six sigma
• Improve Phase: Design Of Experiments (DOE); Analysis
of Variance (ANOVA); Failure Mode and Effects Analysis
Case Studies/Projects
(FMEA); house of quality and Quality Function
Deployment (QFD); simulation software;
action/communication plan; etc.
CI Definition
‘A comprehensive management philosophy

that focuses on continuous improvement by
applying scientific methods to gain
knowledge and control over variation in work
processes’
Tindill & Stewart (1993)
3
1/1/2018
Knowledge Hierarchy Kaizen

Japanese expression
Importance Wisdom In English = continuous improvements
Kaizen
Morals
• Changes to better, improve
Understanding
• Finding lots of “little” ideas
Principles
Knowledge
Patterns
Information
Relationships
Data
Learning
Healthcare Quality managements. (2008) Health Administration Press.
Kaizen (cont’d) CI Gurus/ Pioneers

Aspects
• Eliminating waste (muda)
• Improve efficiency
• Improve processes
• Improve moral
Philosophy
• Standardize
• Make improvements, changes
• Standardize/measure again
. McLaughlin & Kaluzny (2004)
4
1/1/2018
W. Edwards Deming Joseph M. Juran

Father of the quality revolution Quality trilogy
1970s energy crisis • Quality planning
Common cause versus special variation • Control
System of profound knowledge • Improvement
• Appreciation for a system
• Knowledge about variation
• Theory of knowledge
• Knowledge of psychology
McLaughlin & Hays (2008) McLauhlin & Hays (2008)
CI – PDSA Cycle CI – Quality Circles

Derived from PDSA
Group of volunteers
Under supervision
Identify, analyze and solve work related problems
Improve performance
Langley et al. (1994)
5
1/1/2018
Important Events in Performance Improvement
McLaughlin & Hays (2008)
Appearance of Lean In Healthcare Introduction to Lean
Definition of lean
Types of waste
Tools
Kaizen
Value stream mapping
Luciano Brandao de Souza, 2009. Trends and approaches in lean healthcare. Leadership in Health Services. 22(2): 121–139.
6
1/1/2018
What’s Lean? Value Added vs. Non-value Added

Elimination of waste
• Toyota Production System (TPS)
• An action that a • An activity that
customer is willing to consumes resources
Philosophy
pay for without creating value
• Produce only what is needed, when it is needed, with no
waste for the customer
• Introduces two key terms – ‘Value’ & ‘Waste’ • An activity that • An activity that is
Methodology transforms a product unpredictable in
• Determination of value added in the process or service creating value
Tools • An activity that
• Takt time, throughput time, five Ss, spaghetti diagrams, • An activity done requires more time,
standardized work, jidoka, andon, kanban, SMED, flow and correctly the first time effort or resources than
pull, heijunka, advanced access necessary
Seven Types of Wastes (Muda)

MUDA is the Japanese word for WASTE. An 8th waste
is the wasted
potential
5 7 2 of people
1 4 3 6 Overproduction To produce sooner,faster
Seek it out and get rid! or in greater quantities
than customer demand.
Over Processing 1 Inventory
Processing beyond
the standard
required by the
7 2 Raw material,
work in progress
customer. or finished goods
which is not having
value added to it.
Rework
Non right Waiting
first time.
Repetition 6 3 People or parts
or correction that wait for
of a process. a work cycle to
be completed.
Transportation 5 4 Motion
Unnecessary movement
of people, parts or
Unnecessary movement of people machines within
or parts between processes.
a process.
7
1/1/2018
Tools & Techniques Essential Flowcharting or Imagineering Symbols
• Brainstorming • Pareto Principles &

Analysis
• Flow Charting &
“Imagineering” • Run Charting
• Cause & Effect • Control Charting

Diagramming
Basic Symbols in Process Mapping Other Essential Flowcharting Symbols

Start & End: the materials, information or action (inputs) to
start the process or to show the results at the end (output)
of the process
Activity: a task or activity performed in the process
Direction: flow arrows show the order in which activities

are completed
Decision: points in the process where a yes/no question is Yes

being asked or a decision is required
No
Delay: wait or hold until next step occurs
Inventory: can also indicate customer waiting
8
1/1/2018
Types of Process Maps

High-level flow chart
• Brief overview of a process only
highlighting major events
Detailed flow chart

• Mark every step in a process
Swimlane diagram
• Displays processes carried out
for multiple roles across
multiple stages
• AHRQ (2013). Mapping and Redesigning Workflow: A Guide on Workflow Mapping. Available at:
http://www.ahrq.gov/professionals/prevention-chronic-care/improve/system/pfhandbook/mod5appendix.html#sl1
Systems Science and Industrial Engineering

Binghamton University 33 34
A Short Video – Quality A Short Video – Systems Approach
Systems Science and Industrial Engineering 35 Systems Science and Industrial Engineering 36
9
1/1/2018
Pareto Principle Pareto Charts

•Based on 80-20 concept, where the majority of effects are
caused by a minority of issues
Isolate “vital few causes” from “trivial many”
Pareto Principle: 80 / 20 Rule
20% Items Cause 80% Problems
De Mast & Trip (2007)
Pareto Analysis:
What Makes a Product More Difficult to Remanufacture How to Create a Pareto in Minitab
Hammond et al. (1998)
10
1/1/2018
How to Create a Pareto in Minitab (cont’d) Example 1 – Error Type While Downloading File
Error Type Frequency

Can't Change Background 5
Can't Find the File 30
Can't Open the File 3
Can't Save Changes 1
Doesn't Work in OpenOffice 1
Don't Have Excel 1
Found a Bug 2
Opens as Read-Only 15
Unable to Download 50
Example 2 – Late Arrival to Work Cause and Effect Diagram (Ishikawa)
Ishikawa developed with Deming to study

processes in Japan
Reason for Lateness Frequency
Child Care 8 • Identify • Display
Emergency 2
Overslept 3
Public Transportation 4 • Isolate • Demonstrate
Traffic 12
Weather 6
• Explore • Understand
4Ps vs. 4Ms
11
1/1/2018
Ishikawa (Fishbone) or Cause and Effect Diagram Example: Inner/Outer Rings of Ball Bearings
Reddy & Reddy (2010)
Eliminate Bottlenecks
Bottlenecks:
• A point in a process where Demand > Capacity
• A chokepoint or problem area where the available

Optimizing Process Flows resources (e.g., staff/workers, supplies,) are not
enough to keep up with demand
• De-bottlenecking a process is a systematic process to

identify and eliminate obstacles
12
1/1/2018
Identifying Bottleneck Capacity
Available recourses, supplies, equipment, space, etc.
Labor of key providers (depends on industry/service
type)
Availability and uptime of key technologies
Supplies and other resources
Facility, inventory, services
Types of Capacity
Design (or Theoretical) Capacity:

• Maximum stated capacity of a resource (for example, a
piece of technology might be capable of operating 24
hours per day, or a labor be able to operate 8 hours per
day); also known as Cd
Effective (or Actual) Capacity:

• Actual capacity that can be expected over normal
operations, after adjustments for average uptime and
utilization; Ce
Ce = Cd x U
13
1/1/2018
House of Lean Lean Process
Continuous Learning
Highest quality, lowest cost,
shortest lead time by
eliminating wasted time and activity Capabilities & Skills Lean Building Blocks
Decision
Strategic Planning C&E Pareto chart
support Plannng
Diagram systems
Employee Empowerment
Just in Time (JIT) Jidoka Information & Support
PDSA Advanced
• Produce what • Manual or Systems Access
Push & Pull
Financial Excellence
Process Excellence
Client Satisfaction
Results Feed the
you need automatic line Pull Value Organization’s
• Non-batching stop & Streamline Flow Takt Time Kanban
SMED Andon Balanced Score
• Pull • Separate Card
Structured
operator and Jidoka
Problem Solving Standardized work FMEA
machine Plan -Do-Check -Act
activities Spaghetti Diagrams
5S Kaizen
• Error proofing Basic Stability
Events
• Visual control
Basic Knowledge Throughput
Visual Management Hejunka
Time
(Getting Started) Value
Operational Stability
High Performing
Transformational
Stream
5S Level loaded Standard Work Preventative Maintenance Value Leadership Mapping
Lean Tools
Push & Pull Value Stream Map
Visual Control Heijunka
Kanban Time and Motion
Takt Time SIPOC
Throughput and Throughput
Time
Standardized Work Lean Tools
Jidoka & Andon
Mistake Proofing (Poka-Yoke) • Lean tools for identifying opportunity
Single Minute Exchange of Die • Lean tools for improving flow
(SMED) • Lean tools for improving quality/reducing errors
5Ss • Lean tools for implementation
Kaizen Blitz Event
Spaghetti Diagram
14
1/1/2018
Lean Tools for Identifying

Opportunity Lean Tools for Improving Flow
• 5S
• Value stream mapping
• Visual management
• Spaghetti diagram
• Push/Pull (Kanban)
• Time and motion study
• Heijunka
• Single-minute-exchange-of-die
Lean Tools for Improving

Quality Lean Implementation
• Jidoka • Kaizen
• Poka-Yoke • PDCA
• Root Cause Analysis • A3 reporting
15
1/1/2018
Takt Time Takt Time (cont’d)

The speed with which customers
Cycle time
must be served to satisfy demand for the service
Two quantities are required: • Time to accomplish a task in the system
D = Customer Demand
W = Available Work Time System cycle time
• Longest task cycle time in the system—the rate at which
customers or products exit the system, or “drip time”
Or,
T = W / D (time unit)
Throughput Time River View Clinic Example
Time for an item to complete the entire process, which

includes:
• Waiting time
• Transport time
• Actual processing time
Given that, Calculate:

No of Physicians: 8 • System cycle time
Working hours: 5 hours/day • Throughput time
Patients arrival rate: 100 per day • Takt time
McLaughlin & Hays (2008) Healthcare quality management. (2008) Health Administration Press
16
1/1/2018
Heijunka (cont’d) Heijunka (cont’d)

• Leveling workload based on Takt time • Leveling workload based on Takt time
• Distribute work activities across the value stream to meet takt time • Example: New patient visit in outpatient clinic
• Example: New patient visit in outpatient clinic
20
Step Activity Cycle Time 15 mins

15
Registration Set up a new file in system 2 minute Current Takt time = 12 mins
10
6 mins
Input patient demographic 10 minutes 5 mins
4 mins
information
5
Filling insurance information 3 minutes
0
Medical Assistant Prep Record patient chief complaint 2 minutes Registration MA Prep Nurse Assess MD Assess
Prepare patient and report to nurse 2 minutes Re-design

Nurse Assessment Take vital signs 3 minutes 15
12 mins
Takt time = 12 mins
Take medical history 2 minutes
10
Physician assessment Diagnosis and treatment 6 minutes 7 mins
6 mins
5 mins
Individual Step cycle time: 15mins, 4 mins, 5 mins, 6mins 5
Average new patient demand per day: 35 patients
Available working time per day: 420 minutes
Takt time: 420/35 = 12mins 0
Binghamton University 65 Registration MA Prep Nurse
Binghamton Assess
University MD Assess 66
SIPOC SIPOC – Definitions

Suppliers: The individuals, departments, or organizations that
Capture all the relevant information about the process that is provide the materials, information, or resources in the process
being reviewed Inputs: The information or materials provided by the suppliers;
Inputs are transformed, consumed, or otherwise used by the process
Construction:
Process: The macro steps (typically 4-6) or tasks that transform the
Step 1: Create a simple, high-level process map of the process and record it in Process column inputs into outputs: the final products or services
Step 2: Identify the outputs of this process and record them in Outputs column Outputs: The products or services that result from the process
Step 3: Identify the customers who will receive the outputs and record them in Customers Customers: The individuals, departments, or organizations that
column receive the outputs, the products or services, generated by the process
Step 4: Identify the inputs needed for the process to create the outputs and record them in
inputs column Input Metric: Quantitative measurements used to define quality,
Step 5: Identify the suppliers of the inputs and record them in Suppliers column
time or cost of the inputs
Step 6: Clean up the lists by analyzing, rephrasing, combining, moving, etc. Process Metric: Quantitative measurements used to define quality,
Step 7: Create SIPOC diagram and modify it if necessary time or cost of the process
Output Metric: Quantitative measurements used to define quality,
time or cost of the output
SIPOC: Suppliers-Inputs-Process-Outputs-Customers
17
1/1/2018
SIPOC – Template
Suppliers Inputs Process Outputs Customers
(Specifications) (Name) (Requirements) Excellence Through Innovative Research
Step/Activity # 1
Step/Activity #2
Six Sigma
.
.
.
Step/Activity #n
(Pyzdek, 2003) Binghamton University 69
Agenda
Six Sigma Introduction
Lean Six Sigma
Project Management
DMAIC
Case Studies
18
1/1/2018
Do You Know What is Six Sigma? What is Six Sigma?

Sigma is Greek letter representing the standard deviation of a
population of data
• Sigma is a measure of variation (the data spread)
The term “Sigma” is a measurement of how far a given process
deviates from perfection – a measure of number of “defects”
• A process operating at a “Six Sigma” level produces only 3.4
defects per million opportunities (DPMO) for a defect
“A quality improvement methodology that applies statistics to
measure and reduce variation in processes”
A management system that is “comprehensive and flexible for
achieving, sustaining, and maximizing business success”
A disciplined process using customer requirements, data, and
facts to improve business performance by consistently
meeting customer expectations
Pexton (2005)
Introduction to Six Sigma Introduction to Six Sigma

Target
Six Sigma (SS) aims at reduction and removal of variation Target

Customer Customer
Upper Limit Upper Limit
ZERO
• By application of extensive set of statistical tools (~140)

DEFECTS
Product Performance Product Performance
• It is a strategic company wide approach W I D E Variation Slim Variation
Variation means that a process does not produce the same

• Simultaneous reduction of cost and improvement in result (the “Y”) every time
• Variation directly affects customer experience
customer satisfaction Six Sigma focuses on reduction of variation to aid in improvement of
Overall Quality and Cost
Bendell (2006)
19
1/1/2018
Six Sigma Metrics What is Six Sigma? (cont’d)

Defect (nonconformance)
• Any mistake or error that is passed on to a customer
Defects per unit (DPU) Statistically... Sigma
• Number of defects discovered / number of units produced Level DPMO
Defects per million opportunities (DPMO) Six Sigma refers to a 1 691,462
Goal
• (Number of defects discovered / opportunities for error) * process that produces 2 308,537
1,000,000 only 3.4 Defects Per 3 66,807
Critical to Quality (CTQ): Attributes most important to the Million Opportunities 4 6,210
5 233
customer
6 3.4
Process Capability (PC): What your process can deliver?
Stable Operations: Ensuring consistent, predictable processes to

improve what the customer perceives
What is Six Sigma? (cont’d) How good do we need to be?
Target Customer The Classical View of Quality The Six Sigma View of Quality
Specification “99% Good” (Z = 3.8σ) “99.99966% Good” (Z = 6σ)
BEFORE
3s 6.6% Defects
3s 20,000 lost articles of mail per hour
Unsafe drinking water almost

Seven lost articles of mail per hour
One minute of unsafe drinking water

15 minutes each day every seven months
w i d e v a r i a n c e 5,000 incorrect surgical operations 1.7 incorrect surgical operations
Target per week per week
Customer
Specification 2 short or long landings at most One short or long landing at most
major airports daily major airports every five years
AFTER
6s No Defects
6s 200,000 wrong drug prescriptions
each year
68 wrong drug prescriptions
each year
slim variance No electricity for almost One hour without electricity

7 hours each month every 34 years
Customers don’t feel the averages, they feel the variability
20
1/1/2018
How good do we need to be? (cont’d) Sigma Comparison of Industries

The Classical View of Quality The Six Sigma View of Quality
“99% Good” (Z = 3.8σ) “99.99966% Good” (Z = 6σ)
There are 964 U.S. flight 1 U.S. flight is cancelled every 3
cancellations per day weeks
The police make 7 false There are fewer than 4 false arrests
arrests every 4 minutes per month
In one hour, 47,283 international It would take more than
long distance calls are accidentally 2 years to see the same number
disconnected of dropped international calls
Kachalia et al. (2007)
Six Sigma The Six Sigma Evolutionary Timeline
1818: Gauss uses the normal curve 1924: Walter A. Shewhart introduces
to explore the mathematics of error the control chart and the distinction of
analysis for measurement, probability special vs. common cause variation as
analysis, and hypothesis testing. contributors to process problems.
Project 1736: French 1896: Italian sociologist Vilfredo

Management mathematician Alfredo Pareto introduces the 80/20
Abraham de rule and the Pareto distribution in
Economics Customer Moivre publishes Cours d’Economie Politique.
Justification Focus an article
introducing the
normal curve.
1949: U. S. DOD issues Military
Procedure MIL-P-1629, Procedures
1960: Kaoru Ishikawa
Change for Performing a Failure Mode Effects
introduces his now famous
Statistical Methods 6σ and Criticality Analysis.
Management cause-and-effect diagram.
1941: Alex Osborn, head of 1970s: Dr. Noriaki Kano

BBDO Advertising, fathers a introduces his two-dimensional
widely-adopted set of rules for quality model and the three
“brainstorming”. types of quality.
Strategy Business
Alignment Processes
1986: Bill Smith, a senior
Quality engineer and scientist introduces 1995: Jack Welch
the concept of Six Sigma at launches Six Sigma at GE.
Tools Motorola.
Adapted from IBM Learning (2006) Bossidy (1994)
21
1/1/2018
Lean and Six Sigma – Lean Sigma
Lean Sigma combines Lean and Six Sigma methodologies

• Six Sigma tools and techniques
• Provide the quantitative skills required to identify and eliminate all
sources of variation within a process
Lean Sigma
• Follows the DMAIC methodology for project management
Lean and Six Sigma

• Complementary philosophies
Lean Sigma Lean and Six Sigma are Complementary
Lean and Six Sigma are focused on Six Sigma Methodologies

continuous improvement of systems Product or Service Outputs
Lean Six Sigma Increase efficiency
Eliminate waste Eliminate defects Simplify work flows

Focus on high-value steps
Increase consistency
Reduce variation
• Achieve flow • Reduce

Eliminate waste
Eliminate defects
A Lean enterprise is one that In a Six Sigma Customer-driven
Customer-driven delivers value to its
and pull variation in
enterprise, everyone is
stakeholders with little or focused on identifying
no wasteful consumption of and eliminating
processes resources. defects.
Lean Methodology
McLaughlin & Hays (2008) Adapted from IBM Learning (2006)
22
1/1/2018
Differences: Lean vs. Six Sigma The Lean Six Sigma Process
Program Six Sigma Lean Philosophy 2 ¾Map all of the

¾Define value from Map the steps…value added
Principle Reduce variation Remove waste the customers 1 Value Stream
Specify Value
& non-value
Define Identify value perspective and added…that bring a
express value in product of service to
Measure Define value stream
terms of a specific the customer
Methodology Analyze Determine flow product
3
Improve Create Pull 5 Establish
Work to Flow
Control Focus on perfection ¾ The complete Perfection ¾The continuous
Primary Objective Reduce Error/Defects Improve Flow elimination of waste movement of
so all activities create products, services and
Output of the system Many small value for the
4
Implement information from end
improves when improvements can customer Pull to end through the
Focus/Assumptions variations in the improve overall process
processes are system performance ¾ Nothing is done by the
upstream process until the
minimized
downstream customer
Nave et al. (2002)
signals the need
Implementing Lean Six Sigma through Kaizen Kaizen Event

A kaizen event is an implementation of lean six sigma
through
• 3.5 day, focused burst of continuous improvement activity
that results in an improved work area / process
Data collection/process • Consists of try-storming ideas
Sensing/Gathering VOC observations
• Brainstorm then pilot and observe results
Analyze data to
1 2 3 identify waste
• Kaizen events have specific objectives (usually one main
objective)
Value analysis, identify opportunity
• Reduce turn around time
• Improve workflow
• Reduce A/R days
5 4
Kaizen!
!
23
1/1/2018
Common Lean Six Sigma Methodologies Six Sigma – DMAIC

3 major Lean Six Sigma methodologies
DMAIC (Define, Measure, Analyze, Improve, Control)
• Design for Six Sigma (DFSS)
the project goals and customer (internal
• A.K.A, DMADV (Define, Measure, Analyze, Design and Validate)
and external) deliverables
• Six Sigma DMAIC the process to determine current
performance
• (Define, Measure, Analyze, Improve, Control)
and determine the root cause(s) of the
• A structured statistical problem solving approach using Six Sigma tools defects
and techniques to manufacturing, service, and business processes the process by eliminating defects
future process performance
• Lean Sigma
• Blends Lean principles and Six Sigma methodologies
• Lean methodologies grew out of an industry-wide automotive study
conducted by MIT on the Toyota Production System (TPS)
Six Sigma – DMADV Six Sigma Overview

To be successful as an organization,
DMADV (Define, Measure, Analyze, Design, Verify)
need focus on how to effectively apply three areas
the project goals and customer (internal and
external) deliverables
and determine customer needs and
Process Strategic planning helps
specifications Management to “focus” on key
the process options to meet the customer needs helps to projects to reach
(detailed) the process to meet the customer needs maintain good VISION
the design performance and ability to meet results to
customer needs
perform our Process Improvement
MISSION using the DMAIC
process helps to fix
work problems and
improves
PERFORMANCE
24
1/1/2018
Six Sigma Model DMAIC Model (cont’d)
Existing
Six-Sigma ROAD MAP Model
Existing Manufacturing
Business Suppliers Inputs /Service Outputs Customers
Processes Processes
Define
Six Sigma Measure
Methodology
Control DMAIC
Analyze
Improve
Improved
Business Quality Productivity Cost Profitability
Performance
Evans & Lindsay (2004)
DMAIC Model Key Players in Six Sigma Program
Voice of the Customer Champions/Sponsors: Trained business leaders who lead the deployment
of Six Sigma in a significant business area
Measure Analyze Improve

Master Black Belts: Fully-trained quality leaders responsible for
Six Sigma strategy, training, mentoring, deployment and results
Define Control Black Belts: Fully-trained Six Sigma experts who lead
improvement teams, work projects across the business
and mentor Green Belts
Green Belts: Fully-trained individuals who apply
Six Sigma skills to projects in their job areas
Yellow Belts: Individuals who
Institutionalization receive specific Six Sigma training
and who support/direct/guide
projects in their areas
The DMAIC Model
25
1/1/2018
Green Belt Roles & Responsibilities Black Belt Roles & Responsibilities
Roles Responsibilities
Roles Responsibilities
Leads business process improvement projects where
Six Sigma technical expert
Six Sigma project Six Sigma approach is indicated
Demonstrated mastery of Green Belt body of knowledge
originator Temporary, full-time
Six Sigma project Demonstrated proficiency at achieving results through change agent (will return
originator the application of the Six Sigma approach to other duties after Successfully completes high-impact projects that result
completing a two to three in tangible benefits to the enterprise
Recommends and participates on Six Sigma projects year tour of duty as a
Part-time Six Sigma Leads Six Sigma teams in local improvement projects Black Belt)
change agent. Continues
Works closely with other continuous improvement Demonstrated mastery of Black Belt body of
to perform normal duties
leaders to apply formal data analysis approaches to knowledge
while participating on Six
Sigma project teams projects Demonstrated proficiency at achieving results through
the application of the Six Sigma approach
Teaches local teams, shares knowledge of Six Sigma
Internal Process Improvement Consultant for
Six Sigma champion in Successful completion of at least one Six Sigma project
functional areas
local area every 12 months to maintain their Green Belt certification
Coach/Mentor and recommends Green Belts for
Pyzdek (2003) certification Pyzdek (2003)
DMAIC – Define DMAIC – Measure
DMAIC Phase Steps Tools Used DMAIC Phase Steps Tools Used
D - Define Phase: Define the project goals and customer (internal and external) M - Measure Phase: Measure the process to determine current performance
deliverables
• Histogram
• Pareto Chart
•Define Customers and Requirements •Project Charter • Scatter Plot
•Develop Problem Statement, Goals and •Process Flowchart •Define Defect, Opportunity, Unit and • Cause and Effect/Fishbone Diagram
Benefits Metrics • 5 Whys
•Identify Champion, Process Owner and •SIPOC Diagram •Detailed Process Map
• Process Map Review and Analysis
Team •Stakeholder Analysis •Develop Data Collection Plan
•Validate the Measurement System • Data Collection Plan
•Define Resources
•Evaluate Key Organizational Support •DMAIC Work Breakdown Structure •Collect the Data • Benchmarking
•Determine Process Capability and Sigma
•Develop Project Plan and Milestones •CTQ Definitions Baseline
• Measurement System Analysis/Gauge R&R
•Develop High Level Process Map • Process Sigma Calculation
•Voice of the Customer Gathering
www.isixsigma.com www.isixsigma.com
26
1/1/2018
DMAIC – Analyze DMAIC – Improve
DMAIC Phase Steps Tools Used DMAIC Phase Steps Tools Used
A - Analyze Phase: Analyze and determine the root cause(s) of the defects I - Improve Phase: Improve the process by eliminating defects
•Define Performance Objectives •Perform Design of Experiments

•Develop Potential Solutions • Action/communication Plan
•Identify Value/Non-Value Added • Regression Analysis
•Define Operating Tolerances of Potential • Design of Experiments
Process Steps • Statistical Analysis
System • House of Quality
•Identify Sources of Variation • Hypothesis Testing (Continuous
•Assess Failure Modes of Potential Solutions • Failure Modes and Effects Analysis
•Determine Root Cause(s) and Discrete)
•Validate Potential Improvement by Pilot (FMEA)
•Determine Vital Few x's, Y=f(x) • Non-Normal Data Analysis Studies • Simulation Software
Relationship •Correct/Re-Evaluate Potential Solution
www.isixsigma.com www.isixsigma.com
DMAIC – Control Understand Six Sigma!
DMAIC Phase Steps Tools Used
C - Control Phase: Control future process performance
•Define and Validate Monitoring and

Control System • Control Charts (Variable and
•Implement Statistical Process Control Attribute)
•Determine Process Capability • Cost Savings Calculations (ROI)
•Verify Benefits, Cost • Control Plan
Savings/Avoidance, Profit Growth • Process Capability (new sigma level)
•Close Project, Finalize Documentation
www.isixsigma.com
27
1/1/2018
Lean Six Sigma Program Structure
Program Direction,
Deployment Process
Support,
Team Owner
and Marketing Leadership
Black Belt
Change Agents
and
Process Leaders
Green Belt/Lean Expert
Organizational
“Buy-in”
Program is structured to build a self-sustaining critical mass of

process improvement competencies.
Champion / Functional Leader Role Lean Six Sigma Roles

1. Lead the Six Sigma efforts overall in their Division/Dept. • Black Belt (BB)
• Full-time position
2. Provide Strategic Direction for Lean Six Sigma Project teams
• Black Belt is accountable for
3. Track the Project’s Progress • Reporting usually To The Champion, For The Project / Team Results
• Responsible For The Project / Team’s Progress
4. Help the Black Belt / Green Belt overcome roadblocks, including • Provides Leadership In Planning The Project / Team’s Work
seeking collaboration • Applies Six Sigma Tools And Teaches Team Members How To Apply Them
• Often Leads Team Meetings, And Ensures That Decisions Are Made By The
5. Help find resources for the team as Needed, Allocate resources when Team In A Timely Manner To Meet Its’ Goals
authorized • Green Belts (GB)/Lean Experts
6. Keep Black Belt / Green Belt focused on desired results • Part-time position
• Green Belt is accountable
7. When immovable objects block the road, Redirect Project / Team • Has Good Understanding of Lean Six-Sigma Tools
activities • Often Co-Leads Team Meetings
• Aide BB in Reporting Usually To The Champion, For The Project / Team
8. Serve as the Team’s Champion from Top-To-Bottom Results with Black-Belt
• Aide BB who Is Responsible For The Project / Team’s Progress with Black-Belt
9. Ensure that Project Solutions are well implemented, Gains are • Aide BB Provide Leadership In Planning The Project / Team’s Work
sustained and on-going responsibility transfers to Process Owner
28
1/1/2018
Lean Six Sigma Roles (cont’d) System Value Stream Analysis (VSA)
• Yellow Belt (YB)
• Part-time position
• Yellow Belt
• Understand DMAIC process
• Use methodology such as PDCA, Six Sigma philosophy (stakeholder

involvement)
• Smaller process improvement projects
• Using Median return on each trained Yellow Belt is $15,000- $20,000

per project
System Value Stream Analysis (VSA) Constraint Analysis
Data Sources: Data Data Data Data

-ORSAS
Sources: Sources: Sources: Sources:
-Siemens - Not sure -Bed -Not sure ??
-OR/GI/Cath
system management
lab/ECT computer ??
systems -ER system system
29
1/1/2018
Strategic Gap Analysis Improvement Opportunities by Methodology
Data Sources for Strategic Gap Analysis

• Press Ganey Patient Satisfaction • Quality Indicators (including
Survey HEDIS)
• Culture of Safety Survey • State Dept of Health (National
• Treo Resource Management: Database of Nursing Quality
LOS, Cost and Charges Indicators)
• Joint Commission Periodic • Hospital System Performance

Performance Review Dashboard
• Joint Commission /Centers for • Excellus Scorecard

Medicare & Medicaid Services • Clinical Excellence Measures
(CMS) Core Measures
Project Selection Project Selection Method

Project
Suggestion
Comes In
Feasibility/
Patient Financial Strategic Create a Project Select Team
Likelihood Charter Members
Impact Impact Alignment ME/QM Will
of Success Does the
project
No Manage/Track the
Project
cross Schedule Project
department Identify Team
al lines? Champion(s)
1 - Low 1 - High 1 - Low 1 - Low Report To SC

Yes
5- Send Project to
5 - Medium 5 - Medium 5 - Medium SC for
Medium Prioritization
Create a Project
Charter
Select Team
Members
9 - High 9 - Low 9 - High 9 - High

Schedule Project
Prioritize & Select Identify Team
Criteria Weight 0.45 0.3 0.2 0.2 Total Projects Champion(s)
30
1/1/2018
Lean Six Sigma Implementation Strategy Where To Start?
• Long-time source of
Train….
frustration
Apply…..
Review….. • New regulatory
requirements
• Easiest to implement
UHSH Quality & Patient Services
-Program Management
• Stakeholder delight
--100 Lean Experts
-- 70 Champions • Biggest financial
impact
• Burning platform
Integrate training with metrics performance to maximize the bottom
line impact
Decision Matrix Decision Criteria
• Simultaneously Set of criteria that are applied to all projects

assesses multiple
projects on the same • Weight: Impact on the patient
criteria • Ease of Implementation
• Can use consensus
• Stakeholder Buy-in
methodology or
multi-voting process • Financial Impact
• Allows you to • New Technology
prioritize projects in a
logical manner • Market Impact
• Community Need
31
1/1/2018
Decision Matrix Scoring Scales
Project Weight Ease to Stakeholder Financial Total Using scoring scales helps to simplify the
Implement Buy in Impact
process
• 1,2,3,4…10
• Results in “too little” variation in scoring
1= minimal impact or high difficulty
5= moderate impact or somewhat difficult
9= substantial impact or low difficulty
Decision Matrix Decision Matrix
Weight Ease to Stakeholder Financial Total Weight Ease to Stakeholder Financial Total
Project Implement Buy in Impact
ED TAT ED TAT 9
Pharmacy Pharmacy 1
Ordering Ordering
Billing Accuracy Billing Accuracy 5
Meal Delivery Meal Delivery 5
Outpatient Outpatient 9
Wait Times Wait Times
32
1/1/2018
Decision Matrix Decision Matrix
ED TAT 9 1 ED TAT 9 1 5
Pharmacy 1 5 Pharmacy 1 5 9
Ordering Ordering
Billing Accuracy 5 5 Billing Accuracy 5 5 5
Meal Delivery 5 5 Meal Delivery 5 5 5
Outpatient 9 1 Outpatient 9 1 9
Wait Times Wait Times
Decision Matrix Totaling Scores
ED TAT ED TAT 9 1 5 9 135

9 1 5 9
Pharmacy Ord 1 5 9 5
Pharmacy 1 5 9 5
Ordering Billing Accuracy 5 5 5 5
Billing Accuracy 5 5 5 5 Meal Delivery 5 5 5 1
Outpatient 9 1 9 1
Meal Delivery 5 5 5 1 Wait Times
Outpatient 9 1 9 1 Weight x Sum of Criteria= Total

Wait Times ED TAT: 9 x (1+5+9)= Total
9 x 15= 135
33
1/1/2018
Decision Matrix Multi-Voting
Same process except individual votes are used

Weight Ease to Stakeholder Financial Total
Project Implement Buy in Impact instead of consensus votes
ED TAT 9 1 5 9 135
Weights should be the same and criteria scales
Pharmacy 1 5 9 5 19
Ordering remain the same (1, 5, 9)
Billing Accuracy 5 5 5 5 75
Averages are used in the criteria scoring
Meal Delivery 5 5 5 1 55
Have a process and an expected timeframe
Outpatient 9 1 9 1 99
Wait Times
Project Scoping
How far down should I scope my project?

Why?
High level Why?
problem Initial Why?
Contributor Secondary
Level I Why?
Contributor Project
Level
Level
II ???
Level III
Level IV
Project Level
When you can no longer answer the “Why?” with

confidence, you have arrived at the project level.
34
1/1/2018
Project Scoping Action Type Selection Flow
Why? Problem
CHF Core Identified
Measure Why?
Scores are The process is
poor
too complex Why?
We are not
Level I performing Why? YES Root N
The Complex Cause O
well on ¾ of Known?
Process,
Level II elements Documentation,
follow-up, Med Streamline process & Solution Problem
Level lII Rec documentation
Known? Type?
Level IV YES N
O
VELOCITY VARIATION
Project Level
Our project then becomes in measurable terms: Just Do It!
Rapid Improvement
Workshop
Process VSA
Six Sigma/DMAIC
Project
Improving CHF Core Measure Performance by 25% of base-
line within 3 months post implementation
138
ASQ Project Authorization/Charter

A Basic Project Charter-ASQ
Template
Once you have defined the problem, use this simple charter Key components of a project charter include:
document to communicate and confirm agreement between the (1) Business case - financial impact, why important,
problem-solving team and management.
why now
(2) Problem statement
(3) Project scope – what are the boundaries, or what is
in scope and what is out of scope
(4) Goal statement or success measures
(5) Roles of team members
(6) Milestones and deliverables - what needs to be
accomplished and delivered by when
35
1/1/2018
A Basic Project Charter-ASQ A Basic Project Charter Example - ASQ

How to Use
Step 1. Define the business case. Why is the project important, and
why now?
Step 2. Define the problem statement. Quantify the problem, its
impact on the process, and how long it has existed without
causation.
Step 3. Define the project scope. What are the boundaries of the
project team, what is in scope, and (sometimes even more
important) what is out of scope?
Step 4. Define the goal(s) of the project team. What are the
tangible performance improvements that the team has as a target?
Step 5. Define the resources and the roles that each will play.
Step 6. Define the milestones and deliverables. What needs to be
accomplished and delivered by when?
Step 7. Review the project charter with the entire team to solicit
input. If potential changes are identified, review with the champion
to gain approval before publication.
Six Sigma Project Charter Template Six Sigma Project Charter Template (cont’d)
36
1/1/2018
Another Project Charter Example Another Project Charter Example (cont’d)
Another Project Charter Example (cont’d) Project Management

• How many have been involved in a project?
37
1/1/2018
Project Definition 9 Project Management Areas
What is a Project? Project

Communication Integration
Management Management
• Definite Beginning
Scope Time
• Definite End Management Management
• Create a Unique Product, Service, or Result

Human Resource Cost
Management
Project Management
Procurement
Quality
Management
Management
(Vendors
)
Risk
Management
Important Issues Project Life Cycle

• The Most Important Issue!
• PROJECT PLANNING
• Based on Project Size and Scope
• Why so Important?
38
1/1/2018
Project Life Cycle Project Life Cycle

154
Project Life Cycle Aide to Project Planning

• Work Breakdown Structure (WBS)
• Identify tasks needing completion and perform flow for each
task
39
1/1/2018
Project Plan Communicate

• Creation of Project Plan • Project Communication
• Stakeholders
• Identify WBS • Project Team
• Executive Team
• Identify “Experts” in WBS tasks • Steering Committee
• Have “Experts” identify task definition and length

In A Single Word
• Input tasks and times in Project Plan
CUSTOMERS!!!!!
• Calculate Project Timeline
Project Communication Plan Project Management
• Scheduling techniques for creating a timetable:

Communication
Objective of Communication Medium Frequency Audience Owner Deliverable
Type
Introduce the project team and • Project

• Agenda
Kickoff Meeting
the project. Review project
objectives and management
• Face to Face Once
•
Sponsor
Project Team
Project Manager • Meeting • Gantt Chart
Minutes
approach. • Stakeholders
• Face to Face • Agenda

• PERT Chart (Program Evaluation and Review Technique)
Project Team Review status of the project
• Conference Weekly • Project Team Project Manager • Meeting
Meetings with the team.
Call Minutes
• Project • Agenda
• Shows time and probabilities
Technical Design Discuss and develop technical
• Face to Face As Needed Technical Technical Lead • Meeting
Meetings design solutions for the project.
Staff Minutes
• Face to Face • CPM (Critical Path Method)

Monthly Project Report on the status of the
• Conference Monthly • PMO Project Manager
Status Meetings project to management.
Call
• Shows costs, time, probabilities

• Project
Report the status of the project Sponsor
Project Status • Project Status
including activities, progress, • Email Monthly • Project Team Project Manager
Reports Report
costs and issues. • Stakeholders
• PMO
40
1/1/2018
Project Management Project Management

• Scheduling
• Scheduling
• Gantt Charts continued.
• Gantt Chart
• Must know:
• A record of the events and the progress that has taken
• When
place. What work has been done?
• Rate of accomplishment
• Gantt chart forces a person to have a plan and to keep
• Checks on progress
track of happenings
• A division of space represents both an amount of time
• Promotes the identification and assignment of clear-cut and an amount of work to be done in that time
tasks
• Lines drawn horizontally through that space show the
• Gantt chart enables users to visualize the passing of time, relation of the amount of work actually done in that time to
easy to read the amount of work scheduled to be done.
Project Management Project Management
• Scheduling • Scheduling:
• PERT
• CPM
• Improves on the Gantt chart by showing the relationships
• Builds on PERT
between tasks
• Adds the concept of cost per unit time that a project runs
• project is viewed as an integrated whole
• allows for the determination of the longest series of inter-
• time value for each activity is known
related events that must be completed in the project: the
• perfect for complex projects critical path
• PERT coordinates and synchronizes various parts of the • The times associated with the activities must be well
overall job estimated and costs calculated
41
1/1/2018
Project Management Sample DMAIC Gantt Chart
• Scheduling
• To create a PERT or CPM network:
• Compile a list of activities
• Determine the relationships between the activities
(predecessors, successors)
• Begin at the beginning
Project Management - Summary
• Understand work necessary for project completion
• Identify details and length of project
• Communicate, Communicate, Communicate
• Have integral project components dictate project length
• NEVER arbitrarily select completion date and attempt to
make project fit YOUR timeline
42
1/1/2018
DMAIC Advertisement DMAIC – Define
D - Define Phase:
• Define Customers and Requirements
(CTQs)
• Develop Problem Statement, Goals
and Benefits
• Identify Champion, Process Owner
and Team
• Define Resources
• Evaluate Key Organizational Support
• Develop Project Plan and Milestones
• Develop High Level Process Map
Define Tools Relation Between Basic/New Tools
Define Tools: FACTS

• Project Charter
Data
• Project Planning
• Basic 7 Tools Numerical Data Verbal Data
• New 7 Tools Define problem after Define problem before
collecting numerical data collecting numerical data
• CTQ Trees
• SIPOC The Basic 7 Tools The New 7 Tools
• Analytical approach • Generate Ideas

Organize • Formulate plans
Information
43
1/1/2018
Seven Basic Quality Tools Check Sheet/Checklist

Kaoru Ishikawa developed these seven basic tools Simple form used to collect data
Hatch marks are used to record frequency of occurrence
Frequently used to produce histograms and Pareto chart
Check Sheet Procedure:
Check Sheet Run Chart
Fishbone Diagram • Decide what event or problem will be observed
• Decide when data will be collected and for how long
• Design the form. Set it up so that data can be recorded simply by
making check marks or Xs or similar symbols and so that data do not
have to be recopied for analysis
• Label all spaces on the form
Histogram
• Test the check sheet for a short trial period to be sure it collects the
Pareto Chart Scatter Diagram appropriate data and is easy to use
• Each time the targeted event or problem occurs, record data on the
Flow Chart check sheet
McLaughlin & Hays (2008) The Quality Tools (2005)
Example 1 Example 2
Check sheet depicting telephone interruptions Checklist of a device assembly startup run
Luke (2011)
44
1/1/2018
Histogram Examples
Histograms
Bar graph that shows frequency data which provide the easiest way to Composition in n-3 and n-6 fatty acids of farmed and
evaluate the distribution of data wild rainbow trout and salmon (% of total fatty acids)
Developing a histogram:
• Step 1: Collect data and sort it into categories
• Step 2: Label the data set
Independent variable: The characteristic of data set in each category
Dependent variable: The frequency of data set in that category
• Step 3: Each mark on both axes should be in equal increments
• Step 4: For each category, find the related frequency and make the
horizontal marks to show that frequency
Some applications:
• To determine and compare the distribution of sales, services, quantities
• To understand how the output of a process relates to customer expectations
http://www.dfo-mpo.gc.ca/aquaculture/ref/morin_aaq-eng.htm
Histogram Examples (cont’d)

Scatter Diagrams
Percentages of the toxicological reference value attained for a Plot of 2 variables to identify any potential
contaminant based on the frequency of consumption of meals of a relationship between them or
type of fish
Construction of scatter diagram:
• Step 1: Collect two pieces of data and create a
summary table
• Step 2: Draw a diagram labeling the horizontal
and vertical axes
• Normally “cause” variable be labeled on the X axis
and the “effect” variable be labeled on the Y axis
• Step 3: Plot the data pairs on the diagram
• Step 4: Interpret the scatter diagram for
direction and strength
Applications:
• To study and identify direct relationship between 2 variables (e.g., number
of patient visits to a hospital each hour of the day)
http://www.dfo-mpo.gc.ca/aquaculture/ref/morin_aaq-eng.htm
45
1/1/2018
Scatter Diagram Examples Scatter Diagram Example
Problem:
"As cheese ages, various chemical processes take place that determine the taste
of the final product. In a study of cheddar cheese from the LaTrobe Valley of
Victoria, Australia, samples of cheese were analyzed for their chemical
composition and were subjected to taste tests. Overall taste scores were
obtained by combining the scores from several tasters. This dataset contains
concentrations of various chemicals in 30 samples of mature cheddar cheese,
and a subjective measure of taste for each sample. The variables "Acetic Acid"
and "H2S" are the natural logarithm of the concentration of acetic acid and
hydrogen sulfide respectively. The variable "Lactic Acid" has not been
transformed."
statsci.org
Scatter Diagram Example – Data (cont’d) Scatter Diagram Example (cont’d)

Case
1
Taste
12.3
Acetic Acid
4.543
H2S
3.135
Lactic Acid
0.86
From the data
2 20.9 5.159 5.043 1.53
3 39 5.366 5.438 1.57 • Y variable: test
4 47.9 5.759 7.496 1.81
5
6
5.6
25.9
4.663
5.697
3.807
7.601
0.99
1.09
• X variable: 3 variables
7 37.3 5.892 8.726 1.29
8
9
21.9
18.1
6.078
4.898
7.966
3.85
1.78
1.29
Because we can only plot two variables in a scatter
10
11
21
34.9
5.242
5.74
4.174
6.142
1.58
1.68
diagram, we use a scatter plot matrix to show all options
12 57.2 6.446 7.908 1.9
13 0.7 4.477 2.996 1.06 • e.g.
14 25.9 5.236 4.942 1.3
15 54.9 6.151 6.752 1.52
16 40.9 6.365 9.588 1.74
17 15.9 4.787 3.912 1.16
18 6.4 5.412 4.7 1.49
19 18 5.247 6.174 1.63
20 38.9 5.438 9.064 1.99
21 14 4.564 4.949 1.15
22 15.2 5.298 5.22 1.33
23 32 5.455 9.242 1.44
24 56.7 5.855 10.199 2.01
25 16.8 5.366 3.664 1.31
26 11.6 6.043 3.219 1.46
27 26.5 6.458 6.962 1.72
28 0.7 5.328 3.912 1.25
29 13.4 5.802 6.685 1.08
30 5.5 6.176 4.787 1.25
statsci.org statsci.org
46
1/1/2018
Scatter Diagram Example (cont’d) Scatter Diagram Example (cont’d)

Results: Correlation coefficient table for multiple variables
R/p value Taste Acetic Acid H2S Lactic Acid
Taste - 0.55 0.756 0.704

0.002 0 0
Acetic Acid 0.55 - 0.618 0.604
0.002 0 0
H2S 0.756 0.618 - 0.645

0 0 0
Lactic Acid 0.704 0.604 0.645 -

0 0 0
statsci.org
Run Chart Example

Plot of a process characteristic, in a chronologic sequence Case:
Illustrates patterns/trends of behavior • A company is producing raw material to be used in
Will be discussed in details when we discuss “Control copying machine toner
Charts” • A team is continually monitoring the process to achieve
continuous improvement
Features: Problem:
• Time based • Team suspect that theoretical output quantities does NOT
• Cyclical equal actual results (actual output yield is fluctuating)
• Shows pattern
Action
• Team developed a run chart
http://community.asq.org/statistics/category/Run-Chart
47
1/1/2018
Example (cont’d) New 7 Tools

Results:
Affinity Diagrams
• Run chart shows a real fluctuation in the output
• Further analysis showed that root cause of problem is drop Relations Diagrams
in air pressure at a certain time of process
Run chart Tree Diagrams
Matrix Diagrams
Arrow Diagrams
Process Decision Program Charts
Matrix Data Analysis
Work in conjunction with Basic 7 Tools

http://community.asq.org/statistics/category/Run-Chart
Affinity Diagrams Affinity Diagrams (cont’d)
Definition: •Gathers large amounts of

Advantages:
intertwined verbal data
Pinpointing problem in a chaotic Facilitates breakthrough thinking and stimulate fresh ideas
•Organizes the data into groups •
situation and generating solution
based on natural relationship • Permits the problem to be pinned down accurately
strategies
•Makes it feasible for further • Ensures everyone clearly recognizes the problem
Construction: analysis • Incorporates opinions of entire group
• Step 1: Select a topic • Fosters team spirit

• Step 2: Collect verbal data by brainstorming • Raises level of awareness
• Step 3: Discuss collected data until everyone understands it thoroughly Topic
• Step 4: Write each item on separate data card
• Step 5: Spread out all cards on the table Affinity Statement Affinity Statement Affinity Statement
• Step 6: Move data cards into groups of similar themes Data Card Data Card Data Card Data Card Data Card Data Card
• Step 7: Combine statements on data cards to new Affinity statement step
Data Card Data Card Data Card Data Card Data Card Data Card
• Step 8: Make new card with Affinity statement
• Step 9: Continue to combine until you have less than 5 groups/themes Data Card Data Card
• Step 10: Lay the groups outs, keeping the affinity clusters together Affinity Statement
Data Card Data Card
Data Card
48
1/1/2018
Affinity Diagram Example Relations Diagrams

Development of affinity diagram to study business Definition:
• Finding solution strategies by clarifying relationships with
model components complex interrelated causes
Construction:
• Step 1: Express the problem in the form of “Why isn’t something happening?”
• Step 2: Each member lists 5 causes affecting problem and write each item on a
card
• Step 3: Discuss info collected until everyone understands it thoroughly
• Step 4: Move cards into similar groups
• Step 5: Asking why, explore the cause-effect relationships, and divide the
cards into primary, secondary and tertiary causes
• Step 6: Connect all cards by these relationships
• Step 7: Further discuss until all possible causes have been identified
• Step 8: Review whole diagram looking for relationships among causes and
connect the related groups
It is also known as “Interrelationship” diagrams
Shafer et al., (2004)
Relations Diagrams (cont’d) Example: Why Do Projects Fail?

Advantages:
• Useful at planning stage for obtaining perspective on overall situation
• Facilitatesconsensus among team
• Assists to develop and change people’s thinking
• Enables priorities to be identified accurately
• Makes the problem recognizable by clarifying the relationships
among causes
Tertiary Tertiary
Cause Secondary Cause
Primary Cause Cause
Primary Cause Tertiary

Secondary Cause
Why doesn’t
Cause Secondary
X happen?
Cause
Primary Cause Primary Cause

6th level
Secondary
Tertiary Cause
Cause
Cause
Tertiary
Secondary
Cause
Cause 4th level 5th level
4th level Cause Cause
Cause
49
1/1/2018
Example: Urban Poverty Reasons Relations Tree Diagrams
A tree diagram shows all the possible outcomes of an event
Construction:
• Step 1: Write Tree Diagram topic (Objective card)
• Step 2: Identify constraints on how objective can be achieved
• Step 3: Discuss means of achieving objective (primary means, first level
strategy)
• Step 4: Take each primary mean, write objective for achieving it (secondary
means)
• Step 5: Continue to expand to the fourth level
• Step 6: Review each system of means in both directions (from objective to
means and means to objective)
• Step 7: Add more cards (if needed) and connect all of them
It is also known as systematic diagrams or dendrograms

http://www.skymark.com/resources/tools/relations_diagram.asp
Tree Diagrams (cont’d) Tree Diagram Example

A restaurant manager wants to know more about her customers’ satisfaction
Advantages: factors
• Systematic and logical approach is less likely that items are omitted
• Facilitates agreement among team
• Are extremely convincing with strategies 4th means
3rd means 4th means
Secondary means
3rd means 4th means
Primary means 4th means
3rd means
4th means
3rd means
To Secondary means
4th means
Accomplish 4th means
3rd means 4th means
Secondary means 3rd means 4th means

4th means
Constraints Primary means
3rd means 4th means
Secondary means 3rd means 4th means http://www.syque.com/quality_tools/toolbook/Tree/example.htm
50
1/1/2018
Matrix Diagrams Frequently Used Matrix Diagram Symbols
For Clarifying Problems by Thinking Multi-dimensionally
Construction:
Step 1: Write final-level means from Tree Diagram forming y-axis
Step 2: Write evaluation categories (efficacy, practicability, and rank) on x-axis
Step 3: Examine final-level means to identify whom will implement them
Step 4: Write names along horizontal axis
Step 5: Label group of columns as “Responsibilities” and right-hand end of horizontal axis
as “Remarks”
Step 6: Examine each cell and insert the appropriate symbol
Step 7: Determine score for each combination of symbols, record in rank column
Step 8: Examine cells under Responsibility Columns, insert double-circle for Principal
and single-circle for Subsidiary
Step 9: Fill out remarks column and record meanings of symbol
Matrix Diagrams (cont’d) Matrix Diagram Example

Advantages:
This diagram shows the relations between process
• Enables data on ideas based on extensive experience
• Clarifies relationships among different elements
improvement tools and field function
• Makes overall structure of problem immediately obvious
• Combined from two to four types of diagrams, location of problem is clearer
Michalski (1997)
51
1/1/2018
Arrow Diagrams Arrow Diagrams (cont’d)

Advantages:
• Allowsoverall task to be viewed and potential snags to be identified before
For Working Out Optimal Schedules and Controlling Them Effectively
work starts
Construction: • Leads to discover possible improvements & enables easy monitoring
• Step 1: From strategies on Tree Diagram, select one (Objective of Arrow Diagram) • Deals promptly with changes to plan
• Step 2: Identify constraints to objective and list activities to achieve them • Improves communication among team
• Step 3: Write all essential activities on separate cards • Promotes understanding and agreement among group
• Step 4: Organize cards in sequential order of activities Strategy Constraints
• Step 5: Remove any duplicate activities
• Step 6: Review order of activities, find sequence with greatest amount of activities
4 12
and Arrange parallel activities
• Step 7: Examine path, number nodes in sequence from left to right
1 2 3 5 9 10 13
• Step 8: Record names and other necessary information
• Next, complete the diagram Activity
6 8 11
Applied in PERT (Program Evaluation and Review Technique) and CPM
(Critical Path Method)
7
Arrow Diagram Example Process Decision Program Charts (PDPCs)

Industry production factors For Producing the Desired Result from Many
Possible Outcomes
Construction:
Step 1: Select a highly effective, but difficult strategy from the Tree Diagram
Step 2: Decide on a goal (most desirable outcome)
Step 3: Identify existing situation (starting point) and constraints of objective
Step 4: List activities to reach goal and potential problems with each activity
Step 5: Review list to add extra activities or problems, previously not thought
Step 6: Prepare contingency plan for each step and review what action is needed if step is not
achieved
Step 7: Examine carefully to check for inconsistencies and all important factors are included
Step 8: Examine to make sure all contingency plans are adequate
Next, complete the diagram
Bassi et al. (2012) Finds feasible counter measures to overcome problems

52
1/1/2018
PDPCs (cont’d) PDPCs Example

Advantages:
Incremental sampling method (ISM) used to analyze
• Facilitates forecasting
• Uses past to anticipate contingencies field samples in Laboratories
• Enables problems to be pinpointed
• Illustrates how events will be directed to successful conclusion Multiple criteria available via ISM depending on
• Enables those involved to understand decision-makers intentions
• Fosters cooperation and communication in group contamination of samples
• Easily modified and easily understood
Start PDPCs diagram for ISM processing and analysis is
NO
developed
NO NO
YES YES NO
NO NO
YES
NO
Goal http://www.itrcweb.org/ism-1/6_LABORATORY_SAMPLE_PROCESSING_AND_ANALYSIS.html
PDPCs Example (cont’d) Matrix Data Analysis

To find indicators that quantifies large amount of information
Construction:
Step 1: Set-up goal, its alternatives, and criteria for decision
Step 2: Place selection in order of importance
Step 3: Apply percentage weight to each option (all weights should add up to 1)
Step 4: Sum individual ratings to establish overall ranking (Divide by number of options for
average ranking)
Step 5: Rank order each option with respect to criterion (Average the rankings and apply a
completed ranking)
Step 6: Multiply weight by associated rank in Matrix
Step 7: Result is Importance Score and add up them in each case
Step 8: Rank order the alternatives according to importance
Next, complete the diagram
http://www.itrcweb.org/ism-1/6_LABORATORY_SAMPLE_PROCESSING_AND_ANALYSIS.html
Based solely on numerical data
53
1/1/2018
Matrix Data Analysis Example Summary: New 7 Tools

Four different designs in a manufacturing process are to
Benefits of New Seven Tools
be evaluated using survey results (rank)
Analysis performed using matrix data analysis 1. Provide Training in Thinking
2. Raise People’s Problem Solving Confidence
3. Increase People’s Ability to Predict Future Events
Roles of New Seven Tools
1. Express verbal data diagrammatically

2. Make information visible
3. Organize information intelligibly
4. Clarify overall picture and fine details
5. Get more people involved
Critical to Quality (CTQ) CTQ (cont’d)

To convert customer needs or wants to measurable requirements for CTQ may vary from customer to customer: it may be the timely
the business to implement delivery that someone may consider critical while for others, cost
may be a critical factor
Assumes that the customer can articulate what they want
CTQs in Six Sigma represents the product or service characteristics
that are defined by the internal or external customers
CTQ is a measure of what is important to the customer and one of the
ways to identify it is by knowing the customer needs and
expectations May include the upper and lower specification limits or any other
factors related to the product or service
There are various parameters in a process but all of them may not be
as significant or critical as the ones quoted / expected by the CTQs usually must be interpreted from a qualitative customer
customer statement to an actionable, quantitative business specification
CTQ is a product or service characteristic that must be met to satisfy a

customer specification or requirement
54
1/1/2018
CTQ Example
http://www.whatissixsigma.net/six-sigma-dmaic-define-phase/
Contents
1. Data Collection Methods
2. Basic Statistics
3. Measurement System Analysis Data Collection

4. Basic Probability
5. Probability Distributions
6. Process Capability
55
1/1/2018
Process Data
Attribute data:
• Categorical data that can be counted
• Pass/fail, good/bad, etc.
INPUT PROCESS OUTPUT CUSTOMERS

Key Key Critical Variable data:
Process Process to
Input Output Quality
• Continuous data that is measured
Variables Variables (CTQ)
(KPIV) (KPOV) • Weight, pressure, temperature, etc.
McLaughlin & Hays (2008) McLaughlin & Hays (2008)
Data Classification Data Collection – Key Questions
Type of data What questions are we trying to answer?
• Cross-sectional —data that are collected over a single What type of data will we need to answer the question?
period of time Where can we find the data?
• Time series —data collected over time Who can provide the data?
Number of variables How can we collect the data with minimum effort and with
• Uni-variate —data consisting of a single variable minimum chance of error?
• Multivariate —data consisting of two or more (often

related) variables
56
1/1/2018
Data Collection – Logic Data Collection Map
Why are the data needed?
What will the data be used for?
What questions are going to be asked of the data?
Are the patterns of the past going to be repeated in the
future?
Data Collection Methods Sampling

Definition:
Define and apply methods for collecting different types of data What is the objective of the study?
Reason for data collection:
• Legal, regulatory, or statutory requirements What type of sample should be used?
• Analysis, improvement, and knowledge management Random
• Contractual requirements of customers

What possible error might result from sampling?
Data Collection Methods:
Sampling What will the study cost?
Surveys
Sequential
Face-to-face interviews
Focus groups
Mystery shopping
Customer feedback
Automatic data capture Data Sampling Subgroup
Manual data capture
57
1/1/2018
Selecting A Sampling Plan Sampling Error
“A good sampling plan should select a sample at the lowest cost Sampling error (statistical error)
that will provide the best possible representation of the population, Non-sampling error (systematic error)
consistent with the objectives of precision and reliability that have Factors to consider:
been determined for the study” • Sample size
• Appropriate sample design
Sampling Methods Data Sampling – Example
Simple random sampling
Stratified sampling
Systematic sampling
Cluster sampling
Judgment sampling
McLaughlin & Hays (2008) Munro et al. (2007)
58
1/1/2018
Learning Math!
Basic Statistics
Basic Statistics Basic Statistics (cont’d)

Descriptive Statistics Analytical Statistics Population:
Collect, organize, summarize and Makes inferences, hypothesis tests, Collection of all items and data under consideration
present data and predictions • Traditionally, it is denoted by Greek letters
Use samples to estimates or Sample:

Show mean, median, mode,
inferences about the population It’s a subset of the population which are randomly selected
dispersion, shape, and so on
from which he sample was drawn • Traditionally, it is denoted by Latin letters
Tools such as hypothesis testing, Sample Versus population notations

Graphical tools such as histograms, scatter diagram to determine
pie charts, box plots and so on relationships among variables and Sample Population
predicts using regression equations Size n N
Mean
Standard
Descriptive statistics also called enumerative or summary statistics s
deviation
Analytical statistics also called inferential statistics
Summation ∑
59
1/1/2018
Mathematical Descriptions Mean Mathematical Descriptions - Median and Mode
The mean is the arithmetic average of the population: The median is the middle value of the sample or
population. If the data are arranged into an
array (an ordered data set):
3, 3, 5, 6, 8
The population mean can be estimated from a sample:

5 would be the middle value or median
The mode is the most frequently occurring

value. In the above example, the value 3 occurs
more often (two times) than any other value, so
3 would be the mode. McLaughlin & Hays (2008)
Mathematical Descriptions Mathematical Descriptions

Range and Mean Absolute Deviation Variance, Standard Deviation
The range is the difference between the high and low The variance is the average square difference
values in a data set from the mean
The mean absolute deviation (MAD) is the average of

the absolute value of the differences from the mean. This standard deviation is the square root of the
variance
60
1/1/2018
Mathematical Descriptions
Coefficient of Variation Basic Statistics – Example
Dataset:
The coefficient of variation (CV) is a measure of the 27 17 12 7 21 44 23 3 36 32 21
relative variation in the data. It is the standard deviation
divided by the mean Mean:
Rearrange the data in upper chronological order:

3 7 12 17 21 21 23 27 32 36 44
Mode, M =21
Median, (n+1)/2 th data= (11+1)/2 th data = 21
Measure
Understand causal relationships between process
performance and customer value

Measurement System
Y = f(X)
Analysis (MSA)
where;
Y = customer CTQs
X = critical input variables that influence Y
An Introduction to Six Sigma and Process Improvement (2009)
61
1/1/2018
Sources of Variation in Production Processes Common vs. Special Causes

Measurements
Common Common Special

Measurement Causes or Causes
People Methods
Materials Instruments MEASURE Special MEASURE
?
Investigate all of the variation

INPUTS PROCESS OUTPUTS Measurements Investigate specific data points
ANALYZE
ANALYZE
Tools Human Develop solutions for

Develop solutions for Sustain The
Technology Environment Inspection Improvements
special causes and
the “vital few” process
implement as
Performance and input Xs
appropriate
IMPROVE CONTROL
IMPROVE
An Introduction to Six Sigma and Process Improvement (2009) www.vahimss.org
Process Metrics Measurement System Analysis
Measurement – the act of quantifying the performance A statistical tool to explore the variation in a measurement
system
dimensions of products, services, processes, and other
Due to:
business activities. Calibration: Drift in average measurements of an absolute value
Stability: Drift the absolute value over time
Measures and indicators - numerical information that Repeatability: Variation in measuring equipment when measured by one appraiser in
the same setting at the same time
results from measurement Reproducibility: Variation in measurement when measured by two or more
• Defects/unit appraisers multiples times
Linearity: Accuracy of measurement at various measurement points of measuring
• Errors/opportunity range in the equipment
Bias: It’s the difference between the actual and observed measurement at various
• DPMO measurement points of the measuring
range
All data from the process should be filtered through this system
62
1/1/2018
Measurement System Analysis (cont’d) Components for Measurement Error

Measures quantify: 1. Resolution/Discrimination
2. Accuracy (bias)
– Equipment variation
3. Linearity
– Appraiser (operator) variation 4. Stability (consistency)
– Total variation of a measurement system 5. Precision: Repeatability & Reproducibility (R&R)
Variation:
Precise but not Accurate but not Not accurate or Accurate and
accurate precise precise precise
Precision – Repeatability & Reproducibility (R&R) R&R Studies
Repeatability: Quantify and evaluate the capability of a measurement

variation that occurs when repeated measurements
are made of the same item under identical system
conditions (equipment variation)
Actions: • Select r operators and n parts
• Repair, replace, adjust
• Use SOP • Calibrate the measuring instrument
• Randomly measure each part by each operator for m trials

Reproducibility:
Variation that results when different
conditions are used to make the same
• Compute key statistics to quantify repeatability and
measurements (operator variation) reproducibility
Actions:
• Training
• Use SOP
63
1/1/2018
Gauge R&R Studies R&R Calculations
• A number of appraisers (usually two or three) measure a number of parts (or process r: no. of appraisers (operators)
output) (usually 5 to 20) a number of times (usually two or three)
m: no. of trials
• Results are compared within each appraiser (Repeatability) and between appraisers
(Reproducibility) n: no. of units, parts, samples, etc.
• Randomization is critical for repeat measurements to avoid learning or copying

g = n*r
Method of assessing Repeatability & Reproducibility of a measurement system
R&R Calculations (cont’d) R&R Calculations (cont’d)
r #of appraisers (operators) (Range)

m # of trials
n # of parts/patients/instruments/samples/etc.
R Range of each patient’s trial for individual operator
ഥ‫ܒ‬
‫܀‬ Average range (Repeatability)
ന
‫܀‬ Average range (Reproducibility)
‫ܘ܀‬ Range in Average of observation

‫܆‬ഥܑ Average of observations (Repeatability)
XDiff Difference in average for repeatability
ന
‫܆‬ Average of observation (Reproducibility)
64
1/1/2018
R&R Calculations R&R Calculations (cont’d)

Repeatability
(Sum of the averages)
• EV = Equipment/patient/sample Variation
• K1 = 5.15/d2
• d2 depends on the no. of trials (m) and (g)
• (g)= number of parts * the no. of appraisers
• The value of d2 is obtained from Table 1 in Appendix
Reproducibility
• AV is Appraiser variation
• If a negative value is calculated under the square root sign, the value AV defaults to
zero
• n = No. of parts
• m = No. of trials
• K2 = 5.15/d2 where d2 depends the no. of trials (m) and (g)
R&R Calculations (cont’d) R&R Calculations (cont’d)

R&R Total variation
Part variation
Control limits
Where, K3 = 5.15/d2 is dependent on the no. of parts (m) and g

• A2: Table 2 (Appendix)
65
1/1/2018
R&R Evaluation R&R Example 1
%EV = 100(EV/TV)
%AV = 100(AV/TV) “A Gage Repeatability and Reproducibility (R&R) study
%R&R = 100(R&R/TV) was performed to measure parts specific dimension. The
%PV = 100(PV/TV) analysis was done by different appraisers A, B, C
(reproducibility) and each appraiser repeated the

Under 10% error - OK
measurement 3 times for each part (repeatability)”
10-30% error - may be OK
over 30% error - unacceptable

An Introduction to Six Sigma and Process Improvement (2009) http://www.dmaictools.com/dmaic-measure/grr
R&R Example 1 (cont’d) R&R Example 1 (cont’d)

Results:
http://www.dmaictools.com/dmaic-measure/grr
66
1/1/2018
Example 2 Example 2 (cont’d)

Gauge R&R example The nurses handling the ear thermometer may cause extra
“In a quality improvement project in a hospital, it was variability in the measurements. They were taken along (as
necessary to measure the body temperature of patients. The a factor) in the experiment. Different healthy persons were
temperature measurements were taken with an ear involved in the experiment. They contribute to the
thermometer. A man’s body temperature is under normal observed variation, which is object (read: person) variation,
circumstances in the range of a lower specification limit not part of the measurement variation. The experiment was
(LSL) of 35C and an upper specification limit (USL) of 40C. therefore designed such that it allowed for separation of
To make sure no false conclusions would be drawn during object variability from measurement variability. Thus,
the investigation of the body temperature data, the quality object was taken as a factor. A single ear thermometer was
of the temperature measurement was assessed first by used by all nurses, which was also the case during the
means of a gage R&R experiment. experiment. Hence, it was not a factor during the
experiment. Each person was measured in the right and left
ears. This procedure was done twice.”
[Erdmann et al., 2011] [Erdmann et al., 2011]
Example 2 (cont’d) Results
Repeatability:
Equipment Variation (EV): 0.048916505
Reproducibility:
Appraiser Variation (AV): 1.278321999
R & R:
R&R: 1.279257581
Part variation:
PV: 4.565602837
Total variation: TV: 4.741437464
%EV 1.03 OK
%AV 26.96 May Be Ok
%R&R 26.98 May Be Ok
%PV 96.29 Unacceptable
67
1/1/2018
Process Capability
The range over which the natural variation of a process
occurs as determined by the system of common causes
Measured by the proportion of output that can be
produced within design specifications
A measure of how well the process can produce output
Process Capability that meets desired standards or specifications
Compares process specifications (set by the customer or
management) to control limits (the natural or common
variability in the process)
[An Introduction to Six Sigma and Process Improvement, 2009]
Process Capability Terms Process Capability (cont’d)

Process Capability is a measurement of how the process is performing
Cp = Process Capability with respect to a desired outcome
• A simple and straightforward indicator of process
capability
Cpk = Process Capability Index
• Adjustment of Cp for the effect of non-centered
distribution
Pp = Process Performance
• A simple and straightforward indicator of process
performance
Ppk = Process Performance Index
• Adjustment of Pp for the effect of non-centered
distribution
www.isixsigma.com
68
1/1/2018
Process Capability Steps

Types of Capability Studies
Peak performance study - how a process performs • Measurement System Analysis (MSA): Perform MSA to identify and
remove the source of variation
under ideal conditions
• Control: Identify appropriate rational subgrouping of samples for
Process characterization study - how a process control chart
performs under actual operating conditions • Stability: Plot at least 20 to 30 subgroups of points (2 to 10 samples in
Component variability study - relative each subgroup) on the control chart to check the system stability
contribution of different sources of variation (e.g., • Normality: Distribution of data should be normal. If not, transform the
non-normal data into normal by applying Box-Cox, Johnson or any other
process factors, measurement system)
transformation technique.
• Analysis: Center the process and review the percentage of
nonconformance outside of specification limit (reduce the process
variation smaller than the specifications limit)
Munro et al. (2007)
Test for Stability and Control Short Term (Cp) vs. Long Term (Pp) Capability
69
1/1/2018
Process Performance Vs. Specification Process Capability Illustration
(a) (b)
Meets Specification
specification specification
YES NO
natural variation natural variation

Change process
YES Good
and/or specification
(c) (d)
In statistical
control specification specification
Investigate Out-of-
NO Stop!!
control condition
natural variation natural variation
Process Capability Index Process Capability: Cp and Cpk
The process capability index, Cp (sometimes called the process

potential index), is defined as the ratio of the specification width
to the natural tolerance of the process
Cp relates the natural variation of the process with the design

specifications in a single, quantitative measure
70
1/1/2018
Process Capability: Cp and Cpk (cont’d) Process Capability (cont’d)

Process Capability Ratios (Cp, Pp) & Index (Cpk, Ppk)
Process Capability Scenarios Process Capability – Example

Cp = 1.00 Customers’ waiting time in a restaurant after ordering food
Cp >= 1.33,
Process barley
capable (27%
Process is for 20 days
capable
Defects)
Cp=2, Process
Cp >= 1.33 Achieved 6 Sigma
Process is Quality
capable
71
1/1/2018
Process Capability – Example (cont’d) Process Capability – Example (cont’d)
Calculate the process capability of a restaurant process
for orders delivery if management’s target is to have
customers wait no more than 25 min, but with upper
and lower specifications as low as 5 min and as high as
40 min?
Summary
All processes have variation

Process Capability is a measurement of how the process is
performing with respect to a desired outcome
Capability is defined as the voice of the customer over the voice of
the process
Long-term capability is not the same as short-term capability
Convert discrete to continuous and non-normal data to normal
before analyzing capability or use specialist software Benchmarking
The overall capability of a process can be defined by its Sigma
value
72
1/1/2018
Benchmarking Benchmarking (cont’d)
Process of identifying, understanding, and adapting Steps in benchmarking:

outstanding practices and processes to improve • Determine what to benchmark
organizational performance • Determine how to measure it
Benchmarking – “the search of industry best practices that • Gather information and data
lead to superior performance” • Implement the best practice within the organization
Best practices – approaches that produce exceptional
results, are usually innovative in terms of the use of
technology or human resources, and are recognized by
customers or industry experts
Types of Benchmarking Project Review – Measure (1 of 2)

Team members have received any necessary “just-in-
Competitive benchmarking - studying products, time” training
processes, or business performance of competitors in the Key metrics for all CTQ characteristics have been
defined
same industry to compare pricing, technical quality,
The team has determined what aspects of the problem
features, and other quality or performance characteristics need to be measured, including both process and results
of products and services. measures
Operational definitions of all measurements have been
Process benchmarking – focus on key work processes developed
Strategic benchmarking – focus on how companies All appropriate sources of data have been investigated,
and a data collection plan established before data is
compete and strategies that lead to competitive advantage collected
73
1/1/2018
Project Review – Measure (2 of 2)

Data collection forms have been tested and validated
Sample sizes required for statistical precision have been
identified
Data have been collected in an appropriate fashion,
according to plan
The data are accurate and reliable
Measurement systems have been evaluated using R&R
studies or other appropriate tools
Process capability has been addressed as appropriate
Benchmarks and best practice information has been
collected
Contents Analysis
Analysis is the examination of processes, facts, and data
1. Correlation Coefficient to gain an understanding of why problems occur and
where opportunities for improvement exist
2. Regression Analysis
3. Confidence Intervals
4. Hypothesis Testing
5. Analysis of Variance (ANOVA)
74
1/1/2018
Data Analysis in Six Sigma
Examining patterns, trends, and changes over time
Making comparisons relative to other business units,
competitor performance, or best-in-class benchmarks

Correlation Coefficient
Seeking to understand relationships among different
metrics
Correlation Coefficient Correlation Coefficient (cont’d)

Correlation: Procedure:
It finds the relation between two or more sets of data
1. Calculate the mean of all x and y values
It measures the strength and direction of the relationship between variables
2. Calculate the standard deviation of all x (Sx) and y (Sy) values
Correlation Coefficient: 3. For each pair (x,y) calculate the difference from their mean
It indicates the extent to which two variables are related 4. Get the sum by adding all the products of the together
• The data represented as ordered pairs (x,y) where x is the independent, or explanatory 5. Divide the sum of step 4 by (n-1), where n is the number of (x,y) pairs to get
variable and y is the dependent, or response variable the covariance
• It can range from -1.0 to +1.0
6. Divide the covariance by Sx and Sy to get the correlation r
• A positive correlation coefficient indicates a positive relationship, a negative coefficient
indicates an inverse relationship
• Correlation CANNOT be equated with causation
Causation:
A cause that produces an effect which gives rise to any action, phenomenon, or condition
Cause Effect
75
1/1/2018
Correlation Coefficient (cont’d) Correlation Coefficient Example 1
Y Y Y
X X X
r = -1 r = -.6 r=0
Y
Y Y
X X X
r = +1 r = +.3 r=0
http://www.syque.com/quality_tools/toolbook/
Correlation Coefficient Example (cont’d) Correlation Coefficient Example 1 (cont’d)
76
1/1/2018
Example 2 – Length of Stay Case Study – Excel

Case Study – Excel Calculations (cont’d)
Calculations
“Suppose an alliance of four hospitals is interested in Stay Length of Stay (LOS) Cost
understanding the rise in costs and what can be done to 1

2
3
5
2614
4307
control these costs. Three measures of hospital 3 2 2449
4 3 2569
performance that might be considered are length of stay, 5 3 1936
readmission rates, and costs per case” 6

7
5
5
7231
5343
Suppose we are interested in the relationship between 8

9
3
1
4108
1597
two variables: length of stay and hospital costs 10 2 4061
11 2 1762
We will use the data provided in the excel sheet to 12 5 4779
calculate the correlation coefficient using two excel 13

14
1
3
2078
4714
methods: 15 4 3947
16 2 2903
• Excel function: CORREL 17 1 1439
18 1 820
• Data analysis Æ Correlation 19 1 3309
20 6 5476
Regression Analysis
Linear regression:
It is used to described a straight line that best fits a series of ordered pairs (x,y)
• One variable is considered independent (=predictor) variable (x) and the other the
dependent (=outcome) variable y
Procedure for Least Squares method:

Regression Analysis 1.
2.
Calculate xy, x2, and y2 values
Calculate the sums of x, y, xy, x2, y2 and mean of x and y
3. Find out the linear equation that best fits the data by determining the value for a,
the y-intercept, and b the slope, using the following equations:
Y=aX+b y
b x
77
1/1/2018
Regression Analysis (cont’d) Regression and Correlation

Multiple linear regression:
An extension of linear regression to more than one independent variable Regression analysis – a tool for building statistical
models that characterize relationships between a
dependent variable and one or more independent
bi: are the coefficients variables, all of which are numerical
xi : are the independent variables
Non-linear regression: Scatter diagram – a visual indication of the type of

1. Exponential model: relationship between two variables
2. Power model: Correlation – a measure of a linear relationship between
3. Saturation growth model: two variables
4. Polynomial model:
Exercise: Simple Linear Regression Exercise: Simple Linear Regression (cont’d)

Relationship between number of dependents and yearly Scatter plot with possible relationship lines
healthcare expense
What can you get from the information above?
Which is the best line?
78
1/1/2018
Exercise: Simple Linear Regression (cont’d) Exercise: Simple Linear Regression (cont’d)
Errors associated with the various linear models Sum of squared errors associated with the various linear
models
The best line = 1.3(X) + 2.4 has the lowest squared error
of all the models
LOS Case Study – Excel Calculation

Let’s revisit the case study about the relationship
between LOS and Cost, we think that LOS causes costs–a
longer LOS results in higher costs
• Note that with correlation there are no claims made about
what causes what – just that they move together
Here we are taking it further by “modeling” the
direction of the relationship
Regression Analysis
Confidence Intervals
79
1/1/2018
Confidence Intervals (CIs) Confidence Intervals Summary Table
Confidence interval for the true value of the Confidence intervals:

Confidence intervals
population mean:
Mean
Proportion p + Zα/2 √(P(1-P)/n )
Variation
σ Population standard deviation

Point estimate of variation
Sample average
Values from table for (n-1) degree of freedom
n Sample size
Zα/2 Normal distribution value for a desired confidence level
p Population proportion estimate
n-1 Degrees of freedom
Hypothesis Testing
Make statement(s) regarding unknown population parameter values
based on sample data
Conjectures Î
A … or … B (Hypotheses)
Consequences
Motivation:
Examine two opposing
conjectures (hypotheses), H0
and HA
Zone of Belief
Hypothesis Testing
(Test Method)
Evaluation
hypotheses are mutually exclusive
and exhaustive
collect and analyze sample
information - for the purpose of
determining which of the two
hypotheses is true and vice versa
Beyond the issue of truth,
addressed statistically, is the issue
of justice which is beyond the
scope of statistical investigation
Gather & Evaluate
An assumption about the population parameter
Facts
80
1/1/2018
Hypothesis Testing (cont’d) Hypothesis Testing (cont’d)

Elements of a hypothesis test : Types of Errors:
Null hypothesis (H0): Statement regarding the value(s) of unknown Type I error (α): This type of error occurs when null hypothesis is rejected
parameter(s)(will always contain an equality) when it is actually correct, often referred as the producer’s risk
Alternative hypothesis (H1): Statement contradictory to the null hypothesis (will Example: Incoming products are good but were labeled as defective
always contain an inequality)
Test statistic: Quantity based on sample data and null hypothesis used to test Type II error (β): This type of error occurs when null hypothesis is not
between null and alternative hypotheses rejected when it actually should have been rejected
Rejection region: Values of the test statistic for which the null is rejected in favor of Example: Incoming products are defective but were labeled as good
the alternative hypothesis
Error Matrix False True

Null Hypothesis Alternative Hypothesis P=1-β, correct P=α, Type I α = significance level
H0: µ=x H1: µ≠x
Reject H0 1- β = power
outcome error
H0: µ≤x H1: µ≻x
H0: µ≥x H1: µ≺x Do not reject P=β, Type II P=1-α, correct
H0 error outcome
Goal: Keep α and β reasonably small

One-tail test: Specific interest in only one direction Required sample size:
Not scientifically relevant/interesting if reverse situation true The sample size n needed for hypothesis testing depends on:
• The desired Type I (α) and Type II (β) risk
Two-tail test: No a priori reason 1 group should have stronger effect
• Used for most tests • The minimum value to be detected between the population means
(µ- µ0)
• The variation in the characteristic being measured (s or σ)
Hypothesis tests for Means:
Z test Student’s t test
Hypothesis
Confidence
interval
σ Population standard deviation
s Sample standard deviation
One-tail test Two-tail test Sample average
(Munro et al., 2007) Student‘s t test: sample size, n<30

81
1/1/2018
Hypothesis Testing (cont’d)

Belief or null hypothesis, Ho: μ = b
Alternate belief or hypothesis, Ha: μ ≠ b
Decision rule: If z ≥ z* , reject the null
hypothesis. Where
Tests of Hypothesis
Based on a Single
Sample
Tests About
a
Population Mean
82
1/1/2018
Case I: A Normal Population With Known σ Recommended Steps in Hypothesis Testing

Null hypothesis: Test statistic value: 1. Identify the parameter of interest and describe it in the context of
the problem situation.
2. Determine the null value and state the null hypothesis.
3. State the alternative hypothesis.
4. Give the formula for the computed value of the test statistic.
Alternative Hypothesis Rejection Region for 5. State the rejection region for the selected significance level.
Level α Test
6. Compute any necessary sample quantities, substitute into the
formula for the test statistic value, and compute that value.
7. Decide whether H0 should be rejected and state this conclusion in
the problem context.
or Note: The formulation of hypotheses (steps 2 and 3) should be done

before examining the data.
Example Example Solution

A manufacturer of sprinkler systems used for fire protection in office
buildings claims that the true average system-activation
temperature is 130o. A sample of n = 9 systems, when tested,
yields a sample average activation temperature of 131.08oF. If the
distribution of activations times is normal with standard deviation
1.5oF, does the data contradict the manufacturer's claim at
significance level α = 0.05?
83
1/1/2018
Case II: Large-Sample Tests Example

When the sample size is large, the z tests for case I are A dynamic cone penetrometer (DCP) is used for measuring
modified to yield valid test procedures without requiring material resistance to penetration (mm/blow) as a cone is driven
either a normal population distribution or a known σ. into pavement or subgrade. Suppose that for a particular
application, it is required that the true average DCP value of a
certain type of pavement be less than 30. The pavement will not
be used unless there is conclusive evidence that the
For large n (n>40), s is close to σ.
specification has been met. Let’s test the appropriate
hypotheses using the following data (52 observations): …..
Test Statistic:
The use of rejection regions for case I results in a test

procedure for which the significance level is approximately α.
Example Solution Case III: A Normal Population Distribution

If X1,…,Xn is a random sample from a normal distribution,
the standardized variable
has a t distribution with n – 1 degrees of freedom.
The One Sample t Test:
Null hypothesis:
Test statistic value:
84
1/1/2018
The One-Sample t Test Example

A well-designed and safe workplace can contribute greatly to
Alternative Hypothesis Rejection Region for
increased productivity. It is especially important that workers not
Level a Test be asked to perform tasks such as lifting, that exceed their
abilities. The accompanying data on maximum allowable weight
of lift (MAWL in kg) for a frequency of four lifts/min was reported
in a study where subjects were randomly selected from the
population of healthy males age 18-30. Assuming that MAWL is
normally distributed, does the following data suggest the
population mean MAWL exceeds 25?
or
Hypothesis Testing
Example Solution Type I (α) and Type II (β) Errors
Ho: μ1=μ2 Ha: μ1≠μ2
Type I and Type II Error—Clinic Wait Time Example

Reality
Wait times at Wait times at the
the two clinics two clinics are
are the same NOT the same
μ1=μ2 μ1≠μ2
Wait times at the Type II or
two clinics are the μ1=μ2
Assess- same β error
ment or
guess Wait times at the Type I or
two clinics are μ1≠μ2
NOT the same α error
85
1/1/2018
Equal Variance t-Test Proportions

t-tests are used to test hypotheses about two means. Ho: π1= π2
Ho: μ1=μ2 Ha: μ1≠μ2 Ha: π1≠π2
Decision rule: If t ≥ t*, reject Ho Decision rule: If z ≥ z*, reject Ho
Confidence interval Confidence interval

Paired-comparison t tests: F test:
Two-means, Two-means,
equal variance unequal variance
Null Hypotheses H0:µ1 = µ2 H1:µ1 ≠ µ2
Alternative
H0:µ1= µ2 H1:µ1 ≠ µ2
Hypotheses
Procedure:
Test statistic
df
Pooled standard deviation
(Munro et al., 2007)
86
1/1/2018
Hypothesis Testing (cont’d) Hypothesis Testing Summary – Parametric

Summary: Normal/Large Sample Data?
Yes
Inference on means?
Yes No
Independent? Inference on variance?
No
Yes Yes
Variance Paired t F test for
known? variances
No
Yes Variances equal?

Z test
Yes No
T test w/ T test w/
equal unequal
variance variance
Hypothesis Testing Summary – Non-Parametric Example – Hypothesis Testing

If the wait time at two different clinics were of interest,
wait time for a random sample of patients from each
clinic might be measured. If wait time for a sample of 10
patients (for explanatory purposes only) from each clinic
were measured and it was determined that Clinic A had
a mean wait time of 12 minutes, Clinic B had a mean
wait time of 10 minutes, and both had a standard
deviation of 1.5 minutes.
• Perform a hypothesis test to determine if the wait time at two clinics

differs at a 95% CL.
87
1/1/2018
Example – Hypothesis Testing (cont’d) Project Review – Analyze (1 of 2)

Answer to Q1: Team members have received any necessary “just-in-
time” training
Team members understood how to use analysis tools
appropriately and effectively
The data collected in the Measure Phase have been fully
understood and studied
Appropriate statistical tools have been used to conduct
the analyses of data
Variation is thoroughly understood
Therefore, this test would reject Ho, the belief that the Root causes and hypotheses that explain problems have
mean wait time at the two clinics is the same. been identified
Project Review – Analyze (2 of 2)

Data provide confirmation of key conclusions and
validation of root causes
Process maps are accurate and representative of actual
Define
or desired process flow (in the case of a re-design
activity) Measure
The process has been studied to identify bottlenecks,
sources of error, and non-value added activities
Preliminary improvement or re-design goals have been Analyze
set
Improve Tools
Control
88
1/1/2018
Contents
1. Design of Experiments and Design of Experiments

ANOVA
(DOE)
2. FMEA
3. House of Quality ANalysis Of VAriance

(ANOVA)
Design of Experiments (DOE) Design of Experiments (cont’d)

The Blind Man and the Elephant
Definitions:
• A statistics-based approach to design experiments
• A methodology to achieve a predictive knowledge of a complex, multi-variable
process with the fewest trials possible
Purpose:
• To find out the effect of independent variables (control factors) on dependent
variables (noise factors)
• Identify optimal conditions and factors that most influence the results
• To find out interactions and synergies between factors
Noise factors
Signal Measured
factors System response
What we learn from an experiment may depend on WHERE we look,
HOW we look, and the SCOPE of our view!!!
Control factors
89
1/1/2018
Design of Experiments (cont’d) Design of Experiments (cont’d)

Basic Terms: Basic Terms:
Factor: It’s the variable controlled by the experimenter; can be viewed as a stimulus Variables:
Level: Possible values of a factor in an experimental design which could be • Dependent: Depends on another variable
quantitative or qualitative in nature • Independent: Affect or determine a dependent variable. Its commonly a
Ex.: Quantitative (three different temperatures), and input and does not have interaction with other variables
• Response: Shows the observed results (measures output values)
Qualitative (on-off, high-low)
Treatment: A treatment is a single level assigned to a single factor or experimental Replication: It’s the repetition of the set of all the treatment combinations in an
unit during e experimental run experiment (re-run an experiment with same input)
Ex.: Temperature 70ºF • Each repetition is called replicate or replication and is done to increase
Block: It’s a portion of the experimental material or environment that is common to reliability
itself and distinct from other portions • Used to determine impact of measurement error
Ex.: Samples from the same batch Experimental Error: It’s the variation in the response variables when levels and
Interaction: It’s Effect of one input factor that depends on level of another input factors are held constant
factor Experimental Design: It’s the formal experiment plan that includes the responses,
factors, levels, blocks, treatments, replication and so on
Munro et al. (2007) Munro et al. (2007)
Design of Experiments (cont’d)

DOE checklist:
Example of a Simple Factorial Design
• Define objective of experiment
Apply an
• Learn fact about the process iterative Driving while talking on the cell phone example
• Brainstorm a list of dependent and independent variables DOE
• Run “dabbling experiments” to debug equipment approach
• Assign levels to each independent variable
• Select or develop a DOE plan Hypothesis?
• Run experiments in random order and analyze periodically
• Draw conclusions and verity with replication
What is the criterion measure in this case?
Experimental design?
Residuals: How many groups?
Valid & Successful Should be
DOE normally and
Hypothetical scenario
Measurement system
capability:
independently Conclusions
distributed with
Confirm the measurement
Process stability: zero mean and
system before & throughout
Baseline needs to be constant variance
the life of the experiment
defined whether the
process is under control
90
1/1/2018
Example of a Simple Factorial Design Example of a Simple Factorial Design

Interaction between cellular phone use and driving conditions.
Table 1: Hypothetical Data for Driving Study: Average Number of Lane Deviations

Main effect: One-factor ANOVA experiment:
It’s an estimate of the effect of a factor independent of other means
Separates total variation observed in a set of
Interaction effect: measurements into:
It occurs when the effect of one input factor on the output depends on the level of 1. Variation within one system
another input factor
• Notification: High level Æ + (plus)
• Due to random measurement errors
Low level Æ - (minus) 2. Variation between systems
• Due to real differences + random error
Is variation(2) statistically > variation(1)?
One-factor experimental design
Main
effect Interaction
effect
91
1/1/2018

One-factor ANOVA Two-factor ANOVA:
• Each individual measurement is composition of Factor A – a input levels
• Overall mean Factor B – b input levels
• Effect of alternatives n measurements for each input combination
• Measurement errors abn total measurements
• Each individual
Factor A
measurement is 1 2 … j … a
composition of 1 … … … … … …
• Overall mean 2 … … … … … …
• Effects
Factor B
… … … … … … …
• Interactions i … … … yijk … …
n
• Measurement errors replications
… … … … … … …
b … … … … … …
Lilja (2004) Two Factors, n Replications Lilja (2004)

Two-Factor ANOVA: Full-factorial experiments:
• It considers every possible combination in order to complete a full study of
interactions
• It can be retained and converted into different experimental designs
• It requires a large number or runs if there are several factors or levels
• Number of runs =LF where L=# of levels, & F=# of factors
Two-level fractional factorial experiments:

• It saves money and time by not examining every possible combination
Sum of Square total: • It is used for quick exploratory tests, when interactions are insignificant, and many
tests are needed rapidly
SST = SSA + SSB + SSAB + SSE • Procedure:
Degrees of freedom • Select process; identify output factors of concern and input factors and level to
• df(SSA) = a – 1 be investigated
• Select a design, conduct the experiment under predetermined conditions, and
• df(SSB) = b – 1 analyze the data
• df(SSAB) = (a – 1)(b – 1) • Analyze the data and draw conclusion
• df(SSE) = ab(n – 1)
• df(SST) = abn - 1 Lilja (2004) Munro et al. (2007)
92
1/1/2018
Example 1 – DOE for Printed Circuit Board Quality Example 1 – DOE for Printed Circuit Board Quality
Improvement Improvement
1. Define the experiment objective: 3. Factors and levels
¾ The objective is to determine the optimal settings of the
critical-to quality factors in the screening process of PCB.
Any defect in the solder joint can lead to circuit failure, the
screening process is regarded as the most critical process
in PCB manufacturing. According to the current process
capability study of a PCB company, the capability of the
screening process is under 1.33—the company
requirement.
2. Identify the key variables and their levels:

¾ The variables are Solder paste viscosity Speed of squeegee,
Pressure of squeegee and Side of stencil, and Blade type.
¾ Levels are shown in table
4. Design the experiment 5. Experiment Run
¾ Experimental conditions
(1) Room temperature: 25C ¾ “A full factorial experiment was carried out, and the whole
(2) Room humidity: 56%
(3) Machine number: 12 experiment was completed in about two hours. There were
(4) Number of operators: 1
48 types of printing, and two PCBs were measured for
(5) Model: Neptune
(6) Snap off distance: nearly zero
each type of printing. In total, 96 PCBs were measured,
(7) Squeegee pressure: 28 bar
(8) Squeegee speed: 0.7 inch/sec and 480 solder paste height data were collected for
(9) Point locations: J1, U1, U1, U2, U2, and
(10) Specification: (4.5–7) mil analysis.”
93
1/1/2018
6. Results (Main effect plots) 7. Normal effect plots for height average
¾ “The pressure of squeegee, blade type, and side of stencil
were significant factors for the height average. The solder
paste viscosity, speed of squeegee, blade type, and side of
stencil were significant factors for the height variation”
8. Normal effect plots for height variation 9. Data Analysis—Interactions.
94
1/1/2018
Example 1 – DOE for Printed Circuit Board Quality

Improvement
9. Data Analysis—Interactions (cont’d).
ANalysis Of VAriance
(ANOVA)
ANOVA/MANOVA/MANCOVA One-Way ANOVA
One-way ANalysis Of VAariance (ANOVA) is It’s a method for comparing several population means and a factor in ANOVA
describes the cause of the variation in the data.
used to test hypotheses about three or more • Conditions for applying ANOVA:
levels of treatment. A t-test will give the same a. Population should be normally distributed
b. The samples must be independent of each other
information as an ANOVA when there are only c. Each population must have the same variance
two treatment levels of interest. Procedure:
Two-way and higher ANOVAs are used when
there is more than one type of treatment variable
of interest.
MANOVA/MANCOVA are used when there is
more than one outcome or dependent variable of
interest.
Munro et al. (2007)
95
1/1/2018
ANOVA Table ANOVA Example 1
Source of
Sum of squares DF Mean squares F-statistic
“Daily product defect rates are collected on three different methods
variation
Between treatment K-1 of production.
Within treatment N-k
Total N-1
• The Null Hypothesis (H0) is stated: There is no statistically
N # reading T Grand total of readings
significant difference in the daily product defect rates and the
n # per level (or treatment)
k # levels (or treatments)

production methods used measured at .05 alpha (level of
C
significance) using a one-tailed F-test (ANOVA).”
Individual measurements
Michalski (1997)
Munro et al. (2007)
ANOVA Example 1 (cont’d) ANOVA Example 1 (cont’d)

Calculate SSB:
Michalski (1997) Michalski (1997)
96
1/1/2018
ANOVA Example 1 (cont’d) ANOVA Example 1 (cont’d)

Calculate df: Calculate the mean squares:
Calculate F-ratio
Michalski (1997) Michalski (1997)
ANOVA Example 1 (cont’d) ANOVA Example 2

ANOVA table: Imagine you manufacture paper bags and you want to improve
the tensile strength of the bag. You suspect that changing the
concentration of hardwood in the bag will change the tensile
strength. You measure the tensile strength in pounds per square
inch (PSI). So, you decide to test this at 5%, 10%, 15% and 20%
hardwood concentration levels. These "levels" are also called
"treatments."
Hardwood Concentration %
Sample #
5% 10% 15% 20%
1 7 12 14 19
2 8 17 18 25
3 15 13 19 22
4 11 18 17 23
5 9 19 16 18
Michalski (1997)
6 10 15 18 20
97
1/1/2018
ANOVA Healthcare Example ANOVA Healthcare Example (cont’d)

Suppose an alliance of four hospitals is interested in understanding
the rise in costs and what can be done to control these costs. Three
measures of hospital performance that might be considered are
length of stay, readmission rates, and costs per case. One initial
question might be whether there are real differences between the
four hospitals in these three –cost-related measures of performance
ANOVA Healthcare Example (cont’d)

Result
Failure Mode and Effects

Analysis (FMEA)
98
1/1/2018
Failure Mode and Effects Analysis (FMEA) FMEA (cont’d)

A Failure Mode is: What is FMEA:
• The way in which the component, subassembly, product, input, or A methodology to evaluate failure modes and their effects in designs and in
process could fail to perform its intended function processes. A structured approach to:
• Failure modes may be the result of upstream operations or may cause
• Identifying the ways in which a product or process can fail
downstream operations to fail • Estimating risk associated with specific causes
• Prioritizing the actions that should be taken to reduce risk
• Things that could go wrong
• Evaluating design validation plan (product) or current control plan
Failure Likeliness Detectability Severity Risk (process)
Specific Cause Effect of Failure
Mode of Failure of Failure of Failure Priority
Inputs Outputs
Patient Patients under
Injury 3 5 10 List of actions
Falls sedation Brainstorming
C&E Matrix to prevent
Process Map causes or detect
Likeliness of Failure: 1-10 with 10 presents most likely
Process History FMEA failure modes
Detectability of Failure: 1-10 with 10 representing most difficult Procedures
Severity of Failure: 1-10 with 10 representing most severe Knowledge History of
Experience actions taken
Risk Priority Number (RPN)= (Likeness of Failure).(Detectability of Failure).(Severity of Failure)
FMEA (cont’d) FMEA (cont’d)

Procedure: FMEA form:
1. For each process input (start with high value inputs), determine the ways in
which the input can go wrong (failure mode)
2. For each failure mode, determine effects
Select a severity level for each effect
3. Identify potential causes of each failure mode
Select an occurrence level for each cause
4. List current controls for each cause
Select a detection level for each cause
5. Calculate the Risk Priority Number (RPN)
6. Develop recommended actions, assign responsible persons, and take actions
Give priority to high RPNs
Identify failure modes Identify causes of the Prioritize Determine and assess
MUST look at severities rated a 10 and their effects failure modes actions
7. Assign the predicted severity, occurrence, and detection levels and compare and controls
RPNs Severity X Occurrence X Detection = RPN
99
1/1/2018
FMEA (cont’d) Summary: Steps to Perform FMEA

Definitions: 1. Identify the process to be analyzed. Typically, this is a principal
process for the organization.
Severity: Importance of the effect on customer requirements
2. Assemble and train the team. Processes usually cross functional
• Often can’t do anything about this
boundaries; therefore, the analysis should be performed by a team
Occurrence: Frequency with which a given cause occurs and creates failure modes of relevant personnel. No one person or persons from a single
Detection: The ability of the current control scheme to detect or prevent a given functional area will have the knowledge needed to perform the
cause analysis.
The relationship
between failure
Failure Mode 1 Effect 1 3. Develop a detailed process flowchart, including all steps in the
Rating scales: modes and effects process.
is not always 1 to 1 Failure Mode 2 Effect 2
4. Identify each step (or function) in the process.
Severity 5. Identify potential failures (or failure modes) at each step in the
1 = Not Severe, 10 = Very Severe Failure Mode 1
Effect 1 process. Note that there may be more than one potential failure at
Occurrence
1 = Not Likely, 10 = Very Likely
Failure Mode 2 each step.
Detection
1 = Likely to Detect, 10 = Not Likely to Effect 1
Failure Mode 1
Detect Effect 2
[Healthcare Operations Management, 2008]
Summary: Steps to Perform FMEA (cont’d) Summary: Steps to Perform FMEA (cont’d)
6. Determine the worst potential consequence (or effect) of each 9. Rate the severity of each effect (on a scale of 1 to 10, with 10 being the
possible Failure. most severe). This rating should reflect the impact of any controls
7. Identify the cause(s) (contributory factors) of each potential failure. that reduce the severity of the effect.
An RCA can be helpful in this step. Note that there may be more 10. Rate the likelihood (occurrence score) that each cause will occur (on
than one cause for each potential failure. a scale of 1 to 10, with 10 being certain to occur). This rating should
8. Identify any failure “controls” that are currently present. A control reflect the impact of any controls that reduce the likelihood of
reduces the likelihood that causes or failures will occur, reduces the occurrence.
severity of an effect, or enables the occurrence of a cause or failure 11. Rate the effectiveness of each control (on a scale of 1 to 10, with 1
to be detected before it leads to the adverse effect. being an error-free detection system).
12. Multiply the three ratings by one another to obtain the risk priority
number (RPN) for each cause or contributory factor.
13. Use the RPNs to prioritize problems for corrective action. All causes
that result in an effect with a severity of 10 should be high on the
priority list, regardless of RPN.
14. Develop an improvement plan to address the targeted causes (who,
when, how assessed, etc.).
[Healthcare Operations Management, 2008] [Healthcare Operations Management, 2008]
100
1/1/2018
Summary: Steps to Perform FMEA (cont’d) Summary: Steps to Perform FMEA (cont’d)
Number
Severity to Occurrence of the Detection of Occ: Likelihood of Occurrence (1-10)
Organization Severity Occurrence
Damage without Absolute
10
warning
Very high (1 in 2)
Uncertainty Det: Likelihood of Detection (1-10)
9 Damage with warning Very high (1 in 3) Very remote
8 Very high risk Very high (1 in 8) Remote

Note: 1 = Very likely it will be detected
7 High risk High (1 in 20) Very low
10 = Very likely it will not be detected
6 Moderate risk High (1 in 80) Low
5 Low risk Moderate (1 in 400) Moderate Sev: Severity (1-10)

4 Very low risk Moderate (1 in 2,000) Moderately high
Risk Priority Number (RPN) = Occ*Det*Sev
3 Minor risk Low (1 in 15,000) High
2 Very minor risk Low (1 in 150,000) Very high

Remote (1 in
1 No risk at all Almost certain
1,500,000)
[Healthcare Operations Management, 2008]
Example 1 – FMEA for Administration of

Example 2 – FMEA for Power Relay Assembly
Cisatracurium in Surgery
101
1/1/2018
Example 3 – FMEA for Reactor System

Components
House of Quality/Quality
Function Deployment (QFD)
http://www.isa.org/InTech
Quality Function Deployment Steps of QFD

Identify
Seeks out spoken and unspoken customer needs from Voice of the Service/Product
Specification
Customer (VOC) Rate Relative 3 Relation between
Uncovers “positive” quality that wows the customer Importance Needs & Specification
Translates these qualities into designs characteristics and deliverable 2
actions
4
Builds and delivers a quality product or service by focusing various Customer
business functions toward achieving a common goal – customer Needs 1 Benchmarking
satisfaction 5 Analysis
Provides an opportunity for competitive advantage

Rate a New Design
Set Target & Calculate
A graphic tool for defining the relationship between customer desires Value 10 6 Improvement
and the firm/product capabilities. Ratio
Planning matrix relates customers’ demand with company’s 7

product/service specifications Establish 9
Specification into 8 Sales Impact
Hierarchy
It’s a customer-driven process used to relates Organize
customer requirement to internal specifications
Binghamton University 407 NeedsUniversity
Binghamton into 408
(Eppinger & Ulrich, 2001)
Hierarchy
102
1/1/2018
House of Quality Identify Customer Needs

Tells about critical
Customer
Correlation to the quality of
Requirement the process
matrix
priorities according to your
customer
Technical
requirement
Voice of
Relationship
Customer
matrix
(VOC)
Technical requirement Competitive

priorities evaluation
(Eppinger & Ulrich, 2001) (Eppinger & Ulrich, 2001)

Identify Customer Needs (cont’d) Key Specifications

Deploying Voice of Customer
Establish Product Specification
Technical
Requirements
Customer
Requirements Identify Specification Correlation
Product
Requirements
Technical Identify Correlation between each Need
Requirements
& Specification
Process
Requirements
Product Establish Specification into Hierarchy
Requirements
Control
Requirements
Process
www.thequalitycatalyst.com Requirements (Eppinger & Ulrich, 2001)
103
1/1/2018
QFD Example 1 QFD Example 1 (cont’d)
Goal: Develop a system to ensure

that diabetes patients receive Technical
preventive exams responses
Appointment
appointment
length range
Information
Subsequent
Customer
notification
notification
needs
schedule
On-time
on need
Time to
Initial
Knowledge that it is time for an office visit Time knowledge
Knowledge of why follow-up is needed Why knowledge
Convenient
Convenient to schedule
Appointment length
Known appointment length Appointment time
Appointment on time
QFD Example 1 (cont’d) QFD Example 1 (cont’d)

Importance:
Patient desire Relationships:
Cost D Strong = 5
Competitive Medium = 3
appointmen
Subsequent
U Weak = 1
notification
notification
advantage
Appointme
Informatio
n on need
Subsequent
notification
notification
nt length
Appointme
Informatio
schedule
appointme
On-time
n on need
Time to
nt length
schedule
On-time
Initial
range
Time to
Initial
range
t
nt
Time knowledge 5
Time knowledge 5 D
Why knowledge 3 Why knowledge 3 D
Convenient 4
Appointment length 3
Convenient 4 D
Appointment time 4 Appointment length 3 D
Appointment time 4 D
104
1/1/2018
Replace icons with numbers Multiply by importance and sum
Relationships:
Appointment
appointment
length range
Information
D Strong = 5
Subsequent
notification
notification
appointmen
Subsequent
notification
notification
Appointme
Informatio
Medium = 3
n on need
schedule
On-time
nt length
on need
schedule
Time to
U Weak = 1
On-time
Time to
Initial
Initial
range
t
Time knowledge 5 5 3
Time knowledge 5 5 3
Why knowledge 3 5
Why knowledge 3 5
Convenient 4 5
Convenient 4 3
Appointment length 3 5 3
Appointment length 3 5 3
Appointment time 4 3 5
Appointment time 4 3 5
25 15 15 20 27 29

Technical
Relationships: correlations
+ = Strong positive Target:
+ + D D
= Strong negative 100
diabetics/month;
appointmen
appointmen
Subsequent
Subsequent
notification
notification
notification
notification
Appointme
Appointme
85% compliance
Informatio
Informatio
n on need
n on need
nt length
nt length
schedule
schedule
On-time
On-time
Time to
Time to
Initial
Initial
range
range
t
t
Time knowledge 5 5 3 Time knowledge 5 5 3
Why knowledge 3 5 Why knowledge 3 5
Convenient 4 5 Convenient 4 5
Appointment length 3 5 3 Appointment length 3 5 3
Appointment time 4 3 5 Appointment time 4 3 5
25 15 15 20 27 29 25 15 15 20 27 29
105
1/1/2018
QFD Example 1 – Outcome QFD Example 2

To: Dan McLaughlin QFD for a product design
From: Southview Clinic
Dear Dan,
You had an appointment with Dr. Adams about six months ago, and it is
now time for another visit. We need to check your blood pressure, do
some blood tests, and adjust your prescriptions if needed. We would
like to review these preventive procedures in advance, so please see
www.southview.com/prev22.
We have two openings available next week, on Tuesday at 8:30 am and
Thursday at 2:30, to see Dr. Adams. Click on one of these days to make
the appointment, or e-mail us with dates and times that work for you.
We appreciate you continuing your care with us, and Dr. Adams looks
forward to seeing you.
Contents
Define 1. Control Charts
2. Control Plan
Measure
Analyze
Improve Tools
Control Tools
106
1/1/2018
Statistical Process Control (SPC) SPC (cont’d)
SPC is a statistics-based methodology for determining

when a process is moving “out of control”
All processes have variation in output
Some of the variation is inherent in the process

(common)
Some of the variation is due to assignable (special)

causes
SPC is aimed at discovering variation due to assignable

causes and correcting those causes
SPC
Out-of-Control Situations SPC (cont’d)
50% of
patients 10% of
wait patients
more wait
than 30 more
minutes than 30
minutes
107
1/1/2018
Common Cause Variations Special Cause Variations

If only common causes of variation are present, the output of a process
forms a distribution that is stable over time and is predictable • Variations that can be traced to a specific reason due to unusual
events
• The objective is to discover when special causes are present
; Eliminate the root causes of the special variations
; Incorporate the good causes
If assignable causes are present, the process output is not stable over time
and is not predicable
Prediction
?
?? ? ?
? ?
Frequency
? ?
? ?
? ?
? ??
?? ?
Frequency
Prediction
Weight
Weight
SPC Interpretation Developing Control Charts

• Choose the characteristic to be charted
• Identify the process variables and conditions contributing towards process
characteristics
(a) In statistical control • Consider attribute data (% of defective) and variables data (numerical
and capable of measurements) to diagnose causes and determine actions
producing within
• Determine earliest point in the process to get information on assignable causes
Frequency control limits
• Choose the type of control chart to be used
• Decide the central line and the basis for calculating control limits
Lower Control Limit Upper Control Limit
• Choose the rational subgroup and the appropriate strategy subgroup frequency,
(b) In statistical control size, and so on)
but not capable of
producing within control Selection of variables:
limits
• Key process input variables (KPIVs) to determine their effect on a process
• Key process output variable (KPOVs) to determine process capability
• DOE and ANOVA methods to identify variables significant to process control
(c) Out of statistical control
and incapable of producing
within limits
Size
(weight, length, speed, etc.)
108
1/1/2018
Subgrouping Population and Sampling Distribution
It is applied to enhance randomness and reduce ‘piece-to-

Three population Distribution of
piece’ variation distributions sample means
Mean of sample means = x
• The division of observations into rational subgroups is key Beta
• Success of control charting depends on the selection of subgroups Standard
• Selection should result in groups as homogeneous as possible deviation of σ
• One subgroup should be representative of all of the production over a given Normal the sample = σx = n
period of time means
• More useful information is derived from smaller groups (that is, 5 subgroups of
5, than 1 subgroup of 25) whereas larger subgroups makes the process unstable Uniform
• Attributes control charts are based on Poisson or Binomial distribution and
requires 50 or more samples within subgroups | | | | | | |
-3σx -2σx -1σx x +1σx +2σx +3σx

95.45% fall within ± 2σx
A method for selecting samples for a control chart 99.73% of all x
fall within ± 3σx
Population and Sampling Distribution (cont’d) Variable Control Chart: Xbar and R
• Determine sample size and frequency
• Calculate average and range, and the average of both measures
• Calculate the control limits based on the subgroup sample size
Sampling • Plot the data and analyze the chart
distribution
of means Upper Control Limit Lower Control Limit
Process
distribution Range (R) Chart
of means
x=μ • x-charts are to control the central tendency of the process

(mean) • R-charts are to control the dispersion of the process
109
1/1/2018
Detection of Central Tendency Detection of Dispersion
These (Sampling mean is These

sampling shifting upward but sampling (Sampling
distributions range is consistent) distributions mean is
result in the result in the constant but
charts charts dispersion is
below below increasing)
UCL UCL
(x-chart (x-chart does
x-chart detects shift x-chart not detect the
in central increase in
LCL tendency) LCL dispersion)
UCL UCL
(R-chart does (R-chart detects
R-chart not detect R-chart increase in
change in dispersion)
LCL mean) LCL
Example 1 Example 1 (cont’d)
smpl 1 smpl 2 smpl 3 smpl 4 smpl 5 smpl 6 Avg.

smpl 1 smpl 2 smpl 3 smpl 4 smpl 5 smpl 6 Avg.
Obs 1 7 11 6 7 10 10
Obs 2 7 8 10 8 5 5 Obs 1 7 11 6 7 10 10
Obs 3 8 10 12 7 6 8 Obs 2 7 8 10 8 5 5
x-bar Obs 3 8 10 12 7 6 8
R X-bar 7.3333 9.6667 9.3333 7.3333 7 7.6667 8.0556
R 1 3 6 1 5 5 3.5
X-bar chart R chart
X-bar chart R chart
UCL
UCL 9.69 7
Centerline Centerline 8.0556 3.5
LCL 6.42 0
LCL
110
1/1/2018
Example 1 (cont’d) Example 2

Let us re-visit the restaurant example: observations were
made for 20 days on a sample size of 6 customers
X-bar chart waiting for food after ordering
R chart
Example 2 – Xbar Chart Example 2 – R Chart
111
1/1/2018
Variable Control Chart: Xbar and S Attribute Control Charts

• Its similar to x-bar and R chart except that each range is replaced by the sample • Used to count data where data are generally qualitative in nature but can be counted
standard deviation s recorded and analyzed
• Require large sample size of (50-200) to fit Binomial or Poisson distributions
• Construction is similar to variable control chart except sample size is larger here
Upper Control Limit Lower Control Limit
• Examples: nonconformities, nonconforming units, and % of nonconforming
(defectives)
Std. dev. (S) Chart Defect Vs Defective

• Defective is a superset of defects (non-
p charts: Defectives-sample size varies conforming units)
• There may be n no. of defects to have
np charts: Defectives-sample size fixed one defective product
c charts: Defects-sample size fixed
• A defect does not necessarily mean the
product/service is defective
u charts: Defects-sample size varies Example: If a man is not feeling well, it is
something similar to a defective. This is
http://www.qualitygurus.com/courses/mod/forum/ because of fever or stomach pain or
discuss.php?d=771 diarrhea (defects)etc. All these illness are
defects
Example : Control Chart of Production of

Other Control Charts
Adhesive Labels on Large Rolls
Upper Control Limit Lower Control Limit
p control chart
np control chart
u control chart
C control chart
n= sample size
Holyes et al., (2007)
112
1/1/2018
Example Example – Data
“An accounts department started an improvement

project to try to reduce the number of internal purchase
forms that its users completed incorrectly. As an overall
measure of their success, they used a p-type Control
Chart to measure the proportion of purchase forms that
were not completed correctly. This was chosen, rather
than measuring the actual number of defects, because
any number of defects on a form required about the
same effort to revise”
http://www.syque.com/quality_tools/toolbook/Control/example.htm
Example – P-Chart
http://www.syque.com/quality_tools/toolbook/Control/example.htm
113
1/1/2018
Process Control Plan – ASQ Process Control Plan – ASQ

Step 1. Briefly describe the process being performed.
Step 2. Indicate the immediate benefactor of the process.
Step 3. List the department and/or persons who either initiate or perform
the process steps.
Step 4. Show the process steps performed.
Step 5. For the major steps of the process, determine the critical success
factors and measures (e.g., time required to execute each step).
Step 6. Indicate the negotiated needs and expectations with relevant quality
indicators.
Step 7. Develop control charts for selected control points and indicate
control limits or any other metric to track performance over time (if control
charts are not applicable).
Step 8. List items that need to be checked, frequency, responsible person,
and contingency plans when process necessitates.
Step 9. List dates of changes and what changes were made.
Step 10. List any other pertinent information needed.
Developing a Control Plan – An Example
114
1/1/2018
Case Study I: A Case Study In Small Scale Industry A Case Study In Small Scale Industry (cont’d)
¾ Ball-bearing manufacturing facility ¾ Grinding process plays a crucial role in the making of
¾ Study was conducted in the unit where inner rings (IR)
and outer rings (OR) of ball bearings are processed Rings
¾ Main process in IR and OR is the grinding process ¾ All the grinding machines are computer controlled
¾ Tight tolerances are required in finishing of both IR and
machines
OR up to a level of 3 – 10 microns.
¾ Besides dimension other requirements are ¾ Even though the machines are computer controlled there
¾ Surface finish are still rejections of rings
¾ Free from pin holes
¾ Ovality of the rings ¾ Rejection rate was 2.2%
¾ Demagnetization
Reddy & Reddy (2010) Reddy & Reddy (2010)
A Case Study In Small Scale Industry (cont’d) A Case Study In Small Scale Industry (cont’d)
¾ Inner/outer ring manufacturing process ¾ DMAIC
¾ Bearing steel (SAE 52100) is machined (turned) or forged ¾ Define
into rough cut, basic ring configurations ¾ Defining process stages using SIPOC
¾ Rings are machined to within rough tolerance ¾ Reduction in rejection rate of IR and OR is targeted
specifications
¾ Rings are heat treated to increase the steel’s strength
¾ Ring faces receive the final grinding, removing any rough
spots
¾ Ring OD and ID are finish ground to a smooth surface
¾ Raceways are also finish ground to an even surface
¾ Raceways are honed to a polished finish
¾ Rings are then cleaned and readied for assembly
115
1/1/2018
¾ DMAIC ¾ DMAIC
¾ Measure ¾ Analyze
¾ Measurement system analysis is conducted which includes ¾ Pareto Chart
gauge repeatability and reproducibility studies. ¾ Bearing rings
¾ Gauge R&R study, results indicate repeatability as 1.97% and rejection arises due
reproducibility as 0.18% and the total measurement variation to:
to be 2.15% ¾ Bore diameter
defect (BD)
¾ Track defects (TD)
¾ Face defects (FD)
¾ Outside diameter
defects (OD)
¾ Bore diameter defects
caused the major
portion in rejection of
Reddy & Reddy (2010)
rings Reddy & Reddy (2010)
¾ DMAIC ¾ DMAIC
¾ Analyze ¾ Analyze
¾ Cause & Effect diagram ¾ Parameters which affect the bore diameter based on C&E :
¾ Material composition
¾ Machine settings
¾ Tool life (grinding wheel)
¾ Coolant quality
¾ Loading and unloading (handling)
¾ Power supply quality
¾ Measurement of rings
116
1/1/2018
¾ DMAIC ¾ DMAIC
¾ Improve ¾ Improve
¾ Steps in improve phase: ¾ Steps in improve phase (cont’d):
¾ Supplier quality: maintain consistency in material composition ¾ Coolant quality: recycling and change of coolant at regular
as well as dimensions intervals
¾ Machine settings: operators trained on machine setting as well ¾ Loading and unloading (handling): Loading and unloading
as in statistical methods useful in quality control points in the machines are arranged in the best manner to
¾ Tool life (grinding wheel): for each specific bearing size, avoid the damage to the things
optimum grinding wheel life has been estimated and instructed ¾ Power supply quality: Steps initiated to stabilize the power
operators accordingly supply
¾ Measurement of rings: Instead of measuring immediately after

processing, it is better to measure the diameter of rings after
cooling and cleaning
¾ DMAIC ¾ DMAIC
¾ Control ¾ Control
¾ Sigma levels before and after improvement ¾ Pareto chart after improvement
117
1/1/2018
¾ DMAIC ¾ DMAIC
¾ Control ¾ Control
¾ Process capability after improvement ¾ T-test for 2009 vs.2008
Case Study II: Improving ED Throughput Measure
Project Title: ED Throughput Project Scope: What is the Right Y (CTQ) to Measure? How will it be measured?
In Scope - Treat to Street pts, Staffing patterns (ED MDs
Y = Door to Doc Time. From the time a patient enters through the door until the physician
& RNs), Equip’t, FTEs, Registration, Lab, X-R.
enters the exam room to assess the patient, measured in minutes.
Out of Scope - ED Admits, ED Hold Hours, Bed Control,
Housekeeping, Transport to Floor, MR, US, CT, Pharm.
Customer(s):
Patients, Physicians
What is our goal?
Improve the average ED Throughput Time for Patients by 40%.
Potential Benefits:
• Decrease LWBS
Reduce the weighted average Door-to-Doc time from 65 minutes to 40 minutes.
Project Description : • Increase patient satisfaction (Press Ganey #s) Improve our throughput yield of patients seeing a physician within 60 minutes (USL) from
Moving patients through the ED takes too long. • Reduce ED LOS (Soft Dollars) 67% current to 80%.
One-third of patients wait longer than 60 minutes to This reduction in defect rate of 13% represents over 7,500 customers.
be seen by a physician.
Alignment with Strategic Plan:

• Customer Service
• Growth
• Efficiency What are the specification limits? (LSL, USL) What is the Target?
Based upon our VOC data, we have set a USL of 60 minutes and a Target Mean of 40
minutes.
www.healthcare.isixsigma.com www.healthcare.isixsigma.com
118
1/1/2018
Analyze Analyze
Statistical Analysis
Value Stream Map Opportunities for Performance Improvements:
Door-to-Doc Subcycle Triage

Fax
EKG, Draw
Blood, UA,
X-Ray – written
Front Order X-Ray, In ED report/ED
administer Pain
Desk / QR med
2- RNs Portable
1 Tech
Team Area
ED Lab
Call critical values
Waiting
Patient Flow Room Treatment
People Flow
Tube/blood
(RN, MD, etc.)
E-Info Flow
MD
Other Flow
(blood, etc.)
Phone Call Arr QR QR Triage Triage Bed Bed MD
Patient Wait Time 6.3 min 11.6 min 23.5 min 22.9 min
Current Average Cycle Times

8What do we want to know?

Analyze Improve 8Screen Potential Causes?
8Discover Variable Relationships?
8Establish Operating Tolerances?
What X’s (inputs) are causing most of our variation?
Results for: Historical DOE Door to Doctor Time
Factorial Fit: D2D versus Express Care, X-Ray, Bed Open`
What X’s (inputs) have we chosen to improve?
Estimated Effects and Coefficients for D2D (coded units)
1. Bed Availability
Term Effect Coef SE Coef T P
Constant 87.34 2.547 34.30 0.000 – The Measure Phase data demonstrated that Door-to-Doctor time increased by two to
Express Care 35.56 17.78 2.547 6.98 0.000 three times when there is no bed open for the patient.
X-Ray 36.06 18.03 2.547 7.08 0.000
Bed Open -37.81 -18.91 2.547 -7.42 0.000 2. Ancillary Services
Express Care*X-Ray 33.69 16.84 2.547 6.61 0.000 – The data further showed that the time it takes to perform an X-Ray or Lab testing is
Express Care*Bed Open 32.56 16.28 2.547 6.39 0.000 statistically significant in relation to Door-to-Doctor time.
X-Ray*Bed Open 14.06 7.03 2.547 2.76 0.025
Express Care*X-Ray*Bed Open 5.19 2.59 2.547 1.02 0.338 3. Express Care
S = 10.1865 R-Sq = 96.87% R-Sq(adj) = 94.12% – Lower acuity patients (i.e. Level 3 / Express Care) wait longer to see a physician than
do higher acuity patients (i.e. Level 1).
Analysis of Variance for D2D (coded units)
Source DF Seq SS Adj SS Adj MS F P
Main Effects 3 15979.9 15979.9 5326.6 51.33 0.000
2-Way Interactions 3 9571.7 9571.7 3190.6 30.75 0.000
3-Way Interactions 1 107.6 107.6 107.6 1.04 0.338
Residual Error 8 830.1 830.1 103.8
Pure Error 8 830.1 830.1 103.8
Total 15 26489.4
119
1/1/2018
Improve Improve
Value Stream Map Key Points / Opportunities for Improvement:
Bedside Registration What is the mean and median of our process? What is the standard
deviation?
Triage
EKG, Draw
Measure Phase Control Phase +Δ %
Blood, UA, Registration Non-value added Mean score 64.3 minutes 39.8 minutes 38.1%
Front Order X-Ray,
Desk / QR
administer Pain
med
If rooms ful step removed Median 38.5 minutes 34.0 minutes 11.7%
2- RNs
may reg pt Standard Deviation 44.7 minutes 27.7 minutes 38.0%
while
1 Tech
waiting.
HI/LO 241 / 11 minutes 129 / 4 minutes 46.5% (HI; outliers)
Range 230 minutes 125 minutes 45.7%
ED
What is our process capability (Z score, DPMO, Yield %)?
Waiting Z Short-Term Score = 1.91σ 2.35σ 0.44σ
Patient Flow Room DPMO = 333,333 175,000 <109,523>
People Flow Yield % = 66.7% 82.5% 15.8%
(RN, MD, etc.) Impacts:
1 – Inc. Patient Satisfaction
E-Info Flow 2 – Red. time by 8.7 minutes
Patient Wait Time
3 – Red. variability in process
Control Case Study III: Linen Utilization
What are our financial results? How were they calculated? Project Title: Linen Utilization
Our Financial Impact is $1,120,650 and reflects the improvement in LWBS visits and the
corresponding admissions as well as a conservative (5%) recognition as a result of Project Description: To Identify
throughput improvement. opportunities within the organization
which allows for better linen utilization
without compromising quality or patient
What is the plan for monitoring/ auditing the process? What is the Control care.
Plan?
Problem Statement: Currently, linen usage is
higher than what is expected for a facility of size
and acuity level. We need to look for ways to
better utilize our daily linen supply and lower
our overall pounds per patient day as well as
our cost per patient day.
Project Scope: The use of linen for inpatients.
120
1/1/2018
Define High Level Process Map
What is the Right Y (CTQ) to Measure? How will it be measured?

Y = Pounds Per Patient Day of Linen Used
Pounds Per Patient Day of Linen Used by Service Line Step 1 Step 2 Step 3 Step 4
Inventory of linen Linen order for the Linen is received Exchange carts
is taken in Linen next day is placed the following from previous day
room. with Tartan. morning. are filled.
What are the data sources? How will the data be collected?
Data Sources include the Linen Distribution Program currently in place, as
well as national benchmark data.
Step 7 Step 6 Step 5
Secondary Linen carts are Linen re-stock
deliveries are made exchanged for amounts are
What is our goal? to units as required those already on recorded in Textile
at 12 hour mark. Nursing Units. tracking program.
To reduce the overall linen utilization to between 14 and 16 pounds per
patient day.
Measure
Graphical Analysis
What is a defect, unit, opportunity?

Defects= Missed Delivery and Stock Outs, and any reading <14 or >18 lbs
per patient day
Units = Pounds per Patient Day
Opportunity = monthly data per unit
What are the specification limits? (LSL, USL)

LSL= 14 Pounds per Patient Day Average
USL= 18 Pounds per Patient Day Average
What X’s (inputs) are causing most of our variation?

Usage variations, training, old behaviors.
121
1/1/2018
Graphical Analysis Control Charts
Achieved goal of 14 Pounds per Patient Day. Education and focus on Scrubs,
and ancillary usage will contribute to maintaining this goal.
Control Case Study VI: Supply Chain Improvement
What are our financial results? How were they calculated? • Four hospital system enjoying 50% market
share
Our Per Patient Day costs for linen have decreased by 20% over 2002. From
an average of 20lbs to an average of 16lbs. • Materials management improvements needed
to leverage economies of scale, utilize best
practices, and prevent inefficiencies:
• Pricing structure for orthopedic implants
What is the WWW (Who-What-When) plan for turning the project over to the highly variable
process owner? What is the plan for monitoring/auditing the process?
• Inconsistent orthopedic implant utilization Barry D. Brown Health Education
The process is a permanent one and will be tracked through reports given to • Deficiencies in OR charge master capture Center at Virtua West Jersey Hospital
Voorhees
the units, Executive Sponsor, and the Linen Utilization Committee. • Gap in OR supplies between what patient
pays vs. what hospital is charged
The Linen Utilization Committee will oversee the process and progress. • OR “on hand” inventory management
needed
Process Improvement to Reduce Cost
122
1/1/2018
Reduce Costs Improve Quality
Solutions… Results
• Orthopedic Implant Pricing Cap… Determined • Project results along with data shifted purchases to a primary orthopedic
actual versus lowest and average prices to establish implant vendor, savings of $159,000 were attained.
a fair cap price.
• Annual savings of $239,400 through demand matching template at all
• Orthopedic Implant Demand Matching… hospital sites that do hip and knee replacement surgery.
Examined 132 medical records and compared
implants used against widely accepted industry • Patient billing data review indicated potential loss of greater than $200,000
criteria for implant selection by orthopedist annually due to missing charges, much of which was rectified with the
• Charge Master Review… Reviewed OR charge corrections in the current charge masters.
master systems and identified opportunities for • Project savings attained totaled $63,845 plus shared savings with
improvement and standardization orthopedic cap project.
• Price Point Reduction… Identified price reduction • Conservative inventory reduction by facility: Facility A $187k, Facility B
opportunities $92k, Facility C $47k, and Facility D $18k. Represents an 8% reduction of
• OR Inventory Reduction… HISI contracted to the $4.1MM of baseline inventory on hand.
conduct physical inventories in four ORs and two
surgical centers Sustainable Results With Bottom Line Impact
Case Study: Redefining the Emergency Room Process Project Goals/Deliverables

Project Goal
Project Definition 1) Improve the time from presentation to ED with an “surgical abdomen” to
diagnosis of appendicitis.
2) Early interventions of surgeons while patients are still in ED.
• Part of the overall redesign of the Emergency Department
Key Deliverables
flow 1) Incision time within 6 hours of registration for patients with appendicitis.
2) Decreased average cost/case.
3) Appropriate use and duration of antibiotics.
• Identify clinical problem areas 4) Decrease LOS for diagnosis of appendectomy from 4 to 2 days
Financial and Operational Benefits

• Define Clinical protocols
Improve clinical outcome (ie: less perforations, complications).
Reduce average cost/case by 20% which equals an annual savings of $118,000.
• Apply Six Sigma There is reproducibility for abdominal pain in general, which would yield a
greater savings
www.vahimss.org www.vahimss.org
123
1/1/2018
Project Definition Data Collection/Analysis

Problem Statement
Retrospective chart review performed
Cases presenting to ED with acute abdominal pain
experience a delay in diagnosis of appendicitis resulting Minitab software program utilized for data analysis
in over use of services, misdiagnosis and a delay in
surgical intervention
Process analyzed by multidisciplinary team
Project Scope Protocols developed by team
Patients presenting to ED with ultimate diagnosis of

appendicitis
Descriptive Statistics Descriptive Statistics (cont’d)
124
1/1/2018
Number of Hours to Surgery did not Effect LOS Perforations Did Affect LOS
Solutions Identified Solution Implementation
Standardize protocol for surgeon involvement in case Implement new Abdominal pain protocol that involves
Standardize lab/radiology utilization per case standardized testing and notification of surgeon.
Implement new triage process to see patients sooner Incorporate protocol into new triage process for the
Emergency room
125
1/1/2018
Process Control Plan Results
Process Controls Response Plan
•100% chart audit of all appendectomies

•Clinical indicators (LOS, time, antibiotic •Variance from accepted results to be
use, etc.) monitored and reported out to reported out at team/department level.
the team and PI committee. •Individual physicians informed of outliers
and counseled.
•Results presented and recommendations
Training Plan made at Care Management committee.
All medical directors and medical

leadership informed through 1 on 1
conversations. Medical staff informed
through mailings. Nursing/staff educated
at the department level on an ongoing
basis.
Results (cont’d) Project Benefits
Financial (Hard) Benefits Soft Benefits
•23% reduction of average •Reduction of time from

costs ($153,000 annualized registration to incision from 21
savings) hours to 10 hours (48%
•Average LOS decrease by reduction)
25% (4.5 to 2.7 days) •Standardized protocol
developed for evaluation of
abdominal pain
•These 2 factors combined,
translate into better clinical
quality
126
1/1/2018
ED Process—From Arrival to Seen by Physician New Triage System
Begin Quick Registration Process

Medical Doctor is the 1st person a patient sees
MD, RN and PCA assess patient together in
triage area
Ability to discharge non-emergent patients
from triage
Bedside registration
Registration is now back end of process
Further Successes
• New Process minimizes wait times to treatment

• Registrations completed at bedside
• Better Service to Patients & ED Staff
Thank You!
• Eliminated Initial Waiting
• EMTALA Compliant Process
• Improved Patient Satisfaction Scores
• Number of Patients that Left Without Being Seen
(LWBS) has decreased from 85+/month to
<10/month
• Increased volume 10-12%
www.vahimss.org
127
1/1/2018
Contact Information Appendix – d2 Constant (Table 1)
Dr. Mohammad T. Khasawneh, Ph.D.

Professor & Chair, Systems Science and Industrial Engineering
Associate Director, Watson Institute for Systems Excellence
Director, Healthcare Systems Engineering Center
Graduate Program Director, Executive Master of Science in Health Systems
Thomas J. Watson School of Engineering and Applied Science

State University of New York at Binghamton
Binghamton, New York 13902, U.S.A.
Phone: +1-607-777-4408; Fax: +1-607-777-4094

Email: [email protected]
http://www.ws.binghamton.edu/mkhasawn
Appendix – Control Chart Constants (Table 2)
128
1/1/2018
129
1/1/2018
130
1/1/2018
131
1/1/2018
Thank You!
132

LSSGB SVES January 2018 Final PDF

Uploaded by

Copyright:

Available Formats

LSSGB SVES January 2018 Final PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LSSGB SVES January 2018 Final PDF

Uploaded by

Copyright:

Available Formats

1/1/2018

Excellence Through Innovative Research

All rights reserved. No part of the course materials

Overall Agenda Meet The Instructor

Binghamton University 3 Binghamton University 4

Instructor at a Glance Objectives

Binghamton University 5 Binghamton University 6

Objectives (cont’d) Course Outline

Course Outline (cont’d) Course Outline (cont’d)

• Analyze Phase: Correlation analysis; regression analysis;

‘A comprehensive management philosophy

Tindill & Stewart (1993)

Binghamton University 11 Binghamton University 12

Knowledge Hierarchy Kaizen

Binghamton University 13 Binghamton University 14

Kaizen (cont’d) CI Gurus/ Pioneers

. McLaughlin & Kaluzny (2004)

Binghamton University 15 Binghamton University 16

W. Edwards Deming Joseph M. Juran

McLaughlin & Hays (2008) McLauhlin & Hays (2008)

Binghamton University 17 Binghamton University 18

CI – PDSA Cycle CI – Quality Circles

Langley et al. (1994)

Binghamton University 19 Binghamton University 20

Important Events in Performance Improvement

McLaughlin & Hays (2008)

Binghamton University 21 Binghamton University 22

Appearance of Lean In Healthcare Introduction to Lean

Value stream mapping

Binghamton University 23 Binghamton University 24

What’s Lean? Value Added vs. Non-value Added

Binghamton University 25 Binghamton University 26

Seven Types of Wastes (Muda)

Binghamton University 27 Binghamton University 28

Tools & Techniques Essential Flowcharting or Imagineering Symbols

• Brainstorming • Pareto Principles &

• Cause & Effect • Control Charting

Binghamton University 29 Binghamton University 30

Basic Symbols in Process Mapping Other Essential Flowcharting Symbols

 Activity: a task or activity performed in the process

 Direction: flow arrows show the order in which activities

 Decision: points in the process where a yes/no question is Yes

 Inventory: can also indicate customer waiting

Binghamton University 31 Binghamton University 32

Types of Process Maps

 Detailed flow chart

Systems Science and Industrial Engineering

A Short Video – Quality A Short Video – Systems Approach

Pareto Principle Pareto Charts

 Pareto Principle: 80 / 20 Rule

 20% Items Cause 80% Problems

De Mast & Trip (2007)

Binghamton University 37 Binghamton University 38

Hammond et al. (1998)

Binghamton University 39 Binghamton University 40

Error Type Frequency

Binghamton University 41 Binghamton University 42

Example 2 – Late Arrival to Work Cause and Effect Diagram (Ishikawa)

 Ishikawa developed with Deming to study

Value stream mapping

Activity: a task or activity performed in the process

Direction: flow arrows show the order in which activities

Decision: points in the process where a yes/no question is Yes

Inventory: can also indicate customer waiting

Detailed flow chart

Pareto Principle: 80 / 20 Rule

20% Items Cause 80% Problems

Ishikawa developed with Deming to study

4Ps vs. 4Ms

Available recourses, supplies, equipment, space, etc.

Labor of key providers (depends on industry/service

Availability and uptime of key technologies

Supplies and other resources

Facility, inventory, services

Design (or Theoretical) Capacity:

Effective (or Actual) Capacity:

Time for an item to complete the entire process, which

Six Sigma Introduction

Lean Six Sigma

Six Sigma (SS) aims at reduction and removal of variation Target

Variation means that a process does not produce the same

Stable Operations: Ensuring consistent, predictable processes to

Lean Sigma combines Lean and Six Sigma methodologies

Lean and Six Sigma