The use of Six Sigma
on the Pharmaceutical Industry
Orlando Lozada
Lean-Six Sigma Black Belt
Pfizer Puerto Rico
Pyramid of High Performance Business
Customer
VOC
(Defines Value)
Products
(Service / Goods)
(Business)
Suppliers – Inputs(X) –Process – Outputs (Y)– CTQ
Business needs to focus on CTQ
CTQ
(Defect Control & Reduction)
Six Sigma Tools
DMAIC
DMAIC:
Universal Problem -Solving Methodology
5 Logically Linked Steps
Define
Measure
A l
Analyze
Improve
Control
DMAIC Objective
Practical (Problem / Opportunity)
Statistical Problem (6-sigma tools)
Statistical Solution (6-sigma tools)
Practical Solution / Improvement
Define Measure Analyze Improve Control
1. The Define Phase
Project Charter
Define SIPOC Analysis
Voice of the Customer
Data Collection & Operational Def.
Measure Data Measurement Tools: Funneling, Sampling, Minitab, Gage
R&R, Patterns, Stratification, Process Capability
Data Analysis Tools: Cause & Effect Diagrams,
Stratification, Hypothesis Testing, Regression Analysis,
Design of Experiments
Analyze
Generating, Evaluating, & Selecting Solutions, FMEA,
Pil t IImplementation
Pilots, l t ti Pl
Planning
i
Improve
C
Control
t l Plan,
Pl Standardization,
St d di ti M
Monitoring,
it i K
Key
Control Learning’s, Project Closure
R&S
Define Measure Analyze Improve Control
2. Measure Phase
Project Charter
Define SIPOC Analysis
Voice of the Customer
Data Collection & Operational Def.
Measure Data Measurement Tools: Funneling, Sampling, Minitab,
Gage R&R, Patterns, Stratification, Process Capability
Data Analysis Tools: Cause & Effect Diagrams,
Stratification, Hypothesis Testing, Regression Analysis,
Design of Experiments
Analyze
Generating, Evaluating, & Selecting Solutions, FMEA,
Pil t IImplementation
Pilots, l t ti Pl
Planning
i
Improve
C
Control
t l Plan,
Pl Standardization,
St d di ti M
Monitoring,
it i K
Key
Control Learning’s, Project Closure
R&S
Define Measure Analyze Improve Control
Data Collection Plan
What is the data trend?
Does the data indicate any particular distribution?
Is there any outlier in the available data set?
Data Operational Definitions & Procedures
Measure Type,
Data Sampling
What Unit of How Measured Related Conditions
Type
yp Notes
Measure
Define Measure Analyze Improve Control
I Cha r t o f M e a s ur e me nt v s da t e
C a psule F illing Da te : J uly t o De ce mbe r 2 0 0 8
1
UC L= 0.04514
0.04
Value
0.03
Individual V
_
0.02 X = 0.02071
0.01
0.00
LC L= -0.00373
Jul-08
Oct-08
Oct-08
Oct-08
Nov-08
Nov-08
Nov-08
Dec-08
Dec-08
Dec-08
M fg . M o n th
0 03
0.03 UC L= 0.03002
0 03002
ving Range
0.02
__
0 01
0.01
Mov
M R = 0.00919
0.00 LC L= 0
Jul-08
Oct-08
Oct-08
Oct-08
Nov-08
Nov-08
Nov-08
Dec-08
Dec-08
Dec-08
M fg . M o n th
Define Measure Analyze Improve Control
Gage R&R Analysis
Components of Variation Gage Weight (KG) by GAGE Drum No.
100 % Contribution
56.7
% Study Var
ent
% Tolerance
56 6
56.6
Perce
50
56.5
0
Gage R&R Repeat Reprod Part-to-Part 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
GAGE Drum No.
R Chart by Gage Operator
Angel Burgos Cornelius Williams Elliot Díaz Gage Weight (KG) by Gage Operator
0.10 56.7
Sample Range
0.05 56.6
UCL=0.0343
_
R=0.0133 56.5
0.00 LCL=0
Angel Burgos Cornelius Williams Elliot Díaz
Gage Operator
Xbar Chart by Gage Operator
Angel Burgos Cornelius Williams Elliot Díaz Gage Operator * GAGE Drum No. Interaction
56.70 56.70 Gage Operator
e Mean
Angel Burgos
56.65 _
UCL=56.6499 erage 56.65 Cornelius Williams
Elliot Díaz
X=56.6363
X 56.6363
Sample
Ave
LCL=56.6227
56.60
56.60
1 2 3 4 5 6 7 8 9 10 11 12 13 1 4 15
GAGE Drum No.
Define Measure Analyze Improve Control
Stratified Time Series Plot
Time Series Plot of % Yield
90 Day
Fri
Mon
Thu
85 Tue
Wed
% Yield
d
80
75
70
/05 /05 /05 / 05 / 05 / 05 /05 /05 /05 /05 /06
31 3/
4 /8 13 17 22 26 30 /4 /9 13
1/ 4 5/ 6/ 7/ 8/ 9/ 11 12 1/
Date Produced
Define Measure Analyze Improve Control
Process Capability Report
Using Box-Cox Transformation With Lambda = 0.45 Capsule Filling Date: July to December 2008
I Chart Capability Histogram
0.3 LSL*
UCL=0.2677 Specifications
Individual Value
0.2 _ LSL* 0.127211
X=0.1722
0.1
LCL=0.0766
1
1 20 39 58 77 96 115 134 153 172 0.06 0.09 0.12 0.15 0.18 0.21 0.24
Moving Range Chart Normal Prob Plot
1
0.16 1 AD: 0.411, P: 0.339
Moving Range
UCL=0.1174
0 08
0.08
__
MR=0.0359
0.00 LCL=0
1 20 39 58 77 96 115 134 153 172 0.1 0.2 0.3
Last
ast 25
5 Obse
Observations
at o s Transformed
a s o ed Capa Plot
ot
0.25 Within
Within Overall
StDev 0.0318476 StDev 0.0340961
Values
0.20
Cp * Pp *
O v erall
Cpk 0.47 Ppk 0.44
0.15
Cpm *
S pecs
165 170 175 180 185
Observation
Define Measure Analyze Improve Control
3. Analyze Phase
Project Charter
Define SIPOC Analysis
Voice of the Customer
Data Collection & Operational Def.
Measure Data Measurement Tools: Funneling, Sampling, Minitab, Gage
R&R, Patterns, Stratification, Process Capability
Data Analysis Tools: Cause & Effect Diagrams,
Stratification, Hypothesis Testing, Regression
Analysis, Design of Experiments
Analyze
Generating, Evaluating, & Selecting Solutions, FMEA,
Pil t IImplementation
Pilots, l t ti Pl
Planning
i
Improve
C
Control
t l Plan,
Pl Standardization,
St d di ti M
Monitoring,
it i K
Key
Control Learning’s, Project Closure
R&S
Define Measure Analyze Improve Control
Cause & Effect Diagram
Methods Material
Problem
/Opportunity
Machine Manpower Environment
R&S
Define Measure Analyze Improve Control
Regression Analysis
Define Measure Analyze Improve Control
Experimental Design Results
Surface plot of Response vs.
vs Two Variables
Design-Expert® Software
Transformed Scale
Sqrt(Response)
5.2915
0
3.1
X1 = B: LOD
X2 = D: Vacuum
2.3
A t l Factors
Actual F t
se
Respons
A: Thickness = 0.080 1.5
C: PH = 5.15
0.7
-0.1
350 16.000
312.5 15.325
275 14 650
14.650
237.5 13.975
D: Vacuum B: LOD
200 13.300
Define Measure Analyze Improve Control
Hypothesis
yp testing
g to demonstrate
significance of Change
Two-sample T for Transformed data
Period N Mean StDev SE Mean
Before June_09 22 0.536 0.830 0.18
Post June_09
_ 12 -0.655 0.939 0.27
Difference = mu (Before June_09) - mu (Post June_09)
Estimate for difference: 1.191
1 191
95% CI for difference: (0.516, 1.866)
T-Test of difference = 0 (vs not =): T-Value = 3.68 P-Value = 0.001 DF = 20
Define Measure Analyze Improve Control
4. Improve Phase
Project Charter
Define SIPOC Analysis
Voice of the Customer
Data Collection & Operational Def.
Measure Data Measurement Tools: Funneling, Sampling, Minitab, Gage
R&R, Patterns, Stratification, Process Capability
Data Analysis Tools: Cause & Effect Diagrams,
Stratification, Hypothesis Testing, Regression
Analysis, Design of Experiments
Analyze
Generating, Evaluating, & Selecting Solutions,
FMEA Pilots,
FMEA, Pil t IImplementation
l t ti Planning
Pl i
Improve
Control
C t l Plan,
Pl Standardization,
St d di ti M
Monitoring,
it i K
Key
Control Learning’s, Project Closure
R&S
Define Measure Analyze Improve Control
FMEA Assessment for Identified
Solutions
S O D R
Key Process Step or Potential Failure Potential Failure
E Potential Causes C Current Controls E P Actions Recommended
Input Mode Effects
V C T N
What is the Process In what ways can the What is the impact on How What causes the Key How often What are the existing controls and How well can Risk What are the actions for
Step or Input? Process Step or Input the Key Output Severe is Input to go wrong? does cause procedures that prevent either the Cause you detect the Priority reducing the occurrence of
fail? Variables once it fails the effect to or FM or the Failure Mode? Cause or the Number the cause, or improving
(customer or internal th
the occur? ? 11- M d ? (SEV x
F il Mode?
Failure detection?
requirements)? customer? highly 1-we have OCC x
(1-not unlikely to excellent DET)
severe, 10 ever occur,10- controls,10-we
extremely we expect it have no controls
severe) to happen all or extremely
the time) weak controls
Define Measure Analyze Improve Control
Pilots
– Find flaws in the solution
– Improve the solution before full-scale
i l
implementation
t ti
– Find
Fi d out if you are getting
i the
h results
l you
expected.
Define Measure Analyze Improve Control
5. Control Phase
Project Charter
Define SIPOC Analysis
Voice of the Customer
Data Collection & Operational Def.
Measure Data Measurement Tools: Funneling, Sampling, Minitab, Gage
R&R, Patterns, Stratification, Process Capability
Data Analysis Tools: Cause & Effect Diagrams,
Stratification, Hypothesis Testing, Regression
Analysis, Design of Experiments
Analyze
Generating, Evaluating, & Selecting Solutions, FMEA,
Pil t IImplementation
Pilots, l t ti Pl
Planning
i
Improve
Control
C t l Plan,
Pl Standardization,
St d di ti Monitoring,
M it i Key
K
Control Learning's, Project Closure
R&S
Define Measure Analyze Improve Control
Control Phase Results
Before Implementation: Non-stable process
g
1 1
1
U C L=9.129
9
_
_
Sample M ean
X=8.481
8
LC L=7.833
7
1
Aft IImplementation:
After l t ti St
Stable
bl PProcess
1
1 3 5 7 9 11 13 15 17 19 21
Sample
4
U C L=3.736 Xbar-R Chart After Implementation Control Phase
3 10547C 10557C 07847C
mple Range
8.5
_ U C L=8.339
2 R=2.102
ean
8.0
Sample M e
_
Sam
_
1 X=7.698
LC L=0.469 7.5
0
1 3 5 7 9 11 13 15 17 19 21 LC L=7.057
7.0
Sample
1 5 9 13 17 21 25 29 33 37
Sample
10547C 10557C 07847C
2.4 U C L=2.349
Sample Range
1.8
_
1.2 R=1.111
0.6
0.0 LC L=0
1 5 9 13 17 21 25 29 33 37
Sample
Six Sigma Projects Related to:
• Complaint Investigations
• D i ti IInvestigations
Deviation ti ti
• Environmental Projects
• Laboratory Data / Specifications / Stability
• Lead Time Optimization
• Process Optimization
• Product Optimization
• Safety Assessments & Investigations
Summary
• 6-Sigma tools are embraced by a 5 logically
linked steps, DMAIC, which enables tools
understanding and proper use
• Each year hundreds of projects are
completed using 6-sigma tools
• These projects add significant benefits to the
business performance; related to quality,
safety, process performance and other
End
