See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321977835

Continued Process Verification Statistical Tools for Process Control

and Improvement A Systems Approach

Conference Paper · December 2017

CITATIONS READS

0 10,700

1 author:

Ron Snee
Snee Associates, LLC
432 PUBLICATIONS   8,115 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Statistical Leadership View project

Holistic Improvement View project

All content following this page was uploaded by Ron Snee on 21 December 2017.

The user has requested enhancement of the downloaded file.

 Continued Process Verification
Statistical Tools for Process Control and Improvement
A Systems Approach

Ronald D. Snee, PhD

Snee Associates, LLC
Newark, DE
Statistics in Validation 2017
Institute of Validation Technology
June 20-22, 2017
1 Snee Associates, LLC
 About the Speaker …..
He is also an Adjunct Professor in the
pharmaceutical programs at Temple and Rutgers
Universities. He worked at DuPont for 24 years prior
to starting his consulting career.
Ron received his BA from Washington and
Jefferson College and MS and PhD degrees from
Ron Snee, PhD is Founder of Snee Associates, Rutgers University. He is an academician in the
LLC, a firm dedicated to the successful International Academy for Quality and Fellow of the
implementation of process and organizational American Society of Quality, American Statistical
improvement initiatives. He provides guidance to Association, and American Association for the
pharmaceutical and biotech senior executives in Advancement of Science. He has been awarded
their pursuit of improved business performance ASQ’s Shewhart and Grant Medals, and ASA’s
using Quality by Design, Lean Six Sigma and Deming Lecture, Dixon Consulting and Gerry Hahn
other improvement approaches that produce Quality Achievement Awards. He is a frequent
bottom line results. speaker and has published 6 books and more than
300 papers in the fields of quality, performance
He has authored several articles on how to
improvement, management, and statistics. He
successfully implement QbD, coauthored 3 books
recently received the Institute of Validation
on QbD tools and speaks regularly at
Technology’s Speaker of the Year Award.
pharmaceutical and biotech conferences.
2 Snee Associates, LLC
 Abstract
Stage 3 of the Process Validation Guidance calls for “Continued
Process Verification”. Process monitoring provides the tools for
enabling Continued Process Verification and dealing with
variation. Managing and reducing variation is at the heart of the
FDA Process Validation Guidance. Clearly stated in the guidance
is the need to understand variation, ability to detect variation,
understand the impact of variation and ability to control variation.
This session addresses the concepts, methods and tools used in
continued process verification. It is shown how the tools can be
linked and integrated to create a system that enhances the use of
the tools and the effectiveness of Continued Process Verification.
The issues, tools and systems are illustrated with pharmaceutical
case studies and examples.

3 Snee Associates, LLC

 Agenda

 Today’s Realities – Need to Control and Reduce Variation

 Role of Continuous Process Verification in QbD
 Importance of Process Variation
 Statistical Tools for Stage 3
 Improving Process Stability - Reducing Process Risk
 Test Method Control
 Process Capability
 Continued Process Verification - A Systems Approach
 Approach for Improving Process Stability
 Tips and Traps - What to Watch Out for
 Summary

4 Snee Associates, LLC

Importance of Variation - FDA Process Validation
“A successful validation program depends upon information and
knowledge from product and process development. This knowledge
and understanding is the basis for establishing an approach to
control of the manufacturing process that results in products with the
desired quality attributes. Manufacturers should:
• Understand the sources of variation
• Detect the presence and degree of variation
• Understand the impact of variation on the process and
ultimately on product attributes
• Control the variation in a manner commensurate with the risk
it represents to the process and product
Each manufacturer should judge whether it has gained sufficient
understanding to provide a high degree of assurance in its
manufacturing process to justify commercial distribution of the
product.”
5 Snee Associates, LLC
 Variation Drives Risk,
Quality, Cost and Customer Satisfaction

Risk Quality

Variation Variation

Costs Customer
($$) Satisfaction

Variation Variation

Understanding and Reducing Variation is a Good Thing

6 Snee Associates, LLC
 Building Blocks for Quality by Design
Linked and Sequenced
Process
Capability
Process
Control
Process
Critical Quality Model
Attributes (Ys) Y=f(X) Design
Space Risk
Critical Process Process Level
Parameters (Xs) Robustness

Raw
Materials
(Xs)
Failure Modes and Effects Analysis

“Process” = Process and Analytical

7 Snee Associates, LLC
Contour Plots of Dissolution and Friability as a
Function of Process Parameters 1 and 2

Friability
Spec < 2%

Dissolution
Spec > 80%

Adapted from ICH GUIDELINE PHARMACEUTICAL DEVELOPMENTQ8(R2), August 2009

8 Snee Associates, LLC

 Design Space – Product Meets Specifications
Comprised of the Overlap Region of Spec Ranges
for Friability and Dissolution

Friability
Spec < 2%

Design
Space

How Do We
Dissolution Keep the Process
Spec > 80% On-Target?

Adapted from ICH GUIDELINE PHARMACEUTICAL DEVELOPMENTQ8(R2), August 2009

9 Snee Associates, LLC

FDA Process Validation Guidance (Jan 2011)
“Process validation involves a series of activities taking place
over the lifecycle of the product and process”
“Stage 1 – Process Design: The commercial manufacturing
process is defined during this stage based on knowledge
gained through development and scale-up activities.
Stage 2 – Process Qualification: During this stage, the
process design is evaluated to determine if the process is
capable of reproducible commercial manufacturing.
Stage 3 – Continued Process Verification: Ongoing assurance
is gained through routine production that the process
remains in a state of control.”

How Do We Do Continued Process Verification?

10 Snee Associates, LLC

Building Blocks of Quality by Design

Life Cycle Management and Continuous Improvement

Continued Process Verification
11 Snee Associates, LLC
 Reduce Process Risk
Improve Process Stability and Capability

Control Charts
Process Stability Metrics
Process Variance Components
Measurement Stability and Control Methods
Process Capability Indices

12 Snee Associates, LLC

 Stable Manufacturing Process
Upper Control Limit (UCL)

Process Variable
Average

Lower Control Limit (LCL)

Time or Sequence

 Process that is in a state of statistical control as

 Each batch of tablets is being produced
 Batches of tablets are produced over time.
 A process in a state of statistical control consistently
produces product that varies within the process
control limits; typically
 Upper Limit: Process Average + 3(Process Std Dev)
 Lower Limit: Process Average – 3(Process Std Dev).

13 Snee Associates, LLC

 Capable Process

 Consistently produces
tablets that are within
specifications.
 Process capability
analysis compares the
process variation to the
lower and upper
specification limits
 Broadly used measure of
process capability is the
Ppk index
 Measures what the
customer experiences

14 Snee Associates, LLC

 Common Causes of Variation
If only common causes of variation are present, the
process output is stable over time and is predictable
Common Causes:
 Present all the time
 Influences all process outputs
 Require process changes to reduce
 Minimum possible process variation
Example:
Variation Between
Back-to-Back Samples
Taken from a Vial
Production Process

15 Snee Associates, LLC

 Special Causes of Variation
If special causes of variation are present, the process
output is not stable and is not predictable

Special Causes:
 Due to outside influences
 Affect some of the process outputs
 Can cause data to form non-normal patterns
Examples
 Ambient Temperature
 Raw Material Lot
 Equipment

16 Snee Associates, LLC

 Shewhart Control Chart

Special Cause
Upper Variation
Control Limit (UCL)
Process Variable

Process is:
Common Cause Variation
Average
 Not Stable
Process is Stable and Predictable
 Not Predictable
Lower Control Limit (LCL)
Special Cause Variation

Time or Sequence

Control Limits Are Not Specification Limits!

17 Snee Associates, LLC

 Product A – Tablet Thickness
Unstable Process
Time Series Plot of AVGThickness
Presses 50, 90, 100
0.228

0.226
AVGThickness

0.224

0.222 Press 120

0.220
4 Tablet
0.218
Presses Used
1 6 12 18 24 30 36 42 48 54 60
Index

18 Snee Associates, LLC

Tablet Weight – Stable Process – Two Presses

I Chart of Tablet Weight by Press

Tablet Press A Tablet Press B
1.08

UCL=1.07360
1.07
Individual Value

1.06
_
X=1.05489
1.05

1.04
LCL=1.03619

1.03
1 19 37 55 73 91 109 127 145 163 181
Sample of 10 Tablets

19 Snee Associates, LLC

 Assay – Process Unstable
No Out-of- Spec Product Produced
I Chart of Assay% by Year
Year 1 Year 2 Year 3
102
UCL=101.818

101

100
Individual Value

99 _
X=98.682
98

97

96
LCL=95.546
95
1 14 27 40 53 66 79 92 105 118
Batch

20 Snee Associates, LLC

 Assessing Process Stability
 Q: When Should I worry about process stability?
 A: When Long-Term process variation is too high
 Q: How high is too high?
 A: Long-Term variation becomes a concern when it represents more
than 20% of the total variation
 Total Variation = Long-Term Variation + Short-Term Variation
Process Stability Long-Term Variation
Not a Problem < 20%
May be A Problem 20 – 30%
Corrective Action Needed? > 30%

 This guideline is based on the assumption that

 Well-controlled process will detect a shift of 1.5 short-term
standard deviations

Ref: Snee and Hoerl, Quality Progress, May 2012, 39-41

21 Snee Associates, LLC

Schematic - Between and Within Variation
3 Batches, 2 Samples per Batch

Within-Batch
Short-Term
Variation

Between-Batch
Long-Term
Variation

Batch
Mean

22 Snee Associates, LLC

 Batch Group No. Quality Batch Group No. Quality Batch Group No. Quality
1 1 4.36 10 2 4.24 27 3 4.34
2 1 4.63 11 2 3.96 28 3 4.27
3 1 4.51 12 2 4.20 29 3 4.46
4 1 4.47 13 2 4.13 30 3 4.46
5 1 4.34 14 2 3.82 31 3 4.34
6 1 4.75 15 2 3.97 32 3 4.16
7 1 4.46 16 2 4.04 33 3 4.32
8 1 4.17 17 2 4.07 34 3 4.33
9 1 4.58 18 2 4.04 35 3 4.31
19 2 3.93 36 3 4.49
20 2 4.08 37 3 4.52
21 2 4.05 38 3 4.36
22 2 4.14 39 3 4.53
23 2 3.94 40 3 4.41
24 2 4.02 41 3 4.34
Example 25 2 4.29 42 3 4.35
26 2 4.05 43 3 4.53
Variation in Quality 44 3 4.32
51 Batches 45 3 4.47
46 3 4.26
3 Groups 47 3 4.23
48 3 4.35
49 3 4.31
50 3 4.21
51 3 4.21

23 Snee Associates, LLC

 Example – Variation in Quality
Process Not Stable – Long-Term Variation Large

24 Snee Associates, LLC

 Example – Three Groups
Short-Term and Long-Term Variation
Short-Term: Within Groups
Long-Term: Between Groups
LT Variation = 73% (p=.000)

Group 1
Group 3
Group 2

Group Mean Std Dev

1 4.5 .17
2 4.1 .12
3 4.4 .11
All Data 4.3 .21

25 Snee Associates, LLC

 Example – Variation in Quality
Nested ANOVA: Quality versus Group
Analysis of Variance for Quality Percent of Total Variation Due to Between Groups and Within Groups
90

Source DF SS MS 80
73.1
70

Between Group 2 1.3261 0.6630 60

50

Percent
Within Group 48 0.7276 0.0152 40

30 26.9
Total 50 2.0537 20

10

Variance Components Between Groups (L-T) Within Groups (S-T)

Long-Term Short-Term
Source of Variation

% of Variation Variation
Source Var Comp. Total StDev
Between Group 0.041 73.08 0.203
Long-Term
Within Group 0.015 26.92 0.123
Variation (%)
Total 0.056 100.00 0.237

26 Snee Associates, LLC

Tablet Weight – Stable Process – Two Presses
Long-Term (LT) Variation Analysis
I Chart of Tablet Weight by Press
Tablet Press A Tablet Press B
1.08
LT Variation = 18% LT Variation = 44%
(p=0.112) (p=.001) UCL=1.07360
1.07
Individual Value

1.06
_
X=1.05489
1.05

1.04
LCL=1.03619

1.03
1 19 37 55 73 91 109 127 145 163 181
Sample of 10 Tablets

27 Snee Associates, LLC

 Assay – Process Unstable
No Out-of- Spec Product Produced
I Chart of Assay% by Year
Year 1 Year 2 Year 3
102
UCL=101.818

101

100
Individual Value

99 _
X=98.682
98
LT Variation LT Variation
97 6% 23%
(p=0.41) (p=0.14)
96
LT Variation LCL=95.546
95 42%
1 14 27 40 53 66 79 92 (p=0.000)
105 118
Batch

28 Snee Associates, LLC

 Process Stability Improvement
Where Should I Focus My Effort?

Process Stability Metric

 Provides a quantitative measure of process stability
 Enables management to prioritize which processes
should be worked on first to improve stability
 Enhances the effectiveness of a process control system

Reference:
Snee and Hoerl, “Going on Feel- Monitor and Improve Process Stability to
Make Customers Happy”, Quality Progress, May 2012, 39-41.

29 Snee Associates, LLC

Pharmaceutical Tablet Characteristics
Production Stability

Production Stability is
Poor for:
 Impurity
 Moisture
 Thickness

30 Snee Associates, LLC

 Detecting Non-Random Variation
Using Control Charts

1. Use Tests for Special Cause Variation

 Points Outside of Control Limits
 Trends: Increasing and Decreasing
 Level Shifts
 Cycles
 Uses Variation Zones
 Zone C: Average +/- 1 Standard Deviation
 Zone B: Average +/- 2 Standard Deviation
 Zone A: Average +/- 3 Standard Deviation

2. Trend Identification using Regression Analysis

31 Snee Associates, LLC

 Control
Chart Zones
Average

Rule 1: Points Rule 2: Two out of Three

Outside Lower and Consecutive Beyond 2SD on
Upper Control Limits Same Side of Centerline

32 Snee Associates, LLC

 Control Chart
Tests for
Special Causes
1. 1 point beyond Zone A
2. 9 points in a row in Zone C
or beyond
3. 6 points in a row increasing
or decreasing
4. 14 points in a row
alternating up and down
5. 2 out of 3 points in a row in
Zone A or beyond
6. 4 out of 5 points in a row in
Zone B or beyond
7. 15 points in a row in Zone
C, above or below
centerline
8. 8 points in a row on both
sides of the centerline with
33 Snee Associates, LLC
none in Zone C
 Product Assay – Year 2 – 47 Batches

Long-Term Variation = 23% (p = 0.138)

102
No Special CausesUCL=101.807
Detected
Possible Negative Trend
101
Assay%

_
100 X=99.951

99

LCL=98.095
98
1 6 11 16 21 26 31 36 41 46
Batch

34 Snee Associates, LLC

 Product Assay – Year 2 – 47 Batches
Trend Detection using Regression Analysis
Slope = -0.025 (p=0.000)
101.5 Total Change = 1.15%
AdjR-Square = 0.26
101.0

100.5

100.0
Assay%

99.5

99.0

98.5

98.0
0 10 20 30 40 50
Batch

35 Snee Associates, LLC

 Assay – Process Unstable
Trend in Year 2 Continues
I Chart of Assay% by Year
Year 1 Year 2 Year 3
102
UCL=101.818

101

100
Individual Value

99 _
X=98.682
98

97

96
LCL=95.546
95
1 14 27 40 53 66 79 92 105 118
Batch

36 Snee Associates, LLC

 Test Method Control

Reducing Analytical Risk

Maintaining Stable Analytical Methods
Using
Shewhart Control Charts

Enables Continued Measurement

Process Verification

37 Snee Associates, LLC

Use of Blind Controls (Reference Samples)
 Blind Control – Samples from a common source are
regularly submitted for analysis along with routine
production samples in a way that the analyst can not
determine the difference between the production samples
and the control samples
 “ The use of blind controls is rare in the pharmaceutical
industry. The roots of this probably lie in the compliance
aspects of the study. However there is no better way to
understand the true variability of the analytical method”

B. K Nunnally and J. S McConnell (2007)

Six Sigma in the Pharmaceutical Industry
Understanding, Reducing and Controlling Variation in Pharma and Biologics

38 Snee Associates, LLC

Measurement of Potency (%) on a QC
Control Material

Control Material Potency (%)

 84 Rows of Data
 42 QC Runs
 2 Measurements / Run

39 Snee Associates,VII-39
LLC
 Measurement of Potency (%) on a QC
Control Material (Continued)
Variables Control Chart
XBar of Potency
96.0
95.5
Mean of Potency

95.0 UC L=95.101

94.5
94.0
Av g=94.178  Out-of-Control Runs Detected
93.5
93.0
LCL=93.254  Process Is Not Stable
92.5
 Long-Term Variation = 58%
12
15
18
21
24
27
30
33
36
39
42
3
6
9

Run
Note: T he s ig ma wa s cal c ul ated us in g th e ra nge .
 Method Reproducibility is Poor
R of Potency

UC L=1.604

 Within-Run Variation in Control

1.5
Range of Potency

1.0

0.5 Av g=0.491
 Method Repeatability is Good
0.0 LCL=0.000
12
15
18
21
24
27
30
33
36
39
42
3
6
9

Run 40 Snee Associates, LLC

 Improving Measurement Methods

My measurement method is not adequate. What do I do?

 Improve Measurement Quality – Repeatability and Reproducibility
 Gage R&R Studies to assess measurement quality
 Improve method robustness (ruggedness)
 Repeatability and reproducibility are acceptable,
 Consider a test method robustness (ruggedness) evaluation
 Effects of small variations in the how the method is used.

References:
Schweitzer, etal (2010), Pharma Tech, February 2010, 52-59
Weitzel, Forbes and Snee (2015), Journal of Validation Technology, Vol. 20, Jan 2015

41 Snee Associates, LLC

 Measuring and Reducing
Process Risk Using
Process Capability Indices

• Process Capability
• Short-Term (Cp and Cpk)
• Process Performance
• Long-Term (Pp and Ppk)

42 Snee Associates, LLC

 Capability versus Performance
• Process Capability is the
variation the process would
exhibit if only common cause Process

variation were present: Process Performance

Capability
 “The variation in the
process if the angels ran
the process”
• Process Performance is the
total variation experienced
by the customer; includes
common cause, structural,
and special cause variation:
“The variation when we
mortals run the process”
43 Snee Associates, LLC
 Measuring Process Capability
Cp Index Compares Process Variation to Spec Limits

Tolerance

USL – LSL
Cp 
6 (Short-term Standard Deviation)

44 Snee Associates, LLC

 Cpk Index Compares Process Variation to
Difference Between Process Average and Specs

USL – Average
Cpk = = 1.33 Good!!
3 (Short-term Standard Deviation)

45 Snee Associates, LLC

 Cpk Index Compares Process Variation to
Difference Between Process Average and Specs

Average – LSL
Cpk = = 0.33 Poor
3 (Short-term Standard Deviation)

46 Snee Associates, LLC

 Traditional Performance Indices
Long-Term Variation – What Customers Experience

Two traditional measures of process performance:

USL – LSL
Pp =
6 (Long-term Standard Deviation)
Where:
LSL = Lower Specification Limit
USL = Upper Specification Limit

Min (USL – Average, Average – LSL)

Ppk =
3 (Long-term Standard Deviation)

47 Snee Associates, LLC

 Levels of Process Capability
Marginally Capable
Cp = 1.0
Poor
Capability
Cp < 1.0

Robust Process
Cp = 2.0

48 Snee Associates, LLC

 Process Stability and Capability
Appropriate Actions

Process Capable
Process
Stable Yes No
Use SPC to • Shift Average to Target,
Yes Maintain • Reduce Variation Around Average
Performance • Both
• Reduce Variation, • Increase Process Understanding
No • Improve Control • Identify Appropriate Action

49 Snee Associates, LLC

 How Large A Sample Do I Need?
• It depends on the process, objectives, etc.
• Practical Perspective – General Rule:
 Sample until the total range of process variation has been
observed
• Capability indices are highly variable when estimated from
small samples
• Statistical Perspective:
• Bare Minimum n>30-90
• Preferred: n>60-90
• Conservative Strategy: Report the lower confidence
limit on the capability indices
Example:
 Sample n=30, Ppk = 1.30; Conf. Limits: 1.02 – 1.76
 Report “Ppk = 1.02 (Lower 95% CL)”
50 Snee Associates, LLC
Capability Indices Are Highly Variable

True Sample Lower Upper

Ppk Size Ppk Ppk
1 30 0.76 1.31
1 60 0.83 1.21
1 120 0.88 1.14
1.33 30 1.02 1.76
1.33 60 1.11 1.61
1.33 120 1.17 1.52
1.67 30 1.29 2.19
1.67 60 1.43 2.01
1.67 120 1.47 1.90

Account for Sample Size by Reporting the

Lower Confidence Limit for the Capability Index
51 Snee Associates, LLC
Capability Index Simulation – An Example
Simulation
Data and Specifications Avg = 50, Std Dev = 1.25
Sample Sizes 20, 60 and 120

Capability Indices

Lower Confidence Limit

52 Snee Associates, LLC

Integrating the Statistical Tools into a
Systematic Approach for
Continued Process Verification

53 Snee Associates, LLC

Developing A Systematic Approach for
Process Control and Improvement

Guiding Principle
If you want something to happen on a regular and
sustained basis you need to have a management
system in place to guide and sustain the effort
Management System
Framework of processes and procedures used to
ensure that an organization can fulfill all tasks
required to achieve its objectives. (Wikipedia 2011)

“What Get’s Measured,

and Paid Attention To,
Gets Done”

54 Snee Associates, LLC

 Process Management System Use of Data
Process Control, Improvement and Design Integrated

Customers Process
Design/Redesign

Process Process
The Process Improvements
Performance
Data

Feedback Process Improvement

Adjustments Feedback
Projects

Reports & Periodic Process

Information to Analysis and Improvement
Management Reviews System

55 Snee Associates, LLC

Periodic Reviews of Process Performance
A Team Sport

Review Team Timing

Process Operators Daily/Hourly
Area Managers and Staff Weekly
Site Managers and Staff Monthly
Business Manager and Staff Quarterly

Periodic Reviews Are

Critical to Success

56 Snee Associates, LLC

 Production Monitoring System
Batch
Production
Over Time

Batch Process Process

Models
Quality Data Adjustment Y=f(X)

Periodic
Monitoring Tools Data Analysis Process
•Control Charts and Improvement
•Capability Analysis Review
•Variance
Components
•Histograms DMAIC
•Pareto Charts
57 Snee Associates, LLC
 DMAIC Process Improvement Framework

Sense of Statistical Thinking

and
Urgency Results ($$)
Tools

Data Control

Improve

Analyze
Leadership
Measure Teamwork
Stakeholder Building
Define Project Management
58 Snee Associates, LLC
 Many Process Problems are Due to
Special Cause Variation
Sources of Variation Outside the Process
If special causes of variation are present, the process
output is not stable and is not predictable

Special Causes:
 Due to outside influences
 Affect some of the process outputs
Examples
 Ambient Temperature
 Raw Material Lot
 Equipment

59 Snee Associates, LLC

Tracking Special Cause Variation to
Improve Process Stability (Predictability)
Special Causes – “The Usual Suspects”
 Raw material lot variation
 Differences between pieces of equipment
 Tablet presses
 Bioreactors
 Test instruments
 Sources of Human Intervention
 Operating teams
 Lab Analysts
 Production lines
 Day of Week
 Production Shift
60 Snee Associates, LLC
 BioPharm Process Yield Improvement
Daclizumab Fermentation Yield
Lot 1 Total Grams

Here we see a step 1100 UCL=1083

change in yield 1000 Lot 2
900
What is the cause?
800 Mean=799.9

Grams
700
Raw Material Lot
600
Variation LCL=516.5
500

Better Raw Material 400

Lot 3
Specs Adopted 300
200
Process Yield Subgroup 0 10 20 30 40 50 60
increased 25% A:Batch 62 72 84 94 104

Used DMAIC to Identify Key Drivers of Process Yield

Raw Material Lot Variation Had the Largest Effect

61 Snee Associates, LLC

 Tablet Content Variation Study
Variation Due to:
Time Series Plot of % ACTIVE INGREDIENT
5.3
 Raw Material Lot? No
 Tablet Press? No
5.2  Analyst?
% ACTIVE INGREDIENT

5.1

5.0

4.9

4.8

4.7

4.6
1 12 24 36 48 60 72 84 96 108
Batch

62 Snee Associates, LLC

Tablet Content Variation Study
Results Reported by Analysts A, B, C and D
Time Series Plot of % ACTIVE INGREDIENT
5.3
C
C

5.2
A
B
% ACTIVE INGREDIENT

A B
5.1 CC
B B B
A B B
A BA B B BB B B BBB BBB BB BB B B B
A A BB A B BBB B
5.0 B B BB BB B B B B B B B B B
AA B B B B D
B B C A B B
A B BB B
C B
A B A B B B A A D DD B
4.9 D
B D
A B B BB C
B B
4.8 A
A B CC
A A D
A
A
4.7
A
A
A
Work of
A A
Analyst B
4.6
1 12 24 36 48 60 72 84 96 108
Batch

63 Snee Associates, LLC

 Tablet Content Variation Study
Analyst B Has the Least Amount of Variation
Dotplot of % ACTIVE INGREDIENT vs ANALYST

A
ANALYST

Work of
Analyst B
B
C
D
4.68 4.77 4.86 4.95 5.04 5.13 5.22
% ACTIVE INGREDIENT

64 Snee Associates, LLC

 Chemical Production Process
Box Plot of Yield vs. Team
2200

2150
Yield

2100

2050

2000
A B C D
Team

Team D – Most Experienced Team - Has the Highest Yield

65 Snee Associates, LLC

On-Time Delivery vs Day of the Week

Best On-Time Delivery

Occurs on Fridays

66 Snee Associates, LLC

Special Cause Example - % Off-Spec Product

 Variables Showing Variation

 Shift (1, 2, 3)
 Press (1, 2, 3, 4)
 Pressure Level (1, 2, 3, 4)
 Response Y = % OffSpec
 Data for 72 consecutive Batches

67 Snee Associates, LLC

 % Off-Spec Product
Variation Due to Shift, Press and Pressure Level
Have Significant Effects

Process Improvement
• Shift 3 performs as Shifts 1 and 2
• Presses 1,2,3 perform as Press 4
• Hold Pressure at lowest level

68 Snee Associates, LLC

 % Off-Spec Product - Predicted Performance
Removing Special Causes Reduces Off-Spec Product by 40%
Process Improvement
• Shift 3 performs as Shifts 1 and 2
• Presses 1,2,3 perform as Press 4
• Hold Pressure at lowest level
30
Observed Performance

Predicted Process
20 Performance As Is Impvd
Avg. 14.3 8.8
OffSpec

StdDev 4.9 2.9

10 Process Average and

Process Std Dev
Both Reduced
Approximately 40%
0
Shift 50 100
Batch
69 Snee Associates, LLC
 Creating Stable Operations by
Reducing Special Cause Variation

A System is Needed
Should be Part of the Improvement Review Process

Special
Cause
Problems

70 Snee Associates, LLC

 System for Improving Process Stability
Process Problem List Operations
• Low Yield
• Temperature Deviations
• Equipment Malfunction
• Operator Errors
• SOP Not Followed
• Raw Material Issues

Pareto Analysis by
Process Type
# Deviations

Process to be
Improved

Type of Process

Improvement DMAIC
Implemented and Improvement
Confirmed Process
71 Snee Associates, LLC
 Tips and Traps
Operating Continued Process Verification Systems

 Process understanding - Deep knowledge of critical variables

 Assess process stability and capability at all process steps
 Raw materials, blend uniformity, in-process and finished product
 Magnitudes of the measurement variation are known.
 Repeatability, reproducibility and test method robustness
 Systematic method to keep track of special cause variation to
identify process stability problems – lack of predictability
 Data systems interfaces can be problem source
 CPV continues throughout the year with a summary report in
the Annual Product review
 When problems are identified, action needs to be taken.

Regular Management Reviews are Essential

72 Snee Associates, LLC

 Understanding and Reducing Variation
Tips and Traps
 Variation is all around us and present in everything we do …
……Learn to deal with it to your advantage

 Plot the data, plot the data, plot the data and ….. Think!
 Statistical methods and associated software enable us to deal
with variation …. To “tame” variation.
 Best Practice: Every data analysis includes use of process
information, analytical methods and graphical displays
 Work to understand all the major sources of process variation
 Candidates for causes of process variation include: Raw
material lot, equipment, operators (individuals and teams),
shifts and ambient conditions

73 Snee Associates, LLC

 Think Continued Process Verification,
Data, Variation and Statistical Tools
 Data and its variation are an integral part of Process Validation
and Continued Process Verification
 Statistical thinking, methods and tools are built to help us
understand and deal with variation
 What’s the problem? The Data? The Tools?
 Identify the problem and needed data first.
 Then the appropriate statistical tools can be selected and utilized
 Linking and integrating the tools into a system leads to
effective use of the tools and enhanced Continued Process
Verification

A Systems Approach Enhances the

Effectiveness of Continued Process Verification
74 Snee Associates, LLC
 For Further Information,
Please Contact:
Ronald D. Snee, PhD
Snee Associates, LLC
Newark, DE
(610) 213-5595
[email protected]

Please visit our website at:

www.SneeAssociates.com

75 Snee Associates, LLC

 References
Berridge, J. C. (2009) “PQLI – What Is It?” Pharmaceutical Engineering, May/June 2009, 36-39.
McGurk, T and R. D. Snee (2005) “A Systematic Approach to Deviation Reduction through Six
Sigma,” Pharmaceutical Technology Sourcing and Management, October 2005, 4-18.
Montgomery, D. C. (2009) Statistical Quality Control, 6th Ed, John Wiley & Sons, New York, NY
Schweitzer, M., M. Pohl, M. Hanna-Brown, P. Nethercote, P. Borman, G. Hanson, K. Smith, J.
Larew (2010) “Implications and Opportunities of Applying QbD Principles to Analytical
Measurements”, Position Paper: QbD Analytics, Pharma Tech, February 2010, 52-59
Snee, R. D. (2009) “A Framework for Quality by Design”, Pharma Tech Online, Oct 2009
Snee, R. D. (2010) “Crucial Considerations in Monitoring Process Performance and Product
Quality”, Pharma Technology Analytical Tech and Instrumentation, October 2010, 38-42.
Snee, R. D. (2011) “Using QbD to Enable CMO Manufacturing Process Development, Control and
Improvement”, Pharmaceutical Outsourcing, January/February 2011, 10-18.
Snee, R. D. (2015) “Management Holds the Key to Continued Process Verification”,
Pharmaceutical Manufacturing, January/February 2015, 33-35.
Snee, R. D. and R. W. Hoerl (2003) Leading Six Sigma – A Guide Based on Experience With
General Electric and Other Six Sigma Companies , FT Prentice Hall, New York, NY.
Snee, R. D. and R. W. Hoerl (2012) “Going on Feel: Monitor and Improve Process Stability to
Make Customers Happy”, Quality Progress, May 2012, 39-41
Weitzel, J., R. A. Forbes and R. D. Snee (2015) “The Use of the Analytical Target Profile in the
Lifecycle of an Analytical Procedure: with an example for an HPLC Procedure”, Journal of
Validation Technology, Vol. 20, Issue 4, Jan 2015
76 Snee Associates, LLC
 A Vision
Continued Process Verification - Building Blocks
 We understand our processes
 Core processes identified and mapped
 “Critical Few” process parameters (Xs) are known
 Measurements in place and monitored
 Process inputs, In-process, process output and customer
 Monitoring and reviews are done daily, weekly and monthly
 Process Enhancements are being made
 Process adjustments and improvements
 Training done to develop the needed managing and
operational knowledge and skills
 Good Process characteristics are defined and implemented
 Periodic process management reviews are made

77 Snee Associates, LLC

 Reporting
Continued Process Verification Results

 CPV is a on-going process executed daily to quarterly

depending on the level review
 By Process operator and managers at various levels
 Monthly and quarterly reports should be considered
 Annual CPV summary can be reported as part of the
Annual Product Review (APR)
 CPV annual summary can be built through out the year
 APR is a annual report,
 CPV continues throughout the year

78 Snee Associates, LLC

View publication stats 79 Snee Associates, LLC