Managing

Process Risk through

Application of FMEA
to Batch Records

A Case Study
INTERPHEX
March 17, 2009

Jon Hardy Fred Greulich

Lonza Maxiom Consulting Group, Inc.
Head of Operations – Hopkinton, MA Director of Operational Excellence

Maxiom Consulting Group, Inc.

230 Third Avenue| 3rd Floor| Waltham, MA 02451 | (781) 250-4900 www.maxiomgroup.com
Case Study Topics

• Case Background and Situation

• Risk Discussion
• Introduction to Failure Mode & Effects Analysis
y ((FMEA))
and Root Cause Analysis (RCA)
• FMEA and RCA Project Approach
• Current Status and Results

Proprietary and Confidential 2

Case Background and Situation

Lonza is a worldwide leader in providing development and

manufacturingg services for the pharmaceutical and biotechnology
gy
industries…
• Portsmouth, NH is a large scale mammalian cell culture site
• Strategically important to Lonza
• Operational Excellence, including risk management, are ongoing
business imperatives
• Project objective was to reduce manufacturing process risk
• Lonza had previous experience utilizing Failure Mode and Effects
Analysis and Root Cause Analysis

Lonza partnered with Maxiom Group to utilize FMEA and

RCA applied to batch record steps to reduce risk in the
purification process for one key customer’s
customer s product

Proprietary and Confidential 3

Risk in Biologics Manufacturing

Risk and uncertainty exist across the entire value chain as a result
of many factors …
• Product & process technology
• Commercial process improvements
• Changing product specifications
• Human/operator variability
• Supplier vulnerabilities
• Scale-up and launch uncertainty
• Demand fluctuations
• Raw material & component pricing
• Supply base market dynamics
It is essential to identify and understand risks in order
t effectively
to ff ti l manage the th many sources off uncertainty
t i t

Proprietary and Confidential 4

Categories of Value Chain Risks

Value chain risks can be grouped into four categories…

P d t Security
Product S it Risks
Ri k
• Diversion • Brand security
• Authentication • Physical facility
• Information security
Supply Risks Production Risks Demand Risks
• Supply disruption • Facility & equipment • Demand upside/
• Continuity planning • Process variability downside
• Pricing fluctuations • Process failures • Capacity constraints
• Material variability • Production scale
scale-up
up • Product expiry
p y
• Non-Compliance

FMEA applied to batch records is the key tool for

helping Lonza - Portsmouth manage process failures

Proprietary and Confidential 5

Introduction to FMEA

FMEA is a rigorous method of identifying and preventing process

problems before they occur
occur…

• An established tool for identifying, prioritizing and managing

process and business risk
• Focused on preventing defects, enhancing safety, and increasing
customer satisfaction
• Minimizes “Cost Of Quality” – Focuses on Prevention vs. Detection

Applied to manufacturing batch records, FMEA results in

more robust processes and reduction/elimination of the
need for corrective action

Proprietary and Confidential 6

Introduction to FMEA

There are a few terms that are important to understand…

Failure Mode The manner in which a process fails to meet its intended
purpose

Failure Effects The consequences if the failure occurs

Risk Priority RPN = (Potential Severity) x (Likelihood of Occurrence) x

Number (RPN) (Abilit to
(Ability t Detect)
D t t)
Used for each process step/failure mode combination.
The RPN is used to prioritize failure modes.

High priority failure modes are analyzed to identify their root causes
and solutions are then developed which lead to reduced risk

Proprietary and Confidential 7

 FMEA and RCA Approach

FMEA and RCA work is completed in three phases, starting with

g with implementation…
education and ending
PHASE I. FMEA Assessment & • Conduct team education on FMEA
RPN Scoring • Identify RPN’s for each failure mode
• Determine top priority failure modes

PHASE II RCA & Solution • Conduct RCA for top priority failure modes
Identification • Develop solutions for selected root causes

PHASE III Detailed Design & • Develop detailed implementation designs and
Implementation action plans for solutions identified in Phase II
• Implement action plans

All Phases were completed with the active involvement of

cross-functional teams of Lonza employees to ensure quality
and ownership of results

Proprietary and Confidential 8

Phase I - FMEA Assessment & RPN Scoring

Results from Phase I, FMEA Assessment and Risk Priority Scoring,

were captured in a spreadsheet format…

S O D R
Process Potential Potential Potential Current
E C E P
Step Failure Mode Failure Effects Causes Controls
V C T N
What is In what way What is the What causes the What are the
the could the impact if the failure mode? existing
process process step failure mode controls/ that
step? go wrong? occurs? What is the prevent the
likelihood of this failure mode?
occurring?
How difficult is
it to detect if the
failure mode
occurs?

Proprietary and Confidential 9

 Phase I - FMEA Assessment & RPN Scoring
Templates were customized by the Lonza team to ensure standards
were applied to the assessment and scoring process…
Effect Categories & Severity Rating Scale

Severity = The seriousness of the effect of a failure mode on a downstream operation, equipment,
safety, product, customer, or regulation. Occurrence Rating Scale
Rating Description Definition
Product Impact Operator Safety Impact Compliance Impact Schedule Impact

5 Dangerously High Lose batch Could result in death Notify FDA Months
Detection Rating Scale
Occurrence = How likely a failure mode is to occur for a given cause.
Potentially Would require calling
4 High Major deviation Weeks
salvageable 911 Rating Description Definition

Salvageable with 5 Very High Detectability

Failure is almost inevitable; 5 or more occurrences likely in=10
Difficulty
events of detecting the defect or failure with current process
3 Moderate substantial Would require first aid Minor deviation Days controls.
rework Failure is likely; repeated failures have been observed in similar
4 High processes; 3 or more occurrences likely in 10 events (approximately
Rating 1
Description Definition
Sigma)
Salvageable with
2 Low First aid not required Comment Hours Occasional failures have been experienced in similar processes; 3 or
minimal rework
3 Moderate more occurrences likely in 100 events Very difficult to No known controls available to detect failure mode, or defect is not
5
(approximately 2 Sigma) detect detectable.
1 Minor No impact No impact No impact No impact
Relatively few or isolated failures experienced in similar processes; 3 or
2 Low
NOTE: Take the highest effect rating for the severity score more occurrences likely in 1000 events (approximately 3Somewhat
Sigma) difficult to
4 Remote likelihood current controls will detect failure.
Failure is unlikely; Never been experienced; No failures detect
ever associated
1 Remote
with similar processes
Proprietary & Confidential 19
3 Moderate Moderate likelihood current controls will detect failure.

Somewhat easy to
2 High likelihood current controls will detect failure.
detect
Proprietary & Confidential 20

1 Easy to detect Current controls almost certain to detect the failure mode.

Proprietary & Confidential 21

This customization was done in a project kickoff

session at the beginning of Phase I of the project
Proprietary and Confidential 10
 Phase I - FMEA Assessment & RPN Scoring

Here is a example of the completed FMEA Assessment and Priority

Scoringg Template…
PROCESS STEP POTENTIAL FAILURE MODE (HOW) EFFECT (WHAT) SEV CAUSE (WHY) OCC CURRENT CONTROLS DET RPN

7.2 Starting Material

preparation
7.2.1 Record harvest batch Do not record Lost traceability 1 Oe 2 batch record 3 6
number ; room temperature

7.2.1 Record harvest batch Temperature out of specificxation noncompliance with 2 Probe oot 2 PM Program 2 8
number ; room temperature quality systems
(deviation)

7.2.2 Harvest confirmation Temperature out of specification lost process time (bad 2 probe oot 2 Metrology program 2 8
temp probe
probe, out of spec
temperature)
7.2.2 Harvest confirmation Print,record weight not done noncompliance with 2 Oe 2 batch record 3 12
quality systems
(deviation)

7.2.3 Mix permeate Not able to mix Lost process time 2 equipment failure 2 PM Program 2 8
(mixer problem)
7.2.3 Mix permeate Mix time out of specification noncompliance with 2 Oe 2 batch record 3 12
quality systems
(deviation)

7.2.4 Setup and obtain a bulk Sampler expired lost administration time 2 Oe 3 batch record 3 18
sample of the harvest (comments)
7.2.4 Setup and obtain a bulk Sampler not connected negligible XXXX XXXXX
sample of the harvest

Proprietary and Confidential 11

Phase I - FMEA Assessment & RPN Scoring

The Lonza team began with 8 batch records which eventually

mapped to 25 critical failure modes...

8 Purification Analyzed >2600

266 High Risk 25 Critical
Batch Records Batch Record
RPN Steps Failure Modes
line items

Chrom BRs: Customized Rating Scales Introduced 2-Slope Plots Grouped High Risk RPN
- Chrom A Effect Categories & Severity Rating Scale
y-axis Steps by Failure Mode
- Chrom B Severity = The seriousness of the effect of a failure mode on a downstream operation, equipment,
safety, product, customer, or regulation.
Rating Description
Product Impact
Occurrence Rating Scale
Definition
Operator Safety Impact Compliance Impact Schedule Impact

- Chrom C 5

4
Dangerously High

High
Lose batch

Potentially
salvageable
Could result in death

Would require
911
calling
Occurrence

Rating
= How
Major
Notify FDA

likely
deviation

Description
Months

a failure
Detection Rating Scale
Weeks mode is to occur for a given cause.

Definition
300
Q FF Chrom
- Chrom D
Salvageable with
3 Moderate substantial
rework
Would require first5aid Minor deviation
Very High Days
Failure is almost inevitable; 5 or more occurrences likely in 10 events

Detectability
Failure is likely; repeated failures have=been
Difficulty
observed of detecting
in similar the defect or failure with current process
250
2 Low
Salvageable with 4
First aid not required High
Comment processes; 3 or morecontrols.
Hours occurrences likely in 10 events (approximately 1
minimal rework Sigma)

200
Occasional failures have Rating Description
been experienced in similar processes; 3 or Definition
1 Minor No impact No impact 3 No impact
Moderate No more
impactoccurrences likely in 100 events
(approximately 2 Sigma)
NOTE: Take the highest effect rating for the severity score Very difficult to No known controls available to detect failure mode, or defect is not
5
Relatively few or isolated failures experienced in similar processes; 3 or
2 Low detect detectable.
more occurrences likely in 1000 events (approximately 3 Sigma)

Proprietary & Confidential 19 1 Remote

Failure is unlikely; Never been experienced;
with similar processes 4
Somewhat Nodifficult
failuresto
ever associated
Remote likelihood current controls will detect failure. 150

UF BRs:
detect

3 Moderate Moderate likelihood current controls will detect failure.

100
Proprietary & Confidential 20 2
Somewhat easy to
detect
High likelihood current controls will detect failure.

x-axis
- UF A 50
1 Easy to detect Current controls almost certain to detect the failure mode.

Proprietary & Confidential

0
21

- UF B 1 16 31 46 61 76 91 106121136151166181196211226241256271286

- UF C
300
Determined High Risk RPN 250
Phenyl Chrom
FMEA Table to Calc RPN steps for each batch record
Fill Finish: 200

150
- Bulk Fill A
Process Potential Failure Potential Failure
S
E Potential Causes
O
C Current Controls
D
E
R
P
300
Step/Input Mode Ef f ects
V C T N

What is the In what ways What is the impact What causes the Key What are the 250
process step does the Key on the Key Output Input to go wrong? existing controls 100
and Input Input go wrong? Variables (Customer and procedures
under Requirements)? (inspection and test)
investigation that prevent eith the 200
? cause or the Failure
Mode? 50
150

100 0
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 341 361 381
50

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Proprietary and Confidential 12

 Phase I - FMEA Assessment & RPN Scoring

Of the 25 critical failure modes, 6 were selected to move into Phase II

based on alignment with existing initiatives at Lonza and anticipated
magnitude of impact…

The selected failure modes fell into categories such as -

• Skid or equipment preparation

• Process limitations

• Filter set-up/testing

Proprietary and Confidential 13

 Phase II – RCA & Solution Identification

Root causes were identified and solutions were created and prioritized…

Six Perform Root Develop

Cause Group Root Solutions for Prep for
Failure Causes Implementation
Analysis for Each Each
ac Root
oot
M d
Modes Failure Mode Cause(s)
(from FMEA
Assessment
Phase)

Root Cause Analysis Solution Develop

Workshops Sessions Implementation
•Training
Approach
• Develop
• Root Cause Analysis solution(s) for
• Develop Fishbone each root cause
Diagrams/ 5 Why’s • Rate solutions Management
• Converge on root causes against batch Team Review
• Root cause grouping record
d steps
t
• Identify teams for Solutions

Proprietary and Confidential 14

Phase II – RCA & Solution Identification

RCA was performed for each of the six failure modes…

Cause & Effect

Analysis
Narrowing Selection
• Structured
• Narrows p potential • Selects the root
brainstorming of
root causes down cause(s)( ) ffor
potential root causes
to the likely ones solution
• Utilized “fishbone
• Based on identification
diagrams”
jjudgment
g of the • Based on
5 Why’s
Why s
people who are qualitative/
lit ti /
• Asks “why does that
familiar with the quantitative
happen?”
process “basic data”
• Supports fishbone
diagram by helping Actual Root
to drill deeper Smaller set Causes
Likely Root
Large set
Causes
Potential Root
Causes

As a result, 23 actual root causes were identified

Proprietary and Confidential 15
Phase II – RCA & Solution Identification

The 23 actual root causes were then grouped by affinity to help in the
identification of teams and the approach for solution development ...
UF Operations Chromatography Operations
• Root Cause 1 • Root Cause 11

• Root Cause 2 • Root Cause 12

• Root Cause 3 • Root Cause 13

• Root Cause 14
• Root Cause 4
• Root Cause 15
• Root Cause 5
• Root Cause 16
• Root Cause 6 • Root Cause 17
• Root Cause 7 • Root Cause 18
• Root Cause 19

Documentation • Root Cause 20

• Root Cause 8 Filter Integrity Testing

• Root Cause 9 • Root Cause 21
• Root Cause 10 • Root Cause 22
• Root Cause 23

Proprietary and Confidential 16

 Phase II – RCA & Solution Identification

Solutions for the 23 root causes were then developed and captured
in a standard format...
Solution Name: Install Vent to Improve Draining of Caustic Header in UF
Developed By: Andy, Derek, Stuart, Dan Date: 01Apr2008
Root Cause: XXXXXX
Solution Sketch/Notes:
Failure Mode: XXXXXX
Solution Install block and bleed/vent after liquid filter housings of
Description: caustic tanks (T-35013/T-35014) to improve header draining. New vent & valve
block

Ke
Key • IInstall
t ll block
bl k and
d bl
bleed/vent
d/ t after
ft liliquid
id filt
filter h
housings
i off
Characteristics caustic tanks (T-35013/T-35014)
of Solution: • Engineering and Technical Requirements
− SOP updates
− Validation (I, Q, CSV)
− Controls work, EM’s Additional Benefits:
− Equipment downtime to install
• Will likely require a vent filter Solution can be applied to
• Could be done with ambient or pressurized air chromatography skids as well
Key • Opening bleed/vent valve will provide a high point vent of
Assumptions: drain line q
quickly
y and effectively
y Impact Rating: 750
• No issues with pipe slope, introducing air to filter housing
• Don’t change drain header from T-24030 Feasibility Rating: 3
• Wet test required to better determine feasibility & impact Prioritization Index: 250

Proprietary and Confidential 17

 Phase II – RCA & Solution Identification

Solutions were then prioritized based on estimated impact and implementation feasibility…

Potential Impact to RPNs

For each impacted batch record line item, we will calculate a potential
RPN if the solution were to be implemented…
Feasibility of Implementation
Example: Incorrect Probe Standardization
• 36 batch record line items are impacted 1 If solution
1. l ti were
2. Potential The team used this table to evaluate solution feasibility…
implemented, here is
(Chrom & UF) what RPN we would impact of the
• For each impacted line item, we will assess expect solution
Feasibility Extent of Impact on Capital Cost Time to Resource
solution impact Control Customer to Implement Implement Commitment
Rating

FEASIBLE
Solution X for Incorrect Probe Standardization Within Internal Change
1 Team
Control only (e.g. $0 - $8K Days 0-20 hrs
BR Line Item Baseline RPN RPN To Be ∆RPN SOP Changes)

MORE F
1 USPO-1847 9.25 Ensure conductivity 160 100 60 Wi hi
Within Customer notified,
probes have been cleaned…… 2 Group no approval needed
$8K - $25K Weeks 20-100 hrs
2 USPO-1847 9.25.2 Post standardization 160 50 110
conductivity standard check Within Plant Customer signoff
result… 3 Mgr. required as CC
$25K - $50K Months 100-400 hrs

FDA Notified; CMC

Within NH $50K -
15.7 Standardize UF-27210
4 Site
filing changes
$150K
Quarters 400-1600 hrs
36 USPO-1853 180 60 120 necessary
3. Sum of all conductivity probe…
∆RPN values
l Beyond NH
Total Potential Impact of this Solution: 680 5
New Clinical trials /
approval necessary
>$150K > 1 year > 1600 hrs
Site
8

Low feasibility number is better

Proprietary & Confidential 17

RPN (beginning) – RPN (estimated after action)

= Solution Priority Index
Feasibility Rating

Proprietary and Confidential 18

 Phase III – Detailed Design & Implementation

Detailed Design & Implementation typically includes the following…

Deployment Sustain
Detailed Design Implementation
Planning Improvements

• Charter
Ch t tteams • Develop
D l ddetailed
t il d • Conduct
C d t ttesting/
ti / • Develop
D l metrics
t i and
d
• Develop design design for each validation tracking methods
and implementation solution • Complete SOP/ • Develop processes to
plan for each • Conduct vendor Batch Record sustain improvements
solution selection(s) as rollout and training • Conduct regular
• Ensure integration required • Implement status/update reviews
across the plans • Complete SOP / Batch solution elements • Identify and
• Establish timelines Record updates and • Obtain implement continuous
and milestones obtain approvals management & improvement ideas
• Obtain management customer sign-off
& customer approvals
…that occur in 60-90 day “Sprints” to ensure focus and momentum.

Proprietary and Confidential 19

Phase III – Detailed Design & Implementation

The schedule for these implementation “Sprints” typically looks as follows…

Month 1 Month 2 Month 3 Month 4 Month 5
Deploy-
ment
Planning

Detailed Design
g

Implementation

Sustain Improvements

LEGEND
LEGEND:
Deploy-
Wave 1 Sprint ment
Planning

Detailed Design
Wave 2 Sprint
Implementation

Sustain
Improve-
ments
t
This aligns with Lonza’s overall Operational
Excellence approach of “Plan, Do, Check, Act”
Proprietary and Confidential 20
Current Status and Results

Cross functional teams have begun, and in some cases completed,

implementation of the 23 solutions as targeted projects utilizing the
current systems within Lonza…
• Implementation progress varies due to the nature and scope of the
solution
solution.
• Since most solutions were identified at a high level, Lonza was able to
then take them through detailed design and into the implementation
phase through its change control and document change systems
systems.
• The solutions were mainly championed through Manufacturing with
significant support from Engineering, Controls, Validation, and Quality
Assurance.
Assurance

Benefits in terms of reduced failures are

already being realized

Proprietary and Confidential 21

Maxiom Group Overview

• Maxiom Group is a business and information technology consulting

firm exclusively serving the life sciences industry

• Our clients include emerging, established, and mature Biotechnology,

Pharmaceutical, Diagnostic and Medical Device companies

• Maxiom Group helps life science companies transform their

strategies,
g , business pprocesses,, and business systems
y to achieve
excellence at each stage of their life cycle
– From drug discovery to clinical development
– From clinical development to commercial launch
– From commercial launch to market leadership

• Clients rely on our Focus, Insight and Approach to guide them in

transforming their business and in addressing their ongoing business
challenges
h ll

Proprietary and Confidential 22