managing process risk through application of … process risk through application of fmea to batch...
TRANSCRIPT
ManagingManaging Process Risk through Application of FMEA
to Batch Recordsto Batch Records
A Case StudyINTERPHEX
March 17, 2009
Jon Hardy Fred GreulichLonza 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• Case Background and Situation• Risk Discussion• Introduction to Failure Mode & Effects Analysis (FMEA) y ( )
and Root Cause Analysis (RCA)• FMEA and RCA Project Approach• Current Status and Results
2Proprietary and Confidential
Case Background and Situation
Lonza is a worldwide leader in providing development and manufacturing services for the pharmaceutical and biotechnology
• Portsmouth, NH is a large scale mammalian cell culture site • Strategically important to Lonza
g gyindustries…
Strategically important to Lonza • Operational Excellence, including risk management, are ongoing
business imperatives
• Project objective was to reduce manufacturing process risk• 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 product
3Proprietary and Confidential
purification process for one key customer s product
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• 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 ff ti l th f t i t
4Proprietary and Confidential
to effectively manage the many sources of uncertainty
Categories of Value Chain Risks
Value chain risks can be grouped into four categories…
P d t S it Ri kProduct Security Risks• Diversion• Authentication
• Brand security• Physical facility• Information security
Supply Risks Production Risks Demand Risks• Supply disruption • Facility & equipment • Demand upside/Supply disruption• Continuity planning• Pricing fluctuations• Material variability
Facility & equipment• Process variability• Process failures• Production scale-up
Demand upside/ downside
• Capacity constraints• Product expiryMaterial variability Production scale up
• Non-Compliance
p y
FMEA applied to batch records is the key tool for
5Proprietary and Confidential
helping Lonza - Portsmouth manage process failures
Introduction to FMEA
FMEA is a rigorous method of identifying and preventing process problems before they occur
• An established tool for identifying, prioritizing and managing
problems before they occur…
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
6Proprietary and Confidential
need for corrective action
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 (Abilit t D t t)Number (RPN) (Ability to Detect)
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
7Proprietary and Confidential
and solutions are then developed which lead to reduced risk
FMEA and RCA Approach
FMEA and RCA work is completed in three phases, starting with education and ending with implementation…
PHASE I. FMEA Assessment & RPN Scoring
• Conduct team education on FMEA • Identify RPN’s for each failure mode• Determine top priority failure modes
g
PHASE II RCA & Solution Identification
• Conduct RCA for top priority failure modes• Develop solutions for selected root causes
PHASE III Detailed Design & Implementation
• Develop detailed implementation designs and 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
8Proprietary and Confidential
and ownership of results
Phase I - FMEA Assessment & RPN Scoring
Results from Phase I, FMEA Assessment and Risk Priority Scoring, were captured in a spreadsheet format…
Process Step
Potential Failure Mode
Potential Failure Effects
SEV
Potential Causes
OCC
Current Controls
DET
RPN
were captured in a spreadsheet format…
What is the
process step?
In what way could the process step go wrong?
What is the impact if the failure mode occurs?
What causes the failure mode?
What is the
What are the existing controls/ that prevent the
likelihood of this occurring?
failure mode?
How difficult is it to detect if the failure modefailure mode occurs?
9Proprietary and Confidential
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
DefinitionDescriptionRating
Severity = The seriousness of the effect of a failure mode on a downstream operation, equipment, safety, product, customer, or regulation. Occurrence Rating Scale
Schedule ImpactCompliance ImpactOperator Safety ImpactProduct Impact
Days
Weeks
Months
Minor deviation
Major deviation
Notify FDA
Would require first aid
Would require calling 911
Could result in death
Salvageable with substantial rework
Potentially salvageable
Lose batch
Moderate3
High4
Dangerously High5
Failure is likely; repeated failures have been observed in similar processes; 3 or more occurrences likely in 10 events (approximately 1
Failure is almost inevitable; 5 or more occurrences likely in 10 events
Definition
High4
Very High5
DescriptionRating
Occurrence = How likely a failure mode is to occur for a given cause.Detection Rating Scale
Detectability = Difficulty of detecting the defect or failure with current process controls.
Proprietary & Confidential 19
No impact
Hours
No impact
Comment
No impact
First aid not required
No impact
Salvageable with minimal rework
Minor1
Low2
NOTE: Take the highest effect rating for the severity scoreFailure is unlikely; Never been experienced; No failures ever associated with similar processes
Relatively few or isolated failures experienced in similar processes; 3 or more occurrences likely in 1000 events (approximately 3 Sigma)
Occasional failures have been experienced in similar processes; 3 or more occurrences likely in 100 events(approximately 2 Sigma)
processes; 3 or more occurrences likely in 10 events (approximately 1 Sigma)
Remote1
Low2
Moderate3
High4
Moderate likelihood current controls will detect failure.
Remote likelihood current controls will detect failure.
No known controls available to detect failure mode, or defect is not detectable.
Definition
Moderate3
Somewhat difficult to detect4
Very difficult to detect5
DescriptionRating
Proprietary & Confidential 20
Proprietary & Confidential 21
Current controls almost certain to detect the failure mode.
High likelihood current controls will detect failure.
Easy to detect1
Somewhat easy to detect2
10Proprietary and Confidential
This customization was done in a project kickoff session at the beginning of Phase I of the project
Phase I - FMEA Assessment & RPN Scoring
Here is a example of the completed FMEA Assessment and Priority Scoring Template…
PROCESS STEP POTENTIAL FAILURE MODE (HOW) EFFECT (WHAT) SEV CAUSE (WHY) OCC CURRENT CONTROLS DET RPN
7.2 Starting Material preparation7.2.1 Record harvest batch Do not record Lost traceability 1 Oe 2 batch record 3 6
g
number ; room temperature
7.2.1 Record harvest batch number ; room temperature
Temperature out of specificxation noncompliance with quality systems (deviation)
2 Probe oot 2 PM Program 2 8
7.2.2 Harvest confirmation Temperature out of specification lost process time (bad temp probe out of spec
2 probe oot 2 Metrology program 2 8temp probe, out of spec temperature)
7.2.2 Harvest confirmation Print,record weight not done noncompliance with quality systems (deviation)
2 Oe 2 batch record 3 12
7.2.3 Mix permeate Not able to mix Lost process time (mixer problem)
2 equipment failure 2 PM Program 2 8(mixer problem)
7.2.3 Mix permeate Mix time out of specification noncompliance with quality systems (deviation)
2 Oe 2 batch record 3 12
7.2.4 Setup and obtain a bulk sample of the harvest
Sampler expired lost administration time (comments)
2 Oe 3 batch record 3 18
7.2.4 Setup and obtain a bulk Sampler not connected negligible XXXX XXXXX
11Proprietary and Confidential
sample of the harvest
Phase I - FMEA Assessment & RPN Scoring
The Lonza team began with 8 batch records which eventually mapped to 25 critical failure modes...
8 Purification Batch Records
Analyzed >2600 Batch Record
line items
266 High Risk RPN Steps
25 Critical Failure Modes
Chrom BRs:- Chrom A
line items
Effect Categories & Severity Rating Scale
Customized Rating Scalesy-axis
Introduced 2-Slope Plots Grouped High Risk RPN Steps by Failure Mode
- Chrom B- Chrom C- Chrom D
UF BRs:- UF A
Proprietary & Confidential 19
Effect Categories & Severity Rating Scale
Schedule ImpactCompliance ImpactOperator Safety ImpactProduct Impact
No impact
Hours
Days
Weeks
Months
No impact
Comment
Minor deviation
Major deviation
Notify FDA
No impact
First aid not required
Would require first aid
Would require calling 911
Could result in death
No impact
Salvageable with minimal rework
Salvageable with substantial rework
Potentially salvageable
Lose batch
Definition
Minor1
Low2
Moderate3
High4
Dangerously High5
DescriptionRating
Severity = The seriousness of the effect of a failure mode on a downstream operation, equipment, safety, product, customer, or regulation.
NOTE: Take the highest effect rating for the severity score
Proprietary & Confidential 20
Occurrence Rating Scale
Failure is unlikely; Never been experienced; No failures ever associated with similar processes
Relatively few or isolated failures experienced in similar processes; 3 or more occurrences likely in 1000 events (approximately 3 Sigma)
Occasional failures have been experienced in similar processes; 3 or more occurrences likely in 100 events(approximately 2 Sigma)
Failure is likely; repeated failures have been observed in similar processes; 3 or more occurrences likely in 10 events (approximately 1 Sigma)
Failure is almost inevitable; 5 or more occurrences likely in 10 events
Definition
Remote1
Low2
Moderate3
High4
Very High5
DescriptionRating
Occurrence = How likely a failure mode is to occur for a given cause.Detection Rating Scale
Current controls almost certain to detect the failure mode.
High likelihood current controls will detect failure.
Moderate likelihood current controls will detect failure.
Remote likelihood current controls will detect failure.
No known controls available to detect failure mode, or defect is not detectable.
Definition
Easy to detect1
Somewhat easy to detect2
Moderate3
Somewhat difficult to detect4
Very difficult to detect5
DescriptionRating
Detectability = Difficulty of detecting the defect or failure with current process controls.
y-axis
x-axis 50
100
150
200
250
300
Steps by Failure Mode
Q FF Chrom
- UF A- UF B- UF C
Fill Finish:- Bulk Fill A
Proprietary & Confidential 21
Process Step/Input
Potential Failure Mode
Potential Failure Ef fects
SEV
Potential CausesOCC
Current ControlsDET
RPN
What is the process step
and Input
In what ways does the Key Input go wrong?
What is the impact on the Key Output Variables (Customer
What causes the Key Input to go wrong?
What are the existing controls and procedures
FMEA Table to Calc RPN
250
300
Determined High Risk RPN steps for each batch record
100
150
200
250
300
01 16 31 46 61 76 91 106121136151166181196211226241256271286
Phenyl Chrom
12Proprietary and Confidential
and Input under
investigation?
Input go wrong? Variables (Customer Requirements)?
and procedures (inspection and test) that prevent eith the cause or the Failure Mode?
0
50
100
150
200
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
0
50
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 341 361 381
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 anticipatedbased 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
13Proprietary and Confidential
Phase II – RCA & Solution Identification
Root causes were identified and solutions were created and prioritized…
Six Failure M d
Perform Root Cause
Analysis for EachGroup Root
Causes
Develop Solutions for
Each Root Prep for
ImplementationModes(from FMEA Assessment
Phase)
Analysis for Each Failure Mode
ac ootCause(s)
Root Cause Analysis Workshops
•Training
Solution Sessions
• Develop
Develop Implementation
Approach• Root Cause Analysis
• Develop Fishbone Diagrams/ 5 Why’s
• Converge on root causes• Root cause grouping
Develop solution(s) for each root cause
• Rate solutions against batch
d t
Management Team Review
14Proprietary and Confidential
Root cause grouping• Identify teams for Solutions
record steps
Phase II – RCA & Solution Identification
RCA was performed for each of the six failure modes…
Cause & Effect Analysis
• Structured brainstorming of
Narrowing• Narrows potential
Selection• Selects the root
( ) fbrainstorming of potential root causes
• Utilized “fishbone diagrams”
5 Why’s
proot causes down to the likely ones
• Based on judgment of the
cause(s) for solution identification
• Based on lit ti /5 Why s
• Asks “why does that happen?”
• Supports fishbone diagram by helping
j gpeople who are familiar with the process
qualitative/ quantitative “basic data”
Actual Rootdiagram by helping to drill deeper
Large set Potential Root
Causes
Smaller setLikely Root
Causes
Actual Root Causes
15Proprietary and Confidential
As a result, 23 actual root causes were identified
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
• Root Cause 1
• Root Cause 2
Chromatography Operations
• Root Cause 11
• Root Cause 12
• Root Cause 3
• Root Cause 4
• Root Cause 5
• Root Cause 13
• Root Cause 14
• Root Cause 15
• Root Cause 6
• Root Cause 7
• Root Cause 16
• Root Cause 17
• Root Cause 18
• Root Cause 19
Documentation
• Root Cause 8
• Root Cause 9
Root Cause 19
• Root Cause 20
Filter Integrity Testing • Root Cause 21
16Proprietary and Confidential
Root Cause 9
• Root Cause 10
• Root Cause 21
• Root Cause 22
• Root Cause 23
Phase II – RCA & Solution Identification
Solutions for the 23 root causes were then developed and captured in a standard format...
Root Cause: XXXXXX
Solution Sketch/Notes:
Solution Name: Install Vent to Improve Draining of Caustic Header in UFDeveloped By: Andy, Derek, Stuart, Dan Date: 01Apr2008
Failure Mode: XXXXXX
Solution Description:
Install block and bleed/vent after liquid filter housings of caustic tanks (T-35013/T-35014) to improve header draining.
Ke I t ll bl k d bl d/ t ft li id filt h i f
Solution Sketch/Notes:
New vent & valve block
Key Characteristics of Solution:
• Install block and bleed/vent after liquid filter housings of caustic tanks (T-35013/T-35014)
• 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• Could be done with ambient or pressurized air
Key Assumptions:
• Opening bleed/vent valve will provide a high point vent of drain line quickly and effectively Impact Rating: 750
Additional Benefits:
Solution can be applied to chromatography skids as well
17Proprietary and Confidential
q y y• No issues with pipe slope, introducing air to filter housing• Don’t change drain header from T-24030• Wet test required to better determine feasibility & impact
Impact Rating: 750
Feasibility Rating: 3
Prioritization Index: 250
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…Example: Incorrect Probe Standardization
1 If l ti
Potential Impact to RPNs
For each impacted batch record line item, we will calculate a potential RPN if the solution were to be implemented…Example: Incorrect Probe Standardization
1 If l ti
Feasibility of ImplementationFeasibility of Implementation
Solution X for Incorrect Probe StandardizationBR Line Item Baseline RPN RPN To Be ∆RPN
• 36 batch record line items are impacted(Chrom & UF)
• For each impacted line item, we will assesssolution impact
1. If solution were implemented, here is what RPN we would expect
2. Potential impact of the solution
Solution X for Incorrect Probe StandardizationBR Line Item Baseline RPN RPN To Be ∆RPN
• 36 batch record line items are impacted(Chrom & UF)
• For each impacted line item, we will assesssolution impact
1. If solution were implemented, here is what RPN we would expect
2. Potential impact of the solution
The team used this table to evaluate solution feasibility…
FeasibilityRating
Extent of Control
Impact on Customer
Capital Cost to Implement
Time to Implement
Resource Commitment
1 Within Team
Internal Change Control only (e.g. SOP Changes)
$0 - $8K Days 0-20 hrs
Wi hi FEAS
IBLE
The team used this table to evaluate solution feasibility…
FeasibilityRating
Extent of Control
Impact on Customer
Capital Cost to Implement
Time to Implement
Resource Commitment
1 Within Team
Internal Change Control only (e.g. SOP Changes)
$0 - $8K Days 0-20 hrs
Wi hi FEAS
IBLE
1 USPO-1847 9.25 Ensure conductivity probes have been cleaned……
160 100 60
2 USPO-1847 9.25.2 Post standardization conductivity standard check result…
160 50 110
36 USPO-1853 15.7 Standardize UF-27210 conductivity probe…
180 60 1203. Sum of all ∆RPN l
1 USPO-1847 9.25 Ensure conductivity probes have been cleaned……
160 100 60
2 USPO-1847 9.25.2 Post standardization conductivity standard check result…
160 50 110
36 USPO-1853 15.7 Standardize UF-27210 conductivity probe…
180 60 1203. Sum of all ∆RPN l
2 Within Group
Customer notified, no approval needed $8K - $25K Weeks 20-100 hrs
3 Within Plant Mgr.
Customer signoff required as CC $25K - $50K Months 100-400 hrs
4 Within NH Site
FDA Notified; CMC filing changes
necessary
$50K -$150K Quarters 400-1600 hrs
MO
RE
F2 Within Group
Customer notified, no approval needed $8K - $25K Weeks 20-100 hrs
3 Within Plant Mgr.
Customer signoff required as CC $25K - $50K Months 100-400 hrs
4 Within NH Site
FDA Notified; CMC filing changes
necessary
$50K -$150K Quarters 400-1600 hrs
MO
RE
F
8
Total Potential Impact of this Solution: 680∆RPN values
8
Total Potential Impact of this Solution: 680∆RPN values
17Proprietary & Confidential
Low feasibility number is better
5 Beyond NH Site
New Clinical trials / approval necessary >$150K > 1 year > 1600 hrs
17Proprietary & Confidential
Low feasibility number is better
5 Beyond NH Site
New Clinical trials / approval necessary >$150K > 1 year > 1600 hrs
18Proprietary and Confidential
RPN (beginning) – RPN (estimated after action)
Feasibility Rating = Solution Priority Index
Phase III – Detailed Design & Implementation
Detailed Design & Implementation typically includes the following…
Deployment Planning
Ch t t D l d t il d C d t t ti / D l t i d
Detailed Design Detailed Design Sustain ImprovementsImplementationImplementation
• Charter teams• Develop design
and implementation plan for each
• Develop detailed design for each solution
• Conduct vendor
• Conduct testing/ validation
• Complete SOP/ Batch Record
• Develop metrics and tracking methods
• Develop processes to sustain improvements
solution• Ensure integration
across the plansEstablish timelines
selection(s) as required
• Complete SOP / Batch Record updates and
rollout and training• Implement
solution elements Obtain
• Conduct regular status/update reviews
• Identify and implement continuous• Establish timelines
and milestones Record updates and obtain approvals
• Obtain management & customer approvals
• Obtain management & customer sign-off
implement continuous improvement ideas
19Proprietary and Confidential
…that occur in 60-90 day “Sprints” to ensure focus and momentum.
Phase III – Detailed Design & Implementation
Month 1 Month 2 Month 3 Month 4The schedule for these implementation “Sprints” typically looks as follows…
Month 5Deploy-
mentPlanning
Detailed Design Detailed Design
Month 1 Month 2 Month 3 Month 4 Month 5
gg
Sustain Improvements
Implementation
LEGENDDeploy-
mentPlanning
Detailed DesignDetailed Design
LEGEND:
Wave 1 SprintDetailed Design Detailed Design
Sustain Improve-
t
ImplementationWave 2 Sprint
20Proprietary and Confidential
mentsThis aligns with Lonza’s overall Operational Excellence approach of “Plan, Do, Check, Act”
Current Status and Results
Cross functional teams have begun, and in some cases completed, implementation of the 23 solutions as targeted projects utilizing the
• Implementation progress varies due to the nature and scope of the solution
implementation of the 23 solutions as targeted projects utilizing the current systems within Lonza…
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 systemsphase through its change control and document change systems.
• The solutions were mainly championed through Manufacturing with significant support from Engineering, Controls, Validation, and Quality AssuranceAssurance.
Benefits in terms of reduced failures are already being realized
21Proprietary and Confidential
already being realized
Maxiom Group Overview
• Maxiom Group is a business and information technology consulting firm exclusively serving the life sciences industryfirm exclusively serving the life sciences industry
• Our clients include emerging, established, and mature Biotechnology, Pharmaceutical, Diagnostic and Medical Device companiesPharmaceutical, Diagnostic and Medical Device companies
• Maxiom Group helps life science companies transform their strategies, business processes, and business systems to achieve g , p , yexcellence 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– 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 h ll
22Proprietary and Confidential
challenges