managing process risk through application of … process risk through application of fmea to batch...

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


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


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


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


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


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


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?



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



Current controls almost certain to detect the failure mode.

High likelihood current controls will detect failure.

Easy to detect1

Somewhat easy to detect2


This customization was done in a project kickoff session at the beginning of Phase I of the project


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)


7.2.4 Setup and obtain a bulk sample of the harvest

Sampler expired lost administration time (comments)


7.2.4 Setup and obtain a bulk Sampler not connected negligible XXXX XXXXX


sample of the harvest


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


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



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


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


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


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


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


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


Root cause grouping• Identify teams for Solutions

record steps


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


As a result, 23 actual root causes were identified


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


Root Cause 9

• Root Cause 10

• Root Cause 21

• Root Cause 22

• Root Cause 23


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


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


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


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


…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


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


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


challenges

managing process risk through application of … process risk through application of fmea to batch...

Documents