Professional Documents
Culture Documents
A Case Study
INTERPHEX
March 17, 2009
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
Failure Mode The manner in which a process fails to meet its intended
purpose
High priority failure modes are analyzed to identify their root causes
and solutions are then developed which lead to reduced risk
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
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?
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
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.
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
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)
UF BRs:
detect
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.
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
• Process limitations
• Filter set-up/testing
Root causes were identified and solutions were created and prioritized…
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
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
Solutions were then prioritized based on estimated impact and implementation feasibility…
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
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.
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