2013 process failure modes and effects analysis. risk based thinking is not industry dependent! ...
TRANSCRIPT
Risk based thinking is not industry dependent! Airline (1,895 air traffi c control
errors in 2012) Device (48 recalls in 2012) Pharma (45 recalls in 2012)
BASIC PHILOSOPHY
To control risk, you need to know where it exists
Dynamic tool = controlStatic tool = waste of timeBrainstorm, the more brains
the betterYou can never eliminate all risk
1. Map all ACTUAL steps of your processa) Best practice: Brainstorm with the actual hands on peopleb) Pitfalls: Don’t just go off of SOP steps, may loose detail
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What are the steps in the process?
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
2. List all things that could go wrong with the stepa) Best practice: Put it all down, even if crazy! Include human
factors.b) Pitfalls: Don’t avoid putting it down because of low risk
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What could go wrong with that step?What environmental or human factors could contribute (stress, flow of work,
concentration)?
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
3. List the eff ects of each failure mode identifi eda) Best practice: Put it all down, even if crazy!b) Pitfalls: Don’t avoid putting it down because of low risk
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What effects could result from the failure occurring?
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
SCORING
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Recommended Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTakenSev
Occur
Detec
R.P.N.
How bad could the effect of the failure be?
Severity of Effect
10 May result in safety issue or regulatory violation without warning9 May result in safety issue or regulatory violation with warning8 Primary function is lost or seriously degraded7 Primary function is reduced and customer is impacted6 Secondary function is lost or seriously degraded5 Secondary function is reduced and customer is impacted4 Loss of function or appearance such that most customers would return product or stop using service3 Loss of function or appearance that is noticed by customers but would not result in a return or loss of service
2 Loss of function or appearance that is unlikely to be noticed by customers and would not result in a return or loss of service
1 Little to no impact
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
Note: We only scored regulatory risk since the patient wouldn’t use a
compromised device. You may want to also assess
business risk
4. List the causes of each failure mode identifi eda) Best practice: Remember human factorsb) Pitfalls: Don’t avoid putting it down because of low occurrence
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What caused the failure to occur?
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
How often does this failure mode occur?
Probability of Occurrence
10 1 in 29 1 in 108 1 in 507 1 in 2506 1 in 1,0005 1 in 5,0004 1 in 10,0003 1 in 50,000
2 1 in 250,000
1 1 in 1 Million
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
5. List any current controls you have to prevent that failurea) Best practice: List only what is driven by the procedureb) Pitfalls: Don’t include personal best practices
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What do you currently have in place to prevent the failure?
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
6. List all the ways you can currently detect the failurea) Best practice: List only what is driven by the procedureb) Pitfalls: Don’t include tribal knowledge or personal best
practices
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What do you currently have in place to detect occurrence of the failure?
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What is the probability of detecting the failure?
Likelihood of Detection
10 Absolutely uncertain that failure will be detected9 Very remote chance that failure will be detected8 Remote chance that failure will be detected7 Very low chance that failure will be detected6 Low chance that failure will be detected5 Moderate chance that failure will be detected4 Moderately high chance that failure will be detected3 High chance that failure will be detected
2 Very high chance that failure will be detected
1 Almost certainty that failure will be detected
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
7. Calculate the Risk Product Number(Severity x Occurrence x Detection)
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What is an acceptable threshold for your process?
EXAMPLE: EO CLINICAL BATCH RELEASE PROCESS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/
Mechanism(s) of Failure
Occur
Current Design
Controls Prevention
Current Design
Controls Detection
Detec
R.P.N.
Prep for sterilization by stacking them directly into the
sterilizer or containing them in a
basket in the sterilizer
Compromise integrity of
primary packaging (pouches,
trays)
Unit unsuitable for
use1
Stuff too many units into basket or sterilizer, depends on how many runs client
wants
3
Sponsor could double
pouch or sterilize in secondary
packaging or use
additional sterilizers
Analyst and Sponsor
visual inspection of
units
9 27
EO penetration variance due to
loading configuration
Inconsistent SAL 3
Stuff too many units
into basket/cham
ber and/or stacking of full baskets in chamber
2
EO sterilizer qualification
PQ incorporates a uniformity verification
by Bis throughout fully loaded
chamber with
dunnage. Chamber
size is small.
PCDs placed throughout chamber, then BI
tested for sterility to
verify
1 6
8. Calculate the Risk Product Number(Severity x Occurrence x Detection)
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
What, if any, actions need to be taken to reduce the risk?Who is going to do them and when?
RISK MITIGATION PLAN
R.P.N.
Recommended Action(s) Responsibility & Target Completion Date
Action Results
ActionsTakenSev
Occur
Detec
R.P.N.
27
Sterilization agreement should delineate client's responsibility for final package inspection prior to use
6 none
9. Close the loop from actions taken and calculate your new RPN
STEPS
Item / Function
Potential Failure Mode
Potential Effect(s) of
Failure
Sev
Potential Cause(s)/Mechanis
m(s) of Failure
Occur
Current Design
Controls Preventio
n
Current Design
Controls Detection
Detec
R.P.N.
Recommended
Action(s)
Responsibility & Target
Completion Date
Action Results
ActionsTaken
Sev
Occur
Detec
R.P.N.
Did your actions taken bring you to an acceptable risk?
PFMEA is a useful tool when: You are trying to change an existing process You are designing a new process You are performing root cause analysis (CAPA)
PFMEA must be the horse – not the cart Procedural/Process changes after PFMEA can negate efforts
or add new risks Use dynamically to sync with current procedure
PFMEA is an excellent training tool People performing the process learn the “WHYs” behind
control points and “What” can happen if it is not done right
LESSONS LEARNED
Controls fail – redundancy is better For high risks, don’t just rely on one control. Have a safety net!
Think outside the scope of your process Consider defects of the process prior and how that effects yours
Audit process after actions You typically discover new failure modes you missed the first
time around Use constant input from quality system (complaints, CAPAs, etc.)
Don’t rush, this takes time to do right Some have taken me up to 5 meetings to complete
LESSONS LEARNED