failsafe_fmea
TRANSCRIPT
-
8/3/2019 FailSafe_FMEA
1/7
Quality risk managementprinciples areused eectively in many areas o business and government
including nance, insurance, occupational saety and public
healthand by the agencies regulating these industries.
Risk managements widespread use isnt a surprise because
every product and every process has an associated risk. But
while there are some examples o the use o quality risk man-
agement in the medical product manuacturing industry, they
are limited and do not take ull advantage o what risk manage-
ment has to oer. Ater all, inadequate or ineective quality
risk management can harm patients, product users and com-
pany value.
Fail-Safe
FMEAI 50 WordOr Le Faiintomanaerisk
putsconsumersandentireoranizationsinjeopardy.
Usinfaiuremodeandeffectsanaysis(FMEA)canhepidentifyriskbutisntenouhtoavoidpotentiadisaster.
BycombininFMEAwithothertoos,orani -zationscanensuretheirproductswontharmthepeopetheyresup-posedtohep.
Combinationofquaitytooskeep rik i check
byJosRodruez-PrezandManueE.Pea-Rodruez
-
8/3/2019 FailSafe_FMEA
2/7
January2012QP 31
Risk ManageMent
-
8/3/2019 FailSafe_FMEA
3/7
QPwww.quaityproress.com2
The current ocus o the U.S. Food and Drug Admin-
istration (FDA) on risk-based determination requires
that regulated industries dramatically improve their
understanding and use o hazard control concepts.
The appropriate use o quality risk management
can help organizations comply with regulatory require-
ments, such as good manuacturing practices or good
laboratory practices.
Quality risk management is a valuable component
o an eective quality systems ramework. It can, or
example, help guide the setting o specications and
process parameters or manuacturing, assess and mit-
igate the risk o changing a process or specication,
and determine the extent o deviation investigations
and corrective actions.
An eective quality risk management approach can
ensure a high-quality product by providing a proactive
means to identiy and control potential quality issues
during development and manuacturing. Additionally, it
can improve decision making i a quality problem arises.
exp
Risk combines the probability o occurrence o harm
with the severity o that harm. Quality risk manage-
ment supports a scientic and practical approach to
decision making during the lie cycle o a product. It
provides documented and reproducible methods to ac-
complish the quality risk management process based
on current knowledge about the probability, severity
and detectability o the risk.
But each stakeholder perceives dierent potential
harms, computes dierent probabilities and assigns
dierent severities. In relation to medical products, al-
though there are many dierent stakeholdersinclud-
ing patients, medical practitioners, government and the
industry as a wholeprotecting the patient by manag-
ing risk should always be o the utmost importance.1
The manuacture and use o any medical product
necessarily involves some degree o risk. The risk to
product quality is just one component o overall risk.
Product quality must be maintained throughout the
product lie cycle so the characteristics that are impor-
tant to product quality remain consistent.
Eective quality risk management can also acili-
tate better and more inormed decisions, assure regu-
lators o a manuacturers ability to deal with potential
risks, and positively aect the extent and level o regu-
latory oversight.
Manuacturers o regulated products are required
to have a quality management system and processes
or addressing product-related risks. These processes
or managing risk can evolve into a standalone man-
agement system. While manuacturers may choose to
maintain these two management systems separately, it
may be advantageous to integrate them because doing
so could eliminate redundancies and lead to a more e-
ective management system.
to ch h owRisk management is a complex subject because each
stakeholder places a dierent value on the probability
o harm occurring and its severity. As one o the stake-
ip
Tabets
Bisterpacks
Formedpart
Feeder
Fierbisterpacks
Sensoradjustment
Printed
adjustment
Inkribbon
Temperature
Presstime
Fiedbister
packs
Foi
Poc p Faocpco
P s
Op
Competebisterpacks
Incompetebisterpacks
Acceptedbisterpacks
Rejectedbisterpacks
Printedmateria Seaedbisterpacks
Unseaedbisterpacks
Partial proce map for tablet packagig /FIgURE1
-
8/3/2019 FailSafe_FMEA
4/7
January2012QP 33
holders, the manuacturer makes judgments relating
to the saety and perormance o a medical product,
including the acceptability o risks.
Traditionally, risk is assessed and managed in a
variety o inormal ways based on a compilation o
observations, trends and other inormation. That ap-
proach can provide useul inormation that supports
the handling o complaints, quality deects, deviations
and resource allocation. But with a more ormal ap-
proach, industry and regulators can assess and manage
risk using recognized risk management tools:
Basicriskmanagementfacilitationmethods,such
as fowcharts and check sheets.
Failuremodeandeffectsanalysis(FMEA).
Faulttreeanalysis.
Hazardanalysisandcriticalcontrolpoints(HACCP).
Hazardoperabilityanalysis.
Preliminaryhazardanalysis.
Riskrankingandltering.
These tools need to be adapted or specic uses.
Quality risk management methods and the supporting
statistical tools can be used in combination. The com-
bined use o these tools provides fexibility that can
acilitate the application o quality risk management
principles.
The degree o rigor and ormality o quality risk
management should refect available knowledge and
be commensurate with the complexity or criticality o
the issue being addressed.
sc ppochWithin pharmaceutical and medical products manuac-
turing, ew tools are more useul than FMEA. In one o
itsguidesforindustryICHQ9:QualityRiskManage -
mentthe FDA summarizes some o the most com-
mon risk management tools. Among the tools men-
tioned is FMEA.
The FDA writes in section I.2, FMEA provides or
an evaluation o potential ailure modes or processes
and their likely eect on outcomes and/or product
perormance. Once ailure modes are established, risk
reduction can be used to eliminate, contain, reduce or
control the potential ailures. FMEA relies on product
and process understanding. It methodically breaks
down the analysis o complex processes into manage-
able steps. It is a powerul tool or summarizing the im-
portant modes o ailure, actors causing these ailures
and the likely eects o these ailures.2
Furthermore, it mentions that FMEA can be used
to prioritize risks and monitor the eectiveness o
risk control activities.3 The guide denes FMEA and
discusses areas o application but does not provide a
method to carry out an eective FMEA. For that rea-
son, a systematic approach is needed.
That approach consists o three interrelated tools: a
process map, a cause and eects matrix (also called a
prioritization matrix), and an FMEA. Instead o using
the FMEA template rom the beginning o the process,
lets analyze the ailure modes or the inputs that have
a larger eect on the critical customer requirements.
Process map: As opposed to the commonly used
fowchart, the process map is a high-level tool that
ocuses on process steps, as well as their inputs and
outputs. One o the main uses o the process map is to
identiy process inputs that cause high variability.
The process map usually consists o six to 10 boxes
that represent major steps in the process. For each
step, the inputs and outputs are identied. Figure 1
shows an example o a partial process map or a tablet
packaging acility.
Cause and eects matrix: Ater the process
map is developed, the prioritization process begins.
Many companies start to ll in the FMEA template at
this stage. That isnt a good practice because in many
processes, the most critical steps are located near the
middle or toward the end o the process.
Those who use the beginning-to-end approach are
likely to be exhausted by the time they start to analyze
the critical steps. For that reason, it makes sense to
add this intermediate tool.
Risk ManageMent
a ffcv r mm pproch ur
high-quality product by provd m o
dfy d corol potetial quality iue.
-
8/3/2019 FailSafe_FMEA
5/7
QPwww.quaityproress.com4
A cause and eects matrixnot to be conused
with the cause and eects diagram or shboneis
used to analyze how a process input aects process
outputs. To develop a cause and eects matrix, ollow
these six steps:
1. Identiy the customer requirements and write them
in the upper columns.
2. Assign a rating on a scale with one being the least
important and 10 the most important. These ratings
are assigned rom the customers perspective.
3. Using the process map, list the process steps and
process inputs in each row.
4. For each process input, analyze how it aects each
customer requirement using the ollowing scale:
0=theinputdoesnotaffectthecustomer
requirement.
1=theinputhasahardlynoticeableeffecton
the customer requirement.
3=theinputhasamediumeffectonthecus-
tomer requirement.
9=theinputhasastronganddirecteffecton
the customer requirement.
5. Cross-multiplytheratingsfromstepstwoandfour.
For each row, multiply the value o the process in-
put and its corresponding customer requirement.
Add those cross-products at the end o each row.
6. Sort the values rom step ve in descending order.
The end product o using this tool is a list o the
process inputs that have the biggest eect on the cus-
tomer requirements. In this way, you can ocus the de-
velopment o the FMEA on those process inputs that
really matter to the customer and leave unimportant
issues or later analysis.
Eventually, you should analyze all the process in-
puts and how they could ail. But the key takeaway is
knowing where to ocus your eorts. Table 1 shows
an example o the tablet packaging cause and eects
matrix already sorted in descending order.
FMEA: Ater completing the cause and eects ma-
trix, you can start to ll in the FMEA template. Instead
o the beginning-to-end approach, analyze the ailure
modes o the inputs that are strongly related to cus-
tomer requirements.
Using the cause and eects matrix, select the pro-
cess step or input. For that process step or input, con-
sider the possible ways it could ail (the ailure mode).
Then, or each ailure mode, determine the result i it
ails (the eect).
Ater identiying the eects, make a quantitative
assessment o the severity. Typical scales range rom
one (no severity) to 10 (critical to human saety). A-
ter the severity rating is assigned, identiy the potential
causes o the ailure mode. In this stage, dont think
about causal actors or the immediate causes. Instead,
ocus on potential root causes.
These are labeled as potential causes because
causes have not come into play yet. This is one mis-
conception about an FMEA. Its dened as a proactive
tool to identiy and prevent the way in which a pro-
cess, product or system could ail. But in most cases,
r of poco co
10 9 7 7
1 2 3 4 5
Correct
amoundof
tabletsper
blisterpack
Completely
sealedblister
packs
Nobroken
tabletsin
blisterpacks
Blisterpacks
foilisnot
burned
Tota
Poc p Poc p
9 Seain Temperature 0 9 0 9 144
10 Seain Presstime 0 9 0 9 144
3 Fiin Machinespeed 9 0 3 0 111
4 Fiin Feeder 9 0 3 0 111
6 Automaticinspection Sensoradjustment 9 0 3 0 111
12 Seain Foi 0 9 0 3 102
Caue ad effect matrix /TABlE1
-
8/3/2019 FailSafe_FMEA
6/7
January2012QP 35
FMEAs are used ater the ailure has occurred. That
issue will be addressed later.
Ater you identiy the potential root causes, rate
the probability o occurrence on a scale rom one (low
probability o occurrence) to 10 (high probability o
occurrence). Then determine the current controls used
to detect causes or ailures and rate their eectiveness
on a scale rom one (controls are 100% eective) to 10
(controls are not eective).
At this point, you can calculate the risk priority num-
ber (RPN) by multiplying the severity, occurrence and
detectability indices. But thats not the end o the FMEA
process. The next step is to sort the table in descending
order o RPNs. In that way, the inputs with higher RPNs
(most critical to the customer) will be located at the top
o the table, while those with lower RPNs (least critical
to the customer) will be toward the bottom o the table.
Not sorting RPNs rom highest to lowest is another
error many organizations make. This is the essence o
the FMEAto prioritize eorts to prevent the most crit-
ical inputs to ailure. It does not mean you wont man-
age the least critical inputs; its a matter o prioritizing.
This is aligned with what the FDA expects: that an
organizations actions are commensurate with the risk
to and impact on patient saety.
Ater calculating and sorting the RPNs, the real
work begins. At this stage, its possible to identiy and
implement actions to reduce the RPN. Ater those
actions are implemented, the new RPN must be cal-
culated. This is the undamental nature o the FMEA
processcontinuous improvement. Table 2 shows an
example o an FMEA or a tablet packaging process.
m h
At this point, you can integrate the FMEA process with
one o the systems used most requently by an FDA-
Risk ManageMent
Pocp/p
Pofod
Pofffc
s
eveity
Poc
Oc
cuece
Ccoo
D
etectio
rPn aco
codd
re
poibility
aco
s
eveity
Occuece
D
etectio
rPn
Whatistheprocessstep/input
underinvestigation?
Inwhatwaysdoesthekeyinput
gowrong?
Whatistheimpactonthekey
outputvariables(customer
requirements)orinternal
requirements?
How
severeistheeffecttothe
customer?
Whatcausesthekeyinputtogo
wrong?
How
oftendoescauseoffailure
modeoccur?
Whataretheexistingcontrols
andprocedures(inspectionand
test)thatpreventthecauseof
thefailuremode?should
icludeasOPumbe.
How
wellcanyoudetectcause
orfailuremode?
Whataretheactionsfor
reducingtheoccurrencesofthe
Causeorimprovingdetection?
Whosresponsibleforthe
recommendedaction?
Whatarethecompletedactions
takenwiththerecalculated
RPN?Beuetoiclude
completiomoth/yea.
Seain/temperature
Temp.toohih
Burnedbisterpack
10Wronsettin
6Vericationofbatchrecord
7 420
Provide
infraredtemperaturedevicetooperator
M.Pea
Temperature
deviceimpemented(8/11)
10 4 2 80
Seain/temperature
Temp.toohih
Bisterpacknotseaedcompetey
9Wronsettin
6
Vericationofbatchrecord
7 378
Provideinfraredtemperaturedevicetooperator
M.Pea
Temperaturedeviceimpemented(8/11)
9 4 2 72
Seain/presstime
Toomuchtime
Burnedbisterpack
10Machinenotsetpropery
5
Vericationofbatchrecord
7 350Provideavisuadispaytoseetimeeapsed
J.Rodriuez
Visuadispayimpemented(10/11)
10 3 2 60
Seain/presstime
Notenouhtime
Bisterpacknotseaedcompetey
9
Machinenotsetpropery
5
Vericationofbatchrecord
7 315
Provideavisuadispaytoseetimeeapsed
J.Rodriuez
Visuadispayimpemented
9 3 2 54
RPN=riskprioritynumberSOP=standardoperatinprocedure
Failure mode ad effect aalyi for tablet packagig /TABlE2
-
8/3/2019 FailSafe_FMEA
7/7
QPwww.quaityproress.com6
regulated industry: the corrective action and preven-
tiveaction(CAPA)system.
Many organizations devote valuable resources to
aproactiveFMEA.OthersfocusontheCAPAsystem
to manage the ailures within their processes. A select
ew get the maximum benet o combining both.
The true preventive actionsactions implemented
to eliminate the root causes o nonconormances be-
ore they occurcan be easily identied via an eec-
tive FMEA initiative. That means you can identiy the
potential ailure modes and their root causes, and then
implement the actions required to prevent the occur-
rence o those causes.
But, in an imperect world, ailures occur. When
that happens, an organization immediately activates
itsCAPAsystemtoinitiateaninvestigationgearedto
nding the root causes and implementing the appro-
priateCAPAs.4 In the vast majority o situations, those
systems work independently. But do they need to?
True preventive actions can be identied using the
FMEA tool. Then, those actions are incorporated into
the preventive action portion of the CAPA system.
When a failuredoesoccur, theCAPA system isacti-
vated. But because potential ailure modes and their
causes are part o the FMEA initiative, a disconnect ex-
ists.Forthatreason,theCAPAsystemshouldprovide
eedback to the FMEA tool to revise and update the
FMEA accordingly.
Fmea o wo
Oten, organizations ace situations in which the FDA
alertsthemofthewronguseofsometools.Hereare
excerpts rom two warning letters issued by the FDA
regarding the incorrect application o the FMEA tool
during the past decade:
1. Lack o integration o the FMEA process
with a perormance test. An organization received
customer complaints about its tubing, which was as-
sembled incorrectly. As part o the investigation, the
organization conducted a perormance test, the result
o which showed that the incorrect assembly did not
aect the product.
But the organization did not update its FMEA,
which stated that the incorrect assembly compromised
the unctionality o the device.5 As a result, it let itsel
vulnerable to uture occurrences o the same issue.
2. Lack o defnition o the FMEA indexes or
severity, occurrence and detectability. An orga-
nization received customer complaints about device
ailures, so it rushed to develop a reactive FMEA based
solely on the major cause o complaints: contamination.
When it developed the scales to rank severity, oc-
currence and detectability, it did not provide a ratio-
nale or each scale. The organization only provided the
indexes and the calculated RPN.6
These two examples show how the use o FMEA
alone isnt enough to guarantee product quality. While
FMEA is a proactive tool within an organizations risk
management system, it is a dynamic component that
must be handled correctly using the ollowing tips:
PerformanFMEAusingateamworkapproach.
Use a systematicmethod to complete the FMEA
with help rom a process map, and cause and eects
matrix.
Prioritizetheorderinwhichtheprocessinputswill
be evaluated using the FMEA tool.
SortRPNsindescendingorder.
Identifyallpotentialrootcauses.
IdentifyandimplementCAPAsforeachrootcause.
Recalculate RPNs after CAPAs are implemented,
not beore.
Doing so will ensure your FMEA is as eective as
possible and your products are sae in the hands o the
people theyre meant to help. QP
REFEREnCEs1. irol Orzo for sdrdzo, ISO 14971:2007Mdica
dvicsAppicatio o risk maamt to mdica dvics .2. U.s. Food d Dru admro, gudc for idury: Q9 Quly R
Mm, 2006.
3. ibd.
4. Jo Rodruz-Prz, CAPA or th FDA-Ratd Idstry, asQ QulyPr, 2010.
5. U.s. Food d Dru admro, RD Mdcl Mufcur ic. 2/7/11,
www.fd.ov/cc/forcmco/wrlr/ucm242822.hm.
6. U.s. Food d Dru admro, Vpohrm ic. 12-Ju-06, www.fd.
ov/cc/forcmco/wrlr/2006/ucm075943.hm.
JOS RODRguez-PRez is prsidt o Bsiss ex-cc Costi Ic. i Bayamo, Prto Rico. Hard a doctorat i scic rom th uivrsity ograada i Spai. Rodr-Pr is a sior mmbr
o ASQ ad a ASQ-crtifd qa ity ir, aditor,maar, Six Sima Back Bt, biomdica aditor,HACCP aditor ad pharmactica gMP prossioa.H is aso th athor oCaPa for h FDa-Ruld
idury (ASQ Qaity Prss, 2010).
MAnuel e. PeA-RODRguez is a costat atBsiss excc Costi Ic. H ard a JrisDoctor rom Potifca Cathoic uivrsity i Poc,Prto Rico, ad a mastrs dr i irimaamt rom Cor uivrsity i Ithaca, nY.Pa-Rodr is a sior mmbr o ASQ ad aASQ-crtifd qaity ir, aditor, maar adSix Sima Back Bt.
Risk ManageMent