7/25/2019 Failure Investigation Anselmo
1/27
FA ILUREFA ILURE
INVESTIGATION ANDINVESTIGATION ANDROOT CAUSEROOT CAUSE
ANALYSISANALYSISPresented By
Clay Anselmo, R.A.C.President and C.O.O.
Reglera L.L.C.
Denver, CO
7/25/2019 Failure Investigation Anselmo
2/27
CONFIDENTIALPage 2
Lea rn ing Ob je c tive sLea rn ing Ob je c tive s
Understand the Definitions of Failure Investigation
and Root Cause Analysis (RCA)Introduce a common Failure Investigation and RCA
Process
Define 1271 Regulatory Requirements Related toFailure Investigation and RCA
Understand When and Where to Apply Methods and
ToolsTool Introduction and PurposeBest Practices
Provide Guidance on Documentation of Activit ies
7/25/2019 Failure Investigation Anselmo
3/27
CONFIDENTIALPage 3
Definit ionsDefinit ions
Failure Investigation - The process of
understanding the key attributes of a particularfailure and identifying the likely causes of failure.
Root Cause Analysis The process of conductingan analysis to identify the physical, human, and
latent causes of a particular undesirable event.Root causes are specific underlying causes
Root causes are those that can reasonably be identified
Root causes are those that are controllable and fixable
Root causes are those where prevention is possible
7/25/2019 Failure Investigation Anselmo
4/27
CONFIDENTIALPage 4
RegulatoryRegulatory
RequirementsRequirementsWhen is it REQUIRED to formally investigate (and perform
root cause analysis)?
1271.160 (b)(2) Ensuring that procedures exist for receiving, investigating,
evaluating, and documenting information related to core CGTP requirements,
including complaints
1271.160 (b)(3) Ensuring that appropriate corrective actions relating to core
CGTP requirements Documentation must include, where appropriate; (ii) Thenature of the problem requiring corrective action;
1271.160 (b)(6) Investigating and documenting HCT/P Deviations and trends
of HCT/P deviations relating to core CGTP requirements
1271.230 (a) Procedures. You must establish and maintain procedures for
review, evaluation and documentation of complaintsand the investigation ofcomplaints as appropriate.
1271.230 (b) Complaint File. The complaint file must contain sufficient
information about each complaint for proper reviewand for determining
whether the complaint is an isolated event or represents a trend.
7/25/2019 Failure Investigation Anselmo
5/27
CONFIDENTIALPage 5
RegulatoryRegulatory
RequirementsRequirementsWhen is it REQUIRED to formally investigate (and performroot cause analysis)?
1271.320 (c) Review and evaluation of complaints. As soon as practical, youmust review, evaluate, and investigate each complaint that represents an eventrequired to be reported to FDA You must review and evaluate a complaintrelation to core CGTP requirements that does not represent an event requiredto be reported to determine whether and investigation is necessary When no
investigation is made, you must maintain a record that includes the reason noinvestigation was made.
1271.350 (a) Adverse reaction reports. (1) You must investigate and adversereaction involving a communicable disease related to an HCT/P that you madeavailable for distribution.
1271.350 (a)(3) You must, as soon as possible investigate all adversereactions that are the subject of these 15-day reports
1271.350 (b) Reports of HCT/P deviations. (1) You must investigate all HCT/Pdeviations related to a distributed HCT/P for which you performed amanufacturing step.
7/25/2019 Failure Investigation Anselmo
6/27
CONFIDENTIALPage 6
Pro c e ss Ove rv iewPro c e ss Ove rv iew
Key Process StepsGeneral Problem Identification
Gather / Create Samples
Failure Investigation and Experimentation
Examination / Dissection / Physical Testing
Causal Factor Identification
Brainstorming
Use of tools (fishbone, causal factor chart)
Hypothesis Development
Linking Analysis Data to Causal Factors
Hypothesis TestingDOE / ANOVA / Simple Verification
Statistically Based Methodology
Root Cause Identification
Root Cause Map
7/25/2019 Failure Investigation Anselmo
7/27
CONFIDENTIALPage 7
Pro c e ss Flow Cha rtPro c e ss Flow Cha rt
General Problem
Identification and
Basic ProblemStatement
Gather / Create
Evaluation
Sample(s)
Failure Investigation
and Experimentation
Causal Factor
Identification
Hypothesis
Development
Outcome
Consistent
with
ProblemStatement
Revise / Clarify
Problem Statement
to be Specific
Hypothesis Testing
and Verification
DataConsistent
w/
Hypothesis
Data Analysis
Revise Hypothesis
Final Root
Cause
Identification
Start
7/25/2019 Failure Investigation Anselmo
8/27
CONFIDENTIALPage 8
Fa ilu re Inve stig a tio nFa ilure Inve stig a tio n
Common Tools / ActivitiesSample Dissection
Representative / Worst Case Sample Creation
Physical, Chemical, Human Factors Testing and Analysis
Environmental Stress Screening
Process CharacterizationProcedural Review
7/25/2019 Failure Investigation Anselmo
9/27
CONFIDENTIALPage 9
Roo t Ca use Ana ly sisRoo t Ca use Ana ly sis
Common Tools / AnalysisFishbone Diagrams
Pareto Charts, Run Charts, other Control Chart Methods
ANOVA and Design of Experiments
Causal Factor Charting
Root Cause Map
7/25/2019 Failure Investigation Anselmo
10/27
CONFIDENTIALPage 10
Ba sic Po in tsBa sic Po in ts
Use common senseUse tools when they support a comprehensive analysis, not just because they
are available.Right-Size the Investigation / Analysis be sure not to makethe problem more complicated than necessary
Avoid ANALYSIS PARALYSIS
Analyze the likely causes FIRSTUse feedback loops to refine problem statement and causalanalysis
Most problems have more than one root cause
Dont be afraid to ignore the instructionsUse the tools in a way that make sense for the problem. Tools are generic andmay need to be modified to fit your situation.
The problem dictates the tools, not the other way around
7/25/2019 Failure Investigation Anselmo
11/27
CONFIDENTIALPage 11
ProblemProblem
Ident i f icat ionIdent i f icat ionDraft Problem Statement
Begin the process with a simple statement of the problem.
Do not define the cause in the problem statement. Keep an open
mind.
Create / Gather Samples
Gather evidence showing the existence of the problem.Product Returns
Collecting Documents
Creation of discrepant samples
7/25/2019 Failure Investigation Anselmo
12/27
CONFIDENTIALPage 12
ProblemProblem
Ident i f icat ionIdent i f icat ionSample Evaluation
Samples may be tissue products, packaging, test results, records, etc.
Product Review
Sample Dissection
Physical TestingChemical
MicrobialPhysical
Functional
Records Review
Dates
ChangesRelationship to other relevant records
Chronology / History
Interviews
Duplicates in other locations
7/25/2019 Failure Investigation Anselmo
13/27
CONFIDENTIALPage 13
ProblemProblem
Ident i f icat ionIdent i f icat ionProblem Statement Revision
Inspection / Testing / Review
Are results consistent with Initial Problem Statement?
Updated Problem Statement
Change Scope
Expand Scope
Reduce Scope
BE SPECIFIC based on observable data
7/25/2019 Failure Investigation Anselmo
14/27
CONFIDENTIALPage 14
Id e ntify Lik e lyId e ntify Lik e ly
CausesCausesUtil ize Analysis Tools
Cause and Effect (Fishbone) Diagram
Causal Factor Chart
Process Flow Chart
BrainstormDo not rule out causes or assess likelihood
Look at all elements of the process
When in doubt start with 5 standard categories
Process
Environment
Materials
Human Factors
Equipment
7/25/2019 Failure Investigation Anselmo
15/27
7/25/2019 Failure Investigation Anselmo
16/27
CONFIDENTIALPage 16
HypothesisHypothesis
De ve lo pm ent / Te stingDeve lo pm ent / Te stingCausal Analysis
Identify list of most likely causes (top 3 to 5)
Develop / utilize a rating scale or rating scheme (i.e. hazards evaluation)
Be specific so that the contribution of each cause can be evaluated
Hypothesis Development
Develop one or more scenarios to describe the failure based on thefailure investigation data and the causal analysis
Must be able to be evaluated / tested
Hypothesis Testing / Evaluation
Simple Inspection / Review of DataExperimental Design
Worst Case Analysis
ANOVA
DOE
7/25/2019 Failure Investigation Anselmo
17/27
CONFIDENTIALPage 17
HypothesisHypothesis
De ve lo pm ent / Te stingDeve lo pm ent / Te stingData Analysis
Appropriate Statistical Methods
Does the Data Support the Hypothesis?
Yes Continue to Corrective / Preventive Action
No Revise Hypothesis or Causal Analysis
Maybe Further Experimentation is necessary
Hypothesis RevisionRevise Causal Analysis based on Experimental Results
Revise Hypothesis Statement
Retest as Necessary
7/25/2019 Failure Investigation Anselmo
18/27
CONFIDENTIALPage 18
Pa re to Cha rtsPa re to Cha rts
Calibration Errors
13
2631
44 51
92
0
102030405060
708090
100
Janu
ary
Febr
uary
March
April
May
June
Month
Normalized
Quantity
7/25/2019 Failure Investigation Anselmo
19/27
CONFIDENTIALPage 19
Contro l Cha rtContro l Cha rt
Average Production Hours
40 40 40 40 4044
46 4650 50
59
50
39 4040
25
30
35
40
4550
55
60
65
22 24 26 28 30 32 34 36
Week Number
Hours Mean
UCL
LCL
7/25/2019 Failure Investigation Anselmo
20/27
CONFIDENTIALPage 20
M isc . Cha rtsM isc . Cha rts
Bar Temperature by Position
226
227228
229
230
231
232233
234
235
236
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Position
Temperature
(C)
7/25/2019 Failure Investigation Anselmo
21/27
CONFIDENTIALPage 21
Roo t Ca useRoo t Ca use
Ident i f icat ionIdent i f icat ionSummarize Results of Hypothesis Evaluation
Identify all root causes that significantly contribute to failure
Identify significant interactions between root causes
Prioritize root causes based on impact
Develop recommendations for:
Corrective Action (remedial action to correct items / units / tissues alreadyexhibiting failure)
Preventive Action (action to prevent recurrence of the problem)
Formally Document the Results
7/25/2019 Failure Investigation Anselmo
22/27
CONFIDENTIALPage 22
Roo t Ca useRoo t Ca use
Sum m a ry Ta b leSumm a ry Ta b le
Bag Seal Failure Root Cause Analysis Summary TableRoot
Cause ID Description
Impact
(H/M/L) Siginificant Interactions Recommendations
1 Temperature Variation on Sealer Bar H
Temperature variability increases
dramatically with decreased sealingcycle time Replace sealer bar, increase calibrationfrequency, perform validation
2 Sealer Cycle Time H
Decreasing cycle time increases
temperature variation on sealer bar Specify minimum cycle time as 30s in SOP
3 Material Thickness M
Lower end of current specification does
not seal correctly at highest temperature
seen on sealer bar
Revise specification, implement thickness
measurement at RI
4 Accuracy of Calibration M
Calibration variation accounts for 20%
of overall bar temperature variation
Revise instrumentation specifications for
use during sealer calibration. Improve
instrument accuracy to 1:10 ratio
7/25/2019 Failure Investigation Anselmo
23/27
CONFIDENTIALPage 23
Documenta t ionDocumenta t ion
Failure Analysis / Root Cause Analysis can be
included in:Complaints / Adverse Reactions / HCT/P DeviationsInternal Deviations (Errors, Incidents, Accidents)
Validation Failures
Inspection Failures / Non-Conforming ProductOther Corrective / Preventive Action Projects
Key Elements of Documentation
Specific Problem StatementTraceability / Tracking Information
Failures
Test Samples
7/25/2019 Failure Investigation Anselmo
24/27
CONFIDENTIALPage 24
Documenta t ionDocumenta t ion
Key Elements of Documentation (cont.)Failure Investigation
Observational / Inspection Data
Test InformationProtocol
Data Analysis
Results
Conclusions
Causal Factor Identification
Hypothesis
Hypothesis Testing and Verification
Protocol
Data Analysis
Results
Conclusions
Root Cause Summary (report, table, etc.)
7/25/2019 Failure Investigation Anselmo
25/27
CONFIDENTIALPage 25
Documenta t ionDocumenta t ion
Po in ts to Consid e rPo in ts to Consid e rHow detailed do I need to be in my documentation?
Detailed enough so that someone from the outside can follow your
logic and could replicate your results.
Detailed enough to meet specific regulatory requirements (i.e.
complaints, adverse reactions, HCT/P deviation)
Do some areas need to be more detailed thanothers? If so, what areas should I emphasize?
Yes:
Problem Description
Traceability
Failure Investigation
Root Cause Summary
7/25/2019 Failure Investigation Anselmo
26/27
CONFIDENTIALPage 26
Documenta t ionDocumenta t ion
Po in ts to Consid e rPo in ts to Consid e rDo I have to go through every step for every investigation?
No, use common sense and right-size the activities for the project.
Wherever the term investigation is used in a regulation or standard, a certainlevel of rigor is implied. Again, focus on:
Problem Description
Traceability
Failure Investigation
Root Cause Summary
For Complaints (not reportable to FDA), when can I avoid
doing an Investigation and Root Cause Analysis?
When a similar complaint has been fully investigated and you can refer to such
an investigation.
When the event, based on documented data analysis, is an isolated event and
does not represent a trend (this does not relieve requirements for review and
evaluation for complaints related to core CGTP requirements.)
7/25/2019 Failure Investigation Anselmo
27/27
CONFIDENTIAL
Documenta t ionDocumenta t ion
Po in ts to Consid e rPo in ts to Consid e rWhen in doubt, write it down!
If you were an Auditor, what would you want to seeto ensure activities were done correctly?
If CAPA is done by another group or person, what
will they need to complete the project?If I were gone, would someone be able to follow my
process and reach the same conclusion?