Martin Mayer

GMP Series

Failure Management
in a GMP Regulated
Environment
Effective Root Cause Analysis

PDF Download
Excerpt from the GMP Compliance Adviser
 1

Contents

Failure Management 2
1 Root cause analysis as a quality management system enabler 2
2 Root cause analysis in the product life cycle and pharmaceutical QM system 4
2.1 Pharmaceutical development 5
2.2 Technology transfer 6
2.3 Commercial manufacture and marketing 6
2.4 Cessation of commercial manufacture and marketing 6
3 Reactive and proactive root cause analysis 7
4 Effort, rigour and formality 7
5 Root cause analysis as a process of continual improvement 8
5.1 Failure situation 9
5.2 Team building 10
5.3 Evaluation of the failure situation, data collection 10
5.4 Hypothesis and analysis strategy 11
5.5 Selecting the methods 12
5.6 Determining the root cause(s) 12
6 Corrective actions 14
6.1 Defining actions 14
6.2 Implementing the actions 15
6.3 Monitoring the effectiveness of actions 16
7 Methods and tools of root cause analysis 17
7.1 Principles 17
7.2 Basic tools for facilitating root cause analysis (enablers) 18
7.3 Statistical methods 19
7.4 Questioning technique (Five Whys, 5W) 22
7.5 Causal chain analysis 23
7.6 Change analysis (CA) 25
7.7 Barrier analysis (BA) 26
8 Definitions 27

Contributor 29

Index 30

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 2

Failure Management

Here you will find answers to the following questions:

• What is the role of root cause analysis in a modern quality management system?
• During which phases of the product life cycle should root cause analysis be used, and in what
ways is it applied?
• Can root cause analysis also be used proactively?
• How is the degree of effort, rigour and formality required during the root cause analysis
determined?
• How should the root cause analysis be conducted?
• What are the different causes of failure and how do they differ from one another?
• What types of action are there and how are they selected?
• How should the actions be implemented and who carries the responsibility?
• How is the effectiveness of the implemented actions monitored?
• What are the different methods and tools of root cause analysis?
• Which method best suits a particular situation?
• How are the different methods applied?

1 Root cause analysis as a quality management system enabler

Pharmaceutical companies (authorisation holders & manufacturers) should have processes and sys-
tems in place as part of their quality management system that support the identification of the causes
of detected and potential failures. The processes and systems should encompass the identification
and implementation of corrections, corrective actions and preventive actions as well as the monitor-
ing of the effectiveness of the defined actions.
ISO 90001 defines nonconformity as the non-fulfilment of a requirement2, whereas conformity is
the fulfilment of a requirement. A failure is therefore a non-conformity. In a pharmaceutical company,
the word failure means non-compliance with quality management system requirements, GMP Guide-
lines, etc.
Failures can be indicated or detected as a result of complaints, recalls, internal product rejections,
deviations, process capability analyses, product quality analyses, trend analyses, audits, official inspec-
tions and other indicators.
Effective actions can only be identified if the exact cause(s) of the failure is/are known. Root cause
analysis (RCA) requires a structured and methodical approach, and the rigour and formality of the
approach should be appropriate for the respective failure. Root cause analysis should be based on sci-
entific principles and comprehensive scientific and technical knowledge of the process and product.
If the causes of a detected failure are not analysed or if the causes are not discovered despite close
examination, effective actions cannot be taken and the same failure (or a similar failure) will (re)occur.
The knowledge gained through root cause analysis should deepen the understanding of the pro-
cess and the actions taken should lead to significant product and process improvements. The actions
taken can include corrections, corrective actions or preventive actions.
Root cause analysis is an essential part of the continuous improvement process (CIP). Continuous
improvement is the central requirement and objective of every pharmaceutical quality management
system.
Root cause analysis is closely related to quality risk management/QRM. The methods and tools of
root cause analysis can and should be anchored to the same points of the quality management sys-

1. ISO 9000:2005 Quality management systems – Fundamentals and vocabulary

2. Requirement – need or expectation that is stated, generally implied or obligatory (ISO 9000:2005).

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 3

tem of a pharmaceutical company and integrated into the same processes as quality risk manage-
ment.
As is the case with quality risk management, root cause analysis and the determination of appropri-
ate actions should be used during the entire product life cycle.
The implementation of the root cause analysis process and the use of (formalised) root cause anal-
yses is no longer just common sense or a requirement in the standards affecting quality management
systems (e. g. ISO 9000, ISO 9001 et seqq.). It is now explicitly required by various GMP regulations and
guidelines. Root cause analysis is, for example, listed as a requirement in Section 1.4 (xiv) of the
EU GMP Guidelines and in ICH Q10 (e. g. Section 3.2), which is now included in Part III of the EU GMP
Guidelines. Both documents include a description of the most important elements of a pharmaceuti-
cal quality management system.

Figure 1 shows the key elements, enablers and objectives of a modern quality management system.
Root cause analysis is given a central role as an essential enabler.

Management responsibility
Product life-cycle management
Deviation management
Complaints management
QMS key Change management
Management of corrections, corrective and preventive actions
elements Measurement of the process performance & product/service quality
(key performance indicators)
Management review

Regulatory know-how
Knowledge management
Process & product
understanding
Essential
Management of training and qualification
QMS enabler
Root cause analysis

Quality risk management

Efficient product realisation

QMS objectives Establishment and maintenance of the state of control

Continual improvement

Figure 1 Root cause analysis in a pharmaceutical quality management system – an essential

enabler

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 4

2 Root cause analysis in the product life cycle and pharmaceutical

QM system
The life cycle of a medicinal product can be divided into four phases:
• pharmaceutical development
• technology transfer
• commercial production and marketing
• cessation of commercial production and marketing

As the individual phases of the product life cycle are passed through (see figure 2), the understanding
of the product and process increases and the subsequent cost of a failure is minimised. This is a result
of increased knowledge acquired through practical experience.

Root cause analysis

Cessation of
Commercial
Pharmaceutical Technology commercial
manufacture &
development transfer manufacture &
marketing
marketing

Quality risk management

Process & product understanding

Knowledge management

Figure 2 Root cause analysis in the product life cycle

The earlier a root cause analysis is carried out during a product life cycle, i. e. the earlier the causes are
eliminated, the lower the subsequent cost of a failure (see figure 3).

Cost resulting
from failure
without consequent
use of root cause
Costs resulting from failure

analysis

Reduction in costs
resulting from failure
when root cause
analysis is used
consequently

Product life cycle

Pharmaceutical Technology Commercial Cessation of commercial

development transfer manufacture manufacture & marketing
& marketing
Figure 3 Reducing the subsequent cost of a failure in the product life cycle through consequent
use of root cause analysis

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 8

sary, on whether a less formal and comprehensive root cause analysis is sufficient, or whether a root
cause analysis is unnecessary. Decision trees for the respective procedures can be helpful: the decision
and rationale that led to the decision should be documented, when applicable.

5 Root cause analysis as a process of continual improvement

Root cause analysis as a process is described in the following section (see figure 5). The process flow
can be used for a reactive and proactive determination of (potential) causes of failure.

Recognised/potential
failure situation(s)/pattern

Team building

Evaluation of failure
situation/pattern

Available data
is adequate?

No
Yes
Collection of additional
information/data

)ZQPUIFTJT
BCPVU
DBVTF

PG
GBJMVSF

Analysis strategy & selection

PG
UPPMTNFUIPET
Direct cause(s)

Root cause analysis Contributing cause(s)

Correction(s) Root cause(s)

Corrective action(s) Determine actions

Preventive action(s)

Implement actions
No

Monitor efficiency

Actions
effective?

Yes

End

Figure 5 Root cause analysis as a process of continual improvement

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 19

7.3 Statistical methods

Histograms and Pareto charts

Histograms (figure 7) and Pareto charts (figure 8) are very good tools for the statistical evaluation of
the failure frequencies and causes.
Failure frequency
od on A

n B
Pr ne C

Pr ct A

Pr ct B

Ea c t C

gh ift

ft
La hift

hi
tio e
Pr cti ne

Ni sh
uc lin

ts
s
u

u
u
li
li

te
rly
od

od
od
od n
Pr ctio
u
u
od

Point of occurrence
Pr

Figure 7 Histogram – failure frequencies & points of occurrence

Histograms are also very good for assigning the failure frequencies to the different points of occur-
rence in the company and for visualising other relationships. It is normally very difficult to do this
using lists.
For example, Pareto charts can be used to analyse the distribution of the causes of failure for the
total number of failures inside a specific period of time. Failure bursts in particular areas, processes,
etc. indicate that there are problems in the system. In this case, the causes should be examined using
other methods.

According to the Pareto3 principle (also referred to as the Pareto effect, Pareto distribution or simply
the 80/20 rule), 80% of the failures of a failure population n are caused by 20% of the discovered
causes of failure. This means that 80% of failure situations are removed if only 20% of the causes of
failure are eliminated. The Pareto principle also applies inversely to the effort required to eliminate the
cause of failures. 80% of the causes of failure can be eliminated using 20% of the total effort required
for the removal of all the causes of failure. This also means, by implication, that in order to remove the
other 20% , 80% of the total effort is required.

3. Named after Vilfredo Pareto (1848–1923), the so-called Pareto distribution describes a statistical phenomenon in which the
contribution of a small number of high values in a value set is larger than the contribution of a greater number of the smaller values
in this value set.

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 22

7.4 Questioning technique (Five Whys, 5W)

The so-called Five Why questioning technique is a basic tool of root cause analysis. The questions are
used to determine the cause-effect relations. The questioning technique starts at the failure situation/
failure pattern and asks what caused or may have caused the situation.
The questioning regarding the cause of the situation continues until no further question can be
asked that delivers a meaningful answer about the cause (see figure 10).

Failure

Why did the failure occur?

Direct cause of failure

Why?

Contributing/direct
Sequence

cause of failure
Why?

Contributing/direct
cause of failure
Why?

Contributing/direct
cause of failure
Why?

Root cause

Failure causal chain

Figure 10 Questioning technique (5 Whys)

It is essential, as with any root cause analysis, to identify the failure situation/failure pattern as clearly
as possible, i. e. the problem whose cause is to be discovered must be fully known. The answers to the
questions must be honest and the persons involved must be committed to determining the real root
cause. When individual persons or groups of persons are affected, questions regarding the causes of
failure can be unpleasant, especially when they are related to suspected or identified incorrect behav-
iour. Sensitivity is definitely required. The chain of questions is frequently cut short, i. e. only part of the
causal chain of events is determined and the root cause is not discovered. In cases of suspected or
actual incorrect behaviour, human error is often diagnosed too early and the line of questioning dis-
continued although the causes of incorrect behaviour or human error should be investigated because
they are the real root causes of a large number of deviations in a GMP-regulated environment. These
causes can include inexperience (poor training and induction), overburdening (too few personnel to
carry out the required tasks) and a number of related causes. Incorrect behaviour and human error can
also be caused by poor leadership. The failure culture in a company is decisive in these situations: the
will to learn from failures and to improve the situation in order to prevent further failures.

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 26

Factor of change State of State of Difference/ Cause Impact

(examples) failure control change

What?
• circumstances
• framework conditions
• activities
• equipment
• process
When?
• point of recognition
• point of occurrence
• routine activity
• non-routine activity
• shift
Where?
• point of occurrence
• process/product
• company department
• production line, machine
How?
• point of recognition
• point of occurrence
Who?
• group
• individual person
Figure 14 Change analysis work sheet/question list

7.7 Barrier analysis (BA)

Barrier analysis is a method that is particularly suitable for analysing the failure of physical, technical,
process-related, administrative and procedural barriers or control points. Barriers or control points
are used to prevent or at least detect failures. Process flow charts can be used as an aid.
The method is used to determine why the barriers or control points failed, and what must be done
in order to prevent recurrence of the specific failure situation.

Barrier analysis moves backwards from the state of failure to the previous state of control and checks
whether the failure of one or more control points can have led to the specific failure situation (see fig-
ure 15).

If the failure of a barrier/control point is established, a suitable additional method of root cause analy-
sis is used to determine the direct causes, contributing causes and root causes.
An excellent understanding of the product and process is a prerequisite when using this method
and the certainty that one or more control points for preventing failures or detecting them in good
time were/are available.
If there were no control points in place, the reason for this should be investigated. The HACCP
method (hazard analysis & critical control points) is suitable for identifying the necessary control
points.

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Failure Management 27

Duration/Procedure

control point

control point

control point
Barrier/

Barrier/

Barrier/
Control
Failure
status

"QQMJDBUJPO
"QQMJDBUJPO
"QQMJDBUJPO

NFUIPEUPPM
NFUIPEUPPM
NFUIPEUPPM

PG
SPPU
DBVTF
PG
SPPU
DBVTF
PG
SPPU
DBVTF

BOBMZTJT BOBMZTJT BOBMZTJT

Figure 15 Barrier/control point analysis

8 Definitions
Point of occurrence (PoO) – the point at which a failure situation actually occurs (for the first time).
The point of occurrence can, for example, refer to a specific step in a process chain, a moment in time
or a combination of both. The point of occurrence can also be a company organisation, a function or
something similar.
Please note: the point of occurrence and the point of recognition can, but must not correspond, i.e.
a failure can occur (for the first time) before it is recognised.

Contributing cause – a cause or a combination of causes that contribute(s) to the occurrence of a

specific failure situation, but could not have caused the failure situation (alone).

Failure (non-conformity) – non-compliance with a requirement (ISO 9000:2005).

Cause – a state, circumstance or condition that causes or strengthens an effect/impact.

Root cause – the cause of failure(s) or a combination of causes that have to be corrected to prevent a
failure situation and/or similar related situations (re)occurring.

Correction – an action to remove a detected failure (ISO 9000:2005).

Please note: a correction can be combined with a corrective action.

Corrective action – action taken to eliminate the cause of a detected failure or other undesirable sit-
uation (ISO 9000:2005).
Comment 1: A failure that occurs can have more than one cause.
Comment 2: A corrective action is taken to prevent the recurrence of a (detected) failure, and a pre-
ventive action is taken to prevent the occurrence of a (possible) failure.
Comment 3: Corrections and corrective actions differ.

Direct cause – a cause that leads directly to a specific failure situation, i.e. it causes it directly.

Preventive action – action taken to eliminate the cause of a potential failure or other potential unde-
sirable situation (ISO 9000:2005).
Comment 1: A potential failure can have more than one cause.

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Contributor 29

Contributor
Martin Mayer
martin.mayer@fresenius-kabi.com

Vice President
Fresenius Kabi Deutschland GmbH,
Bad Homburg vor der Höhe

As Vice President of Quality Assurance, Martin Mayer has been responsible for managing the quality
organisations of the Global Operations Generic Drugs & Standard Solutions production sites since
2013. His areas of expertise include quality management standards, GMP compliance, inspection
management and risk management.

Mr Mayer started his professional career in the pharmaceutical industry in 1992. He first worked for
Fresenius in Quality Assurance. After joining Pharmaplan, he worked as a senior project engineer for
several years in a number of different areas, including process optimisation, inspection support, train-
ing, qualification and validation as well as design reviews and facility planning. Between 2007 and
2009, he worked for Fresenius Kabi Product Partnering. As Director of Contract Manufacturing, he was
responsible for contract manufacturing projects. In 2009, Mr Mayer joined the global quality manage-
ment organisation of Fresenius Kabi AG. As Vice President of QM Systems & GMP Compliance, he was
responsible for the global quality management system, GMP compliance standards and internal
auditing for medicinal products and medical devices.

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing
Index 30

Index
C - causal chain analysis 23
cause - change analysis 25
- category 13 - commercial manufacture 6
- contributing cause 13 - continual improvement 8
- determination 12 - control point 26
- direct cause 13 - data collection 10
- root cause 13 - definitions 27
commercial manufacture - effort 7
- root cause analysis 6 - enabler 18
continual improvement - failure situation 9
- root cause analysis 8 - formality 7
control point - hypothesis 11
- root cause analysis 26 - management review 16
corrective action 14 - method selection 12
- action categories 14 - methods 17
- defining actions 14 - pharmaceutical development 5
- implementing actions 15 - proactive use 7
- monitoring actions 16 - process flow 8
- process owner 9
F - product life cycle 4
failure - QRM methods 17
- definition 2 - quality management system 2
- indicator 2 - quality risk management 5
failure cause - questioning technique 22
- assignment 16 - reactive use 7
failure management 2 - report 13
failure situation - required information 11
- assignment 16 - requirement 3
- documentation 9 - statistics 19
- immediate actions 10 - team building 10
- point of occurrence 9 - technology transfer 6
- point of recognition 9 - trend analysis 13
- trigger 7
H
human resource management 29 S
statistics
M - root cause analysis 19
management review
- root cause analysis 16 T
technology transfer
P - root cause analysis 6
pharmaceutical development trend analysis
- root cause analysis 5 - root cause analysis 13
product life cycle
- root cause analysis 4

Q
quality management system
- root cause analysis 2

R
root cause analysis
- analysis strategy 11
- barrier analysis 26

Failure Management in a GMP Regulated Environment – Excerpt from the GMP Compliance Adviser © Maas & Peither AG – GMP Publishing