Académique Documents
Professionnel Documents
Culture Documents
To my wife Sandra
Safety Culture
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system or transmi.ed in any form or by any means, electronic, mechanical, photocopying,
recording or otherwise without the prior permission of the publisher.
John Bernard Taylor has asserted his moral right under the Copyright, Designs and Patents
Act, 1988, to be identified as the author of this work.
Published by
Gower Publishing Limited Ashgate Publishing Company
Wey Court East Suite 420
Union Road 101 Cherry Street
Farnham Burlington,
Surrey, GU9 7PT VT 05401-4405
England USA
www.gowerpublishing.com
5 Epilogue 195
Concluding Remarks 195
Summary 197
Appendix I 199
Appendix II 203
Appendix III 205
Further Reading 209
Index 213
List of Figures
Risky Business
Psychological, Physical and Financial Costs of High Risk Behavior in Organizations
Edited by Ronald J. Burke and Cary L. Cooper
Thanks are given to the companies and organisations noted in this text whose
data enables others to learn from their knowledge and research. The narrative,
commentary and interpretation of data are entirely the authors and do not
necessarily represent the views of the sources.
Thanks are offered to colleagues for their robust comment and critique
at various stages in the development of the text. Further, there have been
numerous discussions with safety and safety-culture specialists nationally
and internationally. Nevertheless, any errors, technical or grammatical, remain
the author’s responsibility. Finally, the book will hopefully assist hard pressed
managers in whose hands we often place our safety as members of their
workforce, members of the public, or visitors to their facilities.
This page has been left blank intentionally
Preface
This safety-culture text draws information from many existing sources and
is presented in self-contained chapters.
J.B. Taylor
This page has been left blank intentionally
1
Organisational Safety-Culture
Theory
Safety-Culture Theory
The concept of safety-culture emerged from the analysis of the 1986 Chernobyl
nuclear power plant accident. Having delivered a severe political and social
shock to Europe it was an imperative that the causes of this event were fully
understood.
From the technical inquiries into the Chernobyl incident with a concentration
on the ‘person dimension’, it emerged that inadequate organisational safety-
culture was a possible major contributor to the accident. Retrospectively, it was
also considered a possible contributor towards many historic accidents where
the root cause was not necessarily due to less than optimum engineering design
or equipment failures, but people’s ‘poor’ human performance.
Safety Culture
The inquiries suggested that the designers and operators ‘good’ safety-
beliefs, attitudes and behaviours act as additional accident barriers. The
human performance element, safety-culture, in accident causation cannot be
ignored. Safety culture emerged from the Chernobyl experience as a complex,
psychological, human behavioural phenomenon that needed to be addressed.
Some safety-culture definitions are given in Appendix I.
2. middle managers
3. supervisors
The employees, or the staff, are the aggregate of the workforce and
management.
• beliefs
• espoused values
• attitudes
• artefacts
• behaviours.
BEHAVIOURS
ATTITUDES
ARTEFACTS
ESPOUSED VALUES
BELIEFS
behaviours
artefacts
espoused values
basic assumptions
Beliefs
Beliefs are emotions and assumptions that something is true. They can become
deep seated to the extent that a person unconsciously subscribes to them.
Because they are deep seated and fundamental they are usually stable. ‘Good’,
shared safety-beliefs, if unconsciously adopted with associated good safety-
behaviours, can deliver business success. On the other hand, ‘poor’ shared
beliefs can give a perception of business success. Usually this is illusionary as
the associated poor safety-behaviours may eventually lead to a severe event or
the progressive deterioration of the business.
We believe:
• The safety of staff, contractors and the public is our number one
priority in all circumstances
Some mature organisations have only one or two founding beliefs with
others arising naturally from these. Beliefs are usually supported by safety-
culture values that are conditions and actions that are held in high esteem by an
organisation because they assist in fulfilling beliefs. As discussed later, values
are regularly espoused or spoken by an organisation.
In general terms, beliefs are not inherent to individuals; they are learnt,
shared and arise from a common experience of organisational ‘survival’.
Although reflected in espoused values, artefacts and personal attitudes, beliefs
become overtly observable through behaviours. Where beliefs are shared and
reinforced with values and artefacts a sense of community and group cohesion
arises.
Artefacts, spoken values and attitudes that are supportive of good safety-
beliefs are usually indicative of an organisational commitment to safety
and observable good organisational safety-behaviours should contribute to
confirming that there are good shared safety-beliefs.
One should not overwork the generic model, if indeed it can be considered
truly generic. There are internal organisational feedback loops and external
effects.
These are fulfilled by various behavioural tactics. These adopted strategies can
support or hinder line managers, peers or their subordinates in shaping the
culture and stabilising beliefs. The employee’s age distribution, experience, the
exercising of perceived organisation status, or the strength of an organisation’s
sub-cultures can influence beliefs and behaviours. An open dialogue culture
is an approach by which an organisation can ensure the human social needs
are managed and the adopted strategies are directed to positively supporting
shared safety-beliefs. Finally, superimposed on the broader feedback loops
is the influence of the senior management through their specific business
requirements, safety expectations, resource control and authority.
An individual’s beliefs are formed from life’s events. They arise, inter alia,
from interactions within the family, through formal education, friendships
and workplace colleagues. An individual’s behaviours are an expression of
accumulated long-term beliefs and supporting values which stabilise and if
delivering ‘success’ become resistant to change.
safety-beliefs arise at the hierarchy’s highest level, are learnt, impressed upon
the staff and become shared. From these adopted beliefs arise the behaviours
that may enhance organisational safety or, if the beliefs are inappropriate, be
detrimental.
Through their beliefs the management may have created a climate of limited
care towards staff safety. Many industrial examples can be given. Because of
the belief that safety-performance will be adequate with just legal compliance,
managers begin to pay scant attention to operational safety and overtly focus
on production. In such a postulated situation if management’s organisational
safety-beliefs are not supporting good safety, a condition can arise where the
management’s perception of the safety status of the plant becomes misaligned
with that of the workforce. The workforce is fully aware of the plant’s hazardous
chemicals and places top priority on safety. In this scenario the workforce safety-
beliefs are not aligned with the management’s beliefs. Productivity is essential
10 Safety Culture
but secondary. For the workforce the priority given to chemical handling safety
has become entrenched possibly by experiencing first-hand the consequences
of mismanaging the hazard. In their own self-interest the workforce may have
established ad hoc safety-behaviours for handling the chemical plant. These
become learnt behaviours and possibly independent of any organisational
management system. The behaviours, good or poor, are adopted for ‘survival’
and passed down to each workforce generation. The workforce acts as a sub-
culture. Perversely, the sense of self-preservation and the sub-culture’s imposed
safety rules may be all that keeps the business from a major event.
Most will recognise a ‘fear culture’ and the following are possible
consequences from adopting a questioning and challenging behaviour within
such a culture:
• Appearing to be stupid/silly
All these issues are natural human frailties. However, if ‘fear’ exists,
matters of safety importance in facility design, build or operation may pass
unchallenged or be suppressed.
The converse of this scenario is true. Here there would be a strong senior
management commitment to integrate safety into all business activities.
Through the business strategy and documented processes the senior team focus
on the fundamental belief in the importance of operational safety as a business
risk to be managed. Emerging from this safety-belief will be observable good
safety-behaviours, safety attitudes and an overt commitment by managers to
an open culture. This scenario is a frequently observed good practice in high-
12 Safety Culture
Espoused Values
Safety values are spoken but they can also appear in documents, an intranet
or posted around a facility. In this way they become embedded organisational
artefacts. Safety values like beliefs need to be specifically generated, owned
and shared by an organisation. Developed with senior management, this is an
important cross-organisational activity in which most employees can engage.
The beliefs and values, as an integral part of a management system, would
be periodically reviewed and rejuvenated to meet the changing needs of the
business.
We value:
• That everyone is responsible for safety, our own, others’ safety and
the protection of the facility
• That respect is given to all safety views as everyone has the right to
question and report safety issues
• That all events and near misses are reported as we recognise that
even minor injuries or events are important
It is noted that the values are typically associated with actions enabling
safety-belief fulfilment and, inter alia, will develop into expected organisational
human performance behaviours.
When stated, values may lead an observer into misunderstanding the real
safety-beliefs and hence the true organisational culture status. Values need to
be analysed with caution and set in the context of the observed employee
behaviours. If there is misalignment between behaviours and the stated
values this may indicate that the deep-seated safety-beliefs are not shared
or fully supported within an organisation. Here, values may be espoused for
a purpose other than a true demonstration of organisational commitment
to safety. They may be to appease a regulator, the public or possibly the
workforce.
ORGANISATIONal SAFETY-CULTURE THEORY 15
The converse is true. Good, shared, owned and understood safety values
will contribute to good safety-behaviours and provide business benefit.
Attitudes
• Heroics
• Invulnerability
• Safety questioning
• Safety challenging
Artefacts
Formal artefacts:
• Safety awards
Informal artefacts:
Artefacts can generate pride and give a visual association with the
organisation’s values. As can be regularly seen with familiar brand names, there
is a psychological power in symbols. They assist in developing organisational
‘team spirit’ and group cohesion.
Some everyday examples of artefacts are given in Figures 1.3 and 1.4.
In Figure 1.3, the artefacts are the people’s uniform, the lines of soldiers and
the weapons. These give clarity that there is a cohesive ‘team’. The artefacts also
demonstrate purpose, tradition and learnt common beliefs; they are military.
The uniforms and the ordered ritual suggest a disciplined organisation. To
achieve this, the artefacts reflect a command and control management regime
with a rigid culture of compliance. By adopting a particular command culture
and being a cohesive and obedient body the group appears successful.
outcome of this could be plant that is operated, not by complying with the
management system, but by ad hoc local rules, through custom and practice
and with a minimal investment in safety requirements. The artefact suggests
a strong commitment to safety. However, in this case it is a misleading safety-
culture element that is not used as a management tool for lowering safety
risks but produced for other reasons; possibly to pacify legal or regulatory
requirements.
Behaviours
Culture has generally been described as deep, broad, and stable. Also, having
a large psychological and social element it is not a superficial phenomenon.
Further, culture stability infers development over time, with its robustness
tested against many internal and external factors. Because of this, detailed
behaviours at each hierarchical level in an organisation cannot be generically
prescribed as they emerge from the shared beliefs, become tacit rules and lead
to good safety behavioural awareness. Detailed behavioural sets emerge from
within an organisation on a platform of safety-beliefs that meet the business
needs.
Middle managers:
Supervisors:
Workforce:
There are two essential factors that are fundamental to safety behaviour
implementation. First, antecedents have to be in place to enable individuals to
implement agreed role safety-behaviours and second, there has to be feedback
regarding the adequacy of implementation. These two factors can on occasions
be neglected and safety-beliefs remain unfulfilled and safety values not upheld.
In the speeding example, if the sign had a speed camera attached the
consequences for inappropriate behaviour could be early, certain and most
unwelcome. The camera suggests modified behaviour is important to the
local community who value enforcing reduced speed in the interests of safety.
Conversely with no camera present the consequences to a driver of failing to
comply become uncertain and distant. As there are essentially no consequences,
the sign – the antecedent – will be regularly ignored if safety is not a driver’s
priority.
The senior managers are taking action to change middle managers’ current
poor safety behaviour of ‘avoiding behavioural observations’ to the positive
safety behaviour of ‘carrying out regular observations across a sample of all
employees’.
28 Safety Culture
In Table 1.1(a) the current consequences of avoiding visiting the plant arise
early and are not distant. The consequences of this inaction are welcomed by
the middle managers. Because the current antecedents are not enablers and
the visit-avoidance consequences are broadly beneficial, both these factors
contribute to ensuring that the middle managers stay off the plant; the middle
managers avoid doing observations.
Table 1.1(b) is for the same scenario but presents the senior managers’
agreed antecedents and the consequences to bring behavioural change.
That is, a change to where middle managers ‘frequently carrying out safety
behavioural observations’. To change behaviours the middle managers need
senior management leadership through upholding the antecedents and
consequences.
In Table 1.1(b), the senior managers have assessed the original antecedents
and consequences of Table 1.1(a), and changed them quite distinctly. Through
this revision the middle manager is now enabled to carry out the changed
behavioural expectation.
Table 1.1(b) could imply that there is an element of pressure from senior
managers on middle managers. This will not be the case. The proposed change
strategy and the change method will have been agreed such that:
• Safety is the first item on the agenda and receives as much attention
as other business agenda items
• A charismatic leader.
examples available indicate that irrespective of the integrity and quality of the
engineering or procedural protective systems, they are only as effective as the
people ‘minding’ them.
Research shows that individuals are generally quite poor at estimating risk
with a tendency to underestimate. On a hazardous facility if non-compliance
persists, serious incidents will occur at an unacceptable frequency.
There were caveats placed on the concept of the layered generic model due
to the socio-technical complexity of culture. However, the generic nature
raises opportunities for how a safety-culture can be observed, ‘measured’ or
reviewed. Originating from Edgar Schein’s business culture studies, a strength
of the model is in the universality of the elements of beliefs, values, artefacts,
attitudes and the shaped behaviours. It is suggested that for most cultures – for
example, political, religious, tribal or company – some or all of the elements
will be identifiable. These culture elements emerge consistently as essential
vehicles to achieve a compelling visions, organisational goals and business
success.
It is evident that not all cultures are the same and are observably different
even though they have the generic model elements. Culture differences emerge
because different social or industrial grouping develop distinctive characteristics
and whose attributes demand certain behaviours to fulfil the founding beliefs and
deliver business success. The characteristics, inter alia, tend to emerge subliminally
from the rationalising of how to effectively achieve goals. They are influenced
ORGANISATIONal SAFETY-CULTURE THEORY 35
by and influence the generic element that act as the visible manifestations of
the organisation’s cultural characteristics. The presence of the organisational
characteristics and their attributes can be elicited from consideration of an
organisation’s generic elements. These elements act as the vehicles to display the
characteristics and hence the culture. Starting with shared beliefs, Figure 1.7
indicates the linkage. This concept forms an integrated safety-culture paradigm.
Subliminal or overt
-Values
Internal and External -Artefacts a: Safety beliefs may be
Influences -Attitudes documented as artefacts
- Behaviours (but not necessarily
upheld)
characteristics as they are driven generally by the same business beliefs and
goals. These beliefs are, perhaps, maximising profits, customer care and public
safety, and so on. These common beliefs generate common delivery culture
characteristics. General experience suggests airlines broadly have the same
common culture ‘type’. Again, this will be visible or can be elicited through the
generic culture elements model. An airline culture due to its specific cultural
characteristics is quite distinctive from a regulatory culture. In both cases the
generic model elements will be in place – beliefs, values, artefacts and attitudes
– but the delivery of different beliefs for business ‘success’ generates dissimilar
culture characteristics. In addition, the resulting, inter alia, organisational staff
behaviours displayed will be culturally shaped and different. This will emerge
as an integrated safety-culture paradigm Figure 1.7.
• Safety is learning-driven.
(Text reproduced with the kind permission of the International Atomic Energy Agency)
Summary
• Beliefs
• Espoused Values
• Attitudes
• Artefacts
• Behaviours.
• beliefs
• espoused values
• attitudes
40 Safety Culture
• artefacts
• behaviours.
To test this, some events are considered. The first two occurred before the
concept of safety-culture emerged pre-1986. The events are:
• Bhopal (1984)
Titanic
The loss of this White Star Liner is well documented as arising from a glancing
blow with a 200,000 tonne iceberg. This occurred at 23.40 on the 14 April 1912
in a calm Atlantic Ocean. Five starboard ‘watertight’ hull compartments were
damaged. Within three hours the ship had slid bow-first under the sea. There
was a loss of nearly 1,500 lives with 705 survivors.
At a British Board of Trade inquiry the Titanic’s owners, the captain, officers
and crew were exonerated. It was recorded that they were operating Titanic at the
time of the incident in accordance with the marine custom and practice of 1912.
were the management, it can be suggested that their poor safety-culture and
associated safety-behaviours caused the ship to founder. It is proposed that
these behaviours arose from inappropriate shared safety-beliefs, leading to a
progressive deterioration in conservative decision-making increasing the risk
of an accident. A key issue perhaps is whether the behaviours and degradation
in the managers’ culture occurred over an extended time or was a collective
aberration due to the ‘excitement’ of the ship’s maiden voyage. This remains a
point of speculation. Yet it should be a matter of some unease if a culture can
decline so rapidly and deliver such fatal consequences.
As noted, the captain may or may not consult with his officers about
operational decisions as ultimately he has total authority. His orders have to
be obeyed, irrespective of subordinates’ beliefs or values. However, the record
gives no evidence as to whether the captain failed to consult his officers on
operational decisions on the April voyage. Similarly, there is no record of any
challenge by the officers to any of the captain’s perverse decisions leading to
the eventual sinking of the ship. This gives a strong suggestion that there were
shared beliefs at this hierarchical level regarding the ship’s management and its
safety in the iceberg field.
The Titanic on her maiden voyage to New York from Cherbourg was ‘state-
of-the-art’ marine engineering for 1912. She was double-hulled with watertight
compartments as protection from severe damage or sinking in the event of a
collision. With a unique three-propeller shaft the ship could achieve a speed of
almost 25 knots.
Some days into the voyage and before the fateful collision, Captain Smith
had received radio messages indicating that there were icebergs in the North
42 Safety Culture
On the new course, Captain Smith maintained the ship’s speed at 21.5
knots. At 46,000 tonnes the ship’s momentum now presented a significant
challenge if rapid speed reduction or an urgent change of course was required.
This behaviour to maintain high speed seems unusual, as the captain and his
officers were aware of the possible iceberg hazard at this latitude.
The ships were at a ‘visual’ distance apart. Shortly after midnight, the watch on
California saw flares and rockets from Titanic. Unknown to the watch, Titanic
was sinking. The periodic lighting of the sky was interpreted as celebration
fireworks for rockets in 1912 carried no emergency significance and were
ignored. On California the radio, an essential safety instrument, was now
unmanned whilst the Titanic was urgently transmitting SOS messages (and
CQD signals, come quick danger, which pre-dated SOS).
Before retiring for the night, Captain Smith had posted an observation
watch. This is some evidence that there remained a distant concern in his
mind regarding the iceberg warnings. Instructions were given to keep a look
out for ‘growlers’. These were small icebergs typically a few metres in length,
depth and height. Being particularly difficult to see, if a ‘growler’ was struck
by the ship travelling at speed, experience suggested there could be local hull
or propeller damage. Although she did have the novelty of three propellers
and damaging all three was probably remote, such an event could potentially
have stranded the ship in mid-Atlantic. It seems odd behaviour, however, to be
wary of striking small, even tiny, icebergs compared with the huge size of the
icebergs on Titanic’s latitude.
The officers may have split the observation tasks. The bridge officers would
keep a watch for large icebergs whilst the on deck the watch crew would look
out for small growlers. However, even with this instruction to look out for
‘growlers’, binoculars were not issued to the posted deck watch. The glasses
were locked in a cupboard and the key had been taken by an officer changing
ships in England before Titanic sailed for Cherbourg. Why the cupboard was
not forced open to equip the seamen is unclear. The captain appeared to be
content that a seaman’s direct vision was sufficient to identify any hazard in
the ship’s path. At this point, the seamen on watch were the only protection the
ship had whilst it steamed at full speed into an iceberg field. This lack of action
to properly equip a watch for its safety task is inappropriate behaviour. The
captain then retired to his bunk for the night.
On the night of 14 April in 1912 the ocean was very calm. There was no moon.
One of the local ships had earlier radioed a signal indicating the observation
of ‘blue’ icebergs. These are almost completely clear walls of ice with no snow
cover and, on dark nights, difficult to see. The ocean being calm, there would
be no waves crashing against the ice wall to give additional visual and possible
audible warning. Such icebergs were virtually invisible, particularly on a
moonless night.
44 Safety Culture
First, for the maiden voyage the owner of the White Star shipping line was
on board, as was the ship’s architect. Clearly, good performance was expected of
crew and officers. The expected good performance may have been to achieve a
fast crossing to New York. There were rumours of an attempt to achieve a faster
crossing than the Titanic’s sister ship Olympic. There was also some suggestion
that if the ship could be pushed towards her maximum speed the crossing from
Europe to New York could be reduced by several days and a record achieved.
However, there were already ships at sea that could achieve over 26 knots and it
was unrealistic that Titanic could achieve this. The attempt at a record is taken as
speculation and probably not a belief that significantly influenced the decision
to maintain a high speed. To make a fast crossing would have been a significant
coup. Technically Titanic could probably cross faster than SS Olympic as the calm
sailing conditions were ideal. It would be an excellent advertisement for the
ship and good for the business in the competitive transport market of the time.
In addition (although speculatively) it would be an impressive demonstration
of performance and seamanship before the ship’s owner. Although strongly
denied by the owner Bruce Ismay it is believed some ‘pressure’ was placed on
the captain to attempt a fast crossing. A speedy crossing on the ship’s maiden
voyage would indeed be ‘good for business’.
A further antecedent were the weather conditions. These were ideal for a
smooth fast crossing. It was a calm clear night and although there had been
numerous iceberg warnings, none had been seen by the crew, officers or
reported by any passengers.
particularly so as it was a clear, calm, cold night. However, the fact that there
was no moon was a crucial element overlooked by the officers and the captain.
The deep belief that they would sight any icebergs before a point of danger
may also have influenced the officers’ lack of urgency about receiving iceberg
warnings from the radio operator.
It has been suggested in many papers that the belief held by the captain
and all his officers was that the ship was unsinkable. This was possibly true
and could be an antecedent to poor safety-behaviours. However, even this
lofty belief would not encourage the behaviour of taking the ship at high speed
into hazardous ice-pack waters. It was not the intent to actually hit an iceberg
whether the ship was unsinkable or not.
Captain Smith was fully aware of the ship-building technology of the age
and when discussing the very latest designed ship, the SS Adriata, he made the
public statement, ‘that he (Smith) could not conceive of any disaster that could
make the ship founder’. He later carried this view onto the Titanic believing
that because of its unique features and the excellence of modern marine design,
the ship was unsinkable.
This belief was held by the ship’s owner Bruce Ismay. His belief that the loss
of the ship was impossible led to a modification of the first-class passenger’s
promenade deck. To enable the modification but against the wishes of the
designers, several lifeboat housings were removed from the deck structure.
When the sinking occurred too few lifeboats (and the poor emergency procedure)
contributed to the large loss of life. The origin of this perverse design decision
was a fundamental belief that the ship was unsinkable. The designers under
pressure compromised passenger safety on the basis of an unfounded belief
in the integrity of the ship in all types of collision. The first-class passengers’
luxury and convenience became the priority before passenger safety.
Neither the ship designers nor the company ever publicly claimed the ship
was unsinkable. The phrase used was that they had arrived at a point that
was ‘as far as possible to design a ship to be unsinkable’. The designers were
aware, however, that the bow compartments on the Titanic were not watertight.
Because of connecting corridors and the bulkhead sealing design, if more than
four compartments were ruptured then the ship would be lost.
As suggested previously, the fact that it was strongly believed that the ship
was unsinkable would not necessarily lead to the behaviour of maintaining
46 Safety Culture
high speed. Even if the ship did not sink following an iceberg collision the
implications for the maiden voyage would have been dramatic. The belief that
the ship was unsinkable was possibly a minor influence on the operational
decisions.
Noting the antecedents, there are shared beliefs overriding safety that were
driving the captain and officers towards perverse safety-behaviours.
At 11.40 p.m. a seaman on watch, in the crow’s nest, saw an iceberg 500
metres ahead of the Titanic. With binoculars it is estimated the seaman would
have seen the iceberg over a thousand metres away. The officers, who had not
seen the iceberg from the bridge, even though it was relatively close to the ship,
rapidly responded by reversing the engines at 21.5 knots and attempted to turn
the ship to port. Her momentum continued to carry her forwards. Thirty-seven
seconds after the sighting, the ship’s starboard bow struck the iceberg a glancing
blow. Five compartments were opened up to seawater ingress by the collision.
The architect, Thomas Andrew, was on board. On learning of the extent of the
damage, witnesses suggest that he informed the captain of the time it would
take for the ship to sink. He was correct almost to within minutes and died on
the ship. The ship began to sink at the bow. By 2.20 a.m. on the morning of 15
April the ship had disappeared with a loss of 1,500 lives.
• they would not see ‘growlers’ from the bridge but the posted watch
crew would see them
Little can be said about the formal culture dimension, that is, the safety
procedures and processes. Ships in 1912 were undoubtedly handled more by
the skill of the officers and crew than by procedure. However, the sea trials
for commissioning the ship were rushed and incomplete. Part of the trial was
the testing of the emergency procedures. The lifeboat emergency drill ended
in chaos, for the crew were unable to follow the lifeboat-launching procedure
and the lifeboat launch-equipment was inadequate. No re-drilling, retraining
48 Safety Culture
The ship sailed with a fire in one of its coal bunkers. The procedural response
would have been for the captain to return to port and have the fire dealt with, as
his own crew were having difficulty bringing it under control. Irrespective of any
fire emergency procedures or the need for a conservative safety decision, the ship
put to sea with the bunker on fire. A company ‘Flag Ship’ on its maiden voyage
with its owner on board, returning to port with a hold ablaze would perhaps
not have made good news. Whether this public-relations point influenced the
poor safety decision to proceed to sea, with this uncontrolled hazard on board,
is, however, speculation. The fire was eventually brought under control but was
still a smouldering menace even on the night of the collision.
There was possibly a construction safety issue that caused the rapid loss
of the ship. Recent research suggests that a decision was made to go from a
high-quality foundry material specification for metal-plate rivets to a reduced
specification. Due to foundry capacity being limited there was a shortage of
plate rivets being delivered to time during the build (about 3 million rivets were
used on the ship). The changed specification enabled additional, less modern
foundries to produce rivets. The reduced specification introduced more slag and
inclusions into the metal, reducing a rivet’s capability to take shear forces across
the head. When the collision occurred it is postulated that the plate rivets failed
in rapid succession under the shear forces. Rivets failed as if the metal was being
unzipped due to these impact forces and from the immediate local high forces
caused by the failing adjacent rivets. The rivets failed in their primary task of
holding the ships hull plates together. There has been much speculation as to
whether this materials substitution did actually occur. However, if it did it is a
keen example of a materials change taking place without consideration of the full
safety implications of such a decision on a high-hazard, low-risk plant.
Beliefs
Espoused values
Attitude
Safety complacency:
• risk denial
Artefacts
Procedures ignored:
• insufficient lifeboats
Resulting behaviours
Bhopal
Madhya Pradesh State is located in the very centre of India. Bhopal is its capital
city and the Bhopal Chemical Plant was located in the city.
The 1984 accident at this plant is perhaps an extreme example of poor safety-
culture. Although the owners argued that the accident was due to sabotage,
the India safety authority’s official report disagreed. In addition, if the plant’s
engineered safety provisions had not been allowed to deteriorate even if
sabotaged, they should have been sufficient to mitigate some of the accident’s
consequences. Bhopal would be virtually unknown outside Madhya Pradesh if
the plant had been in good working order. Unfortunately, this is not the case.
The reported consequences from the plant accident were 2,500 members of
the public killed and 250,000 seriously injured. Some of the injuries persisted
over decades.
were the root cause of the plant’s safety systems being permitted to deteriorate
and the risk of an accident rising beyond what is reasonably tolerable.
The Bhopal plant opened in 1968 and was producing carbamate pesticides
(carbaryl). An intermediate, highly toxic chemical product was methyl
isocyanate which is a significant hazard for operators and the public. The public
was particularly at risk, as close to the site boundary an unofficial ‘shanty town’
had been permitted to arise and expand.
• toxic
• volatile
• boils at 39°C.
tanks at Bhopal had refrigeration units. Noting the flash point of MIC, these
units were safety equipment and used to ensure that the stored material
remained at a low temperature. This and other important engineered safety
features deteriorated over time.
In considering plant failure, a key issue is, inter alia, to identify the root
cause. If acute mechanical failure is a low-probability event on chemical
plants then the quality of the human performance, the safety-behaviours,
requires consideration. This extends from the various managerial levels to the
operator.
With regards to storage, there were two main MIC storage tanks, with a
back-up tank. Each was capable of holding 65 tonne of material. However, as
little as 10 kg was required for a process batch. No attempt had been made
in design or operation to reduce this hazard. The magnitude of this hazard
was not appreciated by the site management. After the accident, inventories in
plants worldwide were reduced by at least 75 per cent.
The official report showed, by its systematic examination of the plant and
its processes, that routes existed into the MIC tanks, and water could enter if
there was mal-operation in plant flushing, cleaning and maintenance. Neither
during the plant design process nor during operational reviews was the plant
subjected to a systematic hazard and operability study. Such a study, even
rudimentary, would have revealed these routes and steps could have been
taken to eliminate them.
The report indicated that the accident was initiated by the flushing through
of a scrubber-vent line with a water wash. Valves linking the vent line to tank
no. 610 (the tank that failed) were not seating correctly. Alternatively they
were not closed before the flushing process was started. This was due to poor
maintenance, equipment failure or they had been left open due to a procedure
failure. Further, blanking plates to isolate the wash line from the tanks were not
in place due to inadequate procedure or supervision.
MIC, as noted, is very volatile and was kept at a low temperature using
refrigeration units. Even though the vessel contained tonnes of MIC, the
refrigeration units on tank 610 had been shut down for maintenance. This
refrigeration equipment was an engineered safety feature for defence in depth
against the MIC overheating and becoming an uncontrolled hazard. There
were no substitution arrangements in place or attempts made to move the MIC
to an alternative refrigerated tank. The refrigeration unit was shut down, with
no substitution arrangements, with the knowledge of line management. This
was an endorsed procedure violation.
It was established in the official report that due to poor maintenance the
instrumentation on the complex was known by the operators to be unreliable.
During plant operation instruments were ignored due to their unreliability.
This was accepted by the management. Degradation of maintenance standards
and the ignoring of instrumentation readings became an accepted behavioural
norm. Poor instrument performance was worked around by the operators who
used judgement, guess work, and experience to keep the facility operational.
54 Safety Culture
When the water entered tank no. 610, via the faulty valve, a slow exothermic
reaction occurred with the MIC. The tank’s temperature and pressure began to
rise. Because of instrument unreliability and such changes not being expected,
the plant operators ignored the instrumentation showing increased temperature
and pressure. However, on this occasion the instrumentation was working
correctly. As the event progressed, a high-level temperature-alarm on the tank,
which had been set too high after previous inadequate maintenance, did not
signal at the design set point. This alarm maintenance failure occurred due to
a breach of procedure and poor supervision. Even though it was a key safety
instrument, the alarm was not tested and signed off after maintenance as being
suitable for service. The consequence of this was that when the instrument did
eventually signal alarm, the tank temperature and pressure were outside the
tank’s designed safe operating envelope.
The tank was also fitted with a relief valve, designed to fail at 3 × 10+5 Pascal
(3 bars) pressure. However, on reaching this pressure the valve did not open.
This was due to poor maintenance or corrosion. At a considerably higher
pressure than the design intent, the valve did eventually release and a two-
phase liquid-vapour mix flowed to a caustic scrubber.
The stack flare was the final line of defence against methyl isocyanate being
discharged directly to atmosphere. It was designated as a 24-hour flare because
of its importance to safety. It was inoperable during the accident. Due to poor
maintenance, pipe corrosion had been identified in the flare’s system. The flare
was shut down several months prior to the accident occurring. There were no
substitution arrangements.
The faults and failures outlined are only some of the safety issues identified
in several Bhopal post-accident reports. Others include failure to follow line-
washing procedures, insufficient gas masks on site, undeclared modifications to
the plant, lack of an emergency plan or its exercise, and a lack of understanding
of the plant hazards by some managers and operators.
safety-culture theory as a predictive model 55
Although there had been several serious accidents since operations began
in 1964, the management appeared not to act effectively to improve the plant.
Audits and reviews did occur but constructive action to reduce risks did not arise.
Identified safety problems prompted the worker’s trade unions to challenge
the facility management on safety-performance. For example, one operator had
been killed (1984) by phosgene gas and there were numerous conventional and
chemical injuries annually. There was a failure of the organisation to learn from
these events and act in the interests of worker and public safety.
Employees were regularly moved from the plant to other sites. By 1983, a
30 per cent reduction in the level of experienced staff had occurred. With the
exception of the unions, this had not been challenged within the company as to
the operational safety implications.
In addition, the senior staff and the site manager were regularly changed
which promoted a general lack of commitment to the plant’s safety. There was
also a loss of senior skills with an associated loss of corporate memory about
the hazardous nature of the materials being processed.
Although this event took place before the concept of safety-culture emerged,
a safety-culture review would have concluded that the culture was in such a
poor state that the facility was a high safety risk. All the expected safety-culture
characteristics both qualitatively and quantitatively were weak. The Bhopal
organisation by 1984 could be defined as a pathological organisation – that
is, one attempting to succeed financially with the absolute minimum of legal
safety compliance.
The Bhopal plant was licensed to produce 5,000 tonnes of pesticide per
year. There had been a continuous decline from the start of the eighties, such
that by 1983 the site was only producing 1,650 tonnes with sales as low as 1,500
tonnes per year. The plant was facing competition from synthetic pesticides
and making no profit. The commercial pressure was to cut financial losses.
56 Safety Culture
• Poor safety training and skills due to declining safety resources and
cost cutting
This accident was not bad luck, nor exceptionally bad design (although
it could have been significantly improved). Using the generic safety-culture
model, it is suggested that the site employees’ safety-behaviours were influenced
directly by the corporate beliefs that over time shaped a poor organisational
safety-culture. The consequences of this were 2,500 deaths and 250,000 long
term injuries, one of the worst chemical plant accidents in history.
Beliefs
Espoused values
Attitude
• Safety complacency
• Hazard denial
Artefacts
Resulting behaviours
Used in the manufacture of poly vinyl chloride, vinyl chloride monomer (VCM)
is a hazardous material. To prevent accidents, where the consequences could
affect operators, the public, or the environment, VCM needs to be managed
with care.
The vapour is a significant health hazard, with the normal means of entry
to the body being by inhalation. It is a heavier than air, colourless gas that is
not readily detected by human senses. Due to its hazardous nature, the short-
term exposure limit (15 minutes) is set at 5 parts per million (ppm) by volume
in air with a normal operational exposure limit of 1ppm. It affects the nervous
60 Safety Culture
The analysis starts with two alternative safety-beliefs that may have been
held by the corporate and site managers. The beliefs appear to be mutually
exclusive.
‘We are committed to the health and safety of our employees and the
communities in which we operate. Through the joint efforts of every employee
we shall keep our environment clean and our workplace free of health and
safety hazards for ourselves and for our communities and for our future
generations.’
These are good principles. If truly believed and shared by all staff, from the
board members down to the sites workforce, this is a good platform to build a
safety-management strategy.
safety-culture theory as a predictive model 61
The Illiopolis plant involved in the 2004 explosion was constructed and
then operated from 1965 by its original designers. In 2001 the plant was sold to
the PVC plant’s corporate owners. One of the first acts of the new owners was to
reorganise the management structure and reduce staff. This was a high-hazard
facility yet, as far as is known, these changes were carried out with no formal
assessment of the safety implications on operations by using a management of
change procedure.
Prior to the reorganisation, the plant was originally divided into local areas
and each area had a highly skilled and respected group leader. This person was
readily available on plant. The leaders had the knowledge and ability to respond
quickly to chemical process issues or plant problems being experienced by the
workforce. The group-leader post was eliminated in the 2001 structural changes
and a single supervisor put in charge of all the plant without the area divisions.
The result of this was that in an emergency or when a worker required advice,
the supervisor might not be available.
The plant had 18 reactor vessels 8 m high and approximately 4.5 m diameter.
They were arranged in the pattern shown in Figure 2.1. Although they formed
‘mirror images’ of each other the vessels were clearly numbered 301 to 318 as
indicated in Figure 2.2.
62 Safety Culture
Following a PVC transfer to the stripper tank the empty reactor vessel,
with its bottom valve and associated drain valve closed, was manually water-
washed using a water-blast unit to remove any residual products.
safety-culture theory as a predictive model 63
There were six persons on a shift including a ‘Poly Operator’ and a ‘Blaster
Operator’. The poly operator was located on the top floor of the tank house
(Figure 2.2). This operator controlled the flow input, the pressure and the
temperature of the reactor vessels. Important to this event was that the only
instrumentation showing the status of any vessel was at the poly operator’s
control panel.
The blaster operator had two main roles. On completion of a cycle the first
task was to transfer PVC product from the reactor to the stripper tank. The
second task was then to open the top of the empty reactor vessel, attach a blaster
unit and wash out the vessel with water. The blaster unit was manually handled
at the top-floor level of the tank house (Figure 2.2). The blaster operator, on
completing the vessel washing, would go to the ground floor, open the vessel’s
‘bottom valve’ and then the vessel’s ‘drain valve’. This was achieved by using a
valve switch-panel local to the base of a reactor vessel. The vessel water flushing
then discharged from the cleaned vessel, through the open air, into a ground-
floor drain. The valve arrangement is shown in Figure 2.3.
For any vessel, its bottom valve and the drain valve were air-interlocked.
This was to prevent a vessel-bottom valve being opened if the drain valve was
open when a vessel was full of VCM/PVC mixture.
The drain valve was free to be manually opened or closed at any time. The
interlock mechanism was on the vessel’s bottom valve. However, the vessel-bottom
valve interlock could be overridden in an emergency to enable a vessel’s content
to be transferred. This was done by a procedure and required a supervisor’s
authorisation. Before doing an override it was expected that the drain valve would
be closed by the operator as part of the emergency procedure. Override was then
achieved by attaching an emergency airline to the vessel’s bottom-valve interlock
mechanism. This opened the vessel bottom-valve and with the drain valve firmly
locked an emergency piped transfer to an empty vessel could take place.
On the ground floor the blaster operator had access to a valve control-
switch panel. One panel served two adjacent reactor vessels and enabled each
vessel’s valves to be manually activated. The operational status of the reactors
was not displayed on the panels. They had no instrumentation status readings.
The blaster operator lacked information on whether vessels were empty, at
pressure, at temperature, full of VCM or containing only washing water. The
operator could only obtain this information by returning to the top floor of the
tank house and asking the poly operator. There were no telephones or intercom
systems.
Figure 2.2 VCM reactor vessels layout – elevation
safety-culture theory as a predictive model 65
Valve Switch
Panel Drain Valve Vessel Bottom
Valve
Figure 2.3 ‘Mirror image’ vessels D306 and D310 layout – ground floor
Source: U.S. Chemical Safety and Hazard Investigation Board, modified by the author
for this text.
On the night of the incident, April 2004, the blaster operator was water
washing reactor vessel 306. All other vessels were full of VCM and reagents,
producing PVC. The blaster operator, on the top floor, had attached the water
blaster to vessel 306. After a short break, he was returning to the ground floor
where, instead of going to the valve switch-panel at vessel 306 to switch the
valve to drain the vessel, he went, in error, to the control panel for vessel 310.
This vessel was full of VCM.
The vessel’s mirror image layout appears to have confused the operator
(Figures 2.2 and 2.3). At the 310 switch-panel the operator opened the drain valve
on 310 so that he could drain the water he knew was in vessel 306. The operator
then attempted to open the vessel-bottom valve using the panel switch. It was,
correctly, interlocked (as the operational vessel 310 was under pressure and at
temperature; these conditions had activated the vessel-bottom valve interlock).
The accident investigators surmise that the operator could not understand
why the vessel-bottom valve was locked; it should not have been since, in the
operator’s mind, he was ‘at vessel 306’. Without questioning his own actions,
analysing the problem or seeking advice, the blaster operator decided against
procedure to override the vessel 310 bottom valve interlock. This was done by
taking the emergency valve-interlock bypass air-supply line that was readily
available, attaching it to the vessel-bottom valve (which the operator believed
was vessel 306), injecting air and overriding the 310 interlock.
66 Safety Culture
The 310 vessel-bottom valve opened and several tonnes of VCM began
to discharge through this valve and the previously open drain valve. The
discharge overwhelmed the ground floor drain and VCM spread rapidly across
the ground-floor surface.
The VCM liquid rose locally to 0.3 m deep and was rapidly vaporising.
Within five minutes it ignited, possibly by static electricity. The explosion and
fire killed four people immediately (one died some days later) and seriously
injured three other workers. Local communities were evacuated to avoid the
consequences of noxious gases and smoke. Several hundred firefighters took
several days to bring the fire under control. The assets were lost and the plant
is unlikely to reopen.
(The notes in italics are the author’s comments and not Chemical Safety Board
information.)
• The original designers and plant operators did not implement the
1992 Process Hazard Analysis recommendations that suggested
revision to hardware for the reactor bottom valve interlock bypass
to reduce the potential for deliberate or accidental misuse.
(Analysis of risks from the bottom valve system and overriding had been
recognised as a possibility. Neither original owners nor the ‘Plastics Plant’
corrected this issue. The safety-management system was not upgraded nor
were physical improvements made to the plant to reduce the risk.)
safety-culture theory as a predictive model 67
(Before the Illiopolis event, there had been numerous incidents with VCM/
PVC internationally in Japan, Europe and the USA. There had been loss of life,
serious injury and asset loss. A similar incident had occurred at the corporation’s
Baton Rouge plant, USA, in April 2004. Here, there had been a failure to follow
procedure yet no action was taken to modify any of the facilities. The safety
behaviour of breaking procedure appeared to have become an accepted social
norm. There was no discipline applied by management. The antecedent to such
behaviour appeared to be management condoning breaches of procedure with
no unwelcomed consequences for individuals doing the breach … Additionally,
the management at corporate and site level appeared not to learn from past
events, this being a significant safety-management failure.)
• Neither the original plant owners nor the ‘corporation’ analysed the
reactor cleaning-procedure to identify hazards and recommend
safeguards to prevent personnel from draining the wrong reactor
during cleaning.
the facility and it was all too easy to override the interlock system with the
emergency air hose. There was also lack of rigour regarding the application of
the interlock override procedure. Staff restructuring may have contributed to
this lack of rigour.)
• Operators on the lower level (the location of the vessel valves) had
no means (indication or communication) to determine the operating
status (of any of the vessels) from the lower level.
(The operator had only the vessel numbering and his memory to guide him
in the draining task. The instruments at the lower level only indicated valve
positions and facilitated valve movement. The blaster operator, on the ground
floor plant level, had no equipment to telephone or radio the poly operator to
determine the status of vessels. This was poor design. Improvements in the
process instrumentation to give vessel status information on the bottom-floor
control panels had not been pursued by the ‘Plastics Plant’s’ management.
(Further, the use of a radio or telephone to determine plant status, as a means
of safety protection, is a weak safety defence on a high-hazard facility.)
Maintained and functioning instrumentation, engineered defence-in-depth,
applied procedures, good training and a design tolerant to operator error
are improvements that could have been sought by managers. If an operator
becomes distracted, the mirror-image layout of the vessels may itself become
a hazard. Further, if the operator becomes absolutely convinced he is at vessel
306 the confusion becomes compounded by noting that the bottom valve is
locked off when ‘he knows he should be able to be open it’. Training to stop,
think, question and challenge ‘the abnormal’ would have come into play if it
had been ingrained into the operators by managers.)
(Previous incidents had occurred due to the ease of access to the vessel-
bottom valve interlock override. Failure to follow procedure had no
consequences to individuals and there was no means of a third party detecting
an interlock bypass. It was also easy to connect the emergency air to the bottom
valve as the normal and emergency valve air connections were common fittings.
This bypass was known to occur by the management. Some of the antecedents
for the operator were therefore:
safety-culture theory as a predictive model 69
• Saving time
(There is evidence that the emergency procedure was not understood in the
plant. More importantly the emergency procedure had not been exercised for
70 Safety Culture
10 years. When the event occurred, instead of immediately evacuating the area
on hearing the incident alarm, operators attempted to stop the leak of VCM by
by using the upper-level control-panel equipment. The operators on the upper
level were fatalities in the explosion.)
(Although a weak safety defence, the plant relied upon written procedure.
However, due to lack of discipline contributing to a poor safety-culture, the
procedures were not a reliable safeguard. Procedure violation did occur, was
condoned and was undetectable. With no consequences for violations this
relaxed attitude towards working with procedures had possibly become a
socially accepted norm.)
Beliefs
c) ‘We know exactly what we are doing with this plant.’ (A belief
arising from over confidence, familiarity or complacency.)
safety-culture theory as a predictive model 71
Espoused values
The espoused public value appeared to support the belief (a) above. That is,
safety of the public and employees was valued There is no evidence of other
values from inquiry reports.
Attitudes
Artefacts
Resulting behaviours
From the incident investigations there appears to have been poor human
performance – poor safety-behaviours, that is, poor safety-culture in the
‘Plastics Plant’s’ organisation:
There were also design errors suggesting a poor safety-culture during the
design process. There was a failure to take action during operation of the plant
to remove hazards identified due to poor design. For example, there was a high
potential for the direct discharge of VCM to the plant floor. There had been
several ‘near misses’ with this event on other plants and it had been identified
in studies as a high frequency event. The hazard could have been engineered
out of the plant.
These behaviours suggest that the true, owned and implemented belief of
the organisation was not belief (a) noted above. There was some other overriding
belief imposed by the corporate body, possibly (b) or a local issue such as (c).
Either of these beliefs shaped a less than adequate safety-culture contributing
to the fatal incident and major asset loss.
safety-culture theory as a predictive model 73
Finally, two additional points are worthy of note. As indicated, the operation
of this facility may have been monotonous. Filling vessels, cleaning vessels,
refilling vessels, cleaning vessels day after day, may lead to complacency and
boredom. On high-hazard plant, managers need to be ready to cycle staff
around jobs to ensure boredom does not lead to loss of concentration. Second,
the Illiopolis plant was a large local employer. There had been job reductions.
This may have introduced an element of fear into the organisation. Workers
may have wanted to show their efficiency and even enhance the throughput
capability of the plant to ensure its future. As such, operational short-cuts may
have been considered on occasions by all staff as acceptable to maintain plant
throughput. There is, however, little evidence of this. Nevertheless, perceived
commercial pressures could lead operators to take chances and managers to
ignore procedural violations.
The location of this fatal event was a construction site. The fatality was a
‘banksman’ guiding a crane driver’s ‘blind lift’ operations. Following the
incident the offending construction company was prosecuted and heavily
fined in a court of law.
During one of the lifts, an error was made and a shutter unit collided
with the apartment building at the ninth floor level. The shutter snagged on
a concrete building structural slab. The impact was of sufficient force for the
74 Safety Culture
shutter to split. The two halves fell from the crane rigging. The banksman who
was inside ‘the personnel exclusion zone’ was struck and instantly killed.
• The crane driver and the banksman may be ‘loan workers’ requiring
additional safety training and precautions
• Inadequate training of the crane driver or the banks man for such
lifts
Nevertheless, this was not the first time the organisation had used a crane
driver and banksman combination. If so, there was perhaps persistent ignorance
of the law. However, there may have been a firmly held belief that the company,
the driver and the banksman actually knew what they were doing and fully
76 Safety Culture
Lifts should be adequately supervised at all The lift appears not to have been supervised by
times by an experienced manager. an experienced manager. The banksman was
possibly a ‘loan worker’.
Lifts should be carried out in a safe manner Due to an inadequate risk assessment or
which includes contingency for incidents. experienced managerial supervision, the
banksman appeared to have been within the
‘personnel crane lift exclusion zone’ when the
event occurred.
During planning, it was unclear whether
alternative lift options had been considered.
The application of the ‘as low as reasonably
practicable’ principle to the risks was not
demonstrated, that is, an examination of
alternatives to a ‘blind lift’ was not recorded.
Learning from experience should be used. The high risk to operators of injury or fatality in
the construction industry is annually reported
and documented by statutory safety bodies.
This appears not to have been assessed by the
employer. (Lifts are notoriously dangerous if
inadequately executed.)
Training of all site personnel on the hazards Such measures appear not to have been taken.
they face on specific sites they are working.
understood the risks. This may account for the inadequate risk assessment,
Table 2.1. A comprehensive assessment may have been considered unnecessary
as the lifts were considered a ‘straightforward’ task, possibly considered routine.
The safety-belief that ‘they knew everything they needed to know about the
safety-culture theory as a predictive model 77
Belief
• Completing the task to time and cost is the priority, not safety
Espoused values
The values are not known. However, they may have been:
• ‘We value our competence and have little to learn from others’
Attitude
• ‘We have done enough’; that is, a poor generic lift assessment was
in place
78 Safety Culture
Artefacts
Resulting behaviours
• The crane driver accepts the concept and carries out ‘blind lifts’,
possibly an accepted norm
For various breaches of health and safety law, the company in question was
heavily fined by the court.
A Criticality Event
The primary role of the site’s plant was to convert uranium hexafluoride
into uranium dioxide fuel for use in Japan’s commercial nuclear power stations.
Once every few years the site’s Conversion Test Building processed purified
triuranium octoxide (U3O8) powder with nitric acid to form uranyl nitrate
solution. The solution was to be used off-site in the manufacture of fuel for
the experimental fast-breeder reactor Joyo. On the 30 September 1999 three
operators were seriously irradiated with neutron and gamma ionising radiation
whilst working on the Joyo triuranium octoxide process. Two received fatal
radiation doses. The third, the supervisor, survived following a period of
prolonged hospitalisation.
isotopes, uranium 235 (at 0.7 per cent by weight) and uranium 238 (at 99.3 per
cent).
The neutron is a tiny particle which forms with protons the structure of
almost every atomic nucleus (the exception being hydrogen which has a single
proton as its nucleus). Uranium isotopes are ‘large heavy’ nuclei and if struck
by a neutron the uranium will split, or fission, into two lighter elements. The
fission process generates energy. This energy is released mainly as the kinetic
energy of the two lighter fission product fragments and as energy associated
with the neutron and gamma radiation emitted in the fission process. (Gamma
radiation is similar to X-rays but of shorter wave length and is more penetrating
through matter).
If the neutrons emitted in one fission strike other local uranium nuclei,
further splitting can occur, resulting with the right conditions, in a chain
reaction of multiple fissions. Such an uncontrolled, unintentional, reaction is
termed a criticality event (Figure 2.5). (Criticality events are not an explosive-
type chain reaction.)
The intense ionising radiation is the major criticality hazard for operators
(and possibly members of the local public). The high energy neutron emission
is the more biologically damaging radiation component in such events. If a
plant is not adequately designed or managed, a criticality event may occur
with fissile materials that are solids, powders or in solutions. Criticality events
take fractions of a second to occur with instantaneous exposure to radiation of
local operators. A criticality can end just as quickly due, inter alia, to the local
dispersion of the uranium material changing its geometrical shape, preventing
further excess fissions and stopping the chain reaction. However, this is not
always the case. There can be multiple bursts of radiation from an oscillating
criticality. Here the material disassembles after the first event, only to reform
sometime later into a critical mass with a favourable geometry and trigger
another criticality. Eventually further material dispersion or some intervention
stops the oscillations.
During a criticality event the intensity of the radiation falls rapidly the
further away a person is from the event; the fall off is one over the square of the
distance. Because of this, it is unusual for there to be life-threatening radiation
exposure of persons outside the immediate location of the event, typically a
few tens of metres away.
82 Safety Culture
Uranium 235 readily fissions with slow (low energy) neutrons, referred to
as thermal neutrons. This concept is exploited in thermal power reactors. Also,
the material used in these reactors is uranium dioxide (UO2) manufactured
from processing U3O8. However, for technical reasons the uranium 235 isotope
in the natural oxide state has to be enriched. Typically, for light water thermal
nuclear reactors, in common use, this would be a 5 per cent by weight U235
enrichment. Enriched uranium was used in the Joyo fast reactor, but at a much
higher level, 18.8 per cent U235 enriched.
the risk of criticality is as low as is reasonably practical. That is, the possible
occurrence is extremely remote; typically about one event in a million years
per plant is adopted as a safety target. This is generally judged by the public
to be an ‘insignificant’ and ‘tolerable’ risk. Fuel facilities worldwide are very
safe in operation due to exemplary design, engineering, diligent operation and
demanding safety targets.
A further safety defence is to restrict the mass of uranium material that can
be placed in various vessels and tanks. If there is insufficient material below a
critical mass nature itself, the physics, will not allow a criticality to occur. This
is uranium mass control. The higher the enrichment of U235 in a process the
lower the permitted mass. On a facility the mass constraint is set at several
times below the critical mass using detailed analysis by highly trained criticality
specialist assessors. Mass constraints become documented requirements for the
operational safety case and procedures.
local to vessels have been studiously designed out, reducing the possibility of
neutron reflection or multi-collision moderation.
The design of process equipment and procedures for safe plant operations,
in any high-hazard industry no matter what the level of excellence, are only as
robust as the people that use and maintain them. In support of any plant design
and safety-management system there has to be a good organisational safety-
safety-culture theory as a predictive model 85
A key factor leading to the event was that the Tokaimura site management
had been under economic pressure for several years following electricity
deregulation within Japan. They were facing intense competition in the fuel
market requiring cost reductions and improved efficiency. Restructuring of the
organisation had been carried out resulting in around one-third of the staff
losing their jobs. This restructuring occurred without due regard to the safety
implications of removing staff from facilities. Any such safety implications
would have been revealed by the application of a management of change
procedure implemented through a robust documented management system.
The drive for efficiency caused the site management to become focused on
product quality, the cleanliness of process equipment (for a quality product), and
86 Safety Culture
• had for a facility an official nuclear safety case and unofficial safety
amendments
Even with this level of design, engineering and procedures in place, such
an infrequent, hazardous operation to produce Joyo, enriched nitrate solution,
should, if following good safety-management practice, have been subjected,
inter alia, to:
• crew training and testing for knowledge and suitability for the
task
Such rigour should have been mandated through the site’s safety-
management system.
The work crew handling the Joyo task were significantly disadvantaged.
First, they were in another team’s building. The usual building team had
other tasks to complete in the building. This put some pressure on the Joyo
task crew to move their uranium nitrate solution production forwards quickly.
In addition, the Joyo work had been delayed for several hours. This was of
concern to the crew as they were expected to have nitrate solution samples
available for quality control analysis on the first day of the task. The crew also
wished to complete the campaign as soon as possible. They considered the task
as an additional burden on top of a high workload from routine missions. The
crew believed there was a need to accelerate the chemical process and make up
time to meet these various pressures.
Another disadvantage for the crew was that the last Joyo campaign, some
years before, was known to have taken several days. This seemed excessive to
the crew and time had to be recovered. Further, none of the crew had worked
on this previous campaign and they had no experience of the Joyo task process.
They routinely worked on automated facilities elsewhere on site producing
5 per cent U235-enriched material. They therefore had a mindset regarding
equipment use. It was automated and so tended to ‘look after’ the operator
regarding errors and contributed to ensuring safety. The plants were tolerant
to human error. The Joyo uranium nitrate solution process in the Conversion
Test Building was essentially ‘hands on’. The Conversion Test Building process
equipment was not automated, it had to be worked. This was novel for the
crew.
The crew’s supervisor had himself only recently had experience on plant
handling intermediate enriched material which was typically 18.8 per cent
U235 enrichment. The other crew members had no experience of working with
such material.
None of the crew had been trained in criticality safety other than to be aware
not to exceed a given batch mass of triuranium octoxide. The crew, however,
believed this was for quality control. They were unaware that it was a criticality
hazard control which if exceeded had serious safety consequences for them.
The management considered that the concept of criticality was too complex
to be part of operator training. They simply did not have this knowledge. The
crew had no concept of the dangers of accumulating uranium mass in batches
safety-culture theory as a predictive model 89
Finally, because of the pressure on time the team did not read the official
procedure. It was difficult to understand. Nevertheless, the official, regulatory
approved, procedure was not followed anyway in the Conversion Building.
The equipment was operated by what had become custom and practice. In
summary, the crew were neither suitably qualified nor experienced to carry
out the task they had been directed to do. They themselves were ignorant of
this point.
With this background, the crew started the Joyo task. Although handling
hazardous materials, the task for the team was relatively straight forwards. This
was to mix 16.6 kg uranium as triuranium octoxide powder enriched to 18.8 per
cent by weight uranium 235 with a specific volume of nitric acid. This was to
be done in seven batches each with a criticality mass control of 2.4 kg U. They
were to ensure a homogeneous mix which was a product quality issue. Batches
would be transferred to a batch storage column. After a satisfactory clearance
of the samples by the laboratories the nitrate solution was to be drained from
the final buffer storage column into the shipping containers.
Each batch mixed was not to exceed 2.4 kg U which the criticality mass
control limit. It is not possible for this mass of material under the worst
geometrical arrangement, a sphere, that is fully water-reflected to go critical.
The equipment to be used was safe with this mass limitation. Further,
the Conversion Building equipment had been specifically designed to be
geometrically safe, up to the site licensed limit of 20 per cent enriched material.
That is, pipes and vessels were geometrically designed such that spontaneous
fission neutrons could leak from the system. The plant was design to protect
the operators; through mass control, through being geometrically safe by
design, with neutron reflection eliminated and with the operators supported
by approved operational procedures.
control. Scrutiny should be mandated for each task through the management
system with independent oversight. This was not the case at Tokaimura.
The Conversion Test Building equipment that should have been used with
the regulatory approved procedure, to generate the Joyo nitrate solution, is
shown in Figure 2.6. The equipment in the figure has been reduced in content
for clarity. The process was to begin with the handling of the 18.8 per cent
U235 enriched triuranium octoxide powder. With appropriate radiological
protection against inhalation of uranium, the powder was to be measured as
single 2.4 kg U batches (the criticality mass control) and placed in the dissolver
tank. This was then to be followed by the addition of a measured amount of
nitric acid. The batch of solution then moved slowly though several columns.
There were solvent extraction columns, followed by stripping columns. These
columns were normally used for other uranium oxide purification tasks but
were ideal, by design, for ensuring a satisfactory quality mixing of the Joyo
product. The enriched uranyl nitrate was eventually pumped to one of two
buffer storage columns for further mixing by air sparge. This overall process
formed a single batch of 18.8 per cent enriched uranyl nitrate solution of
density 370 kg/m3. The design of the storage buffer columns was such that each
could accommodate three batches. A geometrically safe transport flask was
then positioned beneath the vessel. The uranium nitrate, when at the accepted
quality, was manually drained off into the transport flasks and eventually taken
off site. Several transport flasks would need to be filled.
An engineering difficulty with the buffer columns was that due to the
confined space, it was difficult to place the flask under vessel and enable
drainage. It could be done, but it was a time-consuming exercise.
All the process vessels were designed as tall columns with small diameters.
This ensured they were geometrically safe against criticality up to 20 per cent
U235 enriched. This was the envelope of the site licence. Although the equipment
had other functions for purifying uranium oxides, it was designed for safety
and could be used for the relatively straight forward task of producing 18.8
per cent enriched uranyl nitrate. Column contents agitation systems and the
pumping sequences gave good homogenisation of the nitrate solution product.
These were safe mass conditions and the process carried through safe geometry
vessels.
safety-culture theory as a predictive model 91
Solvent extraction
Dissolution column
tank Extraction stripping Buffer
column column
U3O8 HNO3
Powder (nitric acid)
material
UO2(NO3)2
solution
Figure 2.6 A schematic, the approved process in the Conversion Test
Building
Source: U.S. Nuclear Regulatory Commission, modified by the author for this text.
In the building, as part of the equipment, there was an additional vessel. This
was a large diameter precipitation tank. It was for use on other Conversion Test
Building processes and was not part of the approved procedure for generating
the Joyo product.
The documented safety case for the use of the facilities equipment had
been approved many years earlier by the regulatory authority. However, if
the approved procedures were applied and the equipment used to process
16.8 kg U of triuranium octoxide powder, it would take several days. This was
a time frame commensurate with other Joyo campaigns. However, it did not
align with the requirement for more effective and efficient production.
One of the first steps towards improving the efficiency of oxide work in the
Conversion Test Building was to change the approved procedure for uranium
92 Safety Culture
oxide and nitric acid mixing. It was agreed by managers and initially carried
by ‘word of mouth’ that the dissolver tank mixing stage could be bypassed
and the uranium powder mixed with nitric acid in a 10-litre stainless-steel
bucket placed on the building floor. The solution could then be poured into the
dissolver tank ready for pumping forwards into the remainder of the process.
This would reduce a 90-minute dissolver tank batch mixing task to about 20
minutes. The approach was applied over time, saving considerable operational
time for various types of oxide mixing tasks.
What became common practice, the by-passing of the dissolver tank powder
mixing step, was put into an informal procedure and printed for use on all
mixing tasks in the building. It was adopted over several years. This violation
of the process safety case was management-condoned. The regulator was not
informed and the change never approved. (This, of course, brings into question
the effectiveness of the site’s internal or external procedures for auditing.)
A criticality incident did not occur following this change. The various
task criticality mass controls imposed on material handling in the building
continued to be followed by the operators. The approved equipment used after
any mixing operation in a bucket was geometrically safe.
The Joyo task crew were aware of the dissolver tank bypass practice.
nitrate solution had to pass through the solvent extraction columns, stripping
columns, and through pumps and pipe work (Figure 2.6).
A decision was made to take the solution from the stainless-steel bucket,
in small aliquots using the glass beaker, and pour it directly into the buffer
storage columns. This bypassed all the engineered equipment except the buffer
storage columns. It was judged that sufficient additional product mixing could
be achieved in the buffer columns using the vessels’ air sparge systems. With
experience in implementing this second process change, the sampled solutions
appeared to be homogeneous, be of good quality and going directly into the
buffer storage columns saved up to three hours on a nitrate batch run. This was
a significant increase in productivity meeting the site management’s drive for
efficiencies.
This column bypass and direct solution discharge into the buffer columns
became part of the informal procedure that was already bypassing the
dissolver tank. The safety case was not changed. Again an independent nuclear
safety committee was not engaged nor the regulator informed. The site safety
committee was aware and condoned this additional violation. Although the
change was adopted over several years for uranium powder mixing tasks, a
criticality event did not occur. The operators continued to comply with the safe
mass instruction in the informal procedures for mixing uranium and nitric acid
material in the bucket. The buffer columns remained by design, criticality safe
up to 20 per cent U235 enriched material.
At this point, and developed over several years, there were several internally
known, management condoned, safety case violations:
The crew carrying out the Joyo task on 29 September 1999 were aware of all
these bypass options. They were documented in the informal building procedure
used for the Joyo task. They were used on other tasks in the Conversion Test
Building and the procedure was easy to understand and easy to implement.
(The plant engineering was such that the bypasses were relatively easy to
implement. This is a design safety issue but outside the scope of this text.)
On the fatal days of 29 and 30 September the Joyo task crew added a final
contravention to the approved operating procedure, causing a major criticality
event. The engineered equipment and procedural defence in depth, put in
place to reduce the risk of a criticality event for the operators, the public and
the asset, were finally and completely defeated by the crew.
Following custom and practice, the nitrosyl mixing was to be done using the
unapproved process route and bypassing the plant except the buffer columns.
However, before starting the task the crew had been considering additional
ways to save more time in the task. This proved to be fatal.
as is known, only the task crew and the quality department where involved
in this decision to use the precipitation tank in the revised Joyo process. The
intent was now to bypass in its totality all the building’s engineered equipment
including the geometrically safe buffer storage columns. Each 2.4 kg U nitrate
batch mixed in the stainless-steel bucket, one after the other, would all be
poured into the precipitation tank.
At this point it is important to note that the critical mass of U3O8 in a spherical
geometry and fully water reflected (leaking neutrons bounce back into the
system) is about 6.5 kg U. This is a sphere approximately of 100 millimetres
diameter. One stainless-steel bucket contains 2.4 kg U in solution. In the worst
geometry and fully water reflected the sphere would be about 75 millimetres in
diameter; 2.4 kg U is sub-critical even with optimum reflection that is, totally
surrounded by water. As more batches are accumulated and the total mass
of uranium is accumulating, how and where the batches are stored before
discharge to the transport flasks becomes very important. The buffer storage
columns were acceptable as they were geometrically safe. However, the crew
making the additional procedure violation intended to use the unsafe geometry
precipitation tank for the final mixing and storage of the full 16.8 kg U enriched
mass of uranium, this is just over two and a half critical masses. This decision
was taken with a total lack of criticality safety knowledge.
96 Safety Culture
A small quantity of pure water was carefully measured, added to the powder
and mixed to a smooth paste. This was followed by the careful measuring of
6.5 litres of nitric acid into the bucket and the contents manually stirred. By this
means one batch of 2.4 kg U material was made into an initial mix of uranium
18.8 per cent enriched uranium uranyl nitrate solution. This was poured, using
the funnel to avoid spillage, into the precipitation tank. Following further
mixing with the vessels mechanically driven paddles, the next batch of bucket
solution was added.
The first batch took about one hour to produce and the task was to make
up 16.8 kg U in total. This would require seven batches to be prepared and then
one after the other poured into the precipitation tank for additional mixing.
This task would go through the afternoon and into the next day.
By the end of the days shift on 29 September 1999, four batches, or 9.6 kg
U, of U3O8 had been made and the solution poured into the precipitation tank.
Noting that the critical mass, in a worst geometry and fully reflected, is about
safety-culture theory as a predictive model 97
6.5 kg U, the mass from four batches was outside the safe operating envelope
of the facilities safety case. However, there was no criticality event. The
uranium was dispersed in the solution and not in its worst geometrical shape
arrangement. The crew finished the day’s work after the fourth batch, switched
off the mixer paddle and went home. Switching off the paddle is considered as
a possible physical contributor to the eventual criticality event.
During the following morning, the crew continued working by mixing the
remaining three nitrate solution batches in the stainless-steel bucket and doing
transfers to the precipitation tank. At 10.35 a.m. the seventh and final batch was
poured into the precipitation tank. It now contained 16.8 kg U (over two and
a half critical masses) of U3O8. With the pouring into the vessel of the seventh
batch, a criticality event occurred (Figure 2.5). It instantaneously delivered
98 Safety Culture
huge doses of neutron and gamma radiation to operators ‘A’ and ‘B’ local to
the vessel (Figure 2.9). The supervisor ‘C’, who was about four metres distance
away, received a much lower radiation dose.
The operator ‘A’, holding the funnel, received a fatal dose of neutron
and gamma radiation and died 82 days later. Operator ‘B’ pouring the liquid
received a lower but fatal radiation dose and died 210 days later. The supervisor
‘C’ received a lower but very significant radiation dose and survived. However,
he is likely to receive continuing medical attention and observation for many
years.
The physical causes of the criticality are associated with the unsafe geometry
of the precipitation tank, excess uranium mass and neutron reflection by the
vessel’s water jacket. When the operators switched off the tank’s mixing paddle
and left the material to stand over the night of 29th and early morning of 30th
September, it is believed that solid uranium nitrate precipitated or crystallised
out of solution. This formed a mass geometry conducive to a criticality in
safety-culture theory as a predictive model 99
the domed base of the vessel. Adding more uranium to the vessel from the
seventh batch contributed to the formation of a critical mass. The vessel was
also geometrically unsafe. There was limited neutron leakage from the system
and this was further aggravated by any naturally occurring spontaneous
fission neutrons being reflecting back into the system from the water in the
cooling jacket. Neutrons from any initiated fissions were also reflected back
into the solution causing, under the right uranium mass conditions, the full
development of a chain reaction, the criticality.
From the initial criticality not only were the task crew and some employees
local to the plant irradiated, but also public residents adjacent to the site.
These were low-dose exposures but above what would be accumulated in the
same period of exposure from background radiation, (long-term health effects
would not be expected). People up to 10 km from the plant were advised, by
the authorities, to remain indoors as it was unclear whether fission products
(Figure 2.5), had become airborne. (This may appear in retrospect to be an over
reaction. However, at the time it was a sound, conservative decision in the face
of uncertainty.)
There was no trained emergency procedure in place to deal with the event
and the criticality was eventually brought under control by the local fire service
and drafted in specialists and experts. Some service personnel where exposed
100 Safety Culture
The application of the generic model, Chapter 1 of this text, may demonstrate
that it was inappropriate shared organisational safety-beliefs that shaped
manager’s behaviours. These beliefs were cascaded to the plant operators,
becoming fundamental to the event occurring. The operators were, it may be
suggested, the victims of a poor management safety-culture.
The safety-culture generic model suggests that shared beliefs and values
of an organisation are shaped by the senior management. The culture is
cascaded by other layers of management to the workforce. The safety-beliefs
of the workforce will be influenced by this shaping. They can, in addition, be
modified if a local workforce has a strong subculture. The Joyo task crew had
possibly developed a strong subculture and generated group safety-beliefs that
exacerbated the situation.
There are therefore two groups to consider: the management and the
workforce.
Beliefs (managers)
• There was a strong belief that the workforce were not capable of
being trained in criticality safety
• There was a belief that the workforce need not be engaged in safety
decisions
102 Safety Culture
• There was a belief that it was acceptable to break the legal requirement
to inform the regulator of significant changes to plant or process
• There was a belief at all management levels that they had the correct
model of criticality, that is, mass control is all that is required
Espoused values
Attitudes
• Of contentment arising from the use, for many years and familiarity
with, the site’s plant and their processes
• Arrogant in having what they thought was the correct model for all
criticality scenarios (this attitude was demonstrated in particular by
the safety committee).
Artefacts
The business beliefs of the senior managers were cascaded down the
organisation. The belief was that production and quality are the priority in
the interests of business survival. This overwhelmed any founding beliefs in
nuclear and radiological safety. There was a lack of rigour and discipline in
human behaviours and lack of robustness in the documented risk management
structure. The beliefs were supported by the line management, the safety
committee and supported at the workforce level. This poor culture expressed
safety-culture theory as a predictive model 105
In the Tokaimura event, there is a need to consider the culture of the Joyo
task crew. They were the people who made the final violation causing the
criticality event. A judgement has to be made as to whether this behaviour
was reckless or arising from the organisations and their own shared beliefs in
attempting to do a ‘good job’.
Mass control was a quality issue for each batch of material produced (it was not
recognised as a criticality control, of this they had no knowledge):
• They did not need to know about criticality; this was an issue
addressed by others on their behalf (for example through the
written chemical process procedures)
• Completing the Joyo task quickly would show efficiency and enable
them to return to their routine missions
• They believed they had to look for additional time savings because
of the supervisor’s pressure on task time
safety-culture theory as a predictive model 107
• Firmly believed the use of the precipitation tank would solve many
problems, save time, show improved efficiency, meet the time
pressures, give good homogeneity of product (using the mixer paddle
mechanism), and enable efficient filling of the transport flasks
Espoused values
• Did have concerns about time delays and (valued) the need to
reduce the Joyo task time
• Did not value the Joyo task as they saw it as additional work on an
already full work-load
Attitudes
• Confidence that they knew the safety issues associated with the
task
Artefacts
• Did not read the approved procedure for producing uranyl nitrate
solution
The shared organisational beliefs, and the beliefs and values of the crew
expressed through behaviours, were major factors in the criticality event
occurring. The crew never appeared to challenge the use of the unapproved
procedure, or stop as a group to think about the implications of using the
precipitation tank. However, in mitigation, their own poor safety-culture was
shaped and constrained by the management culture and imposed working
environment. Finally, they were not trained for the task they had to do. This,
and their lack of knowledge of criticality, were antecedents over which they
had no control.
When not in operation there was safe storage location for the source rack.
This was a 3 m long by 2 m wide, 6 m deep water pool. The water acted as a
gamma shield when the sources rack was located on the pool floor. With the
rack at the bottom of the pool, the gamma radiation dose rates in the irradiator
room were negligible, enabling access for maintenance task.
safety-culture theory as a predictive model 111
There were emergency override buttons and internal lock overrides inside the
facility should somebody find themselves inadvertently locked in the plant.
Having completed the preliminary checks the PLC would open the
interim area entry barrier door and the loaded conveyer belt would take the
medical equipment products around the interim area and into the irradiator
room. Once the product was in the room, and with the assurance that all the
access doors were interlocked, the PLC would raise the sources rack. The rack
emerged vertically from the pool into the irradiator room. At this process stage
the rack was located unshielded above the pool surface. The medical products
were panoramically exposed to high levels of intense gamma radiation for an
assigned time. On completion of the task the PLC automatically returned the
source rack to its storage position at the bottom of the pool providing six metres
of water-shielding.
After checking that the various rack position limit switches were signalling the
correct data and being assured that the source rack was safely in the pool and
shielded, the PLC released the interim area exit barrier door lock. The sterilised
product was conveyed through the interim area and into the unloading area
where the product was removed manually from the conveyer and made ready
for shipping off site to clients. The cycle was then complete and all access door
locks released.
This is a simple and effective system but because of the intense gamma
radiation it is hazardous. However, such facilities are engineered with defences
to protect the operators and minimise radiological safety risks.
The Puerto Rico facility has several safety features. The building’s irradiator
room and interim area had a combined length of about 24 m and the rooms were
about 14 m wide. To protect the operators outside these rooms from the gamma
radiation the plant walls and roof were over 2 m thick reinforced concrete. This
ensured gamma dose rates outside the rooms, in the operating areas, were
within design criteria when the product was being irradiated. The doors into
the facility were steel, typically 150 mm thick; this attenuated any radiation
scattering down the maze or into the interim area from the unshielded sources
when raised from the pool in the irradiator room.
The primary role of the six-metre-deep safe storage pool was to water
shield the rack when it was located on the pool floor. With the sources in this
location the irradiator room gamma dose rates were negligible and operators
safety-culture theory as a predictive model 113
and maintenance teams could safely access to the maze, the irradiator room
and the interim area.
If the rack was lifted to the surface of the pool the sources were by design
unshielded so they could do the task of irradiating the medical products. The
sources were highly radioactive and could deliver a massive lethal gamma
radiation to a human being within seconds. With such facilities the protection
of the operator from high-radiation fields has to be the primary focus at all
times in design and operation. The operator has to be prevented from entering
the interim area and the irradiator room when the unshielded rack is present.
To prevent access during irradiator operation, the door interlock system, the
instrumentation and the procedures have to function as designed at all times.
The Puerto Rico plant had engineered ‘defence in depth’ equipment to protect
against inadvertent entry into the irradiator room during normal operations.
The plant had a control room adjacent to the maze entrance (Figure 2.10).
This housed the PLC for controlling the product irradiation processes. The PLC
was fitted with various safety instruments and warning lights.
The PLC automatically initiated door interlocks and initiated the product
conveyor movements. It also automatically raised and lowered the rack from
the pool. As an additional safety defence the sources rack lifting system had
been designed such that if there was any identified fault on the rack position
limit switches, located on the gantry rails, the rack would automatically fail-
safe and be driven down to the pool floor as a precaution. The limit switches
were termed the ‘rack up’ and a ‘rack down’ limit switch. These informed the
PLC of the location of the rack, whether shielded at the bottom of the pool or
raised and unshielded in the irradiator room. The automatic lowering of the
rack under switch fault conditions was a designed ‘inherently safe’ feature of
the rack’s operation. The PLC could interrogate the switch data information
and any failure of switches to close or to open or, if they contradicted the PLC’s
expectations, the rack descended to the safe position on the pool floor. The
managers and operators were fully aware of this safety feature.
The plant was fitted inside the irradiator room with gamma ray detector
heads called the L110 probe which was a set of Geiger counters. The electronics
for interpreting the L110 device signals where in a unit located on the wall
in the control room. This was referred to as the L118 monitor. If the source
was unshielded in the irradiator room the gamma detector heads sent a read
out signal and an audible alarm to the L118 unit. Unfortunately in the design
114 Safety Culture
concept, if the irradiator source was unshielded the L110 detectors could
overload and potentially damage the electronics of the L118 unit. To protect
the L118, when the source was unshielded, the L110 gamma detector was
electronically disengaged from the L118. There was therefore no permanent
irradiator room gamma radiation level information for the operators when the
source was unshielded above the pool. The read out information and audible
warning could however, be activated by a switch located on the L118 unit. The
L110 data could also be accessed and the electronics re-engaged by a switch on
the PLC console panel. This would then give a reading of the gamma levels in
the irradiator room. Due to inadequate training, some operators on the plant
believed that with the sources unshielded the L118 was never functional and
they did not understand that the L118 could be interrogated from the PLC
console or by the switch associated with the L118 unit. The L118 was thought to
only come into action when the racks were shielded. That is, by pressing a test
button on the L118 it would indicate there was no gamma radiation field. This
was a null signal decision system. If the L118 was giving a null signal the maze
interlocked door could then be opened. This system appeared to be of poor
design, inadequate as a positive gamma detection information system and not
fully understood by all the operators.
As a further safeguard the rack gantry was fitted with a ‘source travel
alarm’. This alarm indicated rack movement and positioning. If the rack was
at the bottom of the pool with the bottom switch closed this alarm stopped.
There was also a ‘rack fault indicator’ which warned if the PLC was receiving
a rack limit switch contradiction. This was a warning to the operators that the
rack may not be in its shielded position at the bottom of the pool. A ‘rack fault
indicator’ arose when the PLC interrogated the behaviour of the source rack and
determined whether it was shielded. When shielded the PLC should receive a
rack ‘down switch’ closure signal. Alternatively, if the rack is unshielded, that
is out of the pool, the PLC receives a closed signal from the ‘up switch’ with
the ‘down switch’ now signalling open. Any fault or contradiction between the
signals, for example both switches open yet one should be closed, then the rack
automatically fails safe and is driven down to the storage pool floor.
For the door interlocks to be released the gantry bottom switch has to be
closed and the top switch open. This was one of several indications to the
operators that the source is in the pool and shielded.
In the event of a stuck rack there was a ‘stuck rack’ approved emergency
recovery procedure to be implemented. A signal at the control panel from the
safety-culture theory as a predictive model 115
‘source travel alarm’ (that is, the rack is not moving but is somewhere between
the two switches) plus a signal from the ‘rack fault indicator’ would initiate the
emergency procedure.
The final safeguard was that that before the PLC could initiate the irradiation
processes all access doors were interlocked. This was initiated by the L118 unit
identifying high gamma radiation then, as noted above, it disengaged from
the monitor L110. With the interlocks in place the maze door could only be
physically opened from inside the maze. This was not the case for the interim
room exit barrier door. Its interlocks could be bypassed and the door opened
from the loading bay even if the irradiator was unshielded above the pool.
If entry was made to the irradiator room, operators had available hand-held
gamma-detector monitoring equipment. The use of pocket gamma dosimeters
was mandatory for entry.
The plant procedures generally were ad hoc and not within an integrated
management systems. Manufacturing had separate operational and auditing
procedures from the safety procedures. The safety procedures were not
independently audited nor was there any scrutiny of the plant’s safety-
performance. The radiation safety officer (RSO) was responsible for designing
the safety enhancement programme and safety survey monitoring but there
was no independent oversight of either of these tasks.
The plant did have safety procedures in which the operators were initially
trained. However, refreshers were inadequate as re-qualification courses for safe
plant operations. The safety-performance of staff was not annually reviewed.
Emergency procedures were exercised. However, the same scenario was
repeated every year. There were several different plant emergency possibilities
but these were not exercised. Although perhaps speculative, it would appear
that the minimum documentation was produced and the emergency exercises
done only to meet licensing requirements. Inquiries after the event strongly
suggest that the safety documentation was of poor quality. A robust safety-
management system was not embraced by the management as an important
tool contributing to risk management.
the plant manager, there was a team of about 25 people on each shift. The plant
operated three shifts with a shift handover period. The facility operated seven
days a week as there was a production schedule to meet which was managed
by the manufacturing director.
The Puerto Rico irradiator facility came into operation in 1983. It had been
plagued for many years with source rack position switch problems. Matters
became worse in 2001 when the plant’s original relay control console was
replaced by a computer-based PLC system. There were repeated problems
with the PLC along with wiring, cables and switches. Over a two-year period
there had been many recorded switch repairs. All repairs to the ‘up’ and ‘down’
switches required the irradiator to be closed down, the sources rack to be
shielded in the pool and personnel access made into the irradiator room.
If the PLC identified rack switch faults, the ‘rack stuck’ emergency procedure
was implemented. This arose because the door interlocks remained in place
and entry to the plant was not possible. To gain entry required bypassing
the safety systems. As part of the procedure and before entry was permitted
various external radiation monitoring surveys were carried out. Checking the
rack rope housings for the extent of rope run and interrogating the control
panel was carried out. The L118 radiation readings were checked using the
test button to test if the L118 was engaged as this would indicate low gamma
radiation in the irradiation room. Finally entry would be authorised by the
RSO or assistant RSO. Access was then achieved by over riding the barrier
door interlocks (Figure 2.10). This was permitted by the approved emergency
procedure. Pocket dose-metres were issued to maintenance staff and controlled
entry was made holding radiation monitors as an additional precaution.
Experience showed that after using the procedure the radiation dose rates
were always found to be low from the external surveys and on entry. The
‘inherently safe’ design feature initiated when a switch fault was detected by
the PLC always sent the source rack down to its shielded safe location at the
bottom of the pool.
at the bottom of the safe pool. The radiation dose rates in the interim area
and the irradiator room were ‘always’ found to be negligible under fault
conditions.
Due to the number of faults occurring, access time to the irradiator room
had to be reduced and to the use of the emergency procedure avoided. To
achieve this aim an unapproved undocumented ‘trouble shooting’ procedure
was introduced. This was applied over many years up to the incident in 2004.
It was introduced by the plant management but not approved by the regulator
and it was not part of the formal safety-management system. To implement the
‘trouble shooting’ procedure authorisation had to be obtained from the RSO or
the assistant RSO.
The ‘trouble shooting’ procedure did reduce access times but to achieve
this, managers were party to and condoned the breaking of procedure, the
plant licence conditions and the law.
There had developed a deep-seated shared belief that if the doors were
locked by the PLC under fault conditions and could not be opened then this was
always a switch problem. The belief emerged that the access doors remained
locked because the PLC interpreted the various switch position fault signals as
a contradiction. Further, there was the strong shared belief that the source rack
‘always’ went down into safe pool storage under fault conditions. That is, the
plant was believed to be inherently safe. Due to the strength of this belief the
radiation fields were not always checked using the L118 monitor test button
before entry.
The event, which was almost fatal for two operators, occurred on 21 April
2004.The night shift had handed over to the day shift and indicated they had
118 Safety Culture
been having problems with the sources rack ‘up switch’ throughout the night.
The schedule had been delayed due to the rack going to safe storage at the base
of the pool and halting the process. Entries had been made to the irradiator
room in attempts to fix the switch. When the day team took over the irradiator
room was safe and the facility accessible via the maze door.
The morning shift made various attempts to repair the ‘up switch’. The
switch located on the rack gantry guide rails, was in a difficult position (design
issue) being located close to the irradiator room ceiling. In the first intervention
tools and a ladder were taken into the room to address the problem. The ladder
was placed across the pool and leaned against a wall enabling access to the
switch. On task completion the tools and ladder were removed back into the
control room area. An attempt was made to re-start the irradiator; this failed
and entry was again made. The tools and ladder were brought back into the
irradiator room.
Several further attempts were made to fix the switch and restart the plant.
At 10.45 a.m. the management agreed to stop attempting the switch repair
and to replace it. Entry was again made and by 11.30 a.m. the replacement was
complete. The personnel and maintenance equipment was removed from the
room and an irradiator start-up made. This time, the first steps of the sequence
initiated and the doors automatically interlocked. The new switch then failed
again and the procedure automatically stopped.
The PLC received a switch fault signal and interpreted it as designed, that
the ‘up switch’ was closed and that the source may be unshielded, that is, out of
the pool. As usual the operators ‘knew’ this was incorrect; the problem would
be the switch and the rack would have failed-safe and be shielded. The group’s
mental model was that the rack with its inherently safe design feature would
be on the pool floor and not hazardous. Without investigating further or testing
the L118 gamma radiation monitor the maintenance team requested the use of
the ‘trouble shooting’ procedure to bypass the exit barrier door interlocks and
gains access. They would then open the maze door which remained interlocked
from inside the maze and again take in the equipment for another repair. The
use of the informal procedure was granted by the assistant RSO. The fuses
were removed from the PLC, the exit barrier door interlock bypassed and
entry made. As always the source had gone into the deep pool as expected. The
irradiator room was traversed and the maze door was then opened. The tools
and ladder were again brought into the irradiator.
safety-culture theory as a predictive model 119
The ‘up switch’ was examined and repairs made. The operators left the
irradiator room, took out the tools but forgot the ladder. This remained over
the pool. There was no procedure for checking equipment into or out of the
high-hazard irradiator room.
In the control room attempts were made to restart the PLC control panel.
The PLC had an hour or so earlier been interfered with to gain access to the
irradiator room. The PLC failed to operate correctly and a technician was called
out to bring it back on line.
With the PLC fixed the plant was restarted at 12.50 p.m. The sequence
started and the access doors’ interlocks came into place. As required by the
irradiation process the sources rack emerged from the pool into the irradiator
room. The ‘up switch’ failed again. As its fail-safe design intended the source
rack began to descend automatically into the pool for safe storage.
However, the rack struck the ladder and jammed. The radioactive cobalt-60
pencil sources were above the pool’s water surface. As the design intended, intense
gamma radiation irradiated the room and high levels of scattered radiation was
irradiating the interim area (Figure 2.10). The operators were unaware of this.
A request was again made to use the ‘trouble shooting’ procedure for entry
into the irradiator room to fix what was believed to be the switch problem. This
request was granted by the duty assistant RSO.
On this occasion the PLC was sending additional signals to the operators:
• The ‘rack fault’ indicator was illuminated suggesting that both the
‘up switch’ and ‘down switch’ were open. This can happen if there
is a down switch fault or if the rack is actually not at the bottom of
the pool and the bottom switch is not closed.
When both these signals arise it is mandatory for the managers to implement
the approved ‘stuck rack’ emergency procedure. However, the signals did not
120 Safety Culture
match the group’s mental model and their belief that the rack would be on the
pool floor and safely shielded. Because of this belief the PLC signals and the
approved emergency procedure were ignored.
Because the operators and managers had a model of what was wrong and
believed fully in the inherent safety features of the plant they accepted plant
entry under these conditions as ‘routine’. Without the true knowledge of the
situation and with the irradiator room under intense irradiation, the PLC fuses
were removed and the exit barrier door interlock was bypassed.
Two operators were ready to enter the facility. One was not issued with
the mandatory pocket dosimeter which was against procedure. There was no
radiation survey carried out at the barrier door before entry, but the leading
operator did pick up a hand-held dose meter as he entered. This act saved both
men’s lives.
The operators entered the interim area and unknown to them gamma dose
rates were extremely high from radiation scattering out of the irradiator room
(Figure 2.10). By a stroke of luck the lead operator put his hand-held radiation
meter on the floor to remove some debris obstructing their entry path to the
irradiator room. He noticed the huge dose rate measurements on the monitor
and immediately ordered a retreat to the loading bay and closed the steel
barrier door behind them.
With the unshielded radiation source rack jammed against the ladder the
operators would have been fatally injured, within seconds, if they had entered
the irradiator room or lingered for longer in the interim area.
Louis Harold Gray). An acute dose, a little higher than this, is fatal for all. In
the interim room the dose rate was approximately 150 Gy per hour and in the
irradiation room it was on average about 1,000 Gy per hour. These are huge dose
rates. The leading operator with the monitor received approximately 45 mGy and
the other almost 30 mGy (this latter dose was estimated by radiation specialists
as this operator was not wearing the mandatory pocket dose-meter and his
actual received dose was not known). These acute doses were delivered in about
10 seconds from scattered radiation, near to the barrier door and at a distance of
over 15 metres from the sources rack. Any further steps would have proved fatal.
The doses received are typically 20 times normal annual background radiation
doses. The operators were saved by good luck and the exposures are unlikely to
have long-term health effects, although there is a possibility of a slight increase in
long-term cancer risk. If they had entered the irradiator room in direct line with
the unshielded cobalt-60 isotope sources, they would have received a fatal dose
of radiation within about 20 to 30 seconds.
The evidence from this event indicates that it was good luck that saved
these operator’s lives. There was engineered protection and instrumentation,
although both could have been significantly better. Even though the safety-
management system structure, review and auditing were poor, there were
emergency procedures to follow.
It is suggested that these operators came close to a fatality due to poor safety-
culture within the facility arising from people with long service being complacent
and over-confident that they knew exactly how the plant operated and the
solutions to its ills. If the generic safety-culture model is applied it may perhaps be
confirmed that complacency and over-confidence were a significant factor.
• Believed getting the plant back into operation was a priority when
running a production schedule
Espoused values
• The long safety experience of the RSO and the plant supervisors
was possibly valued.
Attitudes
This is speculative but there are very strong indicators that there was:
The management and operators did believe that they were dealing with
a hazardous facility. They would have been aware of the consequences of
exposure to the sources and strongly believed that exposure had to be avoided.
Nevertheless, these beliefs became overwhelmed by the inherent safety features
of the plant and the belief in their own knowledge and capability. This, with
other pressures, changed the staff’s safety-behaviours.
The attitude of the plant employees may have been shaped by complacency
arising from managers having been on the plant too long. They appeared to
no longer ‘recognise it as a high hazard facility’ but as a production plant
having repeat problems that needed to be fixed. Plant familiarity over many
years could have created complacency. There may also have been a high level
of boredom both at management level (20 years of service on the same small
facility, for example) and also boredom at the workforce level (facing the same
problems time and again without them ever being finally resolved).
Artefacts
• The plant was instrumented for safety but the instrumentation was
not assisting the operators. (PLC signals did not match the group’s
‘experience and reality’ and were ignored or re- interpreted to fit
the group’s mental model.)
Resulting behaviours
The beliefs, values and poor artefacts, it is suggested, produced the following
behaviours leading to the radiation exposure event:
• The safety interlock on the exit barrier door was regularly physically
bypassed
• Signals from the PLC that did not fit the group’s beliefs on the status
of the source rack were ignored or reinterpreted
• The PLC was used only as a means of driving the process and not
as a safety information and control system
• Time and effort was not invested in good training, effective re-
qualification of operators, nor in emergency exercises.
The shared belief and values identified from this event were shaped and
encouraged by management. The workforce appeared to share the beliefs. The
opportunity of the workforce to challenge on safety issues is not reported and it
is unclear if a management ‘fear culture’ existed. The workforce may have just
accepted the knowledge, experience and leadership of the management.
The mental model of the functioning of the sources rack under switch
fault conditions had become entrenched, agreed and shared. The signals from
the PLC that the rack could be unshielded and hazardous were dismissed.
Experience over many interventions had shown that the rack always failed-safe
and was always shielded below the pool water. This behaviour of dismissing
the PLC data became a norm. The PLC was no longer contributing to safety
but was being used only to drive the conveyer and irradiation process. Any
deviation from the mental picture of the racks position (for example, the rack
actually being jammed and unshielded) caused the operators to go into denial
and apply their norms.
and emergency procedure were put in place to protect the operator and assist
in fault diagnosis.
Summary
• Titanic
• Bhopal
characteristics and attributes are in place. If the support is weak in the two
dimensions this may demonstrate that the expected safety characteristics are
not developed or are organisationally weak.
For high-hazard low-risk industries there are broadly agreed good safety-
culture practice characteristics. If found to be in place, they may suggest a
strong organisational safety-culture that is meeting the ethos, the safety beliefs,
expressed in the definition:
E Safety is learning-driven
Source: Reproduced by the kind permission of the IAEA, modified by the author for
this text.
The attributes (for example, Table 3.4) are factors expected to be present
for a good practice safety-culture characteristic to be displayed within an
organisation (Table 3.1). An objective of an organisational safety-culture
review is to assess if observed staff behaviours have been established by the
presence of the attributes and hence support the existence of the good practice
characteristic. If, inter alia, the expected good safety-culture behaviours are
observed, then the supporting attributes will be present and the majority of
common characteristics established. This infers that ‘good’ shared safety-
beliefs and values have been adopted, promoting the required safety-
behaviours. On the other hand, if the characteristics and attributes are weak
this will be reflected in the values, artefacts, attitudes and behaviours. This
may reveal organisational beliefs and values which may not be supportive of
a good safety-culture.
Each of the five safety-culture characteristics (Table 3.1) has typically six to eight
attributes. As may be expected for a complex multi-dimensional concept there
is some variation across the literature. Nevertheless, there is broad consensus
on the majority of attributes and when established in an organisation the
expected supporting safety-behaviours and documented systems should be
present. It is assumed the attributes noted in Table 3.4 could apply to a high-
hazard facility but not necessarily to a design organisation. Although some are
generic, a design organisation’s safety-culture attributes, for example, will not
all be the same as those for organisations physically operating chemical plant,
nuclear power stations or a railway. The safety-culture characteristics, however,
will be common. The linkage between attributes, behaviours, documented
expectations and the characteristics, attributes and behavioural expectations to
support the characteristics are shown in Table 3.4.
We believe:
1. The safety of staff, our contractors and the public is our number one priority in all
circumstances.
2. Accountability for safety rests at all times with managers.
3. A responsibility for safety rests with all employees.
4. In safety, vigilance is needed at all times.
5. Human error is normal and can be expected.
6. Our engagement in safe behaviours is necessary for safe operations.
7. Human errors are a learning opportunity.
8. People are fallible and will make mistakes.
9. In a culture of legal compliance as our minimum requirement.
10. In a ‘just’ safety-culture and that people do come to work intending to do a good job.
Table 3.4 links the beliefs and values to the characteristics through the
supporting expected behaviours. If in a review the observed staff safety-
behaviours are not as expected and not supporting the attributes then
this suggests the stated safety-beliefs and values are not being upheld. If
this is identified in a review then the focus comes upon identifying the real
organisational beliefs causing the non-aligned safety-behaviours. Further, an
observed disconnect between the documented expectations, the formal culture
dimension and observed behaviours may suggest that the safety-management
system is weak, not owned or not used by the organisation.
The contents of Tables 3.4, 3.5 and 3.6 form the components of a review
process for three of the five characteristics. In a full review, similar tables of
attributes and behaviours would be developed for the remaining characteristics
D – ‘Safety is integrated into all activities’ – and E – ‘Safety is learning-driven’.
Their expected attributes for a high-hazard organisation are show in Table 3.7.
assessing organisational saefty culture 135
We value:
1. Our individual attention to safety is a condition of employment.
2. That everyone is responsible for safety, our own, others’ safety and the protection of the
facility.
3. That respect is given to all safety views as everyone has the right to question and report safety
issues.
4. People’s interventions to ensure all potential health and safety incidents are prevented.
5. Everyone has the right to challenge on safety issues.
6. That the organisation strives for an open dialogue culture.
7. That teamwork to resolve safety matters is strongly supported.
8. That all events and near misses are reported as we recognise that even minor injuries or
events are important.
9. Thorough safety training and competence as essential for safe working.
10. That we regularly check and report our safety objectives performance.
Safety-culture Supporting Expected safety behaviour to support Documented expectations to support the The safety-beliefs The safety values
characteristic attributes the attributes and the characteristic attributes and the characteristic being supported by being supported
the presence of the by the presence of
characteristic the characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic A A/1 Safety is given – Safety procedures are complied with – Documented health and safety policy exists 1,2 1,2
‘Safety is a clearly a high priority at all – In the face of uncertainty conservative safety – Meeting-minutes demonstrate safety before
recognised value’ levels decisions are demonstrated at all levels production
– Decisions made show safety was/is – Meeting-minutes indicate conservative
considered above production decision making
– There is evidence of a proactive approach to – Safety resources are defined in the annual
resolving problems financial budget
– Safety has a high profile in the organisations
annual report
Characteristic A A/2 Safety values – Safety booklets are prepared (as a route to – Organisational shared safety-beliefs are 1,2,3,4,5 1,2,6
‘Safety is a clearly are communicated sharing beliefs and values) documented and displayed
recognised value’ – Safety posters are displayed and updated – Organisational shared safety values are
– Safety news letter is made available documented and displayed
– Safety intranet is made available – Safety communications procedures are in
– Regular management safety place and used
communications, including emphasis on – A variety of communication routes are
beliefs and values available, resourced and used
– Safety is proactively given high priority in
internal documents, verbal communications
and decision making
Table 3.4 Continued
Safety-culture Supporting Expected safety behaviour to support Documented expectations to support the The safety-beliefs The safety values
characteristic attributes the attributes and the characteristic attributes and the characteristic being supported by being supported
the presence of the by the presence of
characteristic the characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic A A/3 Safety – Giving and accepting behavioural challenge – Safety-culture behavioural requirements for All 1,2,3,4,5,6,8
‘Safety is a clearly conscious is a norm each hierarchical level are documented
recognised value’ behaviour is – Questioning peers and seniors on safety is – A blame tolerant, ‘just’ culture policy exists
accepted by an accepted norm with appropriate supporting documentation
all within the – Reporting events and near misses is – An events and near-miss reporting system is
organisation rigorously carried out in place
– There is proactive searching in operations – There is a procedure for analysing with
for technical faults and human error trending and feedback on reported events
– Positive feedback is given for ‘good’ safety- – A learning-from-experience procedure is in
behaviours place
– Reporting of events is seen as an act for – Soft-skills training is in place for all employee
learning by managers and workforce levels to enhance interpersonal skills
– Safety and safety-culture observations are
carried out by all and across all the hierarchy
– A ‘just’ culture process is applied by all
managers
Characteristic A A/4 Acceptance – Safety is on all business agendas – Business-meeting minutes demonstrate 1,2 1,2
‘Safety is a clearly that safety and – Safety and production are integrated in safety with production is agenda item
recognised value’ production go hand discussion at local plant level (tool box talks) – Documented guidance exists for conducting
in hand –Safety beliefs and values are regularly local tool box talks
referred to at meetings as ‘touch stones’ and – Guidance that is used for managers to
safety guides for the organisation manage production and safety conflicts
– Safety procedures are never bypassed to – Procedures emphasise safety as a
achieve production goals prerequisite for effective business production
– The safety of people, plant and process are in high-hazard industries
integrated into all production objectives and – Procedures exist that integrate safety
planning considerations considerations into production planning and
scheduling
– The consequences to individuals for safety
procedures bypassing, at any hierarchical level,
are available, trained and understood within
the concept of a ‘just’ culture policy
Table 3.4 Concluded
Safety-culture Supporting Expected safety behaviour to support Documented expectations to support the The safety-beliefs The safety values
characteristic attributes the attributes and the characteristic attributes and the characteristic being supported by being supported
the presence of the by the presence of
characteristic the characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic A A/5 Resources are a – Long-term safety resource requirements – Annual business-planning meeting 1,2,6,8,9 1,2
‘Safety is a clearly priority for safety are discussed at annual business-planning documentation identifies safety resources
recognised value’ meetings – Artefacts, for example personal protective
– Resources are made available for equipment equipment, are of good quality
to be maintained or replaced (equipment not – Documentation does not suggest excessive
permitted to degrade) breakdown of safety equipment due to allowed
– Suitably qualified and experienced people degradation
are identified to competently operate facilities – Safety skill requirements and task skill
and at management levels requirements are documented with training
– A management of change procedure is used strategies, recruitment strategies etc.
for structural changes in staffing at any level – A documented and used (register)
management of change procedure exists
– Safety training is resourced and functioning
Characteristic A A/6 There are – Teamwork to resolve safety issues is strongly – Training is implemented on team working and 1,2,3,4,5,6,8 1,2,3,4,5,6,8,9
‘Safety is a clearly interactions encouraged by all employees interpersonal skills requirements at all levels
recognised value’ at all levels of – Managers actively apply processes to –Guidance is available on dealing with safety
management and obtain feedback from all employees on safety feedback and tool box talk feedback
workforce matters – Feedback questionnaires exist, intranet
– Local ‘tool box talks’ address safety by options are available for feedback, and a fast-
dialogue before tasks; current task safety risks track feedback process is available
and past safety experiences are addressed The systems are used, reviewed and feedback
(this would include in depth discussion of any acted upon through a Learning from Experience
expectations from the task risk assessments) procedure
– Senior managers have organisation
safety briefs relating safety matters to the
organisational safety-beliefs and values
Source: Characteristics and attributes (only) reproduced by kind permission of the IAEA, (attributes modified by the author for this text).
Table 3.5 Safety-culture characteristic B – ‘Leadership for safety is clear’, attributes, safety-behaviours (informal
dimension) and the documented expectations (formal dimension)
Safety- Supporting Expected safety behaviour to support the attributes and Documented expectations to support The safety-beliefs The safety values
culture attributes the characteristic the attributes and the characteristic being supported being supported
characteristic by the presence of by the presence of
the characteristic the characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic B B/1 Senior – Managers behave as safety role models, set examples, mentor – Interpersonal skills and ‘soft skill’ 1,2,3,4 1,2,3,6
‘Leadership for managers, through their behaviours for example challenge and question. training is in place for managers at all
safety is clear’ middle – Managers refer to and use at all opportunities the organisational levels and is refreshed through procedure
managers and safety-beliefs and values and are seen to ‘own’ them requirements
supervisors – Managers do not ignore unsafe behaviours in others at any level – Managers are issued, or have to hand,
are clearly and will skilfully give constructive feedback documented the agreed organisational
committed to – Managers do (or have) overtly put safety before production by beliefs and values
safety their actions or decisions
Characteristic B B/2 Visible – A manager introduces all employee safety-training – Training procedures identify the role of 1,2,4,5,6,7,8 1,2, 3, 9,10
‘Leadership for leadership is – Managers have stopped, or will stop, work tasks in the interests senior managers in employee training
safety is clear’ shown of safety – Emergency exercise procedures are
– Lead emergency exercises available and the role of senior officers
– Demonstrably follow all procedures (e.g., using the correct explained
personal protective equipment, follow facility access, egress – Guidance is in place for managers and
procedures) other employees to enable tasks to be
– Overtly praise and support observed ‘good’ safety-behaviours halted in the interests of safety
– Managers expect their own occasional misaligned safety- – Schedules exist for every manager to
behaviours to be challenged and questioned walk the facilities and discuss safety. As
– Managers have scheduled facility (or office) walks to observe appropriate written guidance is available
and discuss safety and safety-culture – A summary of reports on facility walks,
– Managers are engaged in a schedule of safety behavioural observations and safety issues arising are
observations; observations being at all levels of the hierarchy taken as a CEO senior managers’ meeting
– All members of staff engage in behavioural observations across agenda item
all levels promoting a dialogue culture
– Managers periodically lead mixed employee teams to resolve
safety issues
Table 3.5 Continued
Safety- Supporting Expected safety behaviour to support the attributes and Documented expectations to support The safety-beliefs The safety values
culture attributes the characteristic the attributes and the characteristic being supported being supported
characteristic by the presence of by the presence of
the characteristic the characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic B B/3 Managers – Managers frequently and overtly encourage team solutions by – Training in team skills is available and 3,4,5,8 3,6,7,9,10
‘Leadership for seek the making time available implemented
safety is clear’ involvement of – Actively engage in safety discussion with staff to gain feedback – Facilitation training for managers and
all employees before a safety decision is made others is available and implemented
in improving – Managers will frequently and overtly delegate responsibility (but – Technical safety training with refresher
safety not accountability) for resolving safety issues to suitably qualified courses are carried out and recorded
and experienced teams – Evidence of team safety solutions and
– Managers encourage professionalism regarding safety by reports are available
encouraging additional training, safety skills improvement and – Appropriate routes are available
soft-skills practices amongst reporting staff and used for ‘suggestions’ on safety
– Encourages overtly open dialogue on safety matters improvements
– Encourages employees to search for and feedback possible
safety improvements
– Openly challenges employees on their accepted safety status
quo
Characteristic B B/4 Managers – Actively listen to employees’ concerns and comments on safety – The organisation has a documented 2,5,7,8 3,6
‘Leadership for strive for open – Always purposefully encourage safety feedback at meetings and safety-issues communications policy
safety is clear’ communications plant visits – There are scheduled senior managers
and build trust – Always fulfil promises/actions with feedback on safety issues safety-communication briefs
raised by staff – Managers are trained in the
– Through practised charismatic leadership purposefully build a organisation’s communications
trusting working environment expectations
– Has an ‘open door’ policy on safety matters – people can raise – Managers are trained as appropriate in
immediate concerns without hindrance communications skill, handling conflict and
– Managers apply procedure to address and satisfactorily close dealing with employee safety concerns
out safety issues (by team engagement where possible) – To facilitate trust in the organisation
there is a ‘just’ working policy that is
known, understood and used
– A learning from experience procedure
with close out is available and used
Table 3.5 Concluded
Safety- Supporting Expected safety behaviour to support the attributes and Documented expectations to support The safety-beliefs The safety values
culture attributes the characteristic the attributes and the characteristic being supported being supported
characteristic by the presence of by the presence of
the characteristic the characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic B B/5 Managers – A minimum safety manning levels document is produced – Plant’s minimum manning levels are 1,2,5,8 9
‘Leadership for ensure by managers and implemented for normal and emergency documented and periodically reviewed
safety is clear’ sufficient, operations through procedure
competent – Skill and manning levels for specific tasks are addressed through – Task analysis documents/risk
individuals task analysis on high-hazard tasks and are carried out within an assessments are in place and reviewed by
appropriate (legal requirement) risk assessment procedure
– Managers have and use processes to ensure safety training of – The organisation has a stress policy that
individuals taken with refresher as appropriate. This is also carried is evidently used
out to meet the requirements of reducing risks on specific tasks – A management of change procedure is in
– Infrequent, unique and unusual tasks are always addressed by place and evidently used to consider the
managers, through procedure, to ensure adequate trained and implications on safety of changes at any
experienced skills and manning levels are available for the tasks level in the hierarchy
– Managers implement a stress-policy to ensure employees are
fit for tasks
– Managers always use a management of change procedure
to examine the safety implications of structural change. This is
applied to all structural changes of personnel
Characteristic B B/6 Managers – Managers have proactive and reactive safety-performance – Reactive and proactive key performance 1,2,3 6,10
‘Leadership for set performance indicators indicators are documented and monitored
safety is clear’ targets and – Managers implement a behavioural safety programme within a – Training, resources and time is made
monitor. mature environment available to implement a behavioural
– Give constructive feedback as frequently as necessary on the safety programme (if the organisation is
required safety-performance of staff mature)
– Frequently examine local safety data for safety degradation – Resources and time are committed
trends or deviations from the norm to examining trends and root causes
– Senior managers annually engage actively in a full review of of deviations. A reporting route to
the safety-management system to seek assurance it is delivering management for findings is available and
organisational safety-management and to consider improvements used
to the system – Procedures are in place to audit the
– Overtly support as a learning opportunity all safety system safety a management system There are
audits; mentors and encourages others in this support documented audit reports, actions and
prompt close out of actions is evident
– Procedures are in place to enable
‘independent’ safety inspection to be
carried out by experienced managers
Source: Characteristics and attributes (only) reproduced by kind permission of the IAEA, (attributes modified by the author for this text).
Table 3.6 Safety-culture characteristic C – ‘Accountability for safety is clear’, attributes safety-behaviours (informal
dimension) and the documented expectations (formal dimension)
Safety-culture Supporting Expected safety behaviour to support the Documented expectations to support the The safety- The safety
characteristic attributes attributes and the characteristic attributes and the characteristic beliefs being values being
supported by the supported by the
presence of the presence of the
characteristic characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic C C/1 An – There are scheduled meetings between management – A regulatory, management meetings schedule 2,3 6
‘Accountability appropriate and regulators exists and is implemented
for safety is relationship – The expectations of regulators are discussed agreed – Regulatory meetings are minuted, action
clear’ exists with and always implemented recorded and closure monitored
regulators – Relationships with regulators are overtly professional – The role of the regulator is within the training
as expressed through procedures, attitudes and regime
espoused values. There is a constructive and – Procedures exist to formalise the interactions
cooperative attitude towards regulators with regulators
Characteristic C C/2 Roles and – Individuals know understand and act within defined – Job descriptions with safety accountabilities and 2,3,4,5 2,8
‘Accountability responsibilities safety responsibilities responsibilities exist
for safety is for safety are – Individuals are aware of their line of contact if a safety There is clear linkage between the management of
clear’ defined and issue needs resolving and is outside their accountability change procedure and job roles and responsibilities
understood – All staff overtly demonstrate, in their attitude and – Job descriptions and safety responsibilities are
behaviours, that they are responsible for their own and discussed with employees, are understood and
others’ safety kept under review
– Clear organograms for employees seeking safety
advice and addressing safety issues exist and are
kept under review
Table 3.6 Continued
Safety-culture Supporting Expected safety behaviour to support the Documented expectations to support the The safety- The safety
characteristic attributes attributes and the characteristic attributes and the characteristic beliefs being values being
supported by the supported by the
presence of the presence of the
characteristic characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic C C/3 Regulations – Compliance is overtly accepted norm of behaviour – A safety-management system (integral to the 1,2,3,4,5,7,8,9 1,2,3,5,10
‘Accountability and procedures – Deviation from procedures by individuals or groups is business management system) exists, is owned by
for safety is are complied always challenged by peers and seniors the senior management and is used to manage
clear’ with – Compliance with the law is overtly expressed as a safety risks
minimum – The management system is audited for
– There is clear use of personal protective equipment as compliance, inadequacies and to seek
required by procedure improvement. Audit outcomes are reported to
– Work will be stopped at any hierarchical level (where management There is evidence of a management
safe to do so) if the use of a procedure appears to be response to reports
unsafe or inappropriate – There is no significant (months) overdue actions
– Interventions to halt tasks in the interests of safety from audits
are welcomed and treated as a leaning opportunity – Legal compliance is documented as a minimum
irrespective of the outcome expectation of the organisation – typically
– Staff welcome procedural audits as a mechanism to identified in the company safety policy
improve procedures to help them do there jobs and
support compliance
Characteristic C C/4 Interfaces – Staff behave with an understanding of the safety – Procedures exist to safely manage interfaces 2,3,5,8 1,2,6,7,10
‘Accountability are effectively interfaces between facilities and internal safety services – A contractor safety induction course is available
for safety is managed – The staff reflect understanding and authority at safety and used
clear’ interfaces with contractors or off-site suppliers – Information on the shared organisational
– Staff understand that they are responsible for the safety-beliefs, values, behaviours and safety
safety of others event reporting routes are available and used by
– All contractors have an assigned interface staff contractors
member to address safety issues
– All contractors undertake induction safety before
working on site
– All contractors overtly understand that they are
bound by the beliefs, values and safety standards of the
organisation
– The accountability for contractors’ safety and the
safety of their work is clearly assigned
Table 3.6 Concluded
Safety-culture Supporting Expected safety behaviour to support the Documented expectations to support the The safety- The safety
characteristic attributes attributes and the characteristic attributes and the characteristic beliefs being values being
supported by the supported by the
presence of the presence of the
characteristic characteristic
(Ref. Table 3.2) (Ref. Table 3.3)
Characteristic C C/5 Root cause – An inadvertent non-compliance is considered within a – A non-attributing (if required) reporting process 2,5,6,7,8,10 2,4,5,6,8,10
‘Accountability analysis is ‘just’ culture as a learning opportunity is available and evidently used
for safety is applied to non – All observed and self-non-compliances are reported in – Training is given in root-cause analysis to
clear’ compliance and a trusting, ‘just’ culture identified persons, and root-cause analysis is used
other safety – Managers overtly apply a measured response to non- on events
events compliance (people do make mistakes) – A Learning from Experience process is in place to
– Misaligned safety behaviour at any level is challenged ensure the out come of analysis is acted upon and
– Root cause analysis is used where appropriate to is communicated into the organisation.
ensure problems are solved and not repeated – A ‘just’ culture policy exists
Characteristic C C/6 Staff – A safety improvement programme is implemented by – A safety improvement programme is in place 2,4,5,7,8 2,7,8,9
‘Accountability are actively managers – The programme is identified at senior level as
for safety is involved – Managers encourage staff to be involved in the being owned, monitored and annually reviewed
clear’ in safety programme through team work solutions – There is an ‘award’ scheme for effective team
improvements – Praise and reward is given specifically for safety team working
with managers effort from various levels of management – Training on team working and interpersonal skills
– Safety improvement team tasks span across human is in place and used
performance, procedures, physical processes and plant
Source: Characteristics and attributes (only) reproduced by kind permission of the IAEA, (attributes modified by the author for this text).
assessing organisational saefty culture 145
Source: Reproduced by kind permission of the IAEA, (attributes modified by the author for this
text).
• questionnaires
• on-plant observations
For example, thinking long-term, the senior group are engaged in future
corporate policy and strategy. Middle managers are engaged in carrying out
the policy with a medium term horizon, whilst the supervisors, spend much
of their time on the day-to-day issues of direct safety importance to their work
teams. The focus for supervisors will tend to be on immediate task safety,
assessing organisational saefty culture 147
Quantification of Data
The review output can be qualitative and quantitative with both contributing to
understanding an organisation’s shared safety-beliefs and values. In addressing
this, use is made of the collated quantitative formal and informal review data.
Emphasis is placed on the point that this is only one contribution to the review
team’s essential internal debates. Too much emphasis on the quantitative
dimension has to be avoided as the numerical results have to be qualified and
aligned with the conclusions suggested by the qualitative analysis. This analysis
by a team’s review psychologist and the technical specialists is essential for
assessing organisational saefty culture 149
Table 3.8 Criteria for assessing and scoring the degree of presence of an
attribute
Source: Reproduced by kind permission of the IAEA, modified by the author for this
text.
The real beliefs may emerge from a review as different from the organisation’s
espoused beliefs, employees’ attitudes or what is reflected in the artefacts. The
integrity of the qualitative and quantitative information is therefore essential
for it fundamentally assists an organisation to examine its real beliefs. This
150 Safety Culture
Taking this further, for a ‘poor’ organisational culture the analysis may
show a dislocated set of safety-beliefs at the different organisational levels. On
the other hand, with a good organisational safety-culture there should be no
dislocation, but an aligned set of safety-beliefs throughout the organisation.
These beliefs will be expressed through supporting artefacts, espoused values
and consistent ‘good’ safety-behaviours, the latter demonstrating that good
safety-culture characteristics have been embraced and embedded.
For a specific attribute the reviewers’ numerical data can be derived from
different sources and may show sharp discrepancies. A value of the quantitative
approach is that scoring discrepancies are a source of additional information.
Extreme separations in numerical scores for a particular attribute need to be
justified. The justification to peers of attribute scores is one aspect of driving
review quality. Discrepancies in data are not ignored nor manipulated. When
compared with the overall organisational mean data they may show a particular
cultural characteristic fracture within a department or hierarchy level. The
influence of perceived fracture on the culture locally or on the organisation
generally may be significant and require further consideration.
The numerical data also enable the emergence of a ‘pictorial view’. However,
too much focus on the pictorial data has to be avoided. A pictorial is only a
‘snapshot’ in time and it is the understanding of the beliefs (the psychology of
basic safety assumptions) behind the picture that is important. Nevertheless,
pictorials are useful if considered with qualitative assessment to give context,
balance and perspective.
assessing organisational saefty culture 151
9
D Lack of management A Strong safety culture
attention to artefacts
8
7 A/1
A/2
A/3
Attributes
SC Dimension
5
A/5
4
Informal
A/4
3 A/6
1
1 2 3 4 5 6 7 8 9
Formal SC Dimension
Figure 3.1 Example – the distribution of attributes for the characteristic
– ‘Safety is a clearly recognised value’
Source: Reproduced by the kind permission of the IAEA, modified by the author for
this text.
The reviewers’ median score for an attribute of a characteristic can fall into
any of four sectors: A, B, C, or D. Scores in Sector A indicate a potentially strong
safety-culture where as Sector C suggests there are significant gaps in attributes
giving a weak culture. The vertical axis on the graph shows the scoring for the
informal dimension, for example the behavioural element. The horizontal axis
indicates the formal dimension which is typically the robustness of the safety-
management system.
All of the attribute scores may appear low down in Sector C. If located here
it suggests that there is no workforce or management commitment to good
assessing organisational saefty culture 153
7 A
Informal SC Dimension
6 E C B
5
D
4
3
C Poor safety culture B Possible alienation -
2 management and staff
1
1 2 3 4 5 6 7 8 9
Formal SC Dimension
Figure 3.2 Example – the distribution of safety-culture characteristics
Source: Reproduced by the kind permission of the IAEA, modified by the author for
this text.
panoramic wet source irradiator, it may be possible to place these events within
the context of Figure 3.2. That is, to consider into which Sector the five safety-
culture characteristics for each event may fall.
With Titanic, there appeared to be weak safety leadership. Safety was not a
recognised value. Because of the organisational belief in the officer’s capability
to assess safety risks in adverse sailing conditions the ship’s safety became less
of a priority than the business objective of a fast Atlantic crossing. Safety was
not integrated into the officers’ behaviours. For example, there appeared to be
no challenge or questioning of the captain’s decisions regarding the ship’s speed
through an iceberg field. There was no rigorous pursuit of the radio operator
for iceberg sightings information or particular concern about having no
binoculars for issue to the ‘iceberg watch crew’. This resulted in the only safety
related equipment to protect the ship being the crew’s eyes. The management
were clearly not learning from other ships’ observations as the radio messages
warning of iceberg sightings were ignored. Nor were they falling back on
their own collective maritime experience of the risk of operating ships in
iceberg fields. Compounding the failure to make conservative safety decisions
regarding excessive speed, the fundamental mistake by the management was
not understanding that on a moonless night ‘blue’ icebergs cannot be seen.
Considering these points the informal safety-culture, the behaviours, would
appear to be scoring low on Figure 3.3(a).
For the safety-management system little can be said as this is not fully
revealed in the documented evidence of the event. The emergency training
156 Safety Culture
aspect and parts of the ship’s commissioning appeared to be poor. There was no
attempt at retraining on essential safety procedures when they had failed. This
was particularly noted after the chaos around the application of the emergency
evacuation procedure and the inability to follow the lifeboat launch procedure.
Also, one may suggest (speculative) that with a fire in a ship’s bunker and the
ship close to port, the written procedure would have advised putting into dock
and having the matter effectively dealt with.
9 A Strong Safety
D Lack of management
attention to artefacts Culture
8
7
Informal SC Dimension
5
Dropped
Titanic load event
4
Plastics facility
3 Bhopal
Sector ‘C’, Figure 3.3(a). The caveats are noted. This was a single event, a
‘snapshot’ in time. However, the stable ‘safety’ belief of complete confidence in
the officers understanding of the risks, poor commissioning behaviours, lack
of attention to emergency procedures, (possibly the concept of an unsinkable
ship) and a bunker fire, suggests this was perhaps historically a poor culture,
an accident waiting to happen.
9 A Strong Safety
D Lack of management
attention to artefacts Culture
8
7
Informal SC Dimension
3
C Irradiator Criticality
2 event. event. B Possible alienation -
management and staff
1
1 2 3 4 5 6 7 8 9
Formal SC Dimension
In the panoramic wet source irradiator safety event there was knowledge
by the operators and the management that procedures were being ignored
and replaced by the unapproved ‘trouble shooting’ procedure. This was not
documented and the procedure detail was carried by word of mouth. The safety-
management system generally, the formal culture dimension, was poor and not
applied. Safety behaviours were systematically poor, such as disengaging the
PLC fuses to disable its functions, ignoring alarms, ignoring instruments and
actually bypassing the interlocked safety doors. In Figure 3.3(b) the irradiator
team’s culture resides, it is suggested, firmly in Sector C.
Figures 3.3(a) and 3.3(b) are speculative. However, if rigorous data are
generated from a culture review, the organisational culture can be revealed,
with the pictorial as one (and only one), helpful contributor to the overall
understanding of the organisations’ culture status.
assessing organisational saefty culture 159
This exercise is applied to the panoramic wet source irradiator. There are
caveats on the quality of the official reported information and the exercise is
for illustrative purposes. The event, where two operators were within seconds
of receiving a lethal dose of gamma ionising radiation, was the result of a
degraded safety-culture. The management shaped the culture by, inter alia,
promoting and condoning poor safety practices. These went unchallenged by
the operators.
Although this exercise has used only limited information (and hindsight)
it can be broadly concluded, by linking the behaviours to the organisational
beliefs, that the employees’ safety-culture was poor. If the remaining four
160 Safety Culture
• a belief that that the plant had to be brought back on line quickly
due the production schedule.
If in a review the quantitative approach was applied (Table 3.8) the formal
culture in the panoramic irradiator facility would probably have scored a low
numerical value. Similarly, the informal dimension where behaviours are
recognised as reckless would generate a low score. This numerical evaluation
(which in a review is one and only one part of the assessment) is reflected in the
positioning of the irradiator event in Sector C as a poor organisational safety-
culture (Figure 3.3(b)).
assessing organisational saefty culture 161
Finally it may be argued that the behaviours would not have been displayed
in a safety-culture review. The behaviours could have been ‘hidden’. This may
be so. However, the various tools used by reviewers would have revealed
fundamental weaknesses in the culture. Understanding the plant’s operations
and examining documents would undoubtedly reveal an unusual approach
to gaining entry to the irradiator under emergency conditions. There would
have been a review team conclusion that good safety-culture beliefs were being
overridden. With further inquiry, the real shared beliefs of the organisation
would then eventually have been revealed.
Following the detailed review with its formal executive report, the
organisation has to develop a strategy. What is required is the closure of any
identified safety-culture gaps, a periodic self-assessment to monitor progress
and a plan for continued enhancement.
Each stage on the ladder has a set of safety-culture elements that if present in
an organisation can locate the organisation on its way along the ladder towards
achieving a strong culture, a generative safety-culture. Relating this back to
the five safety-culture characteristics (Figure 3.2), Sector A can be considered
to be a generative culture whilst Sector C may suggest a pathological culture.
Further, due to the quantitative aspect of the five characteristics, the results,
‘measured’ by the initial independent safety-culture review, can be translated
across to the maturity ladder levels.
• self-assessment of safety-culture
Safety-culture Supporting Expected safety behaviour to support Documented expectations to The safety-beliefs being supported by the The safety values being
characteristic attributes the attributes and the characteristic support the attributes and presence of the characteristic supported by the presence
X = safety behaviour not evident on the characteristic (Ref. Table 3.2) of the characteristic
the plant (Ref. Table 3.3)
NT = not tested (data not available)
Characteristic A/1 Safety is – Safety procedures are complied with X – Documented health and safety – ‘The safety of staff, our contractors and – ‘Our individual attention
A ‘Safety given a high – In the face of uncertainty conservative policy exists NT the public is our number one priority in all to safety is a condition of
is a clearly priority at all safety decisions are demonstrated at all – Meeting minutes demonstrate circumstances’– this belief was not observed in employment’. – this value
recognised levels levels X safety before production NT the employees behaviours appeared not to exist
value’ – Decisions made show safety was/is – Meeting minutes indicate – ‘Accountability for safety rests at all times – ‘Everyone is responsible for
considered above production X conservative decision-making NT with managers’ safety, our own, others safety
– There is evidence of a proactive approach – Safety resources are defined in ‘Safety vigilance at all times’ and the protection of the
to resolving problems … X the annual financial budget NT ‘Absolute safety does not exist’ – these beliefs facility’ … this value was not
– Safety has a high profile in the were not reflected in behaviours. Managers applied. Safety was secondary
organisations annual report NT abandoned their accountability and with the to getting the plant back on
operators lacked vigilance line. Overriding the PLC and
bypassing the interlocks on the
doors showed lack of concern
for personal or others safety
Characteristic A/2 Safety – Safety booklets are prepared (as a route – Organisational shared safety- – ‘The safety of staff, our contractors and There was no evidence that at
A ‘Safety values are to sharing beliefs and values) NT beliefs are documented and the public is our number one priority in all any point over several years of
is a clearly communicated – Safety posters are displayed and updated) displayed NT circumstances’ interventions that safety-beliefs
recognised NT – Organisational shared safety ‘Accountability for safety rests at all times with and values were referred to or
value’ – Safety newsletter is made available) NT values are documented and managers’ enforced
– Safety intranet is made available) NT displayed … NT ‘In safety vigilance at all times’ – these beliefs
– Regular management safety – Safety communications were not upheld in communications. There
communications, including emphasis on procedures are in place and was verbal communications on the need to
beliefs and values X used NT use the ‘trouble shooting’ procedure. The
– Safety is proactively given high priority in – A variety of communication procedure was not documented. Managers
internal documents, verbal communications routes are available, resourced in their communications condoned the use of
and decision making X and used … NT the procedures with no reference to the safety
implications
– ‘legal compliance is our minimum
requirement’ – there appeared to be a culture
of non-compliance on the facility. The approved
procedure was ignored and unapproved
procedures used, Interfering with safety
equipment was against procedure, licensing
and illegal
Table 3.9 Continued
Safety-culture Supporting Expected safety behaviour to support Documented expectations to The safety-beliefs being supported by the The safety values being
characteristic attributes the attributes and the characteristic support the attributes and presence of the characteristic supported by the presence
X = safety behaviour not evident on the characteristic (Ref. Table 3.2) of the characteristic
the plant (Ref. Table 3.3)
NT = not tested (data not available)
Characteristic A/3 Safety – Giving and accepting behavioural – Safety-culture behavioural – ‘The safety of staff, our contractors and – ‘Our individual attention
A ‘Safety conscious challenge is a norm X requirements for each the public is our number one priority in to safety is a condition of
is a clearly behaviour is – Questioning peers and seniors on safety is hierarchical level are all circumstances’ – belief not displayed in employment’ – this value was
recognised accepted by an accepted norm X documented NT behaviours not upheld
value’ all within the – Reporting events and near misses is – A blame tolerant, ‘just’ culture – ‘Accountability for safety rests at all times – ‘That everyone is responsible
organisation rigorously carried out X policy exists with appropriate with managers’ – belief not displayed in for safety, our own, others
– There is proactive searching in operations supporting documentation NT behaviours safety and the protection of
for technical faults and human error X – An events and near miss – ‘safety vigilance at all times’– belief not met, the facility’ – this value was
– Positive feedback is given for ‘good’ reporting system is in place X employees appeared, through their behaviours, not upheld. Reckless behaviour
safety-behaviours NT – There is a procedure for disengaged from the safety risks they faced. endangered individuals and
– Reporting of events is seen as an act for analysing with trending and – ‘Absolute safety does not exist’ – this was not others
learning by managers and workforce X feedback on reported events NT an upheld belief. The plant was believed to be – ‘People’s interventions to
– Safety and safety-culture observations – A learning from experience inherently safe ensure all potential health
are carried out by all and across all the procedure is in place NT – ‘Human error is normal and can be expected’ and safety incidents are
hierarchy NT – Soft-skills training is in place for – not upheld, they did not believe they prevented’– this value was not
– A ‘just’ culture process is applied by all all employee levels to enhance could make a mistake due to inherent safety upheld. There is no evidence
managers X interpersonal skills NT and complete familiarity with the facilities of any intervention to stop bad
operation practice
– ‘Our engagement in safe behaviours is – ‘Everyone has the right to
necessary for safe operations’ – not an belief challenge on safety issues’
of the group – value not upheld as there were
– ‘Human errors are a learning opportunity’ no reported challenges to bad
– there were persistent violation as opposed to practice
errors, there was no learning – ‘ all events and near misses
– ‘legal compliance is our minimum are reported as we recognise
requirement’ – not a shared belief in the that there are no minor injuries’
organisation – value not recognised by the
employees. Did not recognise
breaking procedure as ‘an event’
– ‘Thorough safety training
and competence essential for
safe working’– training not
valued, minimum appears to
have been done for regulatory
requirements
Table 3.9 Continued
Safety-culture Supporting Expected safety behaviour to support Documented expectations to The safety-beliefs being supported by the The safety values being
characteristic attributes the attributes and the characteristic support the attributes and presence of the characteristic supported by the presence
X = safety behaviour not evident on the characteristic (Ref. Table 3.2) of the characteristic
the plant (Ref. Table 3.3)
NT = not tested (data not available)
Characteristic A/4 – Safety is on all business agendas NT – Business meeting minutes – ‘The safety of staff, our contractors and – ‘Our individual attention
A ‘Safety Acceptance – Safety and production are integrated in demonstrate safety with the public is our number one priority in all to safety is a condition of
is a clearly that safety and discussion at local plant level (tool box talks production are agenda items NT circumstances’– this belief was not reflected in employment’. this value
recognised production go – Safety beliefs and values are regularly – Documented guidance exists the employees behaviours appeared not to exist
value’ hand in hand referred to at meetings as ‘touchstones’ for conducting local tool box – ‘Accountability for safety rests at all times – ‘Everyone is responsible for
and safety guides for the organisation X talks NT with managers’ safety, our own, others safety
– Safety procedures are never bypassed to – Guidance that is used for ‘Safety vigilance at all times’ and the protection of the
achieve production goals X managers to manage production ‘Absolute safety does not exist’ – these beliefs facility’ … this value was not
– The safety of people, plant and process and safety conflicts X were not reflected in behaviours. Managers applied. Safety was secondary
are integrated into all production objectives – Procedures emphasis safety abandoned their accountability and with the to getting the plant t back on
and planning considerations X as a prerequisite for effective operators lacked vigilance line. Overriding the PLC and
business production in high- It appeared not to be a belief that ‘safety and bypassing the interlocks on the
hazard industries X production go ‘hand in hand’ ’. The primary
doors showed lack of concern
– Procedures exist that focus was to bring the plant back into operation for personal or others safety
integrate safety considerations quickly with the overwhelming belief that the
into production planning and facility was inherently safe at all times
scheduling X
– The consequences to
individuals for safety procedures
bypassing, at any hierarchical
level, are available, trained and
understood within the concept
of a ‘just’ culture policy X
Table 3.9 Concluded
Safety-culture Supporting Expected safety behaviour to support Documented expectations to The safety-beliefs being supported by the The safety values being
characteristic attributes the attributes and the characteristic support the attributes and presence of the characteristic supported by the presence
X = safety behaviour not evident on the characteristic (Ref. Table 3.2) of the characteristic
the plant (Ref. Table 3.3)
NT = not tested (data not available)
Characteristic A/5 Resources – Long term safety resource requirements – Annual business planning – ‘The safety of staff, our contractors and – ‘Thorough safety training and
A ‘Safety are a priority are discussed at annual business planning meeting documentation the public is our number one priority in all competence essential for safe
is a clearly for safety meetings NT identifies safety resources NT circumstances’ – training resources, emergency working’ – this value was not
recognised – Resources are made available for – Artefacts, for example personal exercises, re qualification of operators, were upheld
value’ equipment to be maintained or replaced protective equipment, are of inadequate. This belief could not be upheld with
(equipment not permitted to degrade) X good quality NT this low resource commitment
– Suitably qualified and experienced people – Documentation does not – ‘Accountability for safety rests at all times
are identified to competently operate suggest excessive breakdown of with managers’ – managers appeared not to
facilities and at management levels X safety equipment due to allowed fulfil their accountability for adequate training
– A management of change procedure is degradation X and retraining for operations on a high-hazard
used for structural changes in staffing at – Safety skill requirements facility. Similarly they were content to solve
any level NT and task skill requirements problems as they arose and allowed inadequate
are documented with training equipment (switches) to fail and be repaired
strategies, recruitment strategies repeatedly instead of supplying resources for
etc. X replacements
– A documented and used – ‘Believe in a ‘just’ safety-culture, for people
(register) management of change do come to work to do a good job’ – ironically
procedure exists NT the managers and operators may have thought
– Safety training is resourced and they were doing a ‘good’ job keeping the plant
functioning X operational
Characteristic A/6 There are – Teamwork to resolve safety issues is – Training is implemented on – ‘The safety of staff, our contractors and – ‘That everyone is responsible
A ‘Safety interactions strongly encouraged by all employees NT team working and interpersonal the public is our number one priority in all for safety, our own, others safety
is a clearly at all levels of – Managers actively apply processes to skills requirements at all levels circumstances’ and the protection of the facility’
recognised management obtain feedback from all employees on NT – ‘Accountability for safety rests at all times – ‘That the organisation strives
value’ and workforce safety matters X – Guidance is available on with managers’ for an open dialogue culture’
– Local ‘tool box talks’ address safety in dealing with safety feedback and – ‘Believe in safety vigilance at all times’ – ‘That all events and near misses
dialogue before tasks; current task safety tool box talk expectations NT – ‘Absolute safety does not exist’ are reported as we recognise that
risks and past safety experiences addressed – Feedback questionnaires exist, – ‘Our engagement in safe behaviours is there are no minor injuries’.
(would include in depth discussion of any intranet options are available for necessary for safe operations’ None of these values appear
expectations from risk assessments) X feedback, a fast-track feedback – ‘Believe in a culture of legal compliance is our to have been upheld The
– Senior managers frequently have cross- process is available NT minimum requirement’ opportunities for interaction
organisation safety briefs relating safety The systems are used reviewed –None of these beliefs were upheld. Unacceptable between hierarchical levels did
matters arising to the organisational safety- and feedback acted upon safety-behaviours, against operational procedures not appear to focus on safety
beliefs and values X through a Learning from and regulatory expectations, had developed over issues but upon the means of
Experience procedure X several years and such behaviours had become a putting the plant back into
deep seated norm operation.
assessing organisational saefty culture 167
Summary
A Way Forward
• lowering costs
A business strategy will arise from the compelling safety vision. This is a
long-term action plan to achieve the safety goals and monitor performance in
meeting the business’s defined safety objectives. An aspect of this strategy could
be, where appropriate, a safety-culture-change programme. However, if there
are a lack of safety goals, a strategy can not develop. ‘If you don’t know where
you are going, it doesn’t matter which direction you take’ (Lewis Carroll).
For moral and legal imperatives to achieve tolerable and as low as reasonably
practicable safety risks, all organisations require a compelling safety-culture
vision. Inappropriate safety-performance is, amongst others, a business risk to
be diligently managed in the interests of business success. Safety goals need to
be financially assessed for their business contribution and become an essential
component of any senior team’s business financial portfolio.
alia, a goal is to have a strong culture, then the evidence on the characteristics,
attributes and behaviours that need to change will be in the review. Tasks
required to move toward the organisation’s safety goals will be evident.
Even when reported data indicate there are safety-culture gaps, one choice
is to do nothing, just soldier on. Here the goal is unclear. The status quo on
past safety-performance is to be upheld. This is negative, as it is a ‘no change’
option. A safety-culture review by its nature engages the entire workforce to a
greater or lesser extent. A ‘no change’ choice may lead to a rapid deterioration
in morale as this stance may confirm to the employees that their safety is not
high on the management’s agenda. An already poor safety-performance may
continue to decline.
Chapter 1 suggests that culture evolves through trial and error, building
and learning a set of shared organisational beliefs and basic assumptions.
These can become so deep-seated as to be unconscious and unchallenged. If
the organisation continues to succeed in its objectives its shared beliefs are
seen as having contributed to the survival and security of the group. Being
psychologically satisfying, success reinforces the validity and stability of the
beliefs. In addition shared beliefs confirm individual group belonging. It is
difficult to change established attitudes and behaviours that are founded on
such strongly held stable beliefs even if they have ceased to deliver for the
business in changed circumstances. In a status quo environment, the group’s
beliefs give predictability to the outturns of any actions and this reduces group
uncertainty. This stability and predictability will be disrupted by change and
is resisted. In this case, to enable change individuals and groups have to be
convinced that a change from the present culture’s status quo will increase
group success or improve security and stability in turbulent times.
The managers collectively, in dialogue with the employees, identify the need for
change. Recognition can arise from an independent safety-culture review or a
self-assessment promoted by a CEO, a regulatory intervention or as directed by
the company shareholders. Change may arise due to a variety of circumstances,
for example frequent accidents causing continuing business losses, or public
and media attention promoting a demand for a specific high hazard industry
to adopt a new approach to improve safety-performance.
Resolution
At this stage, the organisation has resolved to internalise its new learning.
This manifests itself-in new safety-culture behaviours becoming ‘natural’
performance within the organisation. If demonstrably delivering the compelling
safety vision the new culture will become stable.
characteristics and attributes will tend to cluster around the centre of the
pictorial (Figure 3.2), usually showing a tendency towards Sector A. If there are
significant difficulties with an organisation’s culture, most characteristics will
appear well down in Sector C. For Sector A attributes a change programme can
be developed to understand their behavioural strengths and look to transfer
them to improve any weaker attributes. With a Sector C position, there may be
a need for a more radical intervention.
The change programme has to have the leadership of the senior management
team as this is where resources are controlled, direction is given and the culture
is shaped. A dynamic commitment is suggested, where overt action is taken,
the programme’s performance is regularly communicated and management
engagement is experienced by all employees. If effective safety-culture
leadership and commitment emerges and is integrated into the management
systems, a sound investment will have been made.
Poor shared safety-beliefs and values (locally or across functions) will inhibit
bringing safety-performance success and inhibit the delivery of improved risk
management. Gaps in safety-culture characteristics and attributes will reveal
this paucity. The employees’ safety-behaviours revealed as culture gaps and
misalignment with good safety-beliefs will need attention. Where documented
processes, artefacts and employee attitudes are not supporting a strong culture,
the change programme addresses these inhibitors. With any programme
setbacks will occur for individuals, groups and possibly the organisation
collectively. However, it is helpful in a psychologically safe environment if
setbacks are taken as ‘time out’ learning opportunities.
Middle managers
Supervisors
Work force.
• Assess from the independent review output data the gaps in the
safety-management system that are not supporting the revised or
reviewed beliefs and values. Changes to the system, implemented
by procedure, can be programmed
For an organisation to meet the model’s safety-beliefs and the vision there
is a set of minimum behavioural expectations associated with each attribute
(the dimensions) and hence the associated characteristics (the elements) These
behaviours have to be achieved by an organisation, a site or plant in order
to meet the compelling vision. The model has a range of possible behaviour
profiles across five maturity states and by an organisation testing its current
behaviours profile it can determine how far it is from achieving the safety
vision of a generative culture.
The maturity model applies the concept that organisations can be considered
to be in one of five culture states. These form a safety-culture maturity ladder.
The culture states are:
182 safety culture
• Pathological
• Reactive
• Calculative
• Proactive
GENERATIVE
HSE is how we do business
round here
ed
PROACTIVE
m
or
inf
ity
cre
bil
CALCULATIVE
In
ta
un
REACTIVE
ru
gT
PATHOLOGICAL
Who cares as long as
we’re not caught
For organisations using the maturity model concept, the safety vision (the
compelling vision) is to achieve a generative culture. To reach this level all employees
need to display a particular safety-behaviours profile. The behaviours expectations
have arisen from extensive research into the safety human performance needs
of high-hazard, low-risk chemical facilities. It is assumed in the model that the
safety management system is well designed. The model addresses the maturity of
behaviours not the maturity of the safety management system.
Low profile
Low accident rate – but there is always bad luck
Active involvement and accountability for all
Workforce initiative in health safety environment and operations
Short and effective feedback line
Procedures under constant scrutiny
Training, cross-training and more training
Benchmarking against others, inside and out
Obsessive planning – variety creates mindfulness
Willing to try new ideas, but accept the risk of failure
Chronic unease
Source: Hearts and Minds Programme, available from The Energy Institute,
reproduced with the kind permission of Shell International Exploration and
Production. For further information visit www.energyinst.org.uk/heartsandminds
The maturity model modules are commercially available and with the
option to use the modules repeatedly in workshops for safety-culture issues-
resolution, they form a long-term investment. The modules can be applied by
an organisation’s internal resources and the standardised modular concept
introduces systematic rigour with continuity of application. Through workshop
application the modules can engage all employees in a change programme
gaining commitment through a team approach.
Source: Maturity model elements, Hearts and Minds Programme, available from The
Energy Institute, reproduced with the kind permission of Shell International Exploration and
Production. For further information visit www.energyinst.org.uk/heratsandminds
The metrics should be such that changes in their quantitative values can be
measured and frequently reported within the organisation. Metrics need to be
representative of change and be a manageable number. It is suggested they
demonstrate change progress, regression or herald stagnation. Progress can be
celebrated whilst regression or stagnation are opportunities to take stock of the
change programme’s content, its focus and implementation strategy.
and the environmental impact from malfunctions. This risk reduction has been
historically strengthened by the adoption of procedures within a formal safety-
management system. As noted (Chapter 1) the aftermath of the Chernobyl
accident caused considerable focus on the importance of the organisational
safety-culture and the human performance of individuals. It became clear
through research into human organisational performance that engineering and
safety systems are essential but not sufficient. That is, organisational safety-
culture forms an essential contributor to risk reduction. Examples proposed
(Chapter 2) demonstrate that if an organisation’s culture is poor the engineering
and procedural systems can degrade, resulting in increasing accident risks. This
is irrespective of their initial high quality or the designers’ intent to provide
defence in depth to protect operators and the public.
• in an office environment
a string of previously established latent errors. The latent error line starts in many
cases with inappropriate management decisions compounded by inappropriate
safety-beliefs, policy, procedures, supervision, training or facility design. When
mishandled, managers may use observational behavioural safety techniques
as a quick fix to problems whilst the root cause, possibly poor management or
inadequate equipment design, is ignored. This fix can be the assigning of incident
causes to the workforce. If, for example, blame is a frequent consequence of
events reporting, a fear culture may emerge. A response of blaming the worker
can alleviate managers from fulfilling the responsibility of seeking the root cause
of events. Investigations can be time-consuming. A quick fix may perhaps avoid
having to address the costly need for plant or procedure improvements, or seek
improvements in management capability. The rush to ‘blame’ is always a potential
short term ‘solution’. However, evidence suggests that if frequently exercised
it sets the scene for repeat events. In such an environment the application of
behavioural safety techniques is a chronic waste of resources, achieving little in
the support of culture improvement.
and maintaining staff motivation. Monitoring has to be regular with both good
news and setbacks discussed openly contributing to a dialogue culture. As part
of the monitoring process the following tools can be brought to bear:
Summary
Concluding Remarks
• agreeing with the workforce and peers the consequences for good
and poor safety-behaviours
Evidence from the field suggests the root cause of many high-hazard
industry accidents arises from manager’s lack of technical understanding of
the risks their workforce have to manage. This is usually compounded by a
lack of leadership in safety-culture. Management can very rarely be considered
to be reckless regarding safety. However, it is suggested they can suffer from
omissions by not recognising or acting upon available evidence such as poor
organisational safety-performance.
a decline. This may require draconian standards and expectations put in place
and enforced. Such management expectations will be directed at changing a
poor safety climate. However, the long-term aim has to be moving the total staff
complement from automaton compliance towards safety involvement by ‘thinking
safety’ at all times and having interdependence. In a state of interdependence,
attention to health, safety and the environment becomes a natural way to work
at every organisational level. When a state of interdependence is achieved the
organisation is possibly a generative dialogue culture, a strong safety-culture.
The senior management presence nevertheless has to continue to be felt ensuring
the safety messages and beliefs are sustained.
It is often proposed that there could be 10 ‘most important’ questions for senior
managers, so as to draw upon their views, beliefs, opinions and attitudes
about organisational safety-culture. It is further conjectured that from the
spontaneous responses to the questions in a one-to-one interview a broad
understanding of the organisational culture can be deduced as this group is
so influential in shaping the culture. By understanding both the action taken
to support culture and possible inaction, the question responses may establish
preliminary indicators about the strength or weakness of a culture that can be
confirmed during an independent assessment or a self-assessment.
The approach is not new and a little technically crude, nevertheless a set
of considered questions asked spontaneously at least raises the immediate
awareness of culture. It may have merit as a personal exercise to give
preliminary comfort or concern about an organisation’s culture well ahead of
an independent review. Table 5.1 indicates 10 possible questions and the task
in presenting them to senior mangers is to capture immediate thoughts about
an organisation’s safety-culture.
Summary
Questions Questions
1. Do you agree the primary accountability for 6. What actions do you periodically take
safety on this establishment rests with the so that you understand the safety-
regulator? culture status of this facility against good
2. Is there a documented displayed health and international practice?
safety policy signed by the CEO? 7. How in the last month have you
3. Is there a published set of safety-culture demonstrated to the organisation your
expectations that are placed on all personal commitment to a strong safety-
employees? culture?
4. What actions do you personally take to 8. Is safety a standing agenda item at every
ensure people understand the safety monthly senior managers meeting?
policy and the safety-culture behavioural 9. Do you agree annually to commit resources
expectations? to safety improvement programmes?
5. How many scheduled times per month do 10. By what means do you test that the safety-
you discuss safety and formally observe management system is assisting you in
safety-culture behaviours on this facility? managing safety risks on this facility?
do not idly lose assets through the neglect of safety. In hazardous industries
safety is a major business risk to be managed. Starting from a poor culture,
the journey of improvement may be perceived as an expensive option. The
alternative can be severe business damage or, in the long term, no business at
all. Once a strong culture has been achieved the cost is minimal to sustain the
position as safety becomes ‘just the way we do things around here … when
nobody is actually watching’.
It has been recognised that it may take some time to establish a strong safety-
culture against world-class performance criteria. However, if an organisation
structurally changes and inappropriate safety expectations and beliefs emerge
from an autocratic management, an established strong culture can be reversed.
An organisation can revert to poor human performance and become a high-risk
enterprise in a matter of months.
Definitions
Reactive
Safety starts to be taken seriously, but only after incidents is there any action.
Calculative
Safety is treated very seriously. There are management systems and much data
collection.
Proactive
People try to avoid accidents and start to take a more bottom-up approach.
204 safety culture
In such organisations there are few safety rules. There is little or no management
commitment to either safety or leadership in setting safety standards. The
workforce is expected to look after itself regarding keeping safe. It is a belief
that accidents are part of doing the job. Safety is not taken as a serious business
risk and there is little awareness of safety technology. The organisation aims for
the absolute legal minimum regarding safety law.
Source: Hearts and Minds Programme, available from The Energy Institute,
reproduced with the kind permission of Shell International Exploration and
Production. For further information visit www.energyinst.org.uk/heartsandminds
Appendix III
Question Good practice shared Good practice shared Expected responses for
safety culture beliefs safety culture values a ‘good’ safety culture
being tested being tested
1. Do you agree that the – Accountability for safety – Our individual attention The accountability
primary accountability rests at all times with to safety is a condition of rests with the Board
for safety on this managers employment. through the CEO. For
establishment rests with – The safety of staff, our organisations that are
the regulator? contractors and the public issued with a legal licence
is our number one priority to operate high-hazard
in all circumstances. plants, emphasis on the
specific accountabilities
of the licensee would be
expected.
2. Is there a documented Accountability for safety – Our individual attention A positive response would
displayed health and rests at all times with to safety is a condition of be anticipated here with
safety policy signed by managers employment the health and safety
the CEO? – The safety of contractors – That everyone is policy being displayed
and the public is our responsible for safety, to the reviewer. The
number one priority in all our own, others safety reviewer would listen
circumstances and the protection of the for information on the
– Our engagement in safe facility. management annual
behaviours is necessary review of the policy and
for safe operations as appropriate the steps
– We believe in a culture taken so that it remains
of legal compliance is our relevant to the business.
minimum requirement.
3. Is there a published set – Belief in safety vigilance – Our individual attention The reviewer should
of organisational shared at all times to safety is a condition of anticipate reference to an
beliefs, values and safety- – Absolute safety does employment intranet site, distributed
culture expectations not exist – That respect is given booklets and possibly a
that are placed on all – Human error is normal to all safety views as periodic poster campaign.
employees? and can be expected everyone has the right to Particular emphasis will
– Our engagement in safe question and report safety be placed on all managers
behaviours is necessary issues carrying and frequently
for safe operations – People’s interventions to referencing back to the
– People are fallible and ensure all potential health organisation’s beliefs,
will make mistakes and safety incidents are values and safety culture
– We believe in a culture prevented expectation. Safety
of legal compliance is our – Everyone has the right behavioural expectations
minimum requirement. to challenge on safety need to be explained as
issues. having been derived and
agreed for all levels, from
the senior management to
the workforce.
206 safety culture
Question Good practice shared Good practice shared Expected responses for
safety culture beliefs safety culture values a ‘good’ safety culture
being tested being tested
4. What actions do you – The safety of staff, our – Our individual attention The reviewer should
personally take to ensure contractors and the public to safety is a condition of anticipate reference
people understand the is our number one priority employment would be made to the
safety policy, beliefs, in all circumstances – We value thorough integration of the safety
values and the safety – Accountability for safety safety training and culture programme into
culture behavioural rests at all times with competence as essential the safety management
expectations? managers for safe working system. Emphasis
– We believe in safety – That everyone is would be expected on
vigilance at all times responsible for safety: the senior managers’
– We believe a culture of our own, others’ safety communications
legal compliance is our and the protection of the strategies. Evidence
minimum requirement. facility. could be presented of a
monitoring strategy.
5. How many scheduled – The safety of staff, our – Our individual attention A schedule, integrated
times per month do you contractors and the public to safety is a condition of into the management
discuss safety and observe is our number one priority employment system, would be
safety-culture behaviours in all circumstances – Teamwork to resolve expected. Reference to
on this facility? – Accountability for safety safety matters is strongly a strategy and schedule
rests at all times with supported for visiting offices
managers. – That respect is given and operational plant
to all safety views as should be tabled with
everyone has the right to implementation evidence.
question and report safety
issues.
6. What actions do you – Accountability for safety – Our individual attention Required action should be
periodically take so that rests at all times with to safety is a condition of demonstrated as a part of
you understand the managers. employment the management system.
safety culture status of – That we regularly check Evidence of a periodic
this facility against good and report our safety self-assessment of the
international practice? objectives performance. whole organisation or a
part of the organisation
as the company strategy
requires.
7. How, in the last month, – The safety of staff, our – That the organisation Possibilities: behavioural
have you demonstrated contractors and the public strives for an open observations, managers
to the organisation your is our number one priority dialogue culture take part in safety walk-
personal commitment to a in all circumstances – Our individual attention about reviews, carry
strong safety-culture? – Accountability for safety to safety is a condition of out cross-organisation
rests at all times with employment presentations, have
managers – That everyone is stopped production due
– We believe in safety responsible for safety: to a safety issue taking
vigilance at all times our own, others’ safety priority, opened several
– Absolute safety does and the protection of the safety-training sessions,
not exist. facility. etc.
8. Is safety a standing – The safety of staff, our – That we regularly check Agendas presented as
agenda item at every contractors and the public and report our safety evidence. Minutes made
monthly senior managers is our number one priority objectives performance available where possible.
meeting? in all circumstances – That all events and near
– Accountability for safety misses are reported as we
rests at all times with recognise that there are
managers. no minor injuries.
appendix iii 207
Question Good practice shared Good practice shared Expected responses for
safety culture beliefs safety culture values a ‘good’ safety culture
being tested being tested
9. Do you agree annually – Accountability for safety – Our individual attention Documented evidence
to commit resources rests at all times with to safety is a condition of of a safety-management
to safety improvement managers employment system review and the
programmes? – We believe a culture of – We value thorough publication of a safety
legal compliance is our safety training and improvements budget
minimum requirement competence as essential would be expected:
– Absolute safety does for safe working minutes of meetings,
not exist. – That everyone is evidence in the budget.
responsible for safety:
our own, others’ safety
and the protection of the
facility.
10. By what means do – Accountability for safety We value that we Evidence of a formal
you test that the safety- rests at all times with regularly check and report safety-management
management system is managers. our safety objectives system annual review
assisting you in managing performance. would be expected. The
safety risks on this facility? review would typically
be procedure driven
and include safety
reactive and proactive
key performance
indicator outturns,
audit report findings,
independent review
findings, behavioural
safety findings, trends.
The review should be
presented as an open
dialogue on positive safety
matters and areas judged
to need improvement. A
particular strength would
be if the management
system review was
chaired by the CEO. As
a minimum it would be
chaired and attended
by senior managers. A
formal report on the
review and actions arising
for the organisation in
general and the safety
management system in
particular would be made
available for the CEO. This
would link to the annual
safety improvement
plan and business risk
reduction plan.
This page has been left blank intentionally
Further Reading
Hudson, P.T.W and Willekes, F.C. 2000. The Hearts and Minds Project in an
Operating Company: Developing Tools to Measure Cultural Factors. Society of
Petroleum Engineers, SPE 61228. SPE International Conference.
International Atomic Energy Agency. Developing Safety Culture in Nuclear
Activities. Safety Series Report No. 11.
International Atomic Energy Agency. 2002. Key Practical Issues in Strengthening
Safety Culture. Report by the International Safety Advisory Group 15.
International Atomic Energy Agency. 2002. Safety Culture in Nuclear Installations.
International Atomic Energy Agency TECDOC-1329.
International Atomic Energy Agency. 2006. Application of the Management System
for Facilities and Activities Safety Guide GS-G-3.1.
International Atomic Energy Agency. 2007. Lessons Learned from the JCO
Criticality Accident Tokaimura.
Kennedy, R. and Kirwan, B. 1995. Safety Culture in Nuclear Installation. The Failure
Mechanism of Safety Culture. International Topical Meeting, International
Atomic Energy Agency, Vienna.
Kletz, T. 1990. Plant Design for Safety a User Friendly Approach. Published by
Hemisphere Publishing Corporation.
Kletz, T. 1994. Learning From Accidents. Published by Butterworth – Heinemann
Ltd.
Komai, J.L. et al. 1992. The Role of Performance Antecedents and Consequences
in Work Motivation. Journal of Applied Psychology, 67(3), 334–340.
Leveson, N. 2004. A New Accident Model for Engineering Safe Systems. Safety
Science, 42, 237–70.
Los Alamos National Laboratory. 2000. A Review of Criticality Accidents 2000
Revision. LA 13638 Approved for public release, distribution unlimited.
Manuele, F.A. and Christensen, W.C. 1999. Safety Through Design. Published by
the National Safety Council, and NSC Press Product (USA).
Marsh, T. 1996. University of Manchester Institute of Technology (UK).
Developing a Training and Communications Strategy for a Positive Safety Culture.
Health and Safety Management Conference, London.
Parker, D. Lawrie, M. and Hudson, P. 2006. A Framework for Understanding the
Development of Organisational Safety Culture. Safety Science, 44, 551–62.
Rasmussen, J. 1997. Risk Management in a Dynamic Society: a Modeling
Problem. Safety Science, 27(2/3), 183–213.
Reason, J. 2007. Human Error. Cambridge University Press.
Schein, E.H. 2004. Organizational Culture and Leadership. Third Edition. Josey-
Bass.
Shell International Exploration and Production. Heart and Minds. For information
visit www.energyinst.org.uk/heartsandminds.
further reading 211
dialogue culture 7, 13, 17, 23–24, 27, Hearts and Minds 181–185, 204
135, 76–177, 180, 186, 188, 193, human-performance 1, 25
197
I
E informal dimension 17, 136, 151
employees 2, 15 ,23 inherent safety 120,124
Energy Institute UK 181, 183, 185, 204 International Atomic Energy
enrichment 82–86 Authority 2
espoused values 3 irradiator 40, 109
see also values and definitions isotopes 81, 109
events
Bhopal 30, 40 J
dropped load event 30, 40 ‘just’ safety culture 5, 134–135, 180,
poly vinyl chloride plant event 188, 201
40, 59
Titanic 40 L
Tokaimura 40, 79 leadership 22–23, 30, 36, 132,
wet source irradiator event 40, 201
109 see also safety culture shaping
executives 23, 174, 191
see also expectations, and safety M
culture shaping managers 2
expectations 6, 17, 23, 134, 136, 163, see also safety culture shaping and
167, 188, 298 expectations
see also behaviours maturity ladder 162, 181, 184,
192
F see also Energy Institute UK
fear culture 10, 57, 103, 106, 108, 126, methyl isocyanate (MIC) 51
191
fission 81–99 N
formal dimension 131, 136, 151, 188 natural uranium 79
neutrons 79
G norms 7–8, 10, 14, 105, 126, 160
gamma rays 80–99, 109–122 nuclear criticality 79
grays 120
group think 103, 123, 126 O
observational techniques 189, 191
H
hazard 3, 201–202 P
see also risk and definitions poly vinyl chloride (PVC) 61
index 215
V W
values 3–8, 14–15, 17–27, 32–39, 49, wet source irradiator 40, 109
58, 71, 77, 102, 107, 122, 134– workforce 2
152, 161, 163–166, 174, 181, 195,
199, 202, 205–207
vinyl chloride monomer (VCM) 59