Recent years have witnessed a growing concern over the issue of safety culture within aviation and other complex, high-risk industries. The purpose of the present review is to summarize and integrate the numerous reports and studies that have been conducted to define and assess safety culture, as well as the highly related concept of safety climate. Results of the review indicate that few formally documented efforts have been made to assess safety culture within the aviation industry (Treadwell, 2005). Furthermore, there exists considerable disagreement among safety professionals, both within and across industries, as to how safety culture should be defined and whether or not safety culture is inherently different from the concept of safety climate. A synthesis of theses different perspectives is conducted and hybrid definitions of both safety culture and climate are offered. A discussion of key organizational indicators of safety culture and the various methods commonly used to assess these factors is provided. Issues that need to be considered when implementing a safety culture assessment program are also presented (Gaba DM, 2004). The hope is that this review will enable researchers and safety professionals to better understand and assess safety culture and that it will facilitate the sharing of information and strategies for improving safety culture across organizations and industries.

     The rapid development of new technology has fundamentally changed the nature of work and has increased the complexity of systems within a variety of industries (Hendrick, 1991). Among these complex systems are those commonly known as “high-risk” systems, such as nuclear power plants, chemical processing facilities, and aviation operations that require a tight coupling between both technical and human subsystems. The failure of either subsystem can often cause a failure of the entire system. Furthermore, catastrophic breakdowns of these high-risk systems pose serious threats, not only for those within the organization, but also for the surrounding public. For some potentially highly dangerous systems, such as nuclear power, this risk can extend far beyond the immediate locality and even “have adverse effects upon whole continents over several generations” (Reason, 1990, pg. 1).

     Given the potential for enormous damage that failures of high-risk system given the potential for enormous damage that failures of high-risk systems can inflict, the investigation of the causes of system failures is extremely crucial to preventing future occurrences. Toward this end, theories of accident causation have progressed through several stages of development over the past several years in an effort to identify the root causes of system failures (Gordon, Flin, Mearns, & Fleming, 1996; Wilpert, 2000). The first stage is often referred to as the technical period, during which developments in new mechanical systems were rapid and most accidents were caused by mechanical malfunctions, particularly in the design, construction, and reliability of equipment (Wiegmann & Shappell, 2001). The second stage is known as the period of human error, where faults of the human operator, rather than catastrophic mechanical malfunctions, were seen as the source of the system breakdown. The accident at Unit 2 of the Three-Mile Island nuclear plant (TMI-2) raised awareness of human error and cognitive shortcomings of operators and shifted the attention of safety analysis from technical aspects to human errors, where blame and responsibility were assigned to the person directly involved in the unsafe act (Rochlin & Von Meier, 1994; Coquelle, Cura, & Fourest, 1995). The third stage is referred to as the sociotechnical period. This view of human error considers the interaction of human and technical factors when exploring the causes of errors and accidents. Finally, recent years have witnessed the development of a fourth stage, which is often called the “organizational culture” period (Gordon et al., 1996; Wilpert, 2000). This approach recognizes that operators are not performing their duties or interacting with technology in isolation, but rather they are performing as a coordinated team of organizational personnel, which is embedded within a particular culture.

>

Introduction

     The management of safety in the aviation industry appears to be near perfect. Airplanes hardly ever crash and a common statement made about the safety of flying is that the most dangerous part of a trip is the drive to the airport. In many ways this is true. Most people will never be in a crash (although I have) and driving regularly leads to accidents that we can see all about us. Despite this optimistic view I want to set down some critical markers about safety in the aviation industry and then go into how it can be improved (Roberts KH, Gargano G, 2003). This paper will look critically at how safety is measured in the industry and argue that it is not a safe industry, at least not for those who work in it, which has consequences for how we view aviation as a whole. I will then go on to discuss safety management systems and the ‘magic’ term safety culture. As most people do not really know exactly what a safety culture is I will attempt to provide a definition and then discuss how to develop and then maintain such a culture within an aviation organization. Finally I will examine the reasons why it is so difficult to achieve what must appear as a totally laudable goal.

>

Aviation Safety Management – the history

     Aviation has traditionally been interested in safety. Flying is inherently dangerous and it has taken us some time to achieve the levels of passenger safety we nowadays take for granted. But the way this is done is quite old fashioned. Essentially the world of civil aviation is regulated from ‘on high’, with ICAO at the top in Montreal, handing down SARPs, Standards And Recommended Practices, to be implemented by national aviation regulators. SARPs traditionally define what to do and how to do it, leaving little room for alternatives. Attention was directed primarily at major accidents and the main measurements were, and still are, hull losses and fatalities. The management process, as regulated, was therefore framed in terms of outcomes and defined in terms of what to do. This is a relatively old-fashioned approach when compared with modern goal-setting regimes such as are found in the petro-chemical industry. The more modern approaches are based upon the existence and effectiveness of processes, rather than in the specification of exactly how those processes should operate. However, one might wish to argue that the system works and commercial aviation is the world’s safest industry. I would argue, on the contrary, that aviation is only safe when measured in very restricted ways and that, as an industry, it is anything but safe or even safety-minded. Safety performance in civil aviation, at least in the United States, as measured by classic lost time injury (LTI) rates, is worse than such apparently unsafe industries as logging and construction! The hull loss statistics, especially for newer types of aircraft and in regions such as Australasia, do show that flying really is very safe, for passengers. The problem is that the LTI data shows that flying is dangerous for employees. For instance, baggage handlers, maintenance engineers, cabin staff and check-in personnel are all groups that are regularly exposed to being injured. That baggage handlers have back injuries, or that maintenance engineers suffer a variety of occupational injuries, is well known, but often reported to different national regulatory bodies and segregated from the safety function within an airline, that will usually only be interested in ‘flight safety’ issues. The problem that this raises is that the culture of safety within many airlines appears to be very one-sided, in favour of passengers, which is good, but all too often at the cost of the staff.

     If this is so, what is the safety culture of such an airline? Experience in the petro-chemical industry is that accidents are indicators of poor management and there are no real distinctions to be made between types of accident. So the question is: How good is the safety culture in aviation? Put another way – do airlines really take safety seriously, or are they only interested in the avoidance of certain extreme outcomes? If they put all their effort into the specific avoidance of hull loss and fatality outcomes, then they may not be doing their best and they cannot claim to be safety cultures.

>

Safety Management Systems and Safety Cases

     One of the major recent developments in industrial safety has been the application, enforced by law in many cases, of safety management systems and safety cases. These are simply the systematic application of management processes to the problem of the hazards an organisation faces and the proof that the management system is in operation. One typical approach, as taken by the Shell Group (Hudson, 2001), involves the discovery and assessment of the hazards of particular operations, which may differ considerably from place to place, the specification of how those hazards are to be managed and, finally, the definition of what will be done if the management fails.   Surprisingly, such a systematic approach is almost completely lacking in aviation operations. The register of hazards, part of the SMS, and a clear understanding of the nature of the defences applied to manage those hazards, as is normal in certain industries, is at best done in an ad hoc way in the aviation industry. The approach to safety management is haphazard and driven typically by responses to specific incidents in the past. There is no systematic analysis and it is unclear whether the effort expended to achieve the goal is sometimes either too much or too little. Surprisingly the top-down regulatory approach has led to a plethora of uncoordinated bottom-up solutions, often relying on the personal abilities of specific individuals, pilots and licensed engineers, to ensure that all goes well.

     The systematic application of safety management principles, culminating in some form of formal assurance that the goals set can and are being achieved, can help significantly. This approach should also do much to improve the lot of the employees as well as the passengers as all the hazards have to be identified and managed, not just those that contribute to hull losses. However, such systems are, by their very nature, paper-based and bureaucratic. They tend to require minimum standards and can result in the achievement of such standards.

     Safety Management Systems define sound systems, practices and procedures, but those are never enough if they are practiced mechanically; they require an effective safety culture to flourish. This is where the Safety Culture comes in. I have argued (Hudson, 1998) that the aviation industry needs to achieve an improvement of two orders of magnitude in safety performance (measured admittedly in terms of hull loss) to overcome the threat identified by Boeing, of one hull loss every 5-8 days. I have set a target of 1 loss, world wide, every 2 years, or one loss every 500-800 days. Safety Management Systems can achieve one order of magnitude improvement, especially as they go naturally hand in hand with the improvements in technology that the latest generations of aircraft bring. But such an approach will throw the human contribution to disaster into an even more stark light, and it is here that safety cultures can provide the second order improvement.

    People involved in aviation, at all levels, need to be ready for disaster, to seek out its roots and cope when things go wrong, as they invariably will.

The Characteristics of a Safety Culture

     Safety cultures take safety seriously. They function well and make money in the process. DuPont, for instance, is widely regarded as the most safety conscious company in the world (See Fig. 1) but it also stands regularly at the top of the Fortune 500 ratings for chemical companies. Shell Group has improved its safety performance spectacularly in the last 15 years and, while subject to the vagaries of the price of oil, has succeeded in returning considerable profits, measured as return on average capital invested. Interestingly, part of Shell’s improved safety performance, in the area of fatal accidents, can be traced directly to its quite aggressive safety management of its aviation contractors; they have had no aircraft crashes for many years despite high exposure hours and operations in very difficult environments.

What are the characteristics of an organisation that takes safety so seriously? The following list was first identified by Jim Reason (1997), and I have added one extra feature. An organisation should be:

·        Informed – managers should know what is going on in their organisation and the workforce should be willing to report their own errors and near misses;

·        Wary – the organisation and its constituent individuals should be on the look out for the unexpected, maintaining a high degree of vigilance;

·        Just – the organisation should operate a ‘no blame’ culture within the constraint that some actions can be agreed by all to be totally unacceptable and worthy of approbation;

·        Flexible – such organisation can operate according to the demands, so they can provide both high tempo and routine modes of operation and can change when required by circumstances;

·        Learning – organisation should be ready to learn in order to improve and be capable of implementing what needs to be done to reform.

>

Safety Management as a Skill

     One way of looking at how both organizations and individuals manage safety is to regard safety management as a skill. Skills have to be developed, and require practice; they do not suddenly appear just because you read the manual. Furthermore applying the skill metaphor suggests that safety management has to be kept up, it is not enough to succeed once, most skills have to be maintained otherwise the practitioner becomes rusty. Typical examples may be from sport, either playing golf or tennis for individuals, or playing a team game for groups. Alternatively you can consider acquiring the skill of riding a bicycle or driving a car. In all these cases it costs considerable effort to acquire the skill, and a period without some practice soon shows, even the proverbial skill of riding a bike, that is supposed never to be lost!

     To acquire a skill requires practicing individual elements and then learning to put the elements together so that sequences of actions become automatic and attention can be devoted to wider issues. The skill can then be extended. When you initially learn to ride a bicycle, it is all you can do not to fall over. Once the balance and basic pedalling skills have been acquired, they can become automatic after a little practice and the development proceeds to learning to bicycle where you want to go to. This can then serve to be able to bicycle in traffic, clearly a skill in its own that requires the presence of all the underlying psychomotor skills. Once someone is sufficiently skilled to ride safely in traffic, that person can use their new-found skill to achieve their aims: Bicycling is a means, rather than the end, but it can be an essential one. The same analysis can be applied to skills in team games like football. Once the primary skills have been acquired, the higher-level goals can be attempted (although what both individual and team games teach us is that coaching of the basic skills has to continue all the time).

     Highly skilled individuals or teams do difficult things well, they know automatically what is and what is not important and they have the correct repertoire of responses available. They look to discover what they do less well and practice in order to improve. The same applies to safety and to safety in particular. Individuals need to acquire abilities to recognize hazards, to know when to be careful, to know what is worth reporting etc. Senior managers can learn how to demonstrate their commitment, not just by saying they are committed, but also by acting in ways that show that commitment in practice. For example, making regular visits to operational sites, such as platforms or isolated fields, is difficult; there are always many matters that seem more pressing than a lengthy and apparently unstructured site visit.

     A safety culture is one that has acquired considerable skill, both as an organization and as a set of individuals, in managing safety. The advantage of this is that they can operate in a more dangerous environment, one where the less skilled can only trust to luck or be closely managed by their regulatory bodies. What this implies is that the development of a safety culture will be characterized by the, possibly gradual, development from unskilled to highly skilled performance at the management of safety. This development has to go through a number of inescapable stages and so the organisation and its constituent individuals will be characterized by all the behaviors and attitudes that those different levels might imply.

>

Creating a Safety Culture

     If being a safety culture is so advantageous, how can you become one? The answer is, unfortunately, not so simple. Just wanting to become one, no matter how great the desire, is not enough. The first lesson that has to be learned is that the steps that need to be taken will depend upon the current culture. In terms of the evolutionary model it will be necessary to progress from the current stage towards the final Generative stage. So a Reactive organisation will have to progress through the Calculative and Proactive stages to become a full safety culture. It might be argued that a strong leader, an inspirational CEO, could shift the company culture single-handedly. The counter-argument is that the characteristics of the Generative organisation include many of those found in the Calculative, especially the systematic approaches that are found in the infrastructure. The good news for smaller organisations is that understanding what those characteristics are helps in their construction and speeding the transition up the ladder.

     The skills approach to safety culture, described above, also places constraints on how quickly an organisation can progress. The skills have to be acquired and then practiced before they become automatic. Two particular skills that have to be acquired are Informedness and Trust. It is not easy to become informed – the news that one is likely to hear is all too often uncomfortable and the messengers can lead a precarious existence. Being informed involves promoting a desire to hear to possibility of bad news before it turns into the unavoidable, it involves actively seeking for the information that allows one to combat the unexpected and survive bad luck. A manager in an informed organisation has to feel comfortable when hearing news that might be unwelcome, but has to feel uncomfortable when not hearing any such news, because life is not that fair and no news is bad news. But Informedness is a two way street, those with access to the information, typically the workforce, have to realise that what they can tell is worth telling and, by telling; they will be rewarded rather than punished. Described this way it is, I think, clear that Informedness is not a skill that will be acquired by some presidential fiat, or even by the actions of one person or a small and committed group. Informedness is an uncomfortable state that has to involve everyone and it takes time, and patience, to learn to live with it. Organisations that are informed, however, have considerable advantages, not just in the area of safety, over those that do not know what is going on!

     The second skill that has to be acquired is Trust. This means that managers have to trust their workforce. But it also means that the workforce have to trust their managers. Trust has to be earned. Managers have to show by their deeds rather than their words, that they do not have to double-check their workforce and do their work for them because they do not trust them to do it.

     Promoting the notion that safety management is a skill, and that safety cultures can be characterized most succinctly by their possession of this skill, enables us to understand the process of becoming a safety culture. Skills need to be developed at the start, and there has to be the expectation that acquiring a skill does not come immediately. We do not learn to ride a bike by reading the manual; we do not learn how to play a team game by listening to the coaches, no matter how good they might be. We learn by doing, by accepting failures and by trying again. We are motivated by the expectation of what we can do once we have the skill.

>

Safety Culture in Aviation

     What would an airline look like that was a true safety culture? The discussion so far has been quite general, and this is not surprising, as many of the characteristics are common to most organisations. Airlines are, nevertheless, relatively unusual. Like universities and hospitals, and unlike most companies, the core of the organisation is formed of large numbers of highly professional individuals together with specialized support staff. Pilots are members of a number of distinct cultures; being part of the company culture is often felt to be secondary to their membership of the world’s pilot community. Flight crew and the cabin crew are frequently socially distant. Aircraft engineers, while highly specialized technicians, form a self-sufficient group, often physically isolated from other members of the organisation. Finally the office and base staff, including check-in, baggage handlers and ancillary staff, is more like a normal industrial organisation.

>

1- The Flight Deck

     One of the major sources of problems in aviation is formed by the history of flight and its consequences for the attitudes of those who fly. Pilots have always been regarded as an elite, capable and self-sufficient, who are not necessary willing to admit to failure. The whole history of aviation has been based on the pilot as hero, often compensating for the shortcomings of the material. The self-sufficiency attitude means that Informedness is hard to achieve. Pilots will be unwilling to admit to having made mistakes; managers, often themselves members of the piloting fraternity, do not wish to hear about errors and regard other pilots who make mistakes as poor pilots, not up to their ‘standard’. The University of Texas LOSA studies (Helmreich et al. 1999; Klinect et al., 1999) have shown an average of two errors per sector, primarily deliberate non-compliance with procedures, so error is, in a sense, the norm that has not been admitted, at least not until we had live evidence of what happens on a day to day basis.

     In a safety culture this would be very different. Errors would be willingly, or at least openly, reported and the reasons analysed. If procedures are not being followed, primarily checklists being skipped, then management and flight crew would be working together to solve the problem. Failure to use checklists is one of the areas where the just culture has to be negotiated so that, once agreed, sanctions will be accepted if it should ever come that far. One example of what an advanced culture might do is to use considered risk management strategies to determine whether to press on or to turn back when problems arise. Instead of blindly following procedures pilots would develop sophisticated decision-making processes, both in advance and then continuously adapted to reality (Hudson & van der Graaf, 1998).

>

2- Cabin Crew

     The very overt dangers of flying speed and altitude, impinges rather obviously on those who actually pay, passengers, so that it is essential to support the message that it is now a safe activity. The cabin crew personifies the safety culture in the way they display their attitudes to the passengers. Passengers do not see the engineers, rarely see the pilots and never see the people in the background like dispatchers and loaders, so the cabin crews are the real public face of the organisation, as opposed to the image propagated by advertising agencies that people will quickly see through if they are given conflicting messages (Manoj S. Patankar and James C. Taylor, 2004).

     In a safety culture cabin crew can be seen to be actively involved. They may adapt the safety briefing to whoever is on board, possibly identifying first time flyers and giving those personal briefings (this is a situation where regulators may prove difficult). Cabin crews are convinced and convincing when safety issues arise in flight. They work together with the flight deck crew in full Crew Resource Management training and feel respected for what they know and do.

>

3- Maintenance

     Engineers work under continuous time pressure, nowhere more so than in overnight repair operations. The consequences of failures by engineers are often more devastating than those by pilots, as the failures are often impossible to fly out of2. Licensed engineers have considerable freedom to operate and under the pressures to get aircraft into the air errors are not infrequent (e.g. Hobbs, 1997). Engineers, like almost everyone in the aviation industry, are all too aware of just how dangerous aviation is and they are rarely willing to take risks. Nevertheless errors certainly occur, most especially omissions that are not caused by intentions to break the rules. Like pilots, engineers have considerable autonomy and this can easily create a culture in which trust and open sharing of information is not as common as one might hope for in a well developed safety culture.

>

4- Ground Staff

     Much of CRM is totally consistent with the practices of a safety culture, but in such a culture, there would be no need to require CRM, it would be a natural part of ‘how we do things round here’. In such an organization the ground staff also forms a part of the whole. A safety culture would recognize that ground staff all fulfils a function in ensuring the safety of flights, as well as realizing that many individuals, such as loaders and even check-in staff, are themselves at jeopardy. In a safety culture we would expect respect for the contributions such people can make, rather than being regarded as the lowest part of the organization. In fact many ground staff are often not employees of the airline, but work for contracting companies, such as ground handling agents. Much of the work of development of safety cultures in industries such as oil and gas involves the integration of company and contractor staff into task-oriented and integrated groups. The classic culture of aviation does not make such commonality of purpose and identity easy.

>

5- Management

     The final part of an organization that needs to be considered is management. Most of the individuals in management, given the commercial imperatives of the modern airline industry, may not have grown up in the industry, as pilots or engineers, and may not have acquired the ‘gut’ imperative for safety thathas kept aircraft operationssafe over the years. Managers, especially senior ones, are more likely to be lawyers or accountants than pilots and, as such, may be regarded as bold ‘pilots’ rather than being ‘old’ ones. Many pilots do perform management functions, but these are generally of a middle-management level rather than setting the scene for the entire organization, as senior managers do.

     While I do not have any empirical data to support the argument, I suspect that in those high hazard organizations that can be qualified as safety cultures, or that are actively attempting to develop into such a culture, senior managers have considerable technical experience. Engineers and scientists with hands-on experience with the risks run organisations with especially good safety performance, such as DuPont in the chemical industry and Shell Group. The difficulty for the aviation industry lies in the glamour of flying, influencing managers’ judgements. As running an airline can be the most cost-efficient way of losing money yet invented, managers have been increasingly required to be experts at running a business, regardless of the attractions. This means that, given the apparently good performance of the industry, as measured by hull loss statistics, managers may well adopt the classical reactive stance, believing in the importance of safety, but seeing it as a solved problem. When problems do arise, failure is typically attributed to failing individuals rather than organisational shortcomings. Managers in a safety culture would be well calibrated about the nature of the hazards and the magnitude of the risks their organisation, its personnel and passengers, face every time they take to the skies. Understanding this, they would support and encourage the development of the organisation, while recognising that while perfection can never be attained, it can be striven for.

     Finally management would demonstrate chronic unease. Board meetings would have safety issues as first on the agenda, not buried below ‘more important’ issues such as finance. Experience with taking this approach at such a level, in a major international company, found that most of the operational problems that were signalled under this item could be brought back to fundamental problems in non-operational departments, such as finance or human resources. Whereas in less advanced cultures management team or board meetings discussing incidents tended to concentrate on the failures of those making the reports, in the more advanced culture operations provided the symptoms, but the other departments had to recognize that they created the underlying causes (as per Reason, 1997).

    What characterises the descriptions of the different parts of the system in an aviation organisation is a high degree of common care. Whereas the traditional airline might be recognizable in the distinctions between those who fly and those who do not, the safe airline has low barriers to communication. Ground crew talk to aircrew and old distinctions will have been broken down. There are not separate safety departments for air safety and other safety issues, because they are integrated. As one airline that has performed this integration described it (Bruce Tesmer, personal communication 2001) “We no longer argue whether the aircraft hit the air bridge or the air bridge hit the aircraft that saves a lot of bother”. The company has well-developed reporting structures and learns from the information that is provided. While reporting of incidents may be confidential, this is not felt to be a major issue in a just culture. An advanced safety culture will be supported by a sensitive regulator. A regulator recognizes that failures can occur to the best, but who is most sensitive to signs that complacency might be setting in.

     In conclusion, this paper has argued that a safe organisation is one that can make money. The fear of incurring extra costs need form no In highly hazardous operations that can provide what people want; in the case of the airline industry getting to a destination rapidly, there is a premium to guaranteeing safety. Most people in aviation are inherently safety-minded, but this is only a necessary and not a sufficient precondition for the development of a safety culture. As I have argued above, the emphasis on extreme outcomes can paint a more positive picture than the wider picture would show and one of the crucial requirements.

     One of the advantages of promoting the processes that underlie a safety culture, the development of trust and Informedness, is that the organisation as a whole is easier to run. Less work has to be done to acquire the information needed and there will be far fewer surprises under normal circumstances. Small organisations may still argue that this will be too hard for them to do, even if the final advantages are clear. This is not true; the smaller organisations can proceed faster and more easily than larger ones. They can get everyone together in the hangar and have basic agreements before lunchtime; larger organisations have to do much more to reach the same level of agreement, and require more effort to maintain progress and ensure there is no backsliding.