How different is Safety Differently?

Dominic Cooper PhD. CFIOSH CPsychol, B-Safe Management Solutions, Inc., Franklin, IN 46131, USA.

Declaration of Bias. Before starting this manuscript, I did not believe Safety Differently had much to offer the EHS world. I have resolutely examined the evidence in a detached and neutral way, in the hope that I will be able to further assist people in improving safety. I make it available so others can make up their own minds. 

Introduction

Leaning heavily on Hollnagel’s Safety -2 resilience engineering approach to process safety, Safety Differently (SD) is concerned with Human Factors. Coined by Sidney Dekker in 2012, the term SD and its associated approach is postulated to overcome perceived problems in tackling occupational safety, by challenging organisations and the EHS profession to view three key areas of Occupational safety in a different way: [1] how safety is defined (Safety-1); [2] the role of people in safety; and, [3] how businesses focus on safety. 

How Safety is Defined

Definitions are important as they provide the lens or framework by which people look at things, which in turn influences the methodologies and resources required to enact the ‘definition’. SD is promoted as a “mindset” or different lens with which to look at safety. SD advocates berate how traditional safety is defined and assert the EHS profession must focus on the positives and escape from company policies, rules and procedures by eliminating them and ceasing to manage safety. 

A dictionary definition of the word ‘safety’ states it is “an absence of negatives (i.e. freedom from danger, risk, or injury)”. Safety-1 is concerned with ensuring there is an absence of injuries, incidents, accidents, risks, hazards, etc. As such, this perspective defines the purpose of safety management as “the prevention of harm”, a view explicitly intended by regulators to ensure companies try to avoid causing harm to their employees while engaged in the company’s activities. 

Conversely, SD defines safety as “the presence of positives”. Dekker’s explicit examples of positives are the ability to say ‘no’ in the face of acute production pressures (i.e. the stop-work authority, a right already underpinned by EHS regulation, and promoted in many companies); the willingness of superiors to hear bad news and the acceptance and encouragement of dissenting views (i.e. Amy Edmondson’s 1999 concepts of ‘Psychological Safety’); the commitment to learning and the restoration of trust and relationships if vulnerabilities and problems have been identified (i.e. aspects of James Reasons 1998 approach to safety culture). Every single one of these ‘positives’ are Safety-1 concepts, and are sometimes used as leading, as opposed to lagging, indicators. 

Purportedly, SD is marked by a presence of successes, and the more there are, the safer the system is thought to be. Advocates maintain that SD is fundamentally different to Safety-1, as it is not just about successful safety outcomes, but about explicitly trying to change the effectiveness of the entire organisation by focusing on all those things the organisations do well, and building on those. This might be great theoretically, but in practice if an organisation is doing 99.99% of its activities right, it still means there is a lot of scope for error. It only takes a few small system things to go wrong, combined with people’s ineffective behaviour, to destroy a facility, or sink a vessel. It only takes one unsafe behaviour for someone to die. History is replete with examples. SD advocates state we should accept this!

SD’s emphasis on the positives is based on the notion that ‘the absence of accidents does not indicate the presence of safety’. I don’t think anyone entirely disagrees with this: most EHS professionals acknowledge that there is sometimes an element of luck involved. But, the absence of accidents does not necessarily indicate there is an absence of safety, either. Safety-1 approaches advocate people proactively look for hazards and risks (negatives), and fix them or put in safety controls before they cause an injury. If there are no accidents, that does rather suggest that the company is doing the right things (assuming adverse events are not being hidden). It’s hard to escape the conclusion that the difference between Safety-1 and SD definitions of safety is essentially nothing more than semantics, rather than something fundamental: i.e. an absence of negatives (e.g. injuries) by definition has to be a positive. 

In comparison to Safety-1, SD advocates suggest they differ by looking at the whole task to understand, [a] where risks come from; [b] what pressures the risk sources exert on workers; and [c] how they manage the complexity in these tasks. In my view, this scope is exactly the same as Safety-1s Risk Appraisal, Assessment and Evaluation approaches promoted by regulators. I also believe it is a positive for workers to identify hazards and risks, report and deal with them.  

SDs positive focus is also about the fact they believe safety must have the capacity to adapt, tolerate change, be resilient and recover from failures. Every adaptation, change, and resilient bounce back from failure (an accident? a deviation?) is viewed as a positive. I believe this aspect is largely meaningless rhetoric. SD states one of its purposes is about freeing people from bureaucracy, but then wants people to monitor and record every aspect of variation in normal working life. By and large, for many organisations that is a pointless, costly, bureaucratic exercise, with limited proven value in terms of improving safety and reducing injuries. Moreover, most organisations EHS functions and systems already have the ability or capacity to adapt to changing circumstances. They do it all the time as markets change, as new materials are invented, as new technology is introduced etc. To presuppose that Safety-1 risk assessment approaches lack this capacity, is quite simply, a denial of the facts. 

SD advocates also point to the fact that many serious incidents are preceded by long periods of accident free operation, implying there are not enough warning signals from focusing on the presence of risks and danger. SD proponents say this is one of the major reasons we should focus on the positive successes. Strangely, SD advocates do not recognise that leading indicators (e.g. visible leadership behaviour, competence and training, employee involvement in safety activities, percent-safe scores, and corrective action rates), in conjunction with lagging indicators (e.g. near-miss reporting, incident statistics, etc) have been around since the mid 1990’s to provide numerous signals about the presence and quality of safety (i.e. freedom from danger, risk and injury). 

Unfortunately, examinations of this sole focus on safety positives with SD advocates have failed. Requests of how to use tens of thousands of records where people were praised for a job well done, because they had worked safely, were met with total silence. It appears SD advocates themselves do not know what to do with positive data, beyond collecting it! When challenged on this, all automatically revert to Sidney Dekker’s mantra “Giving you help from the top-down would be the diametrical opposite of Safety Differently. To ask for guidance on what to do to implement Safety Differently, is to regress to a Safety-1 mindset”. And there was I under the misapprehension that we were all in this together!

In the same data sets, there are also tens of thousands of records of unsafe behaviours. The organisations who own the records and I, have learned so much more about the tasks involved, and associated system faults driving unsafe behaviour, than we ever do from focusing on the positive behaviours done well. These and near-miss records allow us to focus on potential serious injuries and fatalities (SIFs), in attempts to stop actual serious injuries and fatalities.  

The evidence to hand on both sides of this fence, overwhelming points to the value of focusing on negatives, while acknowledging the positives. However, to entirely ignore the negatives as SD postulates is foolish, and could easily be construed as gross negligence (i.e. an indifference to, and a blatant violation of, a legal duty with respect to the rights of others) in a legal case involving a fatality or catastrophe. Indeed, in 2019 we saw the doubling of incident rates from 2.2 to 4.5 in Origin Energy, SD’s Oil & Gas poster child company, who had adopted the espoused SD practices. 

To be fair, other companies such as Mitchell Services reported a reduction of 58% in its TRIF in 2016-17 because of engaging employees in safety, although the actual number of incidents is unknown. However, using the known power of Employee Engagement, where engaged employees are five times less likely than non-engaged employees to have a safety incident, and seven times less likely to have a lost-time safety incident, does not make this SD. Whichever way you look at it, Employee Engagement is a Safety-1 concept. For example, I and many others (e.g. Tim Marsh, Scott Geller, Terry McSween, Tom Krause, etc.) have been helping companies around the globe implement Employee-Led Behaviour-Based safety processes since 1989, with similar incident reductions, year on year. Results of many employee-led implementations have been published in scientific journals so others can benefit. 

The role of people in safety 

Safety-1 asserts that safety starts at the top of a company with its executive leadership team deciding on strategy to fulfil the organisation’s legal responsibilities. Safety-1 states it’s the duty of all employees regardless of rank and/or task to work safely and get involved in managing any risks. To this end, they put in place safety management systems (SMS), hire EHS personnel and resource EHS departments, develop policies and rules, train employees and fund, resource, and monitor the results of any improvement initiatives so performance can be adjusted as appropriate. Without doubt, some organisations are better at EHS than others, but it is very rare these days for an organisation to entirely ignore EHS legislation, and they certainly do so at their peril. 

SD advocates state that “Safety-1 safety management systems provide EHS professionals with a ‘safe’ retreat behind mountains of bureaucracy and paperwork, so that they may be really busy with the ‘work of safety’, rather than the ‘safety of work’. Moreover, the Safety-I approach would seem to have limited effectiveness primarily because it deals poorly with variability in work performance”. These two bold assertions fly in the face of the voluminous evidence showing that the installation of a SMS improves safety performance, the IMO providing a global example in the Maritime industry. SD implies that EHS professionals don’t want to do safety ‘on-the-floor’: they would rather hide behind paperwork. There may be one EHS professional per 1000, who prefer to do nothing but sit in an office, marking time. The grim reality is that most EHS professionals make a difference in safety, and work hard to make it happen. Many suffer burn-out from the excessive hours they put in caring for their colleagues, and many also suffer PTSD after investigating and dealing with the aftermath of serious injuries and fatalities. 

It appears, SD proponents also do not believe that Safety-1 can cope with complexity, and that SD is the solution to this as it recognises the difference between work as done (WAD) and work as planned (WAP). If true, how do the Oil & Gas majors using Safety-1, develop national infrastructures using third-party labour involving tens of thousands of people, in various parts of the globe, in record times, within budget, and with extremely low, world beating incident rates in some cases? Often achieved within the constraints of having to use local resources, while competing for those same resources with other oil & gas majors, they have produced large scale facilities ultra-safely, by adopting Safety-1 methods. I believe the real issue surfacing here, is the way that safety is sometimes managed and executed in some workplaces, rather than the underlying theories of various Safety-1 approaches not being fit-for-purpose (Behavioural-Based Safety being one typical prime example of badly executed processes). i.e. one negative outweighs 10 positives. It is this, I think, that has given rise to SD and its spread.

SD proponents also state Safety-1 treats people as a problem to be controlled, rather than the solution to safety concerns. The SD premise is people [a] create success far more often than failure; [b] are a source of innovation and insight; and, [c] allow mature conversations around risk to occur. Safety-1 does not deny a, b, and c, but actively promotes people’s participation within the structure of the SMS. Dekker asserts that “SD decentralises and devolves decision-power about how to do things safely to the entity’s expertise residing in the ‘sharp-ends’ of the organisation”. In other words, he explicitly advocates companies ceasing to manage safety at all, and turning over all responsibility for safety to the workforce. As the SD acolytes express it “It is time for Safety Anarchists: people who trust people more than process, who rely on horizontally coordinating experiences and innovations, who push back against petty rules and coercive compliance, and who help recover the dignity and expertise of human work”.  

I suspect most EHS professionals reject anarchy, but do advocate and encourage Employee Engagement, a Safety-1 concept. Explicit Employee Engagement practices readers may recognise include: [a] developing and training workers as safety leaders; [b] encouraging the reporting of incidents, [c] helping to investigate & review incidents; [d] identifying and reporting hazards; [e] conducting risk assessments; [f] reviewing rules and procedures; [g] developing and delivering toolbox talks; [h] pro-actively involving people in behavioural safety processes; and [i] sitting on safety steering committees. I have not yet found any new SD employee engagement practices: everything is already done within Safety-1. I do recognise that not all companies believe in employee engagement, or only allow it for certain activities. In other words, it sometimes comes with restraints. That is also the same for other areas of organisation life. Many SD advocates might argue that I have got this wrong, that employees are engaged in looking at the difference between work as done (WAD) and work as planned (WAP). I would respond, risk appraisal, assessment and evaluation processes promoted by the regulators within Safety-1, do exactly the same thing.

In sum, therefore, the arguments put forth by SD proponents about the role of people in safety seem somewhat fallacious. The evidence points to the efficacy of safety management systems reducing incidents and injuries. Similarly, Safety-1 has explicitly promoted Employee Engagement: in Britain and Europe, since at least 1992, as advocated in the MHSWR 1992. In my view, the only difference between Safety-1 and SD, is that SD explicitly calls for companies to cease managing safety and give all decision-making power for safety to employees. Corresponding responsibilities and accountabilities (moral, financial and legal) are neither mentioned nor implied by SD proponents. It appears to be a free-rein SD party where there are no sanctions or consequences for poor performance. I do wonder if the SD anarchy advocates ever spent years physically working ‘on-the-tools’ before passing on their wisdom to the EHS profession. I suspect not, else they would be much more circumspect in what they advocate. 

How businesses focus on safety

SD advocates assert that Safety-1 ensures organisations focus on safety as a bureaucratic activity with people being held to account by their leaders and senior managers. I am relatively neutral with this assertion, but recognise that some level of bureaucracy and accountability are part and parcel of daily organisational life across all arenas, in all countries, even in devolved or matrix type organisations. Structure is absolutely vital for safe performance, and it is not something that is going to go away. 

SD’s solution to perceived heavy-handed bureaucracy is that safety be treated as an ethical or moral responsibility not a bureaucratic activity. SD promotes the idea that safety management systems are used solely to manage safety not liability. That is an interesting concept as the two are inextricably linked: Insurers charge premiums according to how well a company manages its incident and injury rates (i.e. by how safe, or risk averse, it is deemed to be). Regulators hold companies to account for doing safety poorly, and use the legal system and courts to ascertain the company’s liability in terms of compensation to injured parties. At an organisation system level, none of these can actually be separated (SD advocates do say they like dealing with complexity!). However, at task-level, it is certainly possible for employees to develop a separate set of specific and explicit rules and procedures for each “shop-floor” task, which are derived from the organisation’s over-arching managerial rules/and procedures that do contain liability components used to meet certification (e.g. ISO 45001) and legislative requirements.

SD advocates state that lean management systems enable effective risk management (evidence?). So, in one sense, SD wants to simplify and declutter bloated procedures and rules by enlisting employees to help provide the right ideas. Research in New Zealand coal-mining has shown that fewer, high quality procedures are much more effective than many, poorly written bloated ones. But the problem here, is this is a Safety-1 concept. SD did not come up with this, they adopted it, in the same way as they have adopted lean manufacturing or lean engineering principles. Depending on your perspective, ‘lean …..” can be seen as streamlining your processes or adopting institutional short-cuts. So long as it is not going to lead to someone being injured, or result in significant property damage, it might be a good thing (although people might also lose their jobs as a result). Regardless, none of this is new or different than what came before. 

SD advocates also state that systems should be designed to promote relationships and not transactions. There is little guidance on this, but it seems to be about recognising that risk awareness and risk management can be a messy and complex issue, and the solution is to create interdisciplinary teams to overcome any complexity. In essence, this promotes departmental networking and teamwork as the way forward, particularly if you are looking to enhance the entire organisation's performance. SD asserts this requires a different skill set and a different view of what it takes to be a leader. Again, there is nothing new or different here, that has not already been espoused and expanded upon numerous times over the past three decades or so. 

In summary, it appears that SD’s perspective on separating out safety and liability is about reducing the amount of bureaucracy, albeit a difficult thing to actually achieve. In and of itself, reducing bureaucracy might not be a bad thing, provided the baby isn’t thrown out with the bathwater. Similarly, creating good working networks and collegiate team-working across disciplines is a good thing. However, once again, it appears that SD is no different than Safety-1. 

The science underpinning Safety 1 and SD

In December of 2018, SD advocate John Green, in the British publication the Safety & Health Practitioner, asserted that “no-one approaches safety with more scientific rigor than the Safety Differently community” and “if you are relying on Heinrich, triangles or dominoes for your safety programmes then you are the ones building on sand, you are the ones in glass-houses throwing stones. The absence of scientific testing in these approaches is simply breath-taking”. 

I confess, after 30 years researching industrial/occupational safety, I was stunned by these statements. A brief tour of the science behind Safety-1 and SD shows there are literally hundreds of published scientific journal papers on Safety-1 concepts stretching back decades. Familiar aspect of Safety-1 include, but are not limited to, [1] safety culture research; [2] safety climate surveys; [3] behaviour-based safety; [4] safety management systems; [5] risk assessment methods, risk management; risk perception, etc., [6] incident triangles, incident causation and incident investigation; [7] root cause analysis; [8] employee engagement; [9] psychological safety and teamwork; and [10] safety leadership. 

Although, the quality of some of the Safety-1 research leaves much to be desired, there is absolutely no published peer-reviewed scientific evidence to support SD at all: There are a lot of published sociological “should do and will do” texts by SD acolytes and anarchists who mostly sit in the School of Humanities, Languages and Social Science, Griffith University, Australia, but, to my knowledge, no scientific published experiments or evidence demonstrating “we did this, and the results are….”.

The appeal of SD

The inescapable conclusion of unpacking SD, is that the emperor has no clothes: the box is empty. ‘Safety Differently’ is a concept concerned with the ‘art of managing nothing’, because quite simply, it has no substance of its own: apart from the notion to cease managing safety, every idea, method, and strategy are appropriated from Safety-1. Therefore, SD is reduced to nothing more than a banner, a slogan. The real work of improving safety goes on underneath the surface by the dedicated Safety-1 EHS professionals, using proven Safety-1 concepts, systems, and processes. 

Given SD is shown to be an empty slogan, how did it gain a foothold in the EHS landscape? SDs underlying philosophies are based on Dekker’s rejection of the premise behind Vision Zero. He states it lacks intellectual underpinning by the likes of Marx, Freud, Mill, de Tocqueville, Bonhoeffer, Nietzsche, or Kierkegaard. I doubt that choice of philosopher(s), or lack of, has maimed and killed people at work: It is usually a piece of equipment, a system fault, or number of system faults, combined with unsafe behaviour(s) that leads to incidents that maim and kill.  

While debates rage about whether Vision Zero is realistic, I can honestly say I have seen companies achieve zero accidents, sometimes for a few years at a time. But it is hard work to achieve in the first place, and even harder to maintain over the longer term. As such SD is likely born from frustration with the darker side of the EHS world: i.e. suppressed incident reporting, manipulative injury reporting to downplay severity to create a ‘looking good index’ (Dekker’s term). I have a lot of empathy for his views on these dark practices: I don’t like them, and I don’t condone them. 

There are many other unethical dark practices in the EHS world, like using generic risk assessments and permit-to-works, because the company cannot or will not resource these systems properly; making people sign to say they have attended a tool-box talk, or induction, or work permit, without also telling these same people they have now just become legally liable if they get hurt, as the lawyers/attorneys will use that against them in court; making people do drug tests if they are injured or involved in a near-miss in attempts to shift responsibility onto workers. The list of dark arts in safety is endless, but they are more prone to surface when companies put profit before safety, and create / maintain a culture of fear (two aspects of safety culture related to most of the world's safety catastrophes). Importantly, the dark side causes the workforce to withdraw from safety. This then provides the fertile ground for many injuries, and an ethical void for SD acolytes to exploit. 

As academic Sociologists, SD acolytes have examined and learned how to direct ‘the masses’ into doing what they want. SD is about stirring emotions by highlighting all the problems with Safety-1 to get people to care about safety and act. Because SD states that answering any question means regressing to safety-1 thinking, it gives the EHS profession and companies permission to do anything, like ceasing to managing safety, without simultaneously being held responsible or accountable for anything that goes wrong. Entirely shifting responsibility (and legal liability) for safety onto the workforce at the ‘sharp-end’ is viewed as nirvana by many. In a sense, SD offers the EHS profession pure freedom, a very powerful and seductive drug indeed. 

EHS people most likely tempted by SD are those disillusioned with the status quo, and/or disillusioned with safety’s dark arts, accompanied perhaps with rising incident rates, or a serious injury or fatality that induces feelings of vulnerability, guilt and grief. Identifying the vulnerable, and using sociological knowledge of cult formation, recruitment to the SD cause is about deception (“Emperor Safety Differently” does wear clothes); manipulation (the true nature of SD foisting responsibility and legal liability for safety onto the workforce is hidden); seduction (giving the EHS profession permission to deploy strategies, methods and tools it already has at its disposal) and isolation (accusing doubters of Safety-1 thinking when seeking answers). 

Going forward, the answer to the growth of SD and its ilk in the EHS landscape, has to be a much more ethical approach within Safety-1, with practitioners of the dark arts being called out by their colleagues and professional bodies. 

Summary

When the major tenets of SD are unpacked, it becomes evident that SD is an eclectic blend of cherry-picked aspects of Safety-1, re-packaged and sold to EHS professionals as some type of new quasi-religious movement that will fundamentally improve safety. In other words, SD is confounded with a host of Safety-1 interventions. Ironically, this is the same accusation that Dekker levelled at Vision Zero in 2017.  

It also appears that SD is about marketing sociological approaches to safety under the guise of ground-breaking academic research, but.... the evidence-base does not exist. SD is entirely unsupported by any peer-reviewed scientific validation studies whatsoever. Indeed, reflections like this manuscript are in response to the absolute refusal by SD proponents to provide any scientific evidence whatsoever that it impacts and reduces serious injuries and fatalities (SIFs), or indeed provide any answers to any questions at all.

In my view, SD has gained ground because the EHS world has been practising its dark arts, which has led to major levels of disillusionment among ethical EHS professionals who have been seeking a way forward to make the difference. SD was born out of disillusionment, and despite the emperor not wearing any clothes, SD proponents have cynically exploited the ethical void in cult–like fashion. Caveat Emptor!

Reference

Dekker, S. (2017). Zero Vision: enlightenment and new religion. Policy and Practice in Health and Safety, 15(2), 101-107.