2.1. Health and Safety Legislation in NZ
In the NZ case, the relevant legislation is the Health and Safety at Work Act (2015) [4
]. As the date shows, the Act is relatively recent. Moreover, it introduced a radically different legislative intent towards safety. The initiating event for the change in legislation was the Pike River mine catastrophe [5
]. The resulting investigation into the mining catastrophe [6
] identified multiple weaknesses in the legislative landscape prevailing at the time of the accident. For one, the legislation obligated organisations to eliminate/minimise/isolate hazards for their staff. The NZ approach to H&S was, and still is, consistent with the principle that legislation should ‘stipulate the duties of those with primary responsibility for OSH measures in general terms, rather than to attempt to regulate a multitude of hazards in minute detail [7
]. A legislative system comprises regulations in addition to laws, and the mining industry operated under considerable self-regulation. The mine company had the freedom to make decisions with significant adverse safety implications. For example: they proceeded to put the main ventilation fan inside the mine itself, despite the misgivings of the mine inspector [6
]; they defined for themselves a ‘non-restricted’ zone within the gassy coal seam, where they placed electrical apparatus, and made this decision based on engineering convenience [5
]. The investigation found no evidence of any risk assessment prior to placing the electrical services in the coal [6
]. The safety systems were ad hoc, staff from subcontractor firms were in the mine, sometimes with inadequate or no safety induction: “Its health and safety systems were inadequate” [6
]. Hence the self-management of H&S was ineffective and safety was compromised.
It was possible for executives to avoid the legislative duties if they could show, as in this case, that they were uninformed about risks: “the board of directors did not ensure that health and safety was being properly managed and the executive managers did not properly assess [this]” [6
], despite the numerous prior warning signs that methane was regularly at explosive levels. Fundamentally, the managers did not manage the risks competently.
Consequently, the law was changed, resulting in the current act [4
], which placed a much greater duty of care on executive officers (‘officers’) to inform themselves of H&S risks and resource the treatments. It also changed the definition of workers to include anyone at the workplace, irrespective of their employment contract or remuneration. It also created a joint responsibility for H&S for all organisations that had workers at the workplace, as opposed to partitioning the duty. The NZ act created a requirement on executives to show due diligence for H&S in a way complementary to their fiduciary duties under the company act. (For a wider analysis of the effect of catastrophes on the improvement of process safety, see [8
2.2. Intent of Legislative Frameworks for Self-Management—The International Context
The intent of legislative frameworks is that organisations implement ‘voluntary arrangements to strengthen compliance with regulations and standards leading to continual improvement in OSH performance’ [9
]. In practice, this means that organisations are left to develop their own H&S management systems, as evident in ISO 45001 and the ILO standards [10
]. (For a comparison of the prescriptive vs. flexible approach to legislation see [11
However, to some extent that appears to be a naive premise. As evident in the Pike River mine disaster, even in developed countries the self-management can fail [6
]. A similar observation has been made about the Canadian mining sector specifically the need for audit [12
], and for inspection in the United States dam sector [13
There is also a case to be made that smaller organisations, and organisations in developing nations, struggle to successfully implement self-management. Organisations with less access to professional resources may find it difficult to create such systems. There is evidence from Nigeria that the lack of specificity in H&S legislation adversely affects construction safety [14
]. Weaknesses in clarity and structure have also been identified in the United Kingdom H&S law [16
]. Likewise, the Australian experience has been that small and medium enterprises have difficulties in developing specific workplace practices from the abstract and non-prescriptive legislation [17
]. Results from Italy also show that smaller firms are less effective than larger ones at reducing accident rates [18
]. Likewise, the Malaysian experience has been that H&S management systems are often difficult to implement to a level that meets the legislation [19
]. This has also been found to be the experience in Malawi [20
]. It would appear that developing countries generally struggle to develop the systems necessary to give effect to legislative objectives. Presumably, part of this is the lesser availability of H&S expertise to smaller firms, which was specifically identified as the case in Ghana [21
]. Related to this is the need to actively develop a professional body of health and safety experts, and even in developed nations this is something that requires deliberate work and support [22
The case has been made that flexibility in H&S legislation creates greater liability risk in the aviation sector, and that organisations are tempted to respond to this by suppressing the internal reporting of safety risks [23
]. Hence, the quality of safety reporting becomes a key determinant in the effectiveness of H&S management systems. The older NZ legislation created just such a perverse incentive for executives to maintain their ignorance of risks within their organisations. There can be other perverse gaps between that which the law intends and what organisations implement. In Australia, there are greater H&S responsibilities for larger construction projects, where size is measured in financial terms, but there is evidence that the monetary threshold is not ideal, e.g., contracts may split to avoid the requirements [24
2.3. Challenges to Management Systems
Many engineering projects involve multinational collaboration, and workforces of diverse cultural origins. Challenges to H&S management systems have been observed in large construction projects with multinational labour in the United Kingdom, relating to work practices, communication, and culture [25
There are other shared areas where nations operate but without sovereign jurisdiction, and there is a need for coordination of safety. Historically the international ocean has been this domain, to which can now be added Earth orbit and other planets. Safety is one of many considerations in the legislative framework that is evolving for space, currently in an ad-hoc manner [26
]. New technologies are another area where common understanding of risks and how to assess them can be valuable, nanotechnology being an example [27
Food safety is a more down to Earth application for risk assessment [28
], and the global interconnectedness of the supply chain means that working towards commonality of these safety processes is advantageous. Again small firms are more at risk, e.g., in China [29
]. Better coordination of risk assessment processes and documentation between subcontractors, especially the inclusiveness thereof, has been identified as necessary in the chemical industry [30
Another area where a common understanding of risk is needed is mental health and psychosocial risk [31
]. While mental health is specifically included in some H&S legislation, NZ being an example, the methods for assessing and managing this are only weakly developed if at all. Harmonisation of policy across jurisdictions has been found to have positive effects on psychosocial health in Australia, with non-harmonised jurisdictions showing reduced communication efficacy and management commitment [32
2.4. Need for More Systematic Methods
The international expectation is that risk assessment should comply with the regulations of the jurisdiction (International Labour Organisation) [9
]. The ISO 45001 standard for Occupational health and safety management systems
likewise identifies the central need for organisations to have processes to identify hazards (Section 18.104.22.168) and assess risks (Section 22.214.171.124) [33
]. These documents identify multiple factors to be taken into account. In ISO 45001, these include organisational culture, human factors, history of incidents, emergency situations, other people exposed to the hazard, etc. There is much about internal alignment,
the development of an H&S management system that takes into account how the organisation conducts its operations. The concept that organisations may also need to ensure external alignment
with the national legislative framework is weakly developed in all these documents, and absent regarding risk assessment in particular. Instead each organisation develops its own H&S management system and tools, including risk assessment methods.
There is a need for more systematic methods for use as elements within an H&S management system. It is apparent from the literature that there exist many organisations, especially smaller ones and those in developing nations, which lack resources to create management systems ab initio. Furthermore, there is a need to strengthen the communication mechanisms between those doing the risk assessments, and the managers who have the financial budget to support treatment.
Key to this is the need to anticipate, identify, and evaluate hazards and risks [9
]. This is central to the operation of a safety and health management system within an organisation [33
]. If the risk assessment are not being done, or done to a lower level of rigour, or not being communicated to managers, then treatments may not be resourced and continual improvement fails. Effective use of risk assessment processes showed positive benefits in Finland [34
] and a positive correlation with specific preventative treatments [35
]. (For a general background to the Finnish H&S legal systems see [36
2.6. Specific Needs for New Zealand
In the case of NZ the Act implicitly expects practitioners to conduct hazard assessments, but does not specify the methods. Instead the methods are provided in guidance documents and examples provided by the regulator (Worksafe). These show a semi-qualitative risk-assessment method whereby qualitative estimates of consequence and likelihood are converted to ordered scales. In one publication, Worksafe proposed the use of simple linear scales for consequence and likelihood, e.g.,:
Moderate—first aid and/or medical treatment,
These scales are simple linear ones to which numbers are allocated (typically ranging from 1 to 5). Of itself, this is not necessarily an issue, because there are often insufficient data on which to make a more precise determination. Also, the method is intended to be applied by workers and team-leaders, or at least comprehended by them, hence a simple system is advantageous.
Another example of the type of scale in widespread use in NZ is:
Note the variability in the definition of major harm in the two examples above. One defines it as extensive injuries, the other as permanent injury. It will be appreciated that these are not the same thing, and the extent and severity of these injuries is not specified. Furthermore, neither construct appears in the NZ legislation, hence the external alignment is also missing.
The universal premise of ordered scales is that the increments between the categories are of equal value, but this is not validated. Such scales suffer from methodological invalidity when used in a subsequent mathematical product operation with a similarly constructed likelihood scale, which is invariably how they are used. The issue is the ordinal scale is subsequently being used as an integer one. The reasons why this is problematic are apparent when considering the intervals between the categories. In effect the scale asserts that death is five times worse than a minor incident (such as a scratch that needs a sticking plaster), and that death is another 25% worse than major harm. These are naïve assumptions that do not stand up to scrutiny, however one looks at them. If one applied the perspective of economic rationalism, the cost of a sticking plaster is a few cents, and the financial value of a human life (there are multiple measures) is in the order of a hundred thousand dollars, which is certainly not a 1:5 ratio. Furthermore the death of one person might be a random accident, but the death of many in a catastrophe implies a serious failure of the of H&S management system and may attract fines and punitive consequences beyond the mere proportional increase in the number of deaths. If instead one applied a medical rehabilitation lens, one could measure the diminished quality of life caused by various levels of harm. Quality of life scales do exist, e.g., WHODAS (which measures a person’s ability to function in matters of living) [42
], and have been applied experimentally to the risk assessment process [43
], but do not support simplistic 1…5 consequence scales. Such scales fail to recognise the non-linear effect of catastrophic accidents.
Providing an organisation uses the same scales throughout, it should result in a consistent measurement of risk. However there is a lack of consistency between organisations—each uses its own scale constructs. Consequently, the same hazards may be assessed differently, and hence given different treatment priorities, between organisations. This is where the problem arises, because H&S legislation tends to make the best practices the standard, and hence organisations may find that they expose themselves to legislative risk.
The regulations define a ‘prescribed risk management process’ [44
], but this is merely the need to identify hazards and apply the hierarchy of hazard control. The actual risk assessment process is not prescribed, and the process goes straight from identifying hazards to treatment without mentioning the types of risk assessment processes that would ordinarily be understood from ISO 31000.
Furthermore the scales are primarily focussed on preventative activities. Indeed that is the perspective taken by the NZ regulator, in common with many other jurisdictions. Thus the prevention of a hazard by elimination or minimisation is the primary objective. This is evident in the hierarchy of hazard treatment which Worksafe uses, for which they use the term ‘hierarchy of control measures’ [44
]. That the focus is prevention is evident in the term ‘control’ and by scrutiny of the diagram. The concept of recovery mechanisms, to prevent a small accident propagating into a larger catastrophe, is absent (except perhaps tacitly). Worksafe do identify catastrophic hazards, but their definition of ‘low frequency, high consequence (e.g., major industrial, workplace or transport incidents such as a large explosion’ [45
] is focussed on accidents that occur immediately, and ignores the progressive and cascade failures where opportunities to prevent propagation are missed.
For example, the Pike River Mine recorded excessively high methane levels in the days and hours preceding the disaster, which were definitive indicators that the methane risk was out of control. Yet, management ignored that information and continued mining operations that would release further methane. They had previously experienced occasional small localised explosions of methane that were minor and which did not propagate to full conflagrations. The thermodynamic explosion of the mine occurred on 19 November 2010, but the causal path to disaster commenced months before. The real H&S violation at Pike was not so much the final explosion, but the failure to apply recovery mechanisms earlier when it was apparent from the methane readings that the preventative hierarchy of control had failed.
Where catastrophic outcomes can reasonably be foreseen, it is irresponsible to rely only on prevention: thought needs also to be given to recovery. The hazard assessment method is primarily a preventative method, at least in the way it is commonly applied, and is poor at eliciting disaster recovery actions. The bowtie method is much better at soliciting recovery actions that may prevent an accident propagating into a wider disaster. Some of that thinking needs to be included in the H&S risk assessments.
Another issue is that the risk assessment scales bear no resemblance to the definitions in the NZ H&S legislation. This is particularly true of the consequence scale, where legislation may have specific definitions for types of harm, but these are not always carried through to the scales. Consequently there can be a mismatch between the risk assessment scales used by an organisation, and the expectations of legislation. The Act uses a coarse definition of harm, which is notifiable incident, notifiable accident, and death. There are much finer grades of harm that exist, which are important for hazard analysis.
Furthermore, the hazard assessments tend to be ad hoc. The list of risks in the register tends to be in the order in which they were thought up, without a system. Worksafe itself offers no very systematic method for hazard identification [44
]. Obviously systematic methods do exist, e.g., fault tree analysis, but these seem not to be used. Presumably this is because the purpose of hazard assessments is that they be done by workers rather than safety specialists.
A further deficiency of many hazard assessments is the lack of connectedness of the organisational decision making. The Act created a legal duty for executives to ensure they keep themselves informed of risks. Executives are personally liable if they fail in this duty. In this way, the Act explicitly removed the previous defence of ignorance that was widely used by executives. Hence, there is a need to be more explicit about the thresholds where knowledge of risk is escalated. Existing methods appear not to have a coherent communication strategy and hence executives, at least in NZ and other nations that adopt similar legislation, are potentially exposed to significant personal risk. There is a need for a better way to communicate about risk.