The Pathology of The New Zealand Construction Sector: why are we killing so many of our workers and what are we going to do about it?

This article is the first in a series concentrating on the drivers behind construction related Serious Injuries and Fatalities (SIFs) and where the leadership focus and investment needs to be to make significant change.

Three focus areas are singled out specifically that relate to SIF elimination. It is written by myself and my colleague Jon Harper-Slade from CHASNZ.

  1. Awareness, recognition of and respect for the construction risk environment and the underlying work conditions that make up the risk profile.
  2. Improving the level of performance and reduction in the variability of Construction processes.
  3. Investment in appropriate and resilient controls that widen the safety margin.

The problem

To date, this year the Construction sector has had 14 fatal incidents. 14 people who will not be going home to their families, many children now growing up without a parent.

Although construction makes up 8-10% of the workforce its accounts for 25% of workplace fatalities. The New Zealand Construction sector’s rate of fatalities per 100,000 workers is twice that of Australia and almost four times that of the UK.

Is this a momentary slip in performance of an emerging trend?  On average over the past 10 years, 6-7 construction workers have been killed on site through a workplace accident every year.  In 2019, there have been 14 construction related fatalities to date, a significant and alarming trend.  

Instead of improving we are now looking at a situation where the data spells out that the industry is a much more dangerous place to work than it has been in the past.  Added to this are the significant and regularly occurring serious injuries and adverse health effects, which have permanently changed the lives of those impacted.

The Construction industry must change.  This kind of performance has never been acceptable and is certainly abhorrent to any civilized society. Existing strategies are not reducing fatalities or serious injuries; a new focus is needed!

We need a revolution in our approach to dealing with serious injuries and fatalities (SIFs). With the volume of construction activity increasing the number of SIFs is likely to continue to rise unless we as an industry take a leadership stand now to halt this appalling and continued poor performance!

What is it about Construction that makes it so dangerous for our people?

The answer is in part related to the underlying nature of the work activity, the environment that the work is carried out in and the interactions between both.

Construction is about change. From hour to hour and day to day on a construction site, different and varied activity is taking place and the object of the construction activity (i.e. what is being built) is continually changing.  It is not an inherently stable environment when compared to other workplaces where work is regular, and the workplace is environmentally static.

With work activity comes inherent risk. The safest ship never leaves the harbour, and as such it is difficult to totally eliminate workplace risk without eliminating the work it springs from.

The inherent risk profile for construction changes on a rapid basis depending on what work is happening and what stage the construction process is at.  Construction related activities bring different machinery, skillsets, procedures, and materials to bear often near other similarly potentially hazardous activities.  Added to this are dynamic environmental conditions such as weather, light, ground conditions, and interaction with third parties such as members of the public (e.g. traffic).

Task related activities, which are known to be hazardous, interact with other activities and the workplace conditions to create a highly volatile risk environment.

The conceptual graph below shows that over time the inherent risk profile changes depending on what activity is taking place and as working conditions change. The more complex graph showing the interactions of the many activities has been simplified in the lower graph to show the net inherent peak risk profile hourly.

No alt text provided for this image

Inherent risk is a bi product of the work processes being carried out. If the work doesn’t happen, the risk isn’t present. Normally this inherent risk is recognised and taken account of when the work process is defined.

Work as imagined should consider the known inherent risks associated with that process. For example, when undertaking hot work, a normal process would take into account the risk of an uncontrolled fire and do the work in a way that avoids the risk of such an event.  How hot work is conducted, by whom, with what tools on what materials and in what conditions are normal process factors.

When defining standard operating procedures or ad-hoc processes. a fishbone diagram can be useful in conceptualizing the inputs into any given process.

No alt text provided for this image

We therefore describe and expect processes to happen as we have planned and sometimes documented them (work as imagined). We expect them to have considered the inherent risks associated with the work and often spend much time and effort trying to measure and enforce compliance with these processes.

So why does our actual work process performance vary from what we plan and expect?

The reality is that each of the inputs into the work process, the 5 Ms, vary in the real world. This is normal and we should always expect actual work processes to vary from the imagined work process.

It is so normal in fact that it has been studied and recognised for decades in other industries such as process manufacturing.  Statistical Process Control (SPC) is a technique used to understand process variation and is particularly useful in identifying and managing normal verses abnormal variations.

What SPC tells us is that all processes vary from their intended outcomes, and that trying to correct all variations is fruitless and often contradictory to what you are trying to achieve, i.e. by trying to correct all variations you often make things worse!

The rule of thumb is that any variation with 1.5 standard deviations of the average process performance (or what normally happens) is within what we call common cause variation and should be treated as normal process variation.

Any variation outside of 1.5 standard deviations though is treated as special cause variation – i.e. something unexpected has happened. These variations should be taken note of and investigated to make sure the process isn’t out of control. They will happen though!

To bring this back to construction work processes, common cause or normal variation in a work process will be caused by small but regular changes in the 5 M model.  Methods change because of differences in training and experience, people doing the task obviously change, there are differences in machinery and materials used and available as well as environmental changes from hour to hour.

Special cause variations are typically caused by things like mistakes and errors, health issues, machinery breakdowns, material failures, catastrophic design failures, and extreme weather events.

No alt text provided for this image

When special cause variations come close to a risk profile peak, as in the diagram above this is a near miss with the potential for a SIF. Often near misses are just that, because of a separation in time between the work process that went off track and the underlying hazard being active at that time – e.g. The scaffolding collapsed but there was nobody on it at the time.

When the extreme process variation meets the inherent risk profile a fatality occurs.

No alt text provided for this image

Process variation should be expected and anticipated to allow appropriate safety margins to be built into the work environment.

The causes and correlates for SIF’s are often different to non-SIF events and therefore require different control strategies (critical controls) to eliminate them.

The critical controls create the safety margin to allow for work process variation. In simple terms this means having the checks and controls in place to allow for mistakes and other process variation causes to happen without a SIF occurring.

A useful analogy, used often by Todd Conklin the internationally recognised Organisational Psychologist and H&S expert, can be drawn from aviation.  The major risk profile for aviation is obvious; presence of changing terrain below the aircraft, i.e. mountains which should not be flown into.  Flight paths are designed to consider geographic risks as well as optimum fuel and aircraft performance characteristics – the process as imagined. In reality minor variations to flight direction and height happen all the time, pilots change altitude to avoid storms and turbulence as they encounter them.  If there is a major process variation, for an example an issue with the aircraft, the one thing that gives the pilot the ability to keep the aircraft and its passengers safe is altitude (the time and space to respond)!

No alt text provided for this image

Appropriate altitude above and beyond the highest mountain peak – provides the safety margin (control environment) to allow for even special cause variation in the flight path and altitude.

Bringing this back to a construction context, how do we create the appropriate altitude (safety margins) that take into account the work and the inherent risk in that work and environment that it is performed in?

Controls are an investment decision and do not come free.  They have direct costs and opportunity costs associated with them.  The best control from a safety perspective may have a significant opportunity cost, for example other work may need to cease while a work process is underway.

One of New Zealand’s most prevalent sources of construction workplace fatalities is working close to traffic and mobile plant. The diagram below , adapted from a recent presentation by New Plymouth District Council’s David Langford shows the relative value of investing in controls to create the safety margin in a road working context.

No alt text provided for this image

The most effective control in protecting vulnerable road user such as construction workers and members of the public is without doubt closing the road while work activity is carried out. This does have a high opportunity cost that may be politically challenged. However at the other end of the scale we often rely on stop/go type controls which provide a much narrower safety margin – it is somewhat inevitable that over time these are not enough to eliminate the possibility of the work process and inherent risk profile meeting with a fatality as an outcome.

This view of construction accidents is supported by international research into the causes of SIFs.

Research in 2012 by Dr Tom Krause showed most potential SIFs are disproportionally related to unusual or abnormal precursor situations. The percentage of actual SIFs is higher for abnormal events (59%) than routine events (34%), although the greater overall number of potential SIFs (90%) are related to routine, everyday situations.

In 2016 Donald K Martin and Alison Black studied and grouped fatal workplace incident scenarios as follows:

No alt text provided for this image

Scenario 1 focuses on the breakdown of critical controls. Why the control broke down may have been firstly that it wasn’t a strong enough control for the related work activity and inherent risk. The control may also have been compromised because of special cause variation – i.e. a mistake or lack of knowledge of the control.

No alt text provided for this image

The control may also have degraded over time. We know that controls need to be resilient and without constant supervision and maintenance of controls they suffer from entropy or process decay.

Controls need constant monitoring and upkeep to ensure their effectiveness.

Scenario 2 is about situational awareness and experience.  Effectively the conditions on the worksite have changed to an extent that the risk profile has peaked and this may intersect with a work process that is underway.

These two scenarios make up over 70% of potential SIFs and therefore lead to some common approaches to avoiding them.

What can we do to reduce (and perhaps eliminate) Serious Injuries and Fatalities?

We have discussed how fatalities and serious injuries happen in construction and we now need to move to strategies on how to eliminate them.

Driving Awareness, recognition of and respect for the construction risk environment and the underlying work conditions that make up the risk profile.

The inherent and dynamic risk profile of the construction environment comes hand in hand with the work processes and the work environment they are undertaken in. Higher risk activities are known to the industry. Statistics on high fatality potential activity are available through tools such as the CHASNZ critical risk tool and the UKs Construction RIDDOR analytics, both available through the CHASNZ website.

No alt text provided for this image

Elimination strategies have always been the most powerful. Removing humans from inherently risky activities such as working at height can be achieved through mechanisation, offsite manufacturing as well as safety in design decisions early in the construction cycle.  These should be the first focus areas for clients, designers and constructors. The New Zealand forestry industry has made great strides in fatality reduction through mechanisation of its processes.

Separation and segregation strategies also are strong.  Reducing the number of high-risk activities and workplace conditions happening over the same geographic space and time through better scheduling and planning is an essential strategy. Project managers, designers and site controllers need the skills and competencies as well as budget and resources to achieve lower inherent risk profiles.

All workers should be trained and competent in understanding the inherent risks of a construction site. A heightened sense of awareness for all on site, treating high risk activity as it should be is a vital collective mindset that needs to be reinforced by dedicated, respected and skilled supervisors. This is not the role of health and safety professionals.

Supervisors have a role to be constantly monitoring the work environment, recognising the risks that work and site conditions bring and adapting plans and guiding teams to take account of the constantly evolving risk environment.

Improving the level of performance and reduction in the variability of Construction processes.

Work processes as planned (work as imagined) need to be set so that they consider the inherent risk profile of the work.  Safe work practices should be thought through for known hazardous activities.

Strategies to reduce the variation in how these processes are actually performed include:

A greater investment in a trained and competent workforce. Worker competency needs to be assessed and those with knowledge gaps helped with effective training plans which include appropriate supervision while they get up to speed.

No alt text provided for this image

Another of Todd Conklin’s ideas the focus on start-work protocols, where workers are supported by their supervisors not to start work until the right material, manpower, methods, machinery and mother nature (environment) conditions are within a tolerable level.  Start work protocols are more effective as opposed to stop work permissions, which in reality are very hard to do because of the sunk time and resource investment into a process which may already be going badly (don’t start until you are certain vs stop if you are uncertain).

Again this will take an increased investment in leading hands, foremen and supervisors with the right safety mindset to support teams and individuals keep within safe work process limits. These men and women need to be given the time, resource and skillset (gained through experience and training) to undertake more direct and indirect supervision rather than be on the tools themselves.

These workplace leaders are key in directing safe work (scenario 1), identifying when conditions change through constant dynamic risk assessments (scenario 2) and ensuring that controls do not degrade over time by constantly maintaining them. The dynamic risk assessment activity should not take away from the supervisor’s ability to effectively do their jobs. It is an innate competency where they are constantly assessing work and work conditions throughout the work period.

Supervisor health and safety training and competency assessment is currently a weak area for the New Zealand construction industry with minimal uptake of the quality supervisor training available on the market and the ConstructSafe supervisor competency assessment.

When special cause variations (major variations) to process occur, these need to be recognised and investigated collaboratively with the work teams to understand why they happened and to improve the process and control environment in the future. Near misses that could have been a SIF if not for timing should always be looked at using appreciative techniques (such as learning teams).  Opportunities for even better process performance that relate to process outcomes that are far better than expected should also be studied as these often lead to shift change in how processes are delivered in the future.

Again supervisors, leading hands and foremen need to be equipped with the mindsets, skills and competencies as well as be given the time to make this happen.

Investment in appropriate and resilient controls that create high resilience safety buffers.

The health and safety at work act puts special focus on officers and puts special duties on their shoulders.  This is for a reason – officers of an organisation have the greatest ability to influence the investment in the control environment.  They also need to understand the work activity and its related inherent risk profile in order to make the right decision on what the appropriate control strategy and investment should be. Organisations too have responsibilities to create a strong control environment. In particular when there are multiple organisations involved on a worksite, it is often the construction client who has the most ability to influence the control environment – after all they are paying for it. Those with the least influence unfortunately over the control environment are often the workers, who are often blamed for safety incidents but are left to do the best they can with the cards they have been dealt.

Control strategies should take into account what we know about work processes – that they vary! The control environment needs to be strong enough to create enough altitude to alwaysavoid a collision between an abnormally poorly performing work process and a peak in the inherent risk profile. We say always because if this is not the control investment mindset, given enough time, process science tells us that events such as mistakes and breakdowns will happen, and it is only chance timing whether they collide with the inherent risks on site.

Guidance on control environments is readily available.  Generically, elimination and substitution strategies as discussed above, always are the most effective. Investment in safety in design, off site manufacturing and mechanisation needs to increase to reduce the need for highly inherently hazardous activity.  Designing our roads with contra-flow gates and reducing the need for weed control activities are good examples of design decisions and investments that will improve our control environments.

Engineered controls that create fail safes, barriers and increase separation between the person and the hazard require more investment.  Our over- reliance on low value controls such as PPE and form filling activity needs to be rebalanced with controls that are resilient and always effective – i.e. that do not only rely on human compliance.

No alt text provided for this image

A potential resource for all businesses in New Zealand, big and small on what good controls look like are the construction risk cards, developed by ACC and WorkSafe. These are available online here:

Significantly reducing the number of construction fatalities is totally achievable given the right focus, investment and mindset of those who procure construction services, those who design and those who construct. Investment in supervisors is key as they are the gatekeepers of the work processes, schedule and plan work and maintain the control environment.

So often fatal and significant safety events are associated with the need to find and apportion blame, usually with an individual who like all of us may have deviated from the process or not seen and reacted to a change in the risk environment. True accountability in these types of events sits with those with the most influence over the investments in the control environment and the quality of supervision on site.