Skip to Content, Skip to Navigation

Safeguard OSH Solutions - Thomson Reuters

Safeguard OSH Solutions - Thomson Reuters

Safeguard Magazine

Hierarchy of control

It seems reasonable to expect that considering health and safety in the design phase of a project would produce better outcomes, but where is the evidence? Researcher HELEN LINGARD reports.

Advocates for safety in design often cite Szymberski’s time/safety influence curve, which shows the ability to influence safety at its highest at the very early conceptual stages, and steadily declining through detailed engineering and procurement stages, diminishing to pretty low by the time construction actually begins.

But where is the evidence to demonstrate the truth of this curve? Devising a research method to test the proposition that early consideration of health and safety in project decision-making would produce better outcomes was a challenge, but one which we tackled successfully by using an approach based on a leading indicator – the hierarchy of controls – rather than on an unreliable and luck-of-the-draw lagging indicator – the number of accidents.

It’s worth reminding ourselves about the hierarchy of controls (HOC) before we go on to look at how we built the practical HOC measure that underpinned our benchmarking study.

The HOC classifies the ways of dealing with health and safety hazards or risks according to the level of effectiveness of the control applied. In descending order of effectiveness the HOC levels are:

  • • 
    elimination (the most effective control);
  • • 
    substitution (replace something that produces a hazard with something less hazardous);
  • • 
    engineering controls (which physically isolate people from hazards);
  • • 
    administrative controls (safe work procedures, job rotation schemes etc); and
  • • 
    personal protective equipment (thought to be the last line of defence and the least effective form of control).

The higher levels of control (elimination, substitution, and engineering) change the physical work environment to make it safer or healthier; the lower levels (administrative and PPE) rely on human behaviour for their effectiveness. Controls that rely on human behaviour are less reliable because humans make mistakes; although much emphasised and visible on work sites, PPE should be seen as a solution of last resort.


Before looking at the research project itself, let’s consider the construction industry’s health and safety performance, which is still relatively poor. Construction firms have had some success using a management system and audit approach, but the nub of the problem is that most effort has only been put in during the construction period itself – even though there is increasing recognition that the root causes of safety incidents can be traced back to problems inherent in how the industry operates.

Most contemporary models of accident causation recognise the importance of organisational issues and management actions in contributing to workplace accidents. Some construction accidents can be attributed, at least in part, to failures arising before work commences on site, ie: the project planning and design stages.

The construction environment is much more complex than single-organisation environments. Construction projects are delivered by several organisations across many teams. There is growing recognition that many contributors to construction projects either make or influence decisions that have potential to affect the health and safety of construction workers. In some instances, safety risks in the construction environment can be traced back to decisions made before construction work has begun.

The fragmentation of the construction supply chain has been identified as a critical factor in the emergence of H&S problems during construction projects. The development of a unity of purpose with regard to the H&S of construction workers can be challenging as many contributors to the construction design and delivery process are engaged at different times under different contracts.

In particular, the traditional separation between the design and construction functions can be a barrier to sharing information about H&S. Traditional project procurement strategies can limit access to important H&S knowledge in the design phase. Modifications introduced during construction to improve H&S can produce improvements but often fail to eliminate an inherently dangerous activity.

For example, fixing rails or anchor points for fall arrest devices are important H&S improvements that can be implemented during construction, but they do not eliminate the inherently dangerous activity of working at height.


A five-year international benchmarking project involving analysis of 23 construction projects in Australia, New Zealand and the United States has come up with five best practice principles for improving health and safety management. They are:

  • • 
    address H&S as early as possible in the project;
  • • 
    identify and consult all relevant stakeholders;
  • • 
    ensure construction process knowledge is available to decision-makers;
  • • 
    implement the hierarchy of controls in decision-making; and
  • • 
    review and continuously improve performance.

These construction projects – 10 in Australia/NZ and 13 in the USA – were used to explore the relationship between the timing of risk control decision making and the effectiveness of risk control outcomes.

What distinguishes this research is that it was a prospective study, not a retrospective study. Previous studies had looked at completed construction projects and concluded there was merit in considering H&S at the design stage. Our study looked at each of the 43 projects from the time they began until they were finished, so that we could analyse H&S activity before construction work started and then look at H&S outcomes once construction was complete.

In-depth interviews were conducted with stakeholders involved in the planning, design and construction features of work particular to each project. Interviews explored the timing and sequence of key decisions and the influences that were at play as these decisions were taken in the context of the project. A total of 288 interviews were conducted.

For each construction project we worked with project participants to identify an average of 1.9 features of the planned work which posed a particular H&S problem or challenge. A total of 43 work features were included in the research, and an average of 6.7 interviews were conducted for each work feature as each project progressed.

Once each H&S risk had been identified for each work feature, we identified the methods used to control each risk. We obtained this information over time by interviews, examination of plans and drawings, and on-site observations. The methods used were classified using the HOC assessment tool.

The point in the project at which a risk control solution was selected, ie at planning stage, at design stage, or at construction stage, was a key factor in shaping the quality of risk controls implemented (or HOC outcomes).


Another key finding was that the involvement of people with construction process knowledge in design decision making was also statistically linked to better than average HOC scores.

We looked at this because there is evidence that design professionals are well versed in H&S factors associated with the use and operation of a building but lack knowledge of H&S factors which come into play during construction.

We used social network analysis to measure how communications worked in each of the 13 projects, focusing particularly on communications by the construction contractor to measure the extent to which design decision-makers were able to access to construction H&S knowledge.

The results were statistically significant: in projects where more upper level H&S risk controls were applied, the construction contractor was more engaged in the pre-construction stage. This demonstrates a clear need to push construction process knowledge upstream to make it available to decision makers during the design stage.


We found a positive relationship between the consideration of H&S in the early stages of a project and the quality of risk controls implemented in the construction stages. That is, when H&S risks were identified and control decisions were taken before the commencement of construction, it was more likely that H&S risks would be controlled at source, through the use of the higher levels of risk control: elimination, substitution or engineering measures.

Conversely, when decisions were left until the construction stage, it was more likely that measures implemented to control H&S were at the lower levels of risk control: administrative and PPE.

To conclude: the results provide evidence that high level or technological risk controls – which are much more effective at reducing risk – are more likely to be implemented if H&S risks are considered and controls selected in the planning and design phases of construction projects.


  • • 
    Lingard et al (2015),  The relationship between pre-construction decision-making and the effectiveness of risk control (ECAM Vol. 22 No.1, pp108-124).
  • • 
    Research to Practice Report, 2015, RMIT University Centre for Construction Work Health and Safety Research.

Distinguished Professor Helen Lingard is based at the Centre for Construction Work Health and Safety Research at RMIT University in Melbourne.

comments powered by Disqus

From Safeguard Magazine

Table of Contents