Revisiting the hierarchy of controls
By Michael Wood
In recent weeks, I have had several conversations that ended up hinging on the hierarchy of controls – in most cases, the issue was that the hierarchy was being overlooked in the discussion and controls (such as personal protective equipment) that relied upon a high level of worker "compliance" were being incorrectly treated as controls that largely eliminated the hazards in question.
I have written before that I believe the essence of the hierarchy of controls is less focused on a discussion of "engineering controls" versus "administrative controls" versus "work practice controls" and more focused on a simple set of principles:
- First, a control is preferred if it minimizes the need for the worker's active participation to ensure its effectiveness.
For example, the reason that ventilation is superior to respirator use is not because respirators, if properly used, are inherently less effective. It is because they are more vulnerable to improper use by individual workers. In the same way, a guardrail system is superior to individual fall restraint systems not because the fall restraint systems do not work, but because the individual workers do not need to "hook up" the guardrail systems in order to make them work properly.
Simply put, no hazard can be considered to be fully addressed as long as there is a possibility that the control can fail due to "worker error" (whether the result of poor training, poor work practices, contradictory expectations, inattention, or anything else).
- Second, a control is preferred if, to the degree it does rely upon active participation on the part of the worker, the worker's actions reflect natural and largely automatic actions.
This principle is at play in having dials and valves turn the same direction and not using red for "on" and green for "off" on controls (you'd be surprised how often these simple approaches are overlooked for one reason or another). To take an extreme example, there is really no mechanical reason why turning the steering wheel of an automobile to the right also turns the car in that direction. But, if you reverse it, no amount of "worker training" is going to overcome the worker's natural and conditioned inclinations. I haven't run across a worksite that goes to that extreme – but I've encountered a few that have come close to it.
- Third, a control is preferred if it cannot be easily defeated without active and conscious participation on the part of the worker being protected.
This is the principle at play in the use of interlocks and other "passive" guarding systems. It also is at play in covering a fixed ladder that should not normally be used with a screen making it impossible to climb rather than relying upon a "no unauthorized entry" sign The sign will be completely effective if it is followed. But, clearly, the screen over the ladder is a superior control – not because it works better when properly employed, but because it is much less likely to be actively subverted (and, therefore, much more likely to be properly employed).
- Finally, a control is preferred if it eliminates or at least minimizes injury to the employee if a failure does, in fact, occur.
Substituting a less hazardous chemical for a more hazardous one will not eliminate the possibility of a spill. But it is clearly a good idea that reduces the inherent risk of the activity in question, because it reduces the consequences, even though it does not reduce the likelihood of a problem occurring.
As we design our workplace activities, these principles – especially if designed into the basic work processes – can reduce the need for enforcement and progressive discipline, as well as the need to focus excessive attention on shifting worker behavior. They can make our workplaces more intrinsically safe – and that's where we are likely to see the greatest benefits in the long run.
Oregon OSHA Administrator
If you want to receive the Resource Newsletter, sign up for future issues here.
Reprinting, excerpting, or plagiarizing any part of this publication is fine with us. Please send us a copy of your publication or inform the Resource editor as a courtesy. If you have questions about the information in Resource, please call 503-378-3272.
For general information, technical answers, or information about Oregon OSHA services, please call 503-378-3272 or toll-free within Oregon, 800-922-2689.