Root Cause Tip: What is the REAL Hierarchy of Controls?
You probably heard people refer to a “hierarchy of controls” before. In Chapter 3 of the TapRooT® Book, we reference two different versions – one by Dr Julian Christensen and one by William Haddon – that are shown below:
Dr Julian Christensen…
Dr William Haddon…
In our 5-Day TapRooT® Course (copyright 2008), we propose a hybrid combination of these two hierarchy models …
Recently, I saw a new example of a hierarchy of controls for health and safety in the new ANSI Standard for Occupational Health and Safety Management Systems (ANSI/AIHA Z10-2012). It proposes the following …
Most Effective …
3. Engineering Controls
5. Administrative Controls
6. Personal Protective Equipment
… Least Effective
For each of the categories it gave examples. For example, Administrative Controls included Hazard Communication Training. Personal Protective Equipment included Safety Harnesses and Gloves.
The similarities with previous hierarchies are obvious and the standard mentions a reference written by Fred Manuele as a source for work on hierarchy of controls.
But as I read the standard, I wondered about the arrangement of items 4, 5, and 6. Are warnings really more effective than personal protective equipment? Is hazard communication training really more effective than a safety harness? Is a warning sign (Danger! Sharp Edge!) really better than some cut resistant gloves?
I think they got the top of the hierarchy right (items 1-3) but didn’t get the bottom correct (items 4-6).
To the credit of the standards authors, they do mention using multiple controls to reduce risk if the hazard cannot be eliminated.
Next time you are developing corrective actions, remember to consider the hierarchy of controls. I think the TapRooT® version is preferable, but even the ANSI version is better than no guidance at all. All of these hierarchies help investigators consider the most effective controls rather than relying on the frequently used “easiest” controls to implement (often warnings and training).